def _create(name,
pretrained=True,
channels=3,
classes=80,
autoshape=True,
verbose=True,
device=None):
"""Creates a specified YOLOv5 model
Arguments:
name (str): name of model, i.e. 'yolov5s'
pretrained (bool): load pretrained weights into the model
channels (int): number of input channels
classes (int): number of model classes
autoshape (bool): apply YOLOv5 .autoshape() wrapper to model
verbose (bool): print all information to screen
device (str, torch.device, None): device to use for model parameters
Returns:
YOLOv5 pytorch model
"""
from pathlib import Path
from models.yolo import Model, attempt_load
from utils.general import check_requirements, set_logging
from utils.google_utils import attempt_download
from utils.torch_utils import select_device
check_requirements(requirements=Path(__file__).parent / 'requirements.txt',
exclude=('tensorboard', 'pycocotools', 'thop',
'opencv-python'))
set_logging(verbose=verbose)
fname = Path(name).with_suffix('.pt') # checkpoint filename
try:
if pretrained and channels == 3 and classes == 80:
model = attempt_load(
fname,
map_location=torch.device('cpu')) # download/load FP32 model
else:
cfg = list((Path(__file__).parent /
'models').rglob(f'{name}.yaml'))[0] # model.yaml path
model = Model(cfg, channels, classes) # create model
if pretrained:
ckpt = torch.load(attempt_download(fname),
map_location=torch.device('cpu')) # load
msd = model.state_dict() # model state_dict
csd = ckpt['model'].float().state_dict(
) # checkpoint state_dict as FP32
csd = {
k: v
for k, v in csd.items() if msd[k].shape == v.shape
} # filter
model.load_state_dict(csd, strict=False) # load
if len(ckpt['model'].names) == classes:
model.names = ckpt[
'model'].names # set class names attribute
if autoshape:
model = model.autoshape() # for file/URI/PIL/cv2/np inputs and NMS
device = select_device('0' if torch.cuda.is_available() else 'cpu'
) if device is None else torch.device(device)
return model.to(device)
except Exception as e:
help_url = 'https://github.com/ultralytics/yolov5/issues/36'
s = 'Cache may be out of date, try `force_reload=True`. See %s for help.' % help_url
raise Exception(s) from e
np = sum([x.numel() for x in m_.parameters()]) # number params
m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params
logger.info('%3s%18s%3s%10.0f %-40s%-30s' % (i, f, n, np, t, args)) # print
save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
layers.append(m_)
ch.append(c2)
return nn.Sequential(*layers), sorted(save)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')
parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
opt = parser.parse_args()
opt.cfg = check_file(opt.cfg) # check file
set_logging()
device = select_device(opt.device)
# Create model
model = Model(opt.cfg).to(device)
model.train()
# Profile
# img = torch.rand(8 if torch.cuda.is_available() else 1, 3, 640, 640).to(device)
# y = model(img, profile=True)
# Tensorboard
# from torch.utils.tensorboard import SummaryWriter
# tb_writer = SummaryWriter()
# print("Run 'tensorboard --logdir=models/runs' to view tensorboard at http://localhost:6006/")
# tb_writer.add_graph(model.model, img) # add model to tensorboard
def detect(save_img=True):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
elif source == 'own_camera':
pass
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
if source == 'own_camera':
camera_matrix = np.array(
[[3.50379164e+03, 0.00000000e+00, 7.96197449e+02],
[0.00000000e+00, 3.49926605e+03, 6.94741115e+02],
[0.00000000e+00, 0.00000000e+00, 1.00000000e+00]])
dist_coefs = np.array([
-6.93155538e-01, 2.44136043e+00, -8.62122690e-03, 6.01043129e-03,
-1.64531179e+01
])
camera = pylon.InstantCamera(
pylon.TlFactory.GetInstance().CreateFirstDevice())
camera.Open()
converter = pylon.ImageFormatConverter()
# camera = pylon.InstantCamera(pylon.TlFactory.GetInstance().CreateFirstDevice())
# ========== Grabing Continusely (video) with minimal delay ==========
camera.StartGrabbing(pylon.GrabStrategy_LatestImageOnly)
camera.ExposureTimeAbs = 80000
# ========== converting to opencv bgr format ==========
converter.OutputPixelFormat = pylon.PixelType_BGR8packed
converter.OutputBitAlignment = pylon.OutputBitAlignment_MsbAligned
pygame.init()
# clock = pygame.time.Clock()
display_width = 1920
display_height = 1080
screen = pygame.display.get_surface()
screen = pygame.display.set_mode((display_width, display_height),
pygame.FULLSCREEN)
modes = pygame.display.list_modes()
pygame.display.set_mode(max(modes))
black = (0, 0, 0)
white = (255, 255, 255)
# screen.fill(black) # show black image
while camera.IsGrabbing():
# for projector
for event in pygame.event.get():
if event.type == pygame.QUIT:
crashed = True
screen.fill(black) # show black image
grabResult = camera.RetrieveResult(
5000, pylon.TimeoutHandling_ThrowException)
if grabResult.GrabSucceeded():
# image = converter.Convert(grabResult)
image = converter.Convert(grabResult)
img = image.GetArray()
width = int(img.shape[1])
height = int(img.shape[0])
dim = (width, height)
newcameramtx, roi = cv2.getOptimalNewCameraMatrix(
camera_matrix, dist_coefs, (width, height), 1,
(width, height))
dst = cv2.undistort(img, camera_matrix, dist_coefs, None,
newcameramtx)
x, y, width, height = roi
img0 = dst[y:y + height, x:x + width]
frame = img0
# frame = cv2.resize(frame, (int(frame.shape[1]/1.5),int(frame.shape[0]/1.5)))
img0 = frame.copy()
im0s = img0
img_size = 640
img, ratio, (dw, dh) = letterbox(img0, new_shape=img_size)
img = img[:, :, ::-1].transpose(2, 0,
1) # BGR to RGB, to 3x416x416
img = np.ascontiguousarray(img)
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path('0'), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path('0'), '', im0s
save_path = str(save_dir / p.name)
# txt_path = str(save_dir / 'labels' / p.stem) + (
# '_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(img0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
classes = []
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
s1 = '%g' % (n)
classes.append([names[int(c)], s1])
df_cls_info = pd.DataFrame(classes,
columns=['Class', 'number'])
df_cls_info.to_csv('class_num.csv')
print("\n number of object : ", len(det[:, -1]))
print(df_cls_info)
# Write results
box_info = []
box3D = {}
indice_3D = 0
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
# with open(txt_path + '.txt', 'a') as f:
# f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
img0,
label=label,
color=colors[int(cls)],
line_thickness=3)
c1, c2 = (int(xyxy[0]),
int(xyxy[1])), (int(xyxy[2]),
int(xyxy[3]))
# print("name : ",names[int(cls)])
# print("opt",opt.classflash)
if names[int(cls)] == opt.classflash:
box3D[indice_3D] = (int(xyxy[0]), int(xyxy[1]),
int(xyxy[2]), int(xyxy[3]))
elif opt.classflash == "all":
box3D[indice_3D] = (int(xyxy[0]), int(xyxy[1]),
int(xyxy[2]), int(xyxy[3]))
indice_3D += 1
confident = '%g' % (conf)
box_info.append([label, c1, c2, confident])
# print(box3D)
if box3D != {}:
depth, box_min, dists = D435camera(box3D, img0)
# print(depth, box_min, dists)
# screen.fill(black) # show black image
if dists != {}:
img = drawimage(box3D, dists,
(200, 0, 0)) # set color
# _,img = cv2.threshold(img,0,255,cv2.THRESH_BINARY_INV)
imgpro = pygame.image.frombuffer(
img.tostring(), img.shape[1::-1], "RGB")
screen.blit(imgpro, (50, 50))
else:
screen.fill(black)
pygame.display.flip()
pygame.display.update()
# df_box = pd.DataFrame(box_info,columns=['Class','top left','bottom_right','confidence'])
# df_box.to_csv('box_info.csv')
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
im0 = cv2.resize(
im0, (int(im0.shape[1] / 1.5), int(im0.shape[0] / 1.5)))
cv2.imshow("asd", im0)
if cv2.waitKey(1) == ord('q'): # q to quit
break
if save_txt or save_img:
print('Results saved to %s' % save_dir)
print('Done. (%.3fs)' % (time.time() - t0))
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path(path[i]), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path(path), '', im0s
save_path = str(save_dir / p.name)
txt_path = str(save_dir / 'labels' / p.stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3)
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
# Stream results
if view_img:
cv2.imshow("asd", im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
# cv2.imwrite(save_path, im0)
pass
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
print('Results saved to %s' % save_dir)
print('Done. (%.3fs)' % (time.time() - t0))
help='dynamic ONNX axes')
parser.add_argument('--grid',
action='store_true',
help='export Detect() layer grid')
parser.add_argument('--device',
default='cpu',
help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
parser.add_argument('--hyp',
type=str,
default='../yolov5/data/hyp.scratch.yaml',
help='hyperparameters path')
parser.add_argument('--cfg', type=str, default='', help='model.yaml path')
opt = parser.parse_args()
opt.img_size *= 2 if len(opt.img_size) == 1 else 1 # expand
print(opt)
general.set_logging()
t = time.time()
# Load PyTorch model
device = torch_utils.select_device(opt.device)
model = experimental.attempt_load(opt.weights,
map_location=device) # load FP32 model
labels = model.names
# Checks
gs = int(max(model.stride)) # grid size (max stride)
opt.img_size = [general.check_img_size(x, gs)
for x in opt.img_size] # verify img_size are gs-multiples
# Input,picture
img = cv2.imread(opt.img_path)
def test(
data,
weights=None,
batch_size=32,
imgsz=640,
conf_thres=0.001,
iou_thres=0.6, # for NMS
save_json=False,
single_cls=False,
augment=False,
verbose=False,
model=None,
dataloader=None,
save_dir=Path(''), # for saving images
save_txt=False, # for auto-labelling
save_conf=False,
plots=True,
log_imgs=0): # number of logged images
# Initialize/load model and set device
training = model is not None
if training: # called by train.py
device = next(model.parameters()).device # get model device
else: # called directly
set_logging()
device = select_device(opt.device, batch_size=batch_size)
save_txt = opt.save_txt # save *.txt labels
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
# Multi-GPU disabled, incompatible with .half() https://github.com/ultralytics/yolov5/issues/99
# if device.type != 'cpu' and torch.cuda.device_count() > 1:
# model = nn.DataParallel(model)
# Half
half = device.type != 'cpu' # half precision only supported on CUDA
if half:
model.half()
# Configure
model.eval()
is_coco = data.endswith('coco.yaml') # is COCO dataset
with open(data) as f:
data = yaml.load(f, Loader=yaml.FullLoader) # model dict
check_dataset(data) # check
nc = 1 if single_cls else int(data['nc']) # number of classes
iouv = torch.linspace(0.5, 0.95,
10).to(device) # iou vector for [email protected]:0.95
niou = iouv.numel()
# Logging
log_imgs, wandb = min(log_imgs, 100), None # ceil
try:
import wandb # Weights & Biases
except ImportError:
log_imgs = 0
# Dataloader
if not training:
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
path = data['test'] if opt.task == 'test' else data[
'val'] # path to val/test images
dataloader = create_dataloader(path,
imgsz,
batch_size,
model.stride.max(),
opt,
pad=0.5,
rect=True)[0]
seen = 0
names = {
k: v
for k, v in enumerate(
model.names if hasattr(model, 'names') else model.module.names)
}
coco91class = coco80_to_coco91_class()
s = ('%20s' + '%12s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R',
'[email protected]', '[email protected]:.95')
p, r, f1, mp, mr, map50, map, t0, t1 = 0., 0., 0., 0., 0., 0., 0., 0., 0.
loss = torch.zeros(3, device=device)
jdict, stats, ap, ap_class, wandb_images = [], [], [], [], []
for batch_i, (img, targets, paths,
shapes) in enumerate(tqdm(dataloader, desc=s)):
img = img.to(device, non_blocking=True)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
targets = targets.to(device)
nb, _, height, width = img.shape # batch size, channels, height, width
whwh = torch.Tensor([width, height, width, height]).to(device)
# Disable gradients
with torch.no_grad():
# Run model
t = time_synchronized()
inf_out, train_out = model(
img, augment=augment) # inference and training outputs
t0 += time_synchronized() - t
# Compute loss
if training: # if model has loss hyperparameters
loss += compute_loss([x.float() for x in train_out], targets,
model)[1][:3] # box, obj, cls
# Run NMS
t = time_synchronized()
output = non_max_suppression(inf_out,
conf_thres=conf_thres,
iou_thres=iou_thres)
t1 += time_synchronized() - t
# Statistics per image
for si, pred in enumerate(output):
labels = targets[targets[:, 0] == si, 1:]
nl = len(labels)
tcls = labels[:, 0].tolist() if nl else [] # target class
path = Path(paths[si])
seen += 1
if len(pred) == 0:
if nl:
stats.append((torch.zeros(0, niou, dtype=torch.bool),
torch.Tensor(), torch.Tensor(), tcls))
continue
# Predictions
predn = pred.clone()
scale_coords(img[si].shape[1:], predn[:, :4], shapes[si][0],
shapes[si][1]) # native-space pred
# Append to text file
if save_txt:
gn = torch.tensor(shapes[si][0])[[1, 0, 1, 0
]] # normalization gain whwh
for *xyxy, conf, cls in predn.tolist():
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh,
conf) if save_conf else (cls,
*xywh) # label format
with open(save_dir / 'labels' / (path.stem + '.txt'),
'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
# W&B logging
if plots and len(wandb_images) < log_imgs:
box_data = [{
"position": {
"minX": xyxy[0],
"minY": xyxy[1],
"maxX": xyxy[2],
"maxY": xyxy[3]
},
"class_id": int(cls),
"box_caption": "%s %.3f" % (names[cls], conf),
"scores": {
"class_score": conf
},
"domain": "pixel"
} for *xyxy, conf, cls in pred.tolist()]
boxes = {
"predictions": {
"box_data": box_data,
"class_labels": names
}
} # inference-space
wandb_images.append(
wandb.Image(img[si], boxes=boxes, caption=path.name))
# Append to pycocotools JSON dictionary
if save_json:
# [{"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}, ...
image_id = int(
path.stem) if path.stem.isnumeric() else path.stem
box = xyxy2xywh(predn[:, :4]) # xywh
box[:, :2] -= box[:, 2:] / 2 # xy center to top-left corner
for p, b in zip(pred.tolist(), box.tolist()):
jdict.append({
'image_id':
image_id,
'category_id':
coco91class[int(p[5])] if is_coco else int(p[5]),
'bbox': [round(x, 3) for x in b],
'score':
round(p[4], 5)
})
# Assign all predictions as incorrect
correct = torch.zeros(pred.shape[0],
niou,
dtype=torch.bool,
device=device)
if nl:
detected = [] # target indices
tcls_tensor = labels[:, 0]
# target boxes
tbox = xywh2xyxy(labels[:, 1:5]) * whwh
scale_coords(img[si].shape[1:], tbox, shapes[si][0],
shapes[si][1]) # native-space labels
# Per target class
for cls in torch.unique(tcls_tensor):
ti = (cls == tcls_tensor).nonzero(as_tuple=False).view(
-1) # prediction indices
pi = (cls == pred[:, 5]).nonzero(as_tuple=False).view(
-1) # target indices
# Search for detections
if pi.shape[0]:
# Prediction to target ious
ious, i = box_iou(predn[pi, :4], tbox[ti]).max(
1) # best ious, indices
# Append detections
detected_set = set()
for j in (ious > iouv[0]).nonzero(as_tuple=False):
d = ti[i[j]] # detected target
if d.item() not in detected_set:
detected_set.add(d.item())
detected.append(d)
correct[
pi[j]] = ious[j] > iouv # iou_thres is 1xn
if len(
detected
) == nl: # all targets already located in image
break
# Append statistics (correct, conf, pcls, tcls)
stats.append(
(correct.cpu(), pred[:, 4].cpu(), pred[:, 5].cpu(), tcls))
# Plot images
if plots and batch_i < 3:
f = save_dir / f'test_batch{batch_i}_labels.jpg' # filename
plot_images(img, targets, paths, f, names) # labels
f = save_dir / f'test_batch{batch_i}_pred.jpg'
plot_images(img, output_to_target(output, width, height), paths, f,
names) # predictions
# Compute statistics
stats = [np.concatenate(x, 0) for x in zip(*stats)] # to numpy
if len(stats) and stats[0].any():
p, r, ap, f1, ap_class = ap_per_class(*stats,
plot=plots,
save_dir=save_dir,
names=names)
p, r, ap50, ap = p[:, 0], r[:, 0], ap[:, 0], ap.mean(
1) # [P, R, [email protected], [email protected]:0.95]
mp, mr, map50, map = p.mean(), r.mean(), ap50.mean(), ap.mean()
nt = np.bincount(stats[3].astype(np.int64),
minlength=nc) # number of targets per class
else:
nt = torch.zeros(1)
# W&B logging
if plots and wandb and wandb.run:
wandb.log({"Images": wandb_images})
wandb.log({
"Validation": [
wandb.Image(str(x), caption=x.name)
for x in sorted(save_dir.glob('test*.jpg'))
]
})
# Print results
pf = '%20s' + '%12.3g' * 6 # print format
print(pf % ('all', seen, nt.sum(), mp, mr, map50, map))
# Print results per class
if verbose and nc > 1 and len(stats):
for i, c in enumerate(ap_class):
print(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))
# Print speeds
t = tuple(x / seen * 1E3
for x in (t0, t1, t0 + t1)) + (imgsz, imgsz, batch_size) # tuple
if not training:
print(
'Speed: %.1f/%.1f/%.1f ms inference/NMS/total per %gx%g image at batch-size %g'
% t)
# Save JSON
if save_json and len(jdict):
w = Path(weights[0] if isinstance(weights, list) else weights
).stem if weights is not None else '' # weights
anno_json = glob.glob('../coco/annotations/instances_val*.json')[
0] # annotations json
pred_json = str(save_dir / f"{w}_predictions.json") # predictions json
print('\nEvaluating pycocotools mAP... saving %s...' % pred_json)
with open(pred_json, 'w') as f:
json.dump(jdict, f)
try: # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb
from pycocotools.coco import COCO
from pycocotools.cocoeval import COCOeval
anno = COCO(anno_json) # init annotations api
pred = anno.loadRes(pred_json) # init predictions api
eval = COCOeval(anno, pred, 'bbox')
if is_coco:
eval.params.imgIds = [
int(Path(x).stem) for x in dataloader.dataset.img_files
] # image IDs to evaluate
eval.evaluate()
eval.accumulate()
eval.summarize()
map, map50 = eval.stats[:
2] # update results ([email protected]:0.95, [email protected])
except Exception as e:
print('ERROR: pycocotools unable to run: %s' % e)
# Return results
if not training:
print('Results saved to %s' % save_dir)
model.float() # for training
maps = np.zeros(nc) + map
for i, c in enumerate(ap_class):
maps[c] = ap[i]
return (mp, mr, map50, map,
*(loss.cpu() / len(dataloader)).tolist()), maps, t
def detect(opt, save_img=False):
out, source, weights, view_img, save_txt, imgsz = \
opt.output, opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source == '0' or source.startswith('rtsp') or source.startswith(
'http') or source.endswith('.txt')
# Initialize
set_logging()
device = select_device(opt.device)
if os.path.exists(out):
shutil.rmtree(out) # delete output folder
os.makedirs(out) # make new output folder
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']) # load weights
modelc.to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(names))]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = path[i], '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = path, '', im0s
save_path = str(Path(out) / Path(p).name)
txt_path = str(Path(out) / Path(p).stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * 5 + '\n') %
(cls, *xywh)) # label format
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3)
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
# Stream results
if view_img:
cv2.imshow(p, im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
cv2.imwrite(save_path, im0)
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
print('Results saved to %s' % Path(out))
if platform.system() == 'Darwin' and not opt.update: # MacOS
os.system('open ' + save_path)
print('Done. (%.3fs)' % (time.time() - t0))
def test(
data,
weights=None,
batch_size=16,
imgsz=640,
conf_thres=0.001,
iou_thres=0.6, # for NMS
save_json=False,
single_cls=False,
augment=False,
verbose=False,
model=None,
dataloader=None,
save_dir=Path(''), # for saving images
save_txt=False, # for auto-labelling
save_conf=False,
plots=True):
# Initialize/load model and set device
training = model is not None
if training: # called by train.py
device = next(model.parameters()).device # get model device
else: # called directly
set_logging()
device = select_device(opt.device, batch_size=batch_size)
save_txt = opt.save_txt # save *.txt labels
# Remove previous
if os.path.exists(save_dir):
shutil.rmtree(save_dir) # delete dir
os.makedirs(save_dir) # make new dir
if save_txt:
out = save_dir / 'autolabels'
if os.path.exists(out):
shutil.rmtree(out) # delete dir
os.makedirs(out) # make new dir
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
# Multi-GPU disabled, incompatible with .half() https://github.com/ultralytics/yolov5/issues/99
# if device.type != 'cpu' and torch.cuda.device_count() > 1:
# model = nn.DataParallel(model)
# Half
half = device.type != 'cpu' # half precision only supported on CUDA
if half:
model.half()
#prune
# torch_utils.prune(model, 0.3)
# Configure
model.eval()
with open(data) as f:
data = yaml.load(f, Loader=yaml.FullLoader) # model dict
check_dataset(data) # check
nc = 1 if single_cls else int(data['nc']) # number of classes
iouv = torch.linspace(0.5, 0.95,
10).to(device) # iou vector for [email protected]:0.95
niou = iouv.numel()
# Dataloader
if not training:
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
path = data['test'] if opt.task == 'test' else data[
'val'] # path to val/test images
dataloader = create_dataloader(path,
imgsz,
batch_size,
model.stride.max(),
opt,
hyp=None,
augment=False,
cache=False,
pad=0.5,
rect=True)[0]
seen = 0
names = model.names if hasattr(model, 'names') else model.module.names
coco91class = coco80_to_coco91_class()
s = ('%20s' + '%12s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R',
'[email protected]', '[email protected]:.95')
p, r, f1, mp, mr, map50, map, t0, t1 = 0., 0., 0., 0., 0., 0., 0., 0., 0.
loss = torch.zeros(3, device=device)
jdict, stats, ap, ap_class = [], [], [], []
for batch_i, (img, targets, paths,
shapes) in enumerate(tqdm(dataloader, desc=s)):
img = img.to(device, non_blocking=True)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
targets = targets.to(device)
nb, _, height, width = img.shape # batch size, channels, height, width
whwh = torch.Tensor([width, height, width, height]).to(device)
# Disable gradients
with torch.no_grad():
# Run model
t = time_synchronized()
inf_out, train_out = model(
img, augment=augment) # inference and training outputs
t0 += time_synchronized() - t
# Compute loss
if training: # if model has loss hyperparameters
loss += compute_loss([x.float() for x in train_out], targets,
model)[1][:3] # box, obj, cls
# Run NMS
t = time_synchronized()
output = non_max_suppression(inf_out,
conf_thres=conf_thres,
iou_thres=iou_thres)
t1 += time_synchronized() - t
# Statistics per image
for si, pred in enumerate(output):
labels = targets[targets[:, 0] == si, 1:]
nl = len(labels)
tcls = labels[:, 0].tolist() if nl else [] # target class
seen += 1
if pred is None:
if nl:
stats.append((torch.zeros(0, niou, dtype=torch.bool),
torch.Tensor(), torch.Tensor(), tcls))
continue
# Append to text file
if save_txt:
gn = torch.tensor(shapes[si][0])[[1, 0, 1, 0
]] # normalization gain whwh
x = pred.clone()
x[:, :4] = scale_coords(img[si].shape[1:], x[:, :4],
shapes[si][0],
shapes[si][1]) # to original
for *xyxy, conf, cls in x:
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, conf,
*xywh) if save_conf else (cls,
*xywh) # label format
with open(str(out / Path(paths[si]).stem) + '.txt',
'a') as f:
f.write(('%g ' * len(line) + '\n') % line)
# Clip boxes to image bounds
clip_coords(pred, (height, width))
# Append to pycocotools JSON dictionary
if save_json:
# [{"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}, ...
image_id = Path(paths[si]).stem
box = pred[:, :4].clone() # xyxy
scale_coords(img[si].shape[1:], box, shapes[si][0],
shapes[si][1]) # to original shape
box = xyxy2xywh(box) # xywh
box[:, :2] -= box[:, 2:] / 2 # xy center to top-left corner
for p, b in zip(pred.tolist(), box.tolist()):
jdict.append({
'image_id':
int(image_id) if image_id.isnumeric() else image_id,
'category_id':
coco91class[int(p[5])],
'bbox': [round(x, 3) for x in b],
'score':
round(p[4], 5)
})
# Assign all predictions as incorrect
correct = torch.zeros(pred.shape[0],
niou,
dtype=torch.bool,
device=device)
if nl:
detected = [] # target indices
tcls_tensor = labels[:, 0]
# target boxes
tbox = xywh2xyxy(labels[:, 1:5]) * whwh
# Per target class
for cls in torch.unique(tcls_tensor):
ti = (cls == tcls_tensor).nonzero(as_tuple=False).view(
-1) # prediction indices
pi = (cls == pred[:, 5]).nonzero(as_tuple=False).view(
-1) # target indices
# Search for detections
if pi.shape[0]:
# Prediction to target ious
ious, i = box_iou(pred[pi, :4], tbox[ti]).max(
1) # best ious, indices
# Append detections
detected_set = set()
for j in (ious > iouv[0]).nonzero(as_tuple=False):
d = ti[i[j]] # detected target
if d.item() not in detected_set:
detected_set.add(d.item())
detected.append(d)
correct[
pi[j]] = ious[j] > iouv # iou_thres is 1xn
if len(
detected
) == nl: # all targets already located in image
break
# Append statistics (correct, conf, pcls, tcls)
stats.append(
(correct.cpu(), pred[:, 4].cpu(), pred[:, 5].cpu(), tcls))
# Plot images
if plots and batch_i < 1:
f = save_dir / f'test_batch{batch_i}_gt.jpg' # filename
plot_images(img, targets, paths, str(f), names) # ground truth
f = save_dir / f'test_batch{batch_i}_pred.jpg'
plot_images(img, output_to_target(output, width, height), paths,
str(f), names) # predictions
# Compute statistics
stats = [np.concatenate(x, 0) for x in zip(*stats)] # to numpy
if len(stats) and stats[0].any():
p, r, ap, f1, ap_class = ap_per_class(*stats,
plot=plots,
fname=save_dir /
'precision-recall_curve.png')
p, r, ap50, ap = p[:, 0], r[:, 0], ap[:, 0], ap.mean(
1) # [P, R, [email protected], [email protected]:0.95]
mp, mr, map50, map = p.mean(), r.mean(), ap50.mean(), ap.mean()
nt = np.bincount(stats[3].astype(np.int64),
minlength=nc) # number of targets per class
else:
nt = torch.zeros(1)
# Print results
pf = '%20s' + '%12.3g' * 6 # print format
print(pf % ('all', seen, nt.sum(), mp, mr, map50, map))
# Print results per class
if verbose and nc > 1 and len(stats):
for i, c in enumerate(ap_class):
print(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))
# Print speeds
t = tuple(x / seen * 1E3
for x in (t0, t1, t0 + t1)) + (imgsz, imgsz, batch_size) # tuple
if not training:
print(
'Speed: %.1f/%.1f/%.1f ms inference/NMS/total per %gx%g image at batch-size %g'
% t)
# Save JSON
if save_json and len(jdict):
w = Path(weights[0] if isinstance(weights, list) else weights
).stem if weights is not None else '' # weights
file = save_dir / f"detections_val2017_{w}_results.json" # predicted annotations file
print('\nCOCO mAP with pycocotools... saving %s...' % file)
with open(file, 'w') as f:
json.dump(jdict, f)
try: # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb
from pycocotools.coco import COCO
from pycocotools.cocoeval import COCOeval
imgIds = [int(Path(x).stem) for x in dataloader.dataset.img_files]
cocoGt = COCO(
glob.glob('../coco/annotations/instances_val*.json')
[0]) # initialize COCO ground truth api
cocoDt = cocoGt.loadRes(str(file)) # initialize COCO pred api
cocoEval = COCOeval(cocoGt, cocoDt, 'bbox')
cocoEval.params.imgIds = imgIds # image IDs to evaluate
cocoEval.evaluate()
cocoEval.accumulate()
cocoEval.summarize()
map, map50 = cocoEval.stats[:
2] # update results ([email protected]:0.95, [email protected])
except Exception as e:
print('ERROR: pycocotools unable to run: %s' % e)
# Return results
model.float() # for training
maps = np.zeros(nc) + map
for i, c in enumerate(ap_class):
maps[c] = ap[i]
return (mp, mr, map50, map,
*(loss.cpu() / len(dataloader)).tolist()), maps, t
def detect(save_img=True):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
elif source == 'own_camera':
pass
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
# work space
if source == 'own_camera':
cap = cv2.VideoCapture(0, cv2.CAP_DSHOW)
cap.set(cv2.CAP_PROP_FRAME_WIDTH, 1440)
cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 1080)
cap.set(cv2.CAP_PROP_FPS, 30)
img_name = 0
while True:
# print(source)
_, frame = cap.read()
print(img_name)
print(img_name % 30)
if opt.collect == 'True' and img_name % 5 == 0:
cv2.imwrite("./data/collect/" + str(img_name) + ".jpg", frame)
img_name += 1
im0 = []
img0 = frame.copy()
im0s = img0
img_size = 640
img, ratio, (dw, dh) = letterbox(img0, new_shape=img_size)
img = img[:, :, ::-1].transpose(2, 0,
1) # BGR to RGB, to 3x416x416
img = np.ascontiguousarray(img)
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
t1 = time_synchronized()
cmd = cv2.waitKey(1)
if opt.command == "True" and cmd == ord("c"):
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path('0'), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path('0'), '', im0s
save_path = str(save_dir / p.name)
# txt_path = str(save_dir / 'labels' / p.stem) + (
# '_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(
img0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
classes = []
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]
) # add to string
s1 = '%g' % (n)
classes.append([names[int(c)], s1])
df_cls_info = pd.DataFrame(classes,
columns=['Class', 'number'])
now = datetime.now()
dt_string = now.strftime("%d/%m/%Y %H:%M:%S")
dt_list = list(dt_string)
print(dt_list)
for dt_i in range(len(dt_list)):
if dt_list[dt_i] == ":" or dt_list[dt_i] == "/":
dt_list[dt_i] = "_"
dt_string = "".join(dt_list)
df_cls_info.to_csv('X:/Desktop/test/class_num_' +
dt_string + '.csv')
print("\n number of object : ", len(det[:, -1]))
print(df_cls_info)
# Write results
box_info = []
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (
xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh,
conf) if opt.save_conf else (
cls, *xywh) # label format
# with open(txt_path + '.txt', 'a') as f:
# f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
img0,
label=label,
color=colors[int(cls)],
line_thickness=3)
c1, c2 = (int(xyxy[0]),
int(xyxy[1])), (int(xyxy[2]),
int(xyxy[3]))
confident = '%g' % (conf)
box_info.append([label, c1, c2, confident])
df_box = pd.DataFrame(box_info,
columns=[
'Class', 'top left',
'bottom_right', 'confidence'
])
df_box.to_csv('box_info.csv')
cv2.imwrite('X:/Desktop/test/img_' + dt_string + '.jpg', im0)
elif opt.command == "False":
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path('0'), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path('0'), '', im0s
save_path = str(save_dir / p.name)
# txt_path = str(save_dir / 'labels' / p.stem) + (
# '_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(
img0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
classes = []
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]
) # add to string
s1 = '%g' % (n)
classes.append([names[int(c)], s1])
df_cls_info = pd.DataFrame(classes,
columns=['Class', 'number'])
df_cls_info.to_csv('X:/Desktop/test/class_num.csv')
print("\n number of object : ", len(det[:, -1]))
print(df_cls_info)
# Write results
box_info = []
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (
xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh,
conf) if opt.save_conf else (
cls, *xywh) # label format
# with open(txt_path + '.txt', 'a') as f:
# f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
img0,
label=label,
color=colors[int(cls)],
line_thickness=3)
c1, c2 = (int(xyxy[0]),
int(xyxy[1])), (int(xyxy[2]),
int(xyxy[3]))
confident = '%g' % (conf)
box_info.append([label, c1, c2, confident])
df_box = pd.DataFrame(box_info,
columns=[
'Class', 'top left',
'bottom_right', 'confidence'
])
df_box.to_csv('box_info.csv')
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
if im0 == []:
im0 = frame.copy()
cv2.imshow("asd", im0)
if cv2.waitKey(1) == ord('q'): # q to quit
break
if save_txt or save_img:
print('Results saved to %s' % save_dir)
print('Done. (%.3fs)' % (time.time() - t0))
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path(path[i]), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path(path), '', im0s
save_path = str(save_dir / p.name)
txt_path = str(save_dir / 'labels' / p.stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3)
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
# Stream results
if view_img:
cv2.imshow("asd", im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
# cv2.imwrite(save_path, im0)
pass
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
print('Results saved to %s' % save_dir)
print('Done. (%.3fs)' % (time.time() - t0))
def load_model(weights=None,
image_size=416,
conf_thres=0.25,
iou_thres=0.45,
device_str=''):
global model, imgsz, device, half, opt, names, colors
if weights is None:
weights = "yolov5s.pt"
parser = argparse.ArgumentParser()
parser.add_argument('--weights',
nargs='+',
type=str,
default=weights,
help='model.pt path(s)')
parser.add_argument('--img-size',
type=int,
default=image_size,
help='inference size (pixels)')
parser.add_argument('--conf-thres',
type=float,
default=conf_thres,
help='object confidence threshold')
parser.add_argument('--iou-thres',
type=float,
default=iou_thres,
help='IOU threshold for NMS')
parser.add_argument('--device',
default=device_str,
help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
parser.add_argument('--classes',
nargs='+',
type=int,
help='filter by class: --class 0, or --class 0 2 3')
parser.add_argument('--agnostic-nms',
action='store_true',
help='class-agnostic NMS')
parser.add_argument('--augment',
action='store_true',
help='augmented inference')
# parser.add_argument('--update', action='store_true', help='update all models')
opt = parser.parse_args([])
print(opt)
# namespace = argparse.Namespace()
with torch.no_grad():
weights, imgsz = opt.weights, opt.img_size
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
# model = experimental.attempt_load(weights, map_location=device) # load FP32 model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(names))]
print("names : ", names)
# warming
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
print("model loaded....")
def detect(save_img=False):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
sub_seq = [0 for _ in range(len(dataset))]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
sub_dict = {"bbox": [], "score": [], "label": []}
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = Path(path[i]), '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = Path(path), '', im0s
save_path = str(save_dir / p.name)
txt_path = str(save_dir / 'labels' / p.stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3)
x1, y1, x2, y2 = int(xyxy[0].item()), int(
xyxy[1].item()), int(xyxy[2].item()), int(
xyxy[3].item())
sub_dict['bbox'].append((y1, x1, y2, x2))
sub_dict['score'].append(conf.item())
label_i = 10 if int(names[int(
cls.item())]) == 0 else int(names[int(cls.item())])
sub_dict['label'].append(label_i)
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
# Stream results
if view_img:
cv2.imshow(str(p), im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
cv2.imwrite(save_path, im0)
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
sub_seq[int(path.split('/')[-1].split('.')[0]) - 1] = sub_dict
with open(str(save_dir) + '/sub.json', 'w') as outfile:
json.dump(sub_seq, outfile)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print('Done. (%.3fs)' % (time.time() - t0))
def main(opt):
set_logging()
print(
colorstr('export: ') + ', '.join(f'{k}={v}'
for k, v in vars(opt).items()))
run(**vars(opt))
def detect(save_img=False):
out, source, weights, view_img, save_txt, imgsz = \
opt.output, opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.startswith(
'rtsp') or source.startswith('http') or source.endswith('.txt')
# Initialize
set_logging()
device = select_device(opt.device)
if os.path.exists(out):
shutil.rmtree(out) # delete output folder
os.makedirs(out) # make new output folder
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']) # load weights
modelc.to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(names))]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0 = path[i], '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = path, '', im0s
save_path = str(Path(out) / Path(p).name)
txt_path = str(Path(out) / Path(p).stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
xml_path = p.replace('JPEGImages/',
'Annotations/').rsplit('.')[0] + '.xml'
if not os.path.exists(xml_path):
continue
tree = parse(xml_path)
node_root = tree.getroot()
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%g %ss, ' % (n, names[int(c)]) # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if names[int(cls)] != 'person':
continue
x_min = int(xyxy[0])
y_min = int(xyxy[1])
x_max = int(xyxy[2])
y_max = int(xyxy[3])
node_object = SubElement(node_root, 'object')
node_name = SubElement(node_object, 'name')
node_name.text = 'person'
node_difficult = SubElement(node_object, 'difficult')
node_difficult.text = '0'
node_bndbox = SubElement(node_object, 'bndbox')
node_xmin = SubElement(node_bndbox, 'xmin')
node_xmin.text = str(x_min)
node_ymin = SubElement(node_bndbox, 'ymin')
node_ymin.text = str(y_min)
node_xmax = SubElement(node_bndbox, 'xmax')
node_xmax.text = str(x_max)
node_ymax = SubElement(node_bndbox, 'ymax')
node_ymax.text = str(y_max)
# tree.write(p.replace('JPEGImages/', 'Annotations3/').replace('.jpg', '.xml'))
xml = tostring(node_root, pretty_print=True) # 格式化显示，该换行的换行
with open(
p.replace('JPEGImages/', 'Annotations2/').rsplit('.')[0] +
'.xml', "wb") as f_out:
f_out.write(xml)
f_out.close()
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
print('Done. (%.3fs)' % (time.time() - t0))
def predict():
file = request.files['image']
file_name = file.filename
path = os.path.join('./upload', file_name)
file.save(path)
source = f'./upload/{file_name}'
weights, view_img, save_txt, imgsz = './model/best-m.pt', opt.view_img, opt.save_txt, 640
webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True,
exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(torch.load(
'weights/resnet101.pt', map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
# run once
_ = model(img.half() if half else img) if device.type != 'cpu' else None
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(
pred, 0.7, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(
), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + \
('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
# normalization gain whwh
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(
img.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f'{n} {names[int(c)]}s, ' # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)
) / gn).view(-1).tolist() # normalized xywh
# label format
line = (
cls, *xywh, conf) if opt.save_conf else (cls, *xywh)
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy, im0, label=label,
color=colors[int(cls)], line_thickness=3)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
vid_writer.write(im0)
time.sleep(2)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
# prepare image for response
# _, img_encoded = cv2.imencode('.png', im0)
# response = img_encoded.tostring()
# return Response(response=response, status=200, mimetype='image/png')
resultPath = str(save_dir)[13:]+'/'+file_name
return jsonify({"imgsrc": resultPath})
def test(
data,
weights=None,
batch_size=16,
imgsz=640,
conf_thres=0.001,
iou_thres=0.6, # for NMS
save_json=False,
single_cls=False,
augment=False,
verbose=False,
model=None,
dataloader=None,
save_dir='',
merge=False,
save_txt=False):
# Initialize/load model and set device
training = model is not None
if training: # called by train.py
device = next(model.parameters()).device # get model device
else: # called directly
set_logging()
device = select_device(opt.device, batch_size=batch_size)
merge, save_txt = opt.merge, opt.save_txt # use Merge NMS, save *.txt labels
if save_txt:
out = Path('inference/output')
if os.path.exists(out):
shutil.rmtree(out) # delete output folder
os.makedirs(out) # make new output folder
# Remove previous
for f in glob.glob(str(Path(save_dir) / 'test_batch*.jpg')):
os.remove(f)
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
# Multi-GPU disabled, incompatible with .half() https://github.com/ultralytics/yolov5/issues/99
# if device.type != 'cpu' and torch.cuda.device_count() > 1:
# model = nn.DataParallel(model)
# Half
half = device.type != 'cpu' # half precision only supported on CUDA
if half:
model.half()
# Configure
model.eval()
with open(data) as f:
data = yaml.load(f, Loader=yaml.FullLoader) # model dict
check_dataset(data) # check
nc = 1 if single_cls else int(data['nc']) # number of classes
iouv = torch.linspace(0.5, 0.95,
10).to(device) # iou vector for [email protected]:0.95
niou = iouv.numel()
# Dataloader
if not training:
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
path = data['test'] if opt.task == 'test' else data[
'val'] # path to val/test images
dataloader = create_dataloader(path,
imgsz,
batch_size,
model.stride.max(),
opt,
hyp=None,
augment=False,
cache=True,
pad=0.5,
rect=True)[0]
seen = 0
names = model.names if hasattr(model, 'names') else model.module.names
coco91class = coco80_to_coco91_class()
s = ('%20s' + '%12s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R',
'[email protected]', '[email protected]:.95')
p, r, f1, mp, mr, map50, map, t0, t1 = 0., 0., 0., 0., 0., 0., 0., 0., 0.
loss = torch.zeros(3, device=device)
jdict, stats, ap, ap_class = [], [], [], []
evaluator = COCOEvaluator(root=DATA_ROOT,
model_name=opt.weights.replace('.pt', ''))
for batch_i, (img, targets, paths,
shapes) in enumerate(tqdm(dataloader, desc=s)):
img = img.to(device, non_blocking=True)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
targets = targets.to(device)
nb, _, height, width = img.shape # batch size, channels, height, width
whwh = torch.Tensor([width, height, width, height]).to(device)
# Disable gradients
with torch.no_grad():
# Run model
t = time_synchronized()
inf_out, train_out = model(
img, augment=augment) # inference and training outputs
t0 += time_synchronized() - t
# Compute loss
if training: # if model has loss hyperparameters
loss += compute_loss([x.float() for x in train_out], targets,
model)[1][:3] # GIoU, obj, cls
# Run NMS
t = time_synchronized()
output = non_max_suppression(inf_out,
conf_thres=conf_thres,
iou_thres=iou_thres,
merge=merge)
t1 += time_synchronized() - t
# Statistics per image
for si, pred in enumerate(output):
labels = targets[targets[:, 0] == si, 1:]
nl = len(labels)
tcls = labels[:, 0].tolist() if nl else [] # target class
seen += 1
if pred is None:
if nl:
stats.append((torch.zeros(0, niou, dtype=torch.bool),
torch.Tensor(), torch.Tensor(), tcls))
continue
# Append to text file
if save_txt:
gn = torch.tensor(shapes[si][0])[[1, 0, 1, 0
]] # normalization gain whwh
x = pred.clone()
x[:, :4] = scale_coords(img[si].shape[1:], x[:, :4],
shapes[si][0],
shapes[si][1]) # to original
for *xyxy, conf, cls in x:
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
with open(str(out / Path(paths[si]).stem) + '.txt',
'a') as f:
f.write(
('%g ' * 5 + '\n') % (cls, *xywh)) # label format
# Clip boxes to image bounds
clip_coords(pred, (height, width))
# Append to pycocotools JSON dictionary
if save_json:
# [{"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}, ...
image_id = Path(paths[si]).stem
box = pred[:, :4].clone() # xyxy
scale_coords(img[si].shape[1:], box, shapes[si][0],
shapes[si][1]) # to original shape
box = xyxy2xywh(box) # xywh
box[:, :2] -= box[:, 2:] / 2 # xy center to top-left corner
for p, b in zip(pred.tolist(), box.tolist()):
result = {
'image_id':
int(image_id) if image_id.isnumeric() else image_id,
'category_id': coco91class[int(p[5])],
'bbox': [round(x, 3) for x in b],
'score': round(p[4], 5)
}
jdict.append(result)
#evaluator.add([result])
#if evaluator.cache_exists:
# break
# # Assign all predictions as incorrect
# correct = torch.zeros(pred.shape[0], niou, dtype=torch.bool, device=device)
# if nl:
# detected = [] # target indices
# tcls_tensor = labels[:, 0]
#
# # target boxes
# tbox = xywh2xyxy(labels[:, 1:5]) * whwh
#
# # Per target class
# for cls in torch.unique(tcls_tensor):
# ti = (cls == tcls_tensor).nonzero(as_tuple=False).view(-1) # prediction indices
# pi = (cls == pred[:, 5]).nonzero(as_tuple=False).view(-1) # target indices
#
# # Search for detections
# if pi.shape[0]:
# # Prediction to target ious
# ious, i = box_iou(pred[pi, :4], tbox[ti]).max(1) # best ious, indices
#
# # Append detections
# detected_set = set()
# for j in (ious > iouv[0]).nonzero(as_tuple=False):
# d = ti[i[j]] # detected target
# if d.item() not in detected_set:
# detected_set.add(d.item())
# detected.append(d)
# correct[pi[j]] = ious[j] > iouv # iou_thres is 1xn
# if len(detected) == nl: # all targets already located in image
# break
#
# # Append statistics (correct, conf, pcls, tcls)
# stats.append((correct.cpu(), pred[:, 4].cpu(), pred[:, 5].cpu(), tcls))
# # Plot images
# if batch_i < 1:
# f = Path(save_dir) / ('test_batch%g_gt.jpg' % batch_i) # filename
# plot_images(img, targets, paths, str(f), names) # ground truth
# f = Path(save_dir) / ('test_batch%g_pred.jpg' % batch_i)
# plot_images(img, output_to_target(output, width, height), paths, str(f), names) # predictions
evaluator.add(jdict)
evaluator.save()
def detect(model, device, frame, imgsz, iou_thresh, conf_thresh):
weights = 'yolov5l.pt'
# Initialize
set_logging()
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
print("IMGSZ:", imgsz)
if half:
model.half() # to FP16
# Set Dataloader
cudnn.benchmark = True
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
img = [letterbox(frame, new_shape=imgsz, auto=True)[0]]
img = np.stack(img, 0)
# Convert
img = img[:, :, :, ::-1].transpose(0, 3, 1,
2) # BGR to RGB, to bsx3x416x416
img = np.ascontiguousarray(img)
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=False)[0]
# Apply NMS
pred = non_max_suppression(pred, conf_thresh, iou_thresh, agnostic=False)
t2 = time_synchronized()
det = pred[0]
s = '%gx%g ' % img.shape[2:]
gn = torch.tensor(img.shape)[[1, 0, 1, 0]]
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], frame.shape).round()
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
bboxes = []
for *xyxy, conf, cls in reversed(det):
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
label = names[int(cls)]
bboxes.append((xyxy, label))
return bboxes
def detect(opt):
save_img = False
out, source, weights, view_img, save_txt, imgsz = \
opt.save_dir, opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.startswith(
('rtsp://', 'rtmp://', 'http://')) or source.endswith('.txt')
# Initialize
set_logging()
device = select_device(opt.device)
flag = False
if os.path.exists(out): # output dir
shutil.rmtree(out) # delete dir
os.makedirs(out) # make new dir
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']) # load weights
modelc.to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
names[0], names[1] = names[1], names[0]
colors = [[0, 0, 255], [0, 255, 0]]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
flag = False
if webcam: # batch_size >= 1
p, s, im0 = path[i], ' ', im0s[i].copy()
else:
p, s, im0 = path, '', im0s
save_path = str(Path(out) / Path(p).name)
txt_path = str(Path(out) / Path(p).stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += '%ss : %g, ' % (names[int(c)], n) # add to string
if (names[int(c)] == 'without_mask'):
flag = True
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, conf, *xywh) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line) + '\n') % line)
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3)
# Print time (inference + NMS)
now = datetime.now()
dt_string = now.strftime("%d_%b_%y %H_%M_%S")
dt_folder = now.strftime("%d_%b_%y")
dt_file = now.strftime("%H_%M_%S")
# if (s != ' '):
# print(dt_string," ",(s))
try:
os.makedirs("inference/data/" + dt_folder)
except FileExistsError:
# directory already exists
pass
# Stream results
if view_img:
# im0 = cv2.resize(im0,(1120,840))
if (flag):
cv2.imwrite(
"inference\\data\\" + dt_folder + "\\" + dt_string +
".png", im0)
cv2.imshow(p, im0)
# if cv2.waitKey(1) == ord('q'): # q to quit
# cv2.VideoCapture(source).release()
# cv2.destroyAllWindows()
# return
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
cv2.imwrite(save_path, im0)
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
def detect(save_img=False):
source, weights, view_img, save_txt, imgsz = (
opt.source,
opt.weights,
opt.view_img,
opt.save_txt,
opt.img_size,
)
webcam = (
source.isnumeric()
or source.endswith(".txt")
or source.lower().startswith(("rtsp://", "rtmp://", "http://"))
)
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)
) # increment run
(save_dir / "labels" if save_txt else save_dir).mkdir(
parents=True, exist_ok=True
) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != "cpu" # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name="resnet101", n=2) # initialize
modelc.load_state_dict(
torch.load("weights/resnet101.pt", map_location=device)["model"]
).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, "module") else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img) if device.type != "cpu" else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(
pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms,
)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], "%g: " % i, im0s[i].copy(), dataset.count
else:
p, s, im0, frame = path, "", im0s, getattr(dataset, "frame", 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / "labels" / p.stem) + (
"" if dataset.mode == "image" else f"_{frame}"
) # img.txt
s += "%gx%g " % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}s, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (
(xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn)
.view(-1)
.tolist()
) # normalized xywh
line = (
(cls, *xywh, conf) if opt.save_conf else (cls, *xywh)
) # label format
with open(txt_path + ".txt", "a") as f:
f.write(("%g " * len(line)).rstrip() % line + "\n")
if save_img or view_img: # Add bbox to image
label = f"{names[int(cls)]} {conf:.2f}"
plot_one_box(
xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=3,
)
# Print time (inference + NMS)
print(f"{s}Done. ({t2 - t1:.3f}s)")
# Stream results
if view_img:
cv2.imshow(str(p), im0)
if cv2.waitKey(1) == ord("q"): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == "image":
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fourcc = "mp4v" # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h)
)
vid_writer.write(im0)
if save_txt or save_img:
s = (
f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}"
if save_txt
else ""
)
print(f"Results saved to {save_dir}{s}")
print(f"Done. ({time.time() - t0:.3f}s)")
def run(weights='yolov5s.pt', # model.pt path(s)
source='data/images', # file/dir/URL/glob, 0 for webcam
imgsz=640, # inference size (pixels)
conf_thres=0.25, # confidence threshold
iou_thres=0.45, # NMS IOU threshold
max_det=1000, # maximum detections per image
device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
view_img=False, # show results
save_txt=False, # save results to *.txt
save_conf=False, # save confidences in --save-txt labels
save_crop=False, # save cropped prediction boxes
nosave=False, # do not save images/videos
classes=None, # filter by class: --class 0, or --class 0 2 3
agnostic_nms=False, # class-agnostic NMS
augment=False, # augmented inference
visualize=False, # visualize features
update=False, # update all models
project='runs/detect', # save results to project/name
name='exp', # save results to project/name
exist_ok=False, # existing project/name ok, do not increment
line_thickness=3, # bounding box thickness (pixels)
hide_labels=False, # hide labels
hide_conf=False, # hide confidences
half=False, # use FP16 half-precision inference
):
save_img = not nosave and not source.endswith('.txt') # save inference images
webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://', 'https://'))
# Directories
save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(device)
half &= device.type != 'cpu' # half precision only supported on CUDA
# Load model
w = weights[0] if isinstance(weights, list) else weights
classify, pt, onnx = False, w.endswith('.pt'), w.endswith('.onnx') # inference type
stride, names = 64, [f'class{i}' for i in range(1000)] # assign defaults
if pt:
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
names = model.module.names if hasattr(model, 'module') else model.names # get class names
if half:
model.half() # to FP16
if classify: # second-stage classifier
modelc = load_classifier(name='resnet50', n=2) # initialize
modelc.load_state_dict(torch.load('resnet50.pt', map_location=device)['model']).to(device).eval()
elif onnx:
check_requirements(('onnx', 'onnxruntime'))
import onnxruntime
session = onnxruntime.InferenceSession(w, None)
imgsz = check_img_size(imgsz, s=stride) # check image size
# Dataloader
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
bs = len(dataset) # batch_size
else:
dataset = LoadImages(source, img_size=imgsz, stride=stride)
bs = 1 # batch_size
vid_path, vid_writer = [None] * bs, [None] * bs
# Run inference
if pt and device.type != 'cpu':
model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters()))) # run once
t0 = time.time()
for path, img, im0s, vid_cap in dataset:
if pt:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
elif onnx:
img = img.astype('float32')
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if len(img.shape) == 3:
img = img[None] # expand for batch dim
# Inference
t1 = time_sync()
if pt:
visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
pred = model(img, augment=augment, visualize=visualize)[0]
elif onnx:
pred = torch.tensor(session.run([session.get_outputs()[0].name], {session.get_inputs()[0].name: img}))
# NMS
pred = non_max_suppression(pred, conf_thres, iou_thres, classes, agnostic_nms, max_det=max_det)
t2 = time_sync()
# Second-stage classifier (optional)
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process predictions
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], f'{i}: ', im0s[i].copy(), dataset.count
else:
p, s, im0, frame = path, '', im0s.copy(), getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
imc = im0.copy() if save_crop else im0 # for save_crop
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if save_conf else (cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or save_crop or view_img: # Add bbox to image
c = int(cls) # integer class
label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f}')
plot_one_box(xyxy, im0, label=label, color=colors(c, True), line_thickness=line_thickness)
if save_crop:
save_one_box(xyxy, imc, file=save_dir / 'crops' / names[c] / f'{p.stem}.jpg', BGR=True)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video' or 'stream'
if vid_path[i] != save_path: # new video
vid_path[i] = save_path
if isinstance(vid_writer[i], cv2.VideoWriter):
vid_writer[i].release() # release previous video writer
if vid_cap: # video
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
else: # stream
fps, w, h = 30, im0.shape[1], im0.shape[0]
save_path += '.mp4'
vid_writer[i] = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
vid_writer[i].write(im0)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
if update:
strip_optimizer(weights) # update model (to fix SourceChangeWarning)
print(f'Done. ({time.time() - t0:.3f}s)')
def trt_detect(save_img=False):
# yolov3-ssp with evolve
# anchor_nums = 4
# nc = 1
# anchors = np.array([
# [[11, 10], [17, 9], [18, 16], [29, 16]],
# [[34, 28], [48, 24], [59, 33], [46, 64]],
# [[69, 45], [86, 59], [96, 80], [150, 106]]
# ])
# output_shapes = [
# (1, anchor_nums, 80, 80, nc + 5),
# (1, anchor_nums, 40, 40, nc + 5),
# (1, anchor_nums, 20, 20, nc + 5)
# ]
# yolov5s
anchor_nums = 3
nc = 1
anchors = np.array([
[[10, 13], [16, 30], [33, 23]], # P3/8
[[30, 61], [62, 45], [59, 119]], # P4/16
[[116, 90], [156, 198], [373, 326]]
])
strides = np.array([8., 16., 32.])
output_shapes = [
(1, anchor_nums, 60, 80, nc + 5),
(1, anchor_nums, 30, 40, nc + 5),
(1, anchor_nums, 15, 20, nc + 5)
# (1, anchor_nums*60*80, nc + 5),
# (1, anchor_nums*30*40, nc + 5),
# (1, anchor_nums*15*20, nc + 5)
]
source, weights, view_img, save_txt, imgsz = \
opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
save_img = not opt.nosave and not source.endswith('.txt') # save inference images
webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://', 'https://'))
# Directories
save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
stride = int(strides.max()) # model stride
print(f"Loading trt engine!")
# imgsz = check_img_size(imgsz, s=stride) # check img_size
# Set Dataloader
vid_path, vid_writer = None, None
bird_transform, pts = False, None
if opt.plot_move_routes:
bird_transform = True
# Coordinates of chessboard region needs to apply brid transform
pts = np.array([(567, 28), (1458, 60), (1890, 639), (638, 1068)])
center_point = queue.Queue()
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreamsTrt(source, img_size=imgsz, bird_transform=bird_transform, pts=pts)
else:
dataset = LoadImagesTrt(source, img_size=imgsz, bird_transform=bird_transform, pts=pts)
# Get names and colors
names = ['Robot']
# colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
colors = [[96, 171, 132]]
# Run inference
t0 = time.time()
img_index = 0
# Create an empty picture and draw the route on it, which is independent from the real picture.
route_mask = None
cover_heatmap_accum = None
processor = Processor(weights[0], anchor_nums, nc, anchors, output_shapes, imgsz)
for path, img, im0s, vid_cap in dataset:
if opt.plot_move_routes and route_mask is None:
# initialize something for ploting move routes
route_img = im0s.copy()
route_mask = np.zeros((im0s.shape[0], im0s.shape[1], 3), np.uint8)
route_mask.fill(255)
route_mask_bg_color = 'white'
if opt.cover_heatmap:
if cover_heatmap_accum is None:
# initialize something for ploting cover heatmap
cover_heatmap_img = im0s.copy()
# The template with a black base color is used to
# continuously accumulate the covered elliptical areas going in.
cover_heatmap_accum = np.zeros((im0s.shape[0], im0s.shape[1], 3), np.uint8)
cover_heatmap_accum.fill(0)
cover_heatmap_accum_bg_color = 'black'
# The same img as img0s used to host cumulative heatmaps.
cover_heatmap_accum_img0s = im0s.copy()
# restore cover_heatmap_tmp every loop
cover_heatmap_tmp = np.zeros((im0s.shape[0], im0s.shape[1], 3), np.uint8)
cover_heatmap_tmp.fill(0)
# Inference
# t1 = time_synchronized()
pred = processor.detect(img)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
det[:, 6:8] = scale_coords(img.shape[2:], det[:, 6:8], im0.shape).round()
# plot move routes
if opt.plot_move_routes and img_index % opt.move_routes_interval == 0: # 设置绘制间隔
assert pred[0].shape[0] == 1, "Only one robot can exist in the scene when drawing movement routes!"
center_point.put(list(map(int, det[:, 6:8].tolist()[0])))
# elements of center_point always Less than or equal to 2
if center_point.qsize() <= 1:
pass
else:
pts1 = center_point.get()
pts2 = center_point.get()
plot_move_routes([pts1, pts2], route_mask, colors[int(cls)], 3)
route_img = one_cover_two_with_mask(route_mask, route_img, bg_color=route_mask_bg_color)
cv2.imwrite('/home/yousixia/project/yolov3/runs/detect/tmp/im0s.jpg', im0s)
center_point.put(pts2)
# plot cover heatmap
if opt.cover_heatmap and img_index % opt.cover_heatmap_interval == 0:
x1, y1, x2, y2 = list(map(int, det[:, :4].tolist()[0]))
# Draws a white ellipse with a center at center_point of bbox,
# two axes is (x2-x1, y2-y1), and a line width of 3.
cv2.ellipse(cover_heatmap_tmp, list(map(int, det[:, 6:8].tolist()[0])),
(int((x2 - x1) * 0.9 / 2), int((y2 - y1) * 0.9 / 2)), 0, 0, 360, (32, 16, 16),
-1) # 画椭圆
cover_heatmap_accum = cv2.addWeighted(cover_heatmap_accum, 1, cover_heatmap_tmp, 1, 0) # 累加覆盖面积
cover_heatmap_accum_colormap = cv2.applyColorMap(cover_heatmap_accum, cv2.COLORMAP_JET)
one_cover_two_with_mask(cover_heatmap_accum, cover_heatmap_img, cover_heatmap_accum_colormap,
bg_color='black')
# cv2.imwrite('/home/yousixia/project/yolov3/runs/detect/tmp/original_img.jpg',
# original_img)
# Print results
for c in np.unique(det[:, 5]):
n = (det[:, 5] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls, center_x, center_y in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)
# if opt.plot_move_routes:
# plot_one_box(xyxy, route_img, label=label, color=colors[int(cls)], line_thickness=3)
# if opt.cover_heatmap:
# plot_one_box(xyxy, cover_heatmap_img, label=label, color=colors[int(cls)], line_thickness=3)
# cv2.imwrite('/home/yousixia/project/yolov3/runs/detect/tmp/cover_heatmap_img.jpg',
# cover_heatmap_img)
# plot_center_point((center_x, center_y), im0, color=colors[int(cls)], line_thickness=3)
# Print time (inference + NMS)
# print(f'{s}Done. ({t2 - t1:.3f}s)')
# drawing cover rate plot
cover_rate_plot_data = []
if opt.cover_rate:
heatmap_non_zero_pixels = cv2.countNonZero(cover_heatmap_accum)
all_pixels = cv2.countNonZero(im0s)
cover_heatmap_img
cover_rate = 1.0 * 100 * heatmap_non_zero_pixels / all_pixels
secend = img_index / vid_cap.get(cv2.CAP_PROP_FPS)
cover_rate_plot_data.append([cover_rate, secend])
l1 = plt.plot(secend, cover_rate, 'b--', label='覆盖率')
plt.xlabel('时间/s')
plt.ylabel('覆盖率/%')
plt.legend()
# plt.show()
# stack bbox, movement routes, cover heatmap, cover rate to one matrix.
# TODO
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
# cv2.imwrite('/home/yousixia/project/yolov3/runs/detect/tmp/123.jpg', im0)
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video' or 'stream'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
if vid_cap: # video
fps = vid_cap.get(cv2.CAP_PROP_FPS)
if opt.plot_move_routes:
w, h = im0.shape[1], im0.shape[0]
else:
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
else: # stream
fps, w, h = 30, im0.shape[1], im0.shape[0]
save_path += '.mp4'
vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
if opt.plot_move_routes:
filename = save_path.split('/')[-1]
movement_routes_vid_writer = cv2.VideoWriter(
os.path.join('/'.join(save_path.split('/')[:-1]), 'movement_routes_' + filename),
cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
if opt.cover_heatmap:
cover_heatmap_vid_writer = cv2.VideoWriter(
os.path.join('/'.join(save_path.split('/')[:-1]), 'cover_heatmap_' + filename),
cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
vid_writer.write(im0)
movement_routes_vid_writer.write(route_img)
cover_heatmap_vid_writer.write(cover_heatmap_img)
img_index += 1
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
def main(opt):
set_logging(RANK)
if RANK in [-1, 0]:
print(
colorstr('train: ') + ', '.join(f'{k}={v}'
for k, v in vars(opt).items()))
check_git_status()
check_requirements(exclude=['thop'])
# Resume
wandb_run = check_wandb_resume(opt)
if opt.resume and not wandb_run: # resume an interrupted run
ckpt = opt.resume if isinstance(
opt.resume,
str) else get_latest_run() # specified or most recent path
assert os.path.isfile(
ckpt), 'ERROR: --resume checkpoint does not exist'
with open(Path(ckpt).parent.parent / 'opt.yaml') as f:
opt = argparse.Namespace(**yaml.safe_load(f)) # replace
opt.cfg, opt.weights, opt.resume = '', ckpt, True # reinstate
logger.info('Resuming training from %s' % ckpt)
else:
# opt.hyp = opt.hyp or ('hyp.finetune.yaml' if opt.weights else 'hyp.scratch.yaml')
opt.data, opt.cfg, opt.hyp = check_file(opt.data), check_file(
opt.cfg), check_file(opt.hyp) # check files
assert len(opt.cfg) or len(
opt.weights), 'either --cfg or --weights must be specified'
opt.img_size.extend(
[opt.img_size[-1]] *
(2 - len(opt.img_size))) # extend to 2 sizes (train, test)
opt.name = 'evolve' if opt.evolve else opt.name
opt.save_dir = str(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok or opt.evolve))
# DDP mode
device = select_device(opt.device, batch_size=opt.batch_size)
if LOCAL_RANK != -1:
from datetime import timedelta
assert torch.cuda.device_count(
) > LOCAL_RANK, 'insufficient CUDA devices for DDP command'
torch.cuda.set_device(LOCAL_RANK)
device = torch.device('cuda', LOCAL_RANK)
dist.init_process_group(
backend="nccl" if dist.is_nccl_available() else "gloo",
timeout=timedelta(seconds=60))
assert opt.batch_size % WORLD_SIZE == 0, '--batch-size must be multiple of CUDA device count'
assert not opt.image_weights, '--image-weights argument is not compatible with DDP training'
# Train
if not opt.evolve:
train(opt.hyp, opt, device)
if WORLD_SIZE > 1 and RANK == 0:
_ = [
print('Destroying process group... ', end=''),
dist.destroy_process_group(),
print('Done.')
]
# Evolve hyperparameters (optional)
else:
# Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
meta = {
'lr0':
(1, 1e-5, 1e-1), # initial learning rate (SGD=1E-2, Adam=1E-3)
'lrf':
(1, 0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf)
'momentum': (0.3, 0.6, 0.98), # SGD momentum/Adam beta1
'weight_decay': (1, 0.0, 0.001), # optimizer weight decay
'warmup_epochs': (1, 0.0, 5.0), # warmup epochs (fractions ok)
'warmup_momentum': (1, 0.0, 0.95), # warmup initial momentum
'warmup_bias_lr': (1, 0.0, 0.2), # warmup initial bias lr
'box': (1, 0.02, 0.2), # box loss gain
'cls': (1, 0.2, 4.0), # cls loss gain
'cls_pw': (1, 0.5, 2.0), # cls BCELoss positive_weight
'obj': (1, 0.2, 4.0), # obj loss gain (scale with pixels)
'obj_pw': (1, 0.5, 2.0), # obj BCELoss positive_weight
'iou_t': (0, 0.1, 0.7), # IoU training threshold
'anchor_t': (1, 2.0, 8.0), # anchor-multiple threshold
'anchors': (2, 2.0, 10.0), # anchors per output grid (0 to ignore)
'fl_gamma':
(0, 0.0, 2.0), # focal loss gamma (efficientDet default gamma=1.5)
'hsv_h': (1, 0.0, 0.1), # image HSV-Hue augmentation (fraction)
'hsv_s': (1, 0.0,
0.9), # image HSV-Saturation augmentation (fraction)
'hsv_v': (1, 0.0, 0.9), # image HSV-Value augmentation (fraction)
'degrees': (1, 0.0, 45.0), # image rotation (+/- deg)
'translate': (1, 0.0, 0.9), # image translation (+/- fraction)
'scale': (1, 0.0, 0.9), # image scale (+/- gain)
'shear': (1, 0.0, 10.0), # image shear (+/- deg)
'perspective':
(0, 0.0, 0.001), # image perspective (+/- fraction), range 0-0.001
'flipud': (1, 0.0, 1.0), # image flip up-down (probability)
'fliplr': (0, 0.0, 1.0), # image flip left-right (probability)
'mosaic': (1, 0.0, 1.0), # image mixup (probability)
'mixup': (1, 0.0, 1.0), # image mixup (probability)
'copy_paste': (1, 0.0, 1.0)
} # segment copy-paste (probability)
with open(opt.hyp) as f:
hyp = yaml.safe_load(f) # load hyps dict
if 'anchors' not in hyp: # anchors commented in hyp.yaml
hyp['anchors'] = 3
assert LOCAL_RANK == -1, 'DDP mode not implemented for --evolve'
opt.notest, opt.nosave = True, True # only test/save final epoch
# ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices
yaml_file = Path(
opt.save_dir) / 'hyp_evolved.yaml' # save best result here
if opt.bucket:
os.system('gsutil cp gs://%s/evolve.txt .' %
opt.bucket) # download evolve.txt if exists
for _ in range(opt.evolve): # generations to evolve
if Path('evolve.txt').exists(
): # if evolve.txt exists: select best hyps and mutate
# Select parent(s)
parent = 'single' # parent selection method: 'single' or 'weighted'
x = np.loadtxt('evolve.txt', ndmin=2)
n = min(5, len(x)) # number of previous results to consider
x = x[np.argsort(-fitness(x))][:n] # top n mutations
w = fitness(x) - fitness(x).min() + 1E-6 # weights (sum > 0)
if parent == 'single' or len(x) == 1:
# x = x[random.randint(0, n - 1)] # random selection
x = x[random.choices(range(n),
weights=w)[0]] # weighted selection
elif parent == 'weighted':
x = (x * w.reshape(
n, 1)).sum(0) / w.sum() # weighted combination
# Mutate
mp, s = 0.8, 0.2 # mutation probability, sigma
npr = np.random
npr.seed(int(time.time()))
g = np.array([x[0] for x in meta.values()]) # gains 0-1
ng = len(meta)
v = np.ones(ng)
while all(
v == 1
): # mutate until a change occurs (prevent duplicates)
v = (g * (npr.random(ng) < mp) * npr.randn(ng) *
npr.random() * s + 1).clip(0.3, 3.0)
for i, k in enumerate(hyp.keys()): # plt.hist(v.ravel(), 300)
hyp[k] = float(x[i + 7] * v[i]) # mutate
# Constrain to limits
for k, v in meta.items():
hyp[k] = max(hyp[k], v[1]) # lower limit
hyp[k] = min(hyp[k], v[2]) # upper limit
hyp[k] = round(hyp[k], 5) # significant digits
# Train mutation
results = train(hyp.copy(), opt, device)
# Write mutation results
print_mutation(hyp.copy(), results, yaml_file, opt.bucket)
# Plot results
plot_evolution(yaml_file)
print(
f'Hyperparameter evolution complete. Best results saved as: {yaml_file}\n'
f'Command to train a new model with these hyperparameters: $ python train.py --hyp {yaml_file}'
)
def detect(save_img=False):
'''
input: save_img_flag
output(result):
'''
# 获取输出文件夹，输入路径，权重，参数等参数
out, source, weights, view_img, save_txt, imgsz = \
opt.output, opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.startswith(
('rtsp://', 'rtmp://', 'http://')) or source.endswith('.txt')
# Initialize
set_logging()
# 获取设备
device = select_device(opt.device)
# 移除之前的输出文件夹,并新建输出文件夹
if os.path.exists(out):
shutil.rmtree(out) # delete output folder
os.makedirs(out) # make new output folder
# 如果设备为gpu，使用Float16
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
# 加载Float32模型，确保用户设定的输入图片分辨率能整除最大步长s=32(如不能则调整为能整除并返回)
'''
model = Model(
(model): Sequential(
(0): Focus(...)
(1): Conv(...)
...
(24): Detect(...)
)
'''
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
# 设置Float16
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']) # load weights
modelc.to(device).eval()
# Set Dataloader
# 通过不同的输入源来设置不同的数据加载方式
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
# 获取类别名字 names = ['person', 'bicycle', 'car',...,'toothbrush']
names = model.module.names if hasattr(model, 'module') else model.names
# 设置画框的颜色 colors = [[178, 63, 143], [25, 184, 176], [238, 152, 129],....,[235, 137, 120]]随机设置RGB颜色
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(names))]
# Run inference
t0 = time.time()
# 进行一次前向推理,测试程序是否正常 向量维度（1，3，imgsz，imgsz）
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
"""
path 图片/视频路径 'E:\...\bus.jpg'
img 进行resize+pad之后的图片 1*3*re_size1*resize2的张量 (3,img_height,img_weight)
img0 原size图片 (img_height,img_weight,3)
cap 当读取图片时为None，读取视频时为视频源
"""
for path, img, im0s, vid_cap in dataset:
print(img.shape)
img = torch.from_numpy(img).to(device)
# 图片也设置为Float16
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
# 没有batch_size的话则在最前面添加一个轴
if img.ndimension() == 3:
# (in_channels,size1,size2) to (1,in_channels,img_height,img_weight)
img = img.unsqueeze(0) # 在[0]维增加一个维度
# Inference
t1 = time_synchronized()
"""
model:
input: in_tensor (batch_size, 3, img_height, img_weight)
output: 推理时返回 [z,x]
z tensor: [small+medium+large_inference] size=(batch_size, 3 * (small_size1*small_size2 + medium_size1*medium_size2 + large_size1*large_size2), nc)
x list: [small_forward, medium_forward, large_forward] eg:small_forward.size=( batch_size, 3种scale框, size1, size2, [xywh,score,num_classes])
'''
前向传播 返回pred[0]的shape是(1, num_boxes, nc)
h,w为传入网络图片的长和宽，注意dataset在检测时使用了矩形推理，所以这里h不一定等于w
num_boxes = 3 * h/32 * w/32 + 3 * h/16 * w/16 + 3 * h/8 * w/8
pred[0][..., 0:4] 预测框坐标为xywh(中心点+宽长)格式
pred[0][..., 4]为objectness置信度
pred[0][..., 5:5+nc]为分类结果
pred[0][..., 5+nc:]为Θ分类结果
"""
# pred : (batch_size, num_boxes, no) batch_size=1
pred = model(img, augment=opt.augment)[0]
# Apply NMS
# 进行NMS
# pred : list[tensor(batch_size, num_conf_nms, [xylsθ,conf,classid])] θ∈[0,179]
#pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
pred = rotate_non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms,
without_iouthres=False)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(
pred
): # i:image index det:(num_nms_boxes, [xylsθ,conf,classid]) θ∈[0,179]
if webcam: # batch_size >= 1
p, s, im0 = path[i], '%g: ' % i, im0s[i].copy()
else:
p, s, im0 = path, '', im0s
save_path = str(Path(out) / Path(p).name) # 图片保存路径+图片名字
txt_path = str(Path(out) / Path(p).stem) + (
'_%g' % dataset.frame if dataset.mode == 'video' else '')
#print(txt_path)
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if det is not None and len(det):
# Rescale boxes from img_size to im0 size
det[:, :5] = scale_labels(img.shape[2:], det[:, :5],
im0.shape).round()
# Print results det:(num_nms_boxes, [xylsθ,conf,classid]) θ∈[0,179]
for c in det[:, -1].unique(
): # unique函数去除其中重复的元素，并按元素（类别）由大到小返回一个新的无元素重复的元组或者列表
n = (det[:, -1] == c
).sum() # detections per class 每个类别检测出来的素含量
s += '%g %ss, ' % (n, names[int(c)]
) # add to string 输出‘数量 类别,’
# Write results det:(num_nms_boxes, [xywhθ,conf,classid]) θ∈[0,179]
for *rbox, conf, cls in reversed(
det): # 翻转list的排列结果,改为类别由小到大的排列
# rbox=[tensor(x),tensor(y),tensor(w),tensor(h),tsneor(θ)] θ∈[0,179]
# if save_txt: # Write to file
# xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
# with open(txt_path + '.txt', 'a') as f:
# f.write(('%g ' * 5 + '\n') % (cls, *xywh)) # label format
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
classname = '%s' % names[int(cls)]
conf_str = '%.3f' % conf
rbox2txt(rbox, classname, conf_str,
Path(p).stem,
str(out + '/result_txt/result_before_merge'))
#plot_one_box(rbox, im0, label=label, color=colors[int(cls)], line_thickness=2)
plot_one_rotated_box(rbox,
im0,
label=label,
color=colors[int(cls)],
line_thickness=1,
pi_format=False)
# Print time (inference + NMS)
print('%sDone. (%.3fs)' % (s, t2 - t1))
# Stream results 播放结果
if view_img:
cv2.imshow(p, im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'images':
cv2.imwrite(save_path, im0)
pass
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
print(' Results saved to %s' % Path(out))
print(' All Done. (%.3fs)' % (time.time() - t0))
def detect(self, save_img=False) -> list:
source, weights, view_img, save_txt, imgsz = self.opt.source, self.opt.weights, self.opt.view_img, self.opt.save_txt, self.opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://'))
# Initialize
set_logging()
device = select_device(self.opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# print(half)
# Load model
t0 = time.time()
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
# if half:
# model.half() # to FP16
# Second-stage classifier
# classify = False
# if classify:
# modelc = load_classifier(name='resnet101', n=2) # initialize
# modelc.load_state_dict(torch.load('weights/resnet101.pt', map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
print(time.time() - t0)
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
return_imgs = []
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=False)[0]
# Apply NMS
pred = non_max_suppression(pred,
self.opt.conf_thres,
self.opt.iou_thres,
classes=self.opt.classes,
agnostic=self.opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
# if classify:
# pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = Path(
path[i]), '%g: ' % i, im0s[i].copy(), dataset.count
else:
p, s, im0, frame = Path(path), '', im0s, getattr(
dataset, 'frame', 0)
save_path = str(self.opt.save_dir / p.name)
txt_path = str(self.opt.save_dir / 'labels' / p.stem) + (
'' if dataset.mode == 'image' else f'_{frame}')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f'{n} {names[int(c)]}s, ' # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh,
conf) if self.opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line +
'\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=1)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
return_imgs.append(im0.copy())
if save_txt or save_img:
s = f"\n{len(list(self.opt.save_dir.glob('labels/*.txt')))} labels saved to {self.opt.save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {self.opt.save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
return return_imgs
def detect(save_img=False):
atama = 0
flag1 = False
flag2 = False
flag3 = False
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(torch.load('weights/resnet101.pt', map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
if device.type != 'cpu':
model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters()))) # run once
t0 = time.time()
for path, img, im0s, vid_cap in dataset: # her frame burda dönüyor
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
sonuc = Tespit()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det): # xyxy koordinatlar imiş
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (cls, *xywh)
# label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
(plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3))
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist()
ab = torch.tensor(xyxy).view(1,4)[0]
ab = ab.numpy()
isim = f'{names[int(cls)]}'
sonuc.label_list.append(isim)
sonuc.koordinat_list.append(ab)
#sol cıkıs ve kalıp label bulma
i = 0
cıkıs_list = []
while i < len(sonuc.label_list):
if('CIKIS' == sonuc.label_list[i]):
cıkıs_list.append(sonuc.koordinat_list[i])
if('KALIP' == sonuc.label_list[i]):
kalıp = sonuc.koordinat_list[i]
i += 1
if(len(cıkıs_list) == 2):
if(cıkıs_list[0][0] < cıkıs_list[1][0]):
sol_cıkıs = cıkıs_list[0]
else:
sol_cıkıs = cıkıs_list[1]
elif(len(cıkıs_list) == 1):
sol_cıkıs = cıkıs_list[0]
else:
sol_cıkıs = [0,0,0,0]
sol_alan = (sol_cıkıs[2] - sol_cıkıs[0]) * (sol_cıkıs[3] - sol_cıkıs[1])
cv2.putText(im0, "SOL ALAN" + str(sol_alan), (800, 500), cv2.FONT_HERSHEY_SIMPLEX, 1,
(209, 80, 0, 255), 3)
try:
x1 = kalıp[0]
x2 = kalıp[2]
y1 = kalıp[1]
y2 = kalıp[3]
print("x1 degeri : " , x1 , " x2 degeri : " , x2 , " y1 degeri : " , y1 , " y2 degeri : " , y2)
if ((640 < x1 < 675) and (855 < x2 < 874) and (290 < y1 < 305) and (835 < y2 < 855)) or flag1:
flag1 = True
print("butun kosullar saglandı")
cv2.putText(im0, "UYGUN KONUM" ,(200,200), cv2.FONT_HERSHEY_SIMPLEX, 1, (209,80, 0 ,255), 3)
print("sol kapak alanı : " , str(sol_alan))
if(2000 < sol_alan < 2400) or flag2:
# print("cıkmıs")
flag2 = True
if (900 < sol_alan < 1200) or flag3:
# print("en aşşa indi")
flag3 = True
if (2000 < sol_alan):
if 'BOS' in sonuc.label_list:
cv2.putText(im0, "SIKINTI YOK", (400, 400), cv2.FONT_HERSHEY_SIMPLEX, 1,
(209, 80, 0, 255), 3)
print("******************sıkıntı yok***********************")
else:
cv2.putText(im0, "KALIP DUSMEDI", (400, 400), cv2.FONT_HERSHEY_SIMPLEX, 1,
(209, 80, 0, 255), 3)
print("bunu bi şekilde halletmemiz gerek")
time.sleep(1)
#print("flag yazdırdık: "+ str(flag1)+ " "+ str(flag2)+ " "+ str(flag3))
else:
print("kosullar saglanmadı")
if ((sol_cıkıs[0] > 800) and flag3):
flag1 = False
flag2 = False
flag3 = False
except IndexError:
pass
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
atama += (t2 - t1)
print("toplam zaman", str(atama))
# Stream results
if True: # önceden view_img idi, şimdi True oldu yani resimleri video gibi oynatma sağlandı
cv2.imshow("Result", im0) # farklı isim olursa ayrı pencerelerde açılır, aynı isimle aynı pencerede açar
cv2.waitKey(1) # 1 millisecond - 0 girilir ise oynaması için imagein input bekler -- 1 kalması yeterli bizim için
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
vid_writer.write(im0)
#print("gecen zaman :" , str(t2-t1))
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
def detect(save_img=False):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir if save_txt else save_dir).mkdir(parents=True,
exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
imgsz = check_img_size(imgsz, s=model.stride.max()) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = False
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(
), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / p.stem) + (
'' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
dic = {"With_Mask": 0, "Without_Mask": 0, "Incorrect_Mask": 0}
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f'{n} {names[int(c)]}s, ' # add to string
dic[names[int(c)]] = int(n)
# Write results
with open(txt_path + '.json', 'a') as f:
json.dump(dic, f)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('*.json')))} labels saved to {save_dir}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
def detect(opt):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
save_img = not opt.nosave and not source.endswith(
'.txt') # save inference images
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://', 'https://'))
# Directories
save_dir = increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
else:
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
if device.type != 'cpu':
model(
torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(
next(model.parameters()))) # run once
t0 = time.time()
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(
), dataset.count
else:
p, s, im0, frame = path, '', im0s.copy(), getattr(
dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + (
'' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or opt.save_crop or view_img: # Add bbox to image
c = int(cls) # integer class
label = None if opt.hide_labels else (
names[c]
if opt.hide_conf else f'{names[c]} {conf:.2f}')
plot_one_box(xyxy,
im0,
label=label,
color=colors[c],
line_thickness=opt.line_thickness)
if opt.save_crop:
save_one_box(xyxy,
im0s,
file=save_dir / 'crops' / names[c] /
f'{p.stem}.jpg',
BGR=True)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video' or 'stream'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
if vid_cap: # video
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
else: # stream
fps, w, h = 30, im0.shape[1], im0.shape[0]
save_path += '.mp4'
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
def detect(source="data/images", weights="best.pt", imgsz=640):
result = []
# Initialize
set_logging()
device = select_device("")
half = device.type != "cpu" # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name="resnet101", n=2) # initialize
modelc.load_state_dict(
torch.load("weights/resnet101.pt",
map_location=device)["model"]).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
names = model.module.names if hasattr(model, "module") else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
if device.type != "cpu":
model(
torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(
next(model.parameters()))) # run once
t0 = time.time()
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment="store_true")[0]
# Apply NMS
pred = non_max_suppression(pred,
0.40,
0.45,
classes=None,
agnostic="store_true")
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
p, s, im0, frame = path, "", im0s, getattr(dataset, "frame", 0)
p = Path(p) # to Path
s += "%gx%g " % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s = f"{n} {names[int(c)]}{'s' * (n > 1)}" # add to string
result.append(s)
return result
def detect(save_img=False):
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
if opt.labels:
names = load_labels(opt.labels)
else:
names = model.module.names if hasattr(model, 'module') else model.names
if opt.color == "same":
# same color for all
colors = [[255, 0, 255] for _ in range(len(names))]
else:
if opt.color == "det":
random.seed(2)
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(names))]
# Run inference
if device.type != 'cpu':
model(
torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(
next(model.parameters()))) # run once
t0 = time.time()
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img
) if device.type != 'cpu' else None # run once
if opt.bboxfilt:
# Detections filter
bbox_filter = BboxFilter(30, 5) # 30x30 pixel grids, 5 deep in time
# Object sizes
# se, comml, jet, heli, drone
OBj_SIZES = (10.0, 30.0, 10.0, 5.0, 0.3)
CAM_FOVH = 60 # degree
quit = False
for path, img, im0s, vid_cap in dataset:
if quit: break
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
#print("pred shape", pred.shape)
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(
pred
): # detections per image, this is always 1 except for maybe (batched images?)
#print("i det",i, det)
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(
), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
imh, imw, _ = im0.shape
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + (
'' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1,
0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
x1, y1, x2, y2 = int(xyxy[0]), int(xyxy[1]), int(
xyxy[2]), int(xyxy[3])
if opt.bboxfilt:
bbox_filter.add((x1, y1, x2, y2), conf.item(),
int(cls))
#print("imw", im0.shape, int(cls))
# if int(cls) > len(OBj_SIZES):
# sz = 1.0
# else:
# sz = OBj_SIZES[int(cls)]
# eul, quat, dist = estimate_yaw_pitch_dist(CAM_FOVH, sz, imw, imh, x1, y1, x2, y2)
else:
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (
cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line +
'\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
# label = '%s %.2f %.1f %.2f %.2f %.2f %.2f' % (names[int(cls)], conf, dist, quat[0], quat[1], quat[2], quat[3])
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=2)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
# if view_img:
# cv2.imshow(p, im0)
# if cv2.waitKey(1) == ord('q'): # q to quit
# raise StopIteration
# Save results (image with detections)
# if save_img:
# if dataset.mode == 'images':
# cv2.imwrite(save_path, im0)
# else:
# if vid_path != save_path: # new video
# vid_path = save_path
# if isinstance(vid_writer, cv2.VideoWriter):
# vid_writer.release() # release previous video writer
# fourcc = 'mp4v' # output video codec
# fps = vid_cap.get(cv2.CAP_PROP_FPS)
# w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
# h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
# vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
# vid_writer.write(im0)
if opt.bboxfilt:
print("-----------------BBOX")
boxes = bbox_filter.get_boxes()
#print("filt boxes",boxes)
for box in boxes:
print("b", box)
xyxy, conf, cls = box
# label = '%s %.2f' % (names[cls], conf)
# print(label, conf)
# plot_one_box(b, im0, label=label, color=colors[cls], line_thickness=3)
if int(cls) > len(OBj_SIZES):
sz = 1.0
else:
sz = OBj_SIZES[int(cls)]
eul, quat, dist = estimate_yaw_pitch_dist(
CAM_FOVH, sz, imw, imh, x1, y1, x2, y2)
if save_img or view_img: # Add bbox to image
# with quaternion
# label = '%s %.2f %.1f %.2f %.2f %.2f %.2f' % (names[int(cls)], conf, dist, quat[0], quat[1], quat[2], quat[3])
label = '%s %.2f %.1f' % (names[int(cls)], conf, dist)
print(label)
print("xyxy", xyxy)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)],
line_thickness=2)
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) /
gn).view(-1).tolist() # normalized xywh
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * 5 + '\n') %
(cls, *xywh)) # label format
if view_img:
cv2.imshow(str(p), im0)
k = cv2.waitKey(1) # 1 millisecond
if k == 27: # ESC
quit = True
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release(
) # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(w, h))
vid_writer.write(im0)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
print(f'Done. ({time.time() - t0:.3f}s)')
def detect(save_img=False):
logging.basicConfig(filename='detect.log', level=logging.INFO)
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.lower().startswith(
('rtsp://', 'rtmp://', 'http://'))
imglist = opt.imlist
print("imglist : ", imglist)
source_list = source.split('\n')
#source_list = source
# Directories
save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
# save_dir = Path(increment_path(Path(opt.project) / opt.name)) # increment run
# (save_dir / 'labels' if save_txt else save_dir).mkdir # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(torch.load('weights/resnet101.pt', map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
else:
save_img = True #이미지 저장을 하지 않기위해 주석처
if imglist:
dataset_list = []
for i in range(len(source_list)):
#print("source_list : ", source_list[i])
try:
dataset_list.append(LoadImages(source_list[i], img_size=imgsz, stride=stride))
except:
print("error!!!!!: ", source_list[i])
else:
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
if device.type != 'cpu':
model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters()))) # run once
t0 = time.time()
#sujin
if imglist:
j = len(dataset_list)
#print("imglist True, j = ", j)
else:
print(dataset.nf)
detect_count =0
time_sum=0.0 #시간의 평균을 구하기위한 변
if imglist:
for k in range(j):
dataset = dataset_list[k] ## 필요없는 부분
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
prevTime = 0
# Process detections
for i, det in enumerate(pred): # detections per image
curTime = time.time() * 1000
sec = curTime - prevTime
prevTime = curTime #이전 시간을 현재시간으로 다시 저장시킴
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(), dataset.count
fps_ = 1/(sec)
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
print("p.name=", p.name)
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
pt_start = time.time()*1000
plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)
pt_end = time.time()*1000
if os.path.exists(result_dir) == False:
os.makedirs(result_dir)
with open(result_dir + p.name[:-4] + '.txt', 'w') as f:
xmin=(int(xyxy[0]))
ymin=(int(xyxy[1]))
xmax=(int(xyxy[2]))
ymax=(int(xyxy[3]))
h, w, bs = im0.shape
print("bs h w = ",bs, h, w)
absolute_x = xmin + 0.5 * (xmax - xmin)
absolute_y = ymin + 0.5 * (ymax - ymin)
absolute_width = xmax - xmin
absolute_height = ymax - ymin
x = str(absolute_x / w)
y = str(absolute_y / h)
width = str(absolute_width / w)
height = str(absolute_height / h)
f.write(str(int(cls))+" "+x + " " + y + " " + width + " " + height)
detect_count += 1
print("detect_count = ",detect_count)
else:
tl = 3 or round(0.002 * (im0.shape[0] + im0.shape[1]) / 2) + 1 # line/font thickness
tf = max(tl - 1, 1) # font thickness
cv2.putText(im0, "0", (0, 100), 0, tl / 3, [225, 255, 255], thickness=tf, lineType=cv2.LINE_AA)
# Print time (inference + NMS)
time_sum += t2-t1
print(f'{s}Done. ({t2 - t1:.3f}s)', ' avg fps=', time_sum/detect_count)
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
print("Save path=", save_path)
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fourcc = 'mp4v' # output video codec
fps = int(vid_cap.get(cv2.CAP_PROP_FPS))
print("fps is : ", fps)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
vid_writer.write(im0)
else:
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float()
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # img.jpg
txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # img.txt
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if opt.save_conf else (cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
if save_img or view_img: # Add bbox to image
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)
# Print time (inference + NMS)
print(f'{s}Done. ({t2 - t1:.3f}s)')
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
cv2.imwrite(save_path, im0)
else: # 'video'
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fourcc = 'mp4v' # output video codec
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
vid_writer.write(im0)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
print(f"Results saved to {save_dir}{s}")
if imglist:
print('detect rate = ',(detect_count),' / ',j-1,' =', detect_count/(j-1))
print('average fps = ', 1/(time_sum/(j-1)))
def do_detect(source = ' ', weights =' ', imlist= True, result = ' '):
print('<<Detection Start>>')
parser = argparse.ArgumentParser()
parser.add_argument('--weights', nargs='+', type=str, default='yolov5s.pt', help='model.pt path(s)')
parser.add_argument('--source', type=str, default='data/images', help='source') # file/folder, 0 for webcam
parser.add_argument('--img-size', type=int, default=640, help='inference size (pixels)')
parser.add_argument('--conf-thres', type=float, default=0.25, help='object confidence threshold')
parser.add_argument('--iou-thres', type=float, default=0.45, help='IOU threshold for NMS')
parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
parser.add_argument('--view-img', action='store_true', help='display results')
parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --class 0, or --class 0 2 3')
parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
parser.add_argument('--augment', action='store_true', help='augmented inference')
parser.add_argument('--update', action='store_true', help='update all models')
parser.add_argument('--project', default='runs/detect', help='save results to project/name')
parser.add_argument('--name', default='exp', help='save results to project/name')
parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
parser.add_argument('--imlist', action='store_true', help='existing project/name ok, do not increment')
global opt
global result_dir
result_dir = result
opt = parser.parse_args()
opt.source = source
opt.imlist = imlist
opt.weights= [weights]
check_requirements()
detect_start = time.time()*1000
print("DETECT START: ", detect_start)
with torch.no_grad():
detect()
detect_end = time.time()*1000-detect_start
logging.info("hfgf"+str(detect_end))
print("DETECT TIME : ", time.time()*1000-detect_start)
print("<<DETECT DONE>>")
default='runs/train',
help='save to project/name')
parser.add_argument('--entity', default=None, help='W&B entity')
parser.add_argument('--name', default='exp', help='save to project/name')
parser.add_argument('--exist-ok',
action='store_true',
help='existing project/name ok, do not increment')
parser.add_argument('--quad', action='store_true', help='quad dataloader')
parser.add_argument('--linear-lr', action='store_true', help='linear LR')
opt = parser.parse_args()
# Set DDP variables
opt.world_size = int(
os.environ['WORLD_SIZE']) if 'WORLD_SIZE' in os.environ else 1
opt.global_rank = int(os.environ['RANK']) if 'RANK' in os.environ else -1
set_logging(opt.global_rank)
if opt.global_rank in [-1, 0]:
check_git_status()
check_requirements()
# Resume
if opt.resume: # resume an interrupted run
ckpt = opt.resume if isinstance(
opt.resume,
str) else get_latest_run() # specified or most recent path
assert os.path.isfile(
ckpt), 'ERROR: --resume checkpoint does not exist'
apriori = opt.global_rank, opt.local_rank
with open(Path(ckpt).parent.parent / 'opt.yaml') as f:
opt = argparse.Namespace(**yaml.load(
f, Loader=yaml.SafeLoader)) # replace
def detect(save_img=False):
global cur_pic
global last_pic
source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
webcam = source.isnumeric() or source.endswith(
'.txt') or source.lower().startswith(('rtsp://', 'rtmp://', 'http://'))
# Directories
save_dir = Path(
increment_path(Path(opt.project) / opt.name,
exist_ok=opt.exist_ok)) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(
parents=True, exist_ok=True) # make dir
# Initialize
set_logging()
device = select_device(opt.device)
half = device.type != 'cpu' # half precision only supported on CUDA
# Load model
model = attempt_load(weights, map_location=device) # load FP32 model
stride = int(model.stride.max()) # model stride
imgsz = check_img_size(imgsz, s=stride) # check img_size
if half:
model.half() # to FP16
# Second-stage classifier
classify = False
if classify:
modelc = load_classifier(name='resnet101', n=2) # initialize
modelc.load_state_dict(
torch.load('weights/resnet101.pt',
map_location=device)['model']).to(device).eval()
# Set Dataloader
vid_path, vid_writer = None, None
if webcam:
print("load from webcam")
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz, stride=stride)
else:
print("load from picture")
save_img = True
dataset = LoadImages(source, img_size=imgsz, stride=stride)
# Get names and colors
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]
# Run inference
if device.type != 'cpu':
model(
torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(
next(model.parameters()))) # run once
t0 = time.time()
for path, img, im0s, vid_cap in dataset:
img = torch.from_numpy(img).to(device)
img = img.half() if half else img.float() # uint8 to fp16/32
img /= 255.0 # 0 - 255 to 0.0 - 1.0
if img.ndimension() == 3:
img = img.unsqueeze(0)
# Inference
t1 = time_synchronized()
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred,
opt.conf_thres,
opt.iou_thres,
classes=opt.classes,
agnostic=opt.agnostic_nms)
t2 = time_synchronized()
# Apply Classifier
if classify:
pred = apply_classifier(pred, modelc, img, im0s)
# Process detections
# 对每一张图片做处理：
for i, det in enumerate(pred): # detections per image
if webcam: # batch_size >= 1
p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(
), dataset.count
else:
p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)
# 保存图片的副本，留作截取目标bbox用
im0_copy = im0.copy()
# 每次循环将cur_pic中的内容给last_pic，并将cur_pic置空
if cur_pic:
last_pic = cur_pic
cur_pic = []
# 定义一个list存储一张图片的所有类别，并一起上传到前端
abnormal_list = []
abnormal_list2 = []
# 定义一个比较数量，将当前图片与last_pic中每一张比较，最后退出检测compare_num若等于last_pic张数，表明是新目标；否则是重复目标
compare_num = 0
new_object_flag = False
# 若图片上检测到的目标不为空
if len(det):
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results
# 打印检测到的类别数量
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
# 保存预测结果，此处是针对一张图上的每个检测框来说的
for *xyxy, conf, cls in reversed(det):
compare_num = 0
# 定义一个cur_pic1，字典类型，保留此张图片一个bbox的信息
cur_pic1 = dict()
cur_pic1['cls'] = torch.tensor(cls).view(-1).tolist()[0]
cur_pic1['xyxy'] = torch.tensor(xyxy).view(
1, 4).view(-1).tolist()
cur_pic1['conf'] = torch.tensor(conf).view(-1).tolist()[0]
cur_pic.append(cur_pic1)
# 当前帧一个bbox的坐标信息
x1, y1, x2, y2 = cur_pic1['xyxy'][0], cur_pic1['xyxy'][1], \
cur_pic1['xyxy'][2], cur_pic1['xyxy'][3]
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy,
im0,
label=label,
color=(255, 0, 0),
line_thickness=3)
# 去重方法：与上一帧相比，判断是否有新类别出现，若有，发送新的一帧图的目标检测结果；若无，不发送
# 具体步骤：
# 1.判断上一帧是否为空：
# 1.1.若为空,发送当前帧的所有目标
# 1.2.若不为空，设置new_object_flag=False
# 1.2.1.此帧中每个结果与上一帧每个结果相比较，比较依据为iou。
# 若当前帧某一个bbox与上一帧所有结果比较后iou均小于阈值，表明这个为新目标，将new_object_flag设置为True，并发送这一帧所有结果至前端
# 否则，break;
if last_pic:
for last_pic1 in last_pic:
# 上一帧某一个bbox的坐标信息
x1_hat, y1_hat, x2_hat, y2_hat = last_pic1['xyxy'][0], last_pic1['xyxy'][1], \
last_pic1['xyxy'][2], last_pic1['xyxy'][3]
if not compare_iou(x1, y1, x2, y2, x1_hat, y1_hat,
x2_hat, y2_hat):
print("两张图片的Iou值没有超过阈值,可以保留:{}".format(
cur_pic1['conf']))
compare_num += 1
else:
print("两张图片的Iou值太大,compare_num:{}".format(
compare_num))
continue
if compare_num == len(last_pic):
new_object_flag = True
break
else:
new_object_flag = True
break
if new_object_flag:
cur_pic.clear()
for *xyxy, conf, cls in reversed(det):
# 在原图上画框
label = f'{names[int(cls)]} {conf:.2f}'
plot_one_box(xyxy,
im0,
label=label,
color=(255, 0, 0),
line_thickness=3)
cur_pic1 = dict()
cur_pic1['cls'] = torch.tensor(cls).view(
-1).tolist()[0]
cur_pic1['xyxy'] = torch.tensor(xyxy).view(
1, 4).view(-1).tolist()
cur_pic1['conf'] = torch.tensor(conf).view(
-1).tolist()[0]
cur_pic.append(cur_pic1)
label = f'{names[int(cls)]} {conf:.2f}'
label_name = label.split(' ')[0]
cropped_image = im0_copy[int(xyxy[1]):int(xyxy[3]),
int(xyxy[0]):int(xyxy[2])]
if cropped_image.shape[0] and cropped_image.shape[1]:
cropped_image = base64.b64encode(
cv2.imencode('.jpg', cropped_image)
[1].tostring()).decode("utf-8")
abnormal_list.append((label_name, cropped_image), )
abnormal_list2.append(label_name, )
s = json.dumps(abnormal_list)
dangerousClient.publish("abnormal", payload=s)
print("------>发布一条abnormal_list的消息")
dangerousClient.publish(
topic="warning",
payload="Hazardous material detected in security area")
abnormal_list.clear()
abnormal_list2.clear()
flow_danger = base64.b64encode(
cv2.imencode('.jpg', im0)[1].tostring()).decode("utf-8")
dangerousClient.publish(topic="flowDangerous",
payload=flow_danger)
# Stream results
if view_img:
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond