def init_net():
net = dn.load_net(b"/app/darknet/cfg/yolov2-tiny.cfg", b"/app/darknet/yolov2-tiny.weights", 0)
meta = dn.load_meta(b"/app/darknet/cfg/coco.data")
return net, meta
def vehicleCount(root,
in_file,
output_dir,
lineState='Horizontal',
thresh=0.5,
roiThick=7,
offset=0):
try:
out_file = in_file.split('/')[-1].split('.')[0] + '_out'
vehicle_threshold = thresh
roi = roiThick
vehicle_weights = 'object_detector/data/vehicle-detector/yolov2.weights'
vehicle_netcfg = 'object_detector/data/vehicle-detector/yolov2.cfg'
vehicle_dataset = 'object_detector/data/vehicle-detector/coco.data'
vehicle_net = dn.load_net(vehicle_netcfg, vehicle_weights, 0)
vehicle_meta = dn.load_meta(vehicle_dataset)
writer = None
cap = cv2.VideoCapture(in_file)
cnt = 0
car_count = 0
while cap.isOpened():
ret, frame = cap.read()
if not ret:
root.statusStrVar.set(
'Done...Video saved at {}'.format(output_dir + '/' +
out_file + '.mp4'))
break
WH = frame.shape[:2]
img = nparray_to_image(frame)
R, _ = detect(vehicle_net, vehicle_meta, img, vehicle_threshold)
if lineState == "Horizontal":
linel = (0, WH[0] - (WH[0] / 4) - offset)
liner = (WH[1], WH[0] - (WH[0] / 4) - offset)
elif lineState == "Vertical":
lineu = (WH[1] - (WH[1] / 4) - offset, 0)
lined = (WH[1] - (WH[1] / 4) - offset, WH[1])
R = [r for r in R if r[0] in ['car', 'bus']]
print 'Processing frame {}'.format(cnt)
root.statusStrVar.set('Processing frame {}..'.format(cnt))
print '\t%d cars found' % len(R)
if len(R):
WH = np.array(frame.shape[1::-1], dtype=float)
for i, r in enumerate(R):
name = r[0]
cx, cy, w, h = (np.array(r[2]) / np.concatenate(
(WH, WH))).tolist()
tl = (int((cx - w / 2.) * WH[0]), int(
(cy - h / 2.) * WH[1]))
br = (int((cx + w / 2.) * WH[0]), int(
(cy + h / 2.) * WH[1]))
print '\t\t{}th car Coodrs : ({}, {})'.format(i, tl, br)
cv2.rectangle(frame, tl, br, (255, 0, 0),
2) #crop_region(Iorig,label)
if lineState == "Horizontal":
cv2.line(frame, linel, liner, (0, 0, 255), 3)
if (cy - h / 2) * WH[1] > liner[1] and (
cy - h / 2) * WH[1] < liner[1] + roi:
car_count += 1
elif lineState == "Vertical":
cv2.line(frame, lineu, lined, (0, 0, 255), 3)
if (cx - w / 2) * WH[0] > lineu[0] and (
cx - w / 2) * WH[0] < lineu[0] + roi:
car_count += 1
cv2.putText(frame, name, tl, cv2.FONT_HERSHEY_SIMPLEX, 1.5,
(0, 0, 255), 3, cv2.LINE_AA)
cv2.putText(frame, 'Vehicles crossed : ' + str(car_count),
(0, int(WH[1])), cv2.FONT_HERSHEY_SIMPLEX, 1,
(0, 0, 0), 2, cv2.LINE_AA)
print '\n'
if writer is None:
fourcc = cv2.VideoWriter_fourcc(*'DIVX')
writer = cv2.VideoWriter(output_dir + '/' + out_file + '.mp4',
fourcc, 30,
(frame.shape[1], frame.shape[0]),
True)
writer.write(frame)
del frame
cnt += 1
except:
traceback.print_exc()
sys.exit(1)
writer.release()
cap.release()
sys.exit(0)
file_log_handler = logging.FileHandler("log")
file_log_handler.setLevel(logging.INFO)
file_log_handler.setFormatter(
logging.Formatter(
'%(asctime)s - %(levelname)s -%(pathname)s- %(filename)s - %(funcName)s - %(lineno)s - %(message)s'
))
logger.addHandler(file_log_handler)
start = time.time()
if len(sys.argv) <= 1:
net = darknet.load_net("darknet/cfg/yolov3-tiny.cfg".encode('utf-8'),
"darknet/yolov3-tiny.weights".encode('utf-8'), 0)
else:
net = darknet.load_net("darknet/cfg/yolov3.cfg".encode('utf-8'),
"darknet/yolov3.weights".encode('utf-8'), 0)
meta = darknet.load_meta("darknet/cfg/coco.data".encode('utf-8'))
end = time.time()
logger.info("Time used for loading model: {}s".format(end - start))
@app.route('/')
def hello_world():
return 'Hello World!'
@app.route('/predict', methods=['POST'])
def predict():
start = time.time()
with open("temp.jpg", 'wb') as image:
image.write(request.stream.read())
end = time.time()
from darknet.python.darknet import detect
if __name__ == '__main__':
input_dir = sys.argv[1]
output_dir = input_dir
ocr_threshold = .4
ocr_weights = 'data/ocr/ocr-net.weights'
ocr_netcfg = 'data/ocr/ocr-net.cfg'
ocr_dataset = 'data/ocr/ocr-net.data'
ocr_net = dn.load_net(ocr_netcfg.encode("utf-8"), ocr_weights.encode("utf-8"), 0)
ocr_meta = dn.load_meta(ocr_dataset.encode("utf-8"))
imgs_paths = glob('%s/*lp.png' % output_dir)
print ('Performing OCR...')
for i,img_path in enumerate(imgs_paths):
print ('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
R = detect(ocr_net, ocr_meta, img_path ,thresh=ocr_threshold)
if len(R):
if __name__ == '__main__':
try:
input_dir = sys.argv[1]
output_dir = sys.argv[2]
vehicle_threshold = 0.5
vehicle_weights = 'data/vehicle-detector/yolo-voc.weights'
vehicle_netcfg = 'data/vehicle-detector/yolo-voc.cfg'
vehicle_dataset = 'data/vehicle-detector/voc.data'
vehicle_net = dn.load_net(vehicle_netcfg.encode('utf-8'),
vehicle_weights.encode('utf-8'), 0)
vehicle_meta = dn.load_meta(vehicle_dataset.encode('utf-8'))
imgs_paths = image_files_from_folder(input_dir)
imgs_paths.sort()
if not isdir(output_dir):
makedirs(output_dir)
print('Searching for vehicles using YOLO...')
for i, img_path in enumerate(imgs_paths):
print('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
import traceback
from src.draw_BB import draw_bb
from WPOD_src.label import Label
from WPOD_src.utils import image_files_from_folder
import darknet.python.darknet as dn
# best:FRNet_YOLOv3_50000.weights
FR_weights = 'data/FRD/FRNet_YOLOv3_50000.weights'
FR_netcfg = 'data/FRD/FRNet_YOLOv3.cfg'
FR_data = 'data/FRD/FRNet_YOLOv3.data'
print 'FRD Net pre-loading...'
FR_net = dn.load_net(FR_netcfg, FR_weights, 0)
FR_meta = dn.load_meta(FR_data)
threshold = 0.5
def fr_detect(img):
print '\t\t\tdetecting front and rear using FRD..., Model:', FR_netcfg
results, wh = dn.detect(FR_net, FR_meta, img, threshold)
# the results will be list according to its probability , high prob -> low prob
if len(results):
print '\t\t\tFR detection completed'
FRs = []
category = []
for i, result in enumerate(results):
WH = np.array(img.shape[1::-1], dtype=float)
cx, cy, w, h = (np.array(result[2]) / np.concatenate(
def main():
cfg = util.parse_cla() #"./config/config.yaml"
#Load config
print("Loading configuration...", end="")
args = config.get_parser(cfg.config)
print("Done")
if args.BOUNDING:
#Initialise the yolo predictor once to hold model for all predictions
print("Loading yolo model...", end="")
net = dn.load_net(bytes(args.YOLO_CFG, 'utf-8'),
bytes(args.YOLO_WEIGHTS, 'utf-8'), 0)
meta = dn.load_meta(bytes(args.YOLO_DATA, 'utf-8'))
print("Done")
else:
#Initialise the segmentation predictor once to hold model for all predictions
print("Loading segmentation model...", end="")
predictor = segmentation.SegmentationPredictor(args)
print("Done")
#Initialise the pose estimator
print("Loading pose model...", end="")
ap = AlphaPose(Args(cfg.config, args.POSE_MODEL))
print("Done")
#Initialise classification model
print("Loading classification model...", end="")
nn_classifier = pickle.load(open(args.CLASSIFICATION_MODEL, 'rb'))
print("Done")
#Load in all the images from the in_dir folder
if (args.IMAGE_INPUT) & (not args.VIDEO_INPUT):
images = util.get_images(args.IN_DIR)
elif (args.VIDEO_INPUT) & (not args.IMAGE_INPUT):
#If working on a video convert the input video into a series of images
#Skipping certain frames in order to make it the correct framerate
#Given the option to reduce framerate as can be slow to process many frames
print("Converting video to input frames...", end="")
images = util.video_to_frames(args.IN_DIR, args.VIDEO_FILE,
args.FRAMERATE)
print("Done")
else:
raise ValueError(
"VIDEO input and IMAGE input can't happen simultaneously")
#Tags object to store data calculated
tags = {}
#Initialise data
i = 0
ruck = scrum = maul = lineout = []
images = list(images)
len_images = len(images)
start = datetime.datetime.now()
#For each image
for image, image_path in images:
#Debugging
util.print_progress_bar(i,
len_images,
suffix="{}/{}".format(i, len_images))
#Setup paths
inpath = os.path.join(args.IN_DIR, image_path)
outpath = os.path.join(args.OUT_DIR, image_path)
#Load Image
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
#If using the bounding box YOLO
if args.BOUNDING:
#Convert to correct formay
im = dn.array_to_image(image)
dn.rgbgr_image(im)
#Get clusters
dimentions = dn.detect2(net,
meta,
im,
thresh=.5,
hier_thresh=.5,
nms=.45)
#If clusters detected crop image to cluster parts
if len(dimentions) > 0:
dimentions = np.array(dimentions)[:, 2]
image_clusters = clusters.create_image_clusters(
image, dimentions)
else:
image_clusters = []
#If using segmentation mask PSPN
else:
mask = predictor.predict(image)
#Convert mask from PIL format to numpy/opencv
mask = np.array(mask) * 255
mask = cv2.cvtColor(mask, cv2.COLOR_GRAY2RGB)
mask = cv2.bitwise_not(mask)
#Get clusters
out = clusters.makemask(image, mask)
image_clusters, dimentions = clusters.extractclusters(out, image)
dimentions = [dimentions]
if len(image_clusters) > 0:
#Just work with single biggest cluster
if 0 in image_clusters[0].shape:
continue
#Run through alphapose to get json of poses
json_data = ap.predict(image_clusters[0], image_path)
json_data = json.loads(json_data)
#Classify
try:
cluster = nn_classifier.predict(json_data)
max_index = np.argmax(cluster[0])
except:
continue
#Convert back to textual format
tag = nn_classifier.le.inverse_transform([max_index])
#Unpack dimentions of cluster
x, y, w, h = dimentions[0]
x = int(x)
y = int(y)
w = int(w)
h = int(h)
#Draw bounding box and add tag annotation to original image
cv2.rectangle(image, (x, y), (x + w, y + h), args.FONT_COLOR,
args.FONT_THICKNESS)
cv2.putText(image, "{0}: {1:.3f}".format(tag[0],
cluster[0][max_index]),
((x + w + 10), (y - 10)), args.FONT, args.FONT_SCALE,
args.FONT_COLOR, args.FONT_THICKNESS)
#Draw predictions graph
if args.PREDITION_GRAPH:
#Create random sample data for now
ruck = util.add_to_const_arr(ruck, cluster[0][0],
args.FRAME_HISTORY)
maul = util.add_to_const_arr(maul, cluster[0][1],
args.FRAME_HISTORY)
scrum = util.add_to_const_arr(scrum, cluster[0][2],
args.FRAME_HISTORY)
lineout = util.add_to_const_arr(lineout, cluster[0][3],
args.FRAME_HISTORY)
#Calculate overlay size based on UI_SCALE parameter
UI_SCALE = 1 / args.UI_SCALE
overlay_dim = (int(image.shape[1] / (UI_SCALE * 10)),
int(image.shape[0] / (UI_SCALE * 7)))
#Create graph
fig = ui.plot_preditions_graph(ruck, maul, scrum, lineout)
#Convert to numpy array (cv2 format)
overlay = ui.get_img_from_fig(fig)
#Resize to correct dimentions
overlay = cv2.resize(overlay, overlay_dim)
#Overlay images
image = overlay_images(overlay, image)
#Update the ouput object with image tag
tags[image_path] = {
"tag": tag[0],
"prob": cluster[0][max_index],
"bbox": {
"x": x,
"y": y,
"w": w,
"h": h
}
}
#Save original image with bounding box and associated tag
plt.imsave(outpath, image)
i += 1
end = datetime.datetime.now()
print("Runtime:", end - start)
#Print once again to show 100%
util.print_progress_bar(len_images, len_images)
#Output the results object
with open(os.path.join(args.OUT_DIR, "results.json"), "w") as file:
json.dump(tags, file)
#If evaluating on the test dataset
if args.EVALUATE:
acc = 0
detections = 0
#Get the test dataset as json
data = preprocessor.import_json(args.TEST_DATASET)
#For each image compare tag to predicted tag
for image, image_path in images:
#If both exist, otherwise add 0 as no cluster/poses of note were found on that image
try:
if image_path in tags.keys():
detections += 1
if data[image_path]['tag'] == tags[image_path]['tag']:
acc += 1
except:
continue
acc = acc / len_images
print("Detections in test dataset:", detections)
print("Acuracy on test dataset:", acc)
#If video input delete frames created and make new video from outputed frames
if args.VIDEO_INPUT:
print("Processing output video...", end="")
#Collate filenames to be made into video
img_file_arr = glob.glob('output/*.png')
img_file_arr.sort(key=util.sort_filenames)
#Open files into array
img_array = []
for filename in img_file_arr:
img = cv2.imread(filename)
height, width, layers = img.shape
size = (width, height)
img_array.append(img)
#Create video output file with parameters
out = cv2.VideoWriter(os.path.join("output", 'output.mp4'),
cv2.VideoWriter_fourcc(*'mp4v'), args.FRAMERATE,
size)
#Write frame array to the output video file
for i in range(len(img_array)):
out.write(img_array[i])
out.release()
print("Done")
from darknet.python.darknet import detect
if __name__ == '__main__':
input_dir = sys.argv[1]
output_dir = sys.argv[2]
vehicle_threshold = .5
vehicle_weights = 'data/vehicle-detector/yolo-voc.weights'
vehicle_netcfg = 'data/vehicle-detector/yolo-voc.cfg'
vehicle_dataset = 'data/vehicle-detector/voc.data'
vehicle_net = dn.load_net(vehicle_netcfg.encode("utf-8"),
vehicle_weights.encode("utf-8"), 0)
vehicle_meta = dn.load_meta(vehicle_dataset.encode("utf-8"))
imgs_paths = image_files_from_folder(input_dir)
imgs_paths.sort()
if not isdir(output_dir):
makedirs(output_dir)
print('Searching for vehicles using YOLO...')
for i, img_path in enumerate(imgs_paths):
print('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
context = zmq.Context()
zmq_socket = context.socket(zmq.PUSH)
zmq_socket.connect("tcp://zmq:5559")
sio = socketio.Client()
sio.connect('http://server:5000')
thresh = 0.4
weights = 'darknet/models/bins/backup/yolo-v4-obj-train_last.weights'
netcfg = 'darknet/models/bins/yolo/yolo-v4-obj-test.cfg'
data = 'darknet/models/bins/yolo/annotate/obj.data'
net = dn.load_net(netcfg.encode('utf-8'), weights.encode('utf-8'), 0)
meta = dn.load_meta(data.encode('utf-8'))
camera = 0 # RTSP IP For ip-cam OR 0 For Default video-cam
accept_cls = ['glass', 'can', 'plastic']
pgm = {'glass': 0, 'can': 1, 'plastic': 2}
video_camera = None
def video_stream():
global video_camera, net, meta, data
count = 0
alert_classes = [] # target classes
alert_classes += accept_cls
current = datetime.now()
if video_camera == None:
video_camera = VideoCamera(camera=camera, alert_classes=alert_classes)
while True:
try:
args = parse_args()
input_dir = args.input_dir
output_dir = input_dir
lp_threshold = args.lp_threshold
lp_weights = b'data/simple-lp-detector/lapi.weights'
lp_netcfg = b'data/simple-lp-detector/yolov3-lp.cfg'
lp_dataset = b'data/simple-lp-detector/yolov3-lp.data'
lp_net = dn.load_net(lp_netcfg, lp_weights, 0)
lp_meta = dn.load_meta(lp_dataset)
imgs_paths = image_files_from_folder(input_dir)
# Filter only cropped car images
imgs_paths.sort()
if not isdir(output_dir):
makedirs(output_dir)
print('Searching for license plates in cropped cars using YOLO...')
for i, img_path in enumerate(imgs_paths):
print('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
from darknet.python.darknet import detect
if __name__ == '__main__':
input_dir = sys.argv[1]
output_dir = input_dir
ocr_threshold = .4
ocr_weights = 'data/ocr/ocr-net.weights'
ocr_netcfg = 'data/ocr/ocr-net.cfg'
ocr_dataset = 'data/ocr/ocr-net.data'
ocr_net = dn.load_net(ocr_netcfg.encode('utf-8'),
ocr_weights.encode('utf-8'), 0)
ocr_meta = dn.load_meta(ocr_dataset.encode('utf-8'))
imgs_paths = glob('%s/*lp.png' % output_dir)
print('Performing OCR...')
for i, img_path in enumerate(imgs_paths):
print('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
R = detect(ocr_net,
ocr_meta,
img_path.encode('utf-8'),
thresh=ocr_threshold)
def load_yolo(dn_net, net_weights, net_meta):
vehicle_net = dn.load_net(dn_net, net_weights, 0)
vehicle_meta = dn.load_meta(net_meta)
return vehicle_net, vehicle_meta
cwd = os.getcwd()
try:
os.chdir("./darknet")
os.system("wget 'https://pjreddie.com/media/files/yolov3.weights'")
os.system("make")
finally:
os.chdir("..")
# process2 = subprocess.Popen(["wget", "https://pjreddie.com/media/files/yolov3.weights"], cwd = cwd + '/darknet')
# process2.wait()
# process = subprocess.Popen(["make"], cwd = cwd + '/darknet')
# process.wait()
from darknet.python import darknet as dn
import pdb
net = dn.load_net("darknet/cfg/yolov3.cfg".encode("utf-8"), "darknet/yolov3.weights".encode("utf-8"), 0)
meta = dn.load_meta("darknet/cfg/coco.data".encode("utf-8"))
image = cv2.imread("darknet/data/dog.jpg")
result = dn.detect(net, meta, image)
print(result)
app = Flask(__name__)
net = 0
meta = 0
def load_model():
pass
"""Load and return the model"""
# TODO: INSERT CODE
# return model
# The request method is POST (this method enables your to send arbitrary data to the endpoint in the request body, including images, JSON, encoded-data, etc.)
if __name__ == '__main__':
try:
input_dir = sys.argv[1]
output_dir = input_dir
ocr_threshold = .4
ocr_weights = 'data/ocr/ocr-net.weights'
ocr_netcfg = 'data/ocr/ocr-net.cfg'
ocr_dataset = 'data/ocr/ocr-net.data'
ocr_net = dn.load_net(ocr_netcfg, ocr_weights, 0)
ocr_meta = dn.load_meta(ocr_dataset)
imgs_paths = sorted(glob('%s/*lp.png' % output_dir))
print 'Performing OCR...'
# --store the image name temporarily---
last_handle = ''
# -------------------------------------
for i, img_path in enumerate(imgs_paths):
print '\tScanning %s' % img_path
bname = basename(splitext(img_path)[0])
R, (width, height) = detect(ocr_net,
output_dir = '/home/yaokun/tmp' #sys.argv[2]
bname = basename(splitext(input_video_path)[0])
# vehicle detect
vehicle_threshold = .5
vehicle_weights = 'data/vehicle-detector/yolo-voc.weights' #.encode('ascii')
vehicle_netcfg = 'data/vehicle-detector/yolo-voc.cfg' #.encode('ascii').encode('ascii')
vehicle_dataset = 'data/vehicle-detector/voc.data' #.encode('ascii')
vehicle_weights = vehicle_weights.encode('ascii')
vehicle_netcfg = vehicle_netcfg.encode('ascii')
vehicle_dataset = vehicle_dataset.encode('ascii')
vehicle_net = dn.load_net(vehicle_netcfg, vehicle_weights, 0)
vehicle_meta = dn.load_meta(vehicle_dataset)
# lp detect
lp_threshold = .5
wpod_net_path = 'data/lp-detector/bak/wpod-net_update1.h5'
print(wpod_net_path)
#wpod_net_path = wpod_net_path.encode('ascii')
wpod_net = load_model(wpod_net_path)
# ocr
ocr_threshold = .4
ocr_weights = 'data/ocr/ocr-net.weights'
ocr_netcfg = 'data/ocr/ocr-net.cfg'
ocr_dataset = 'data/ocr/ocr-net.data'
try:
args = parse_args()
input_dir = args.input_dir
output_dir = args.output_dir
vehicle_threshold = args.vehicle_threshold
vehicle_weights = b'data/vehicle-detector/yolov3.weights'
vehicle_netcfg = b'data/vehicle-detector/yolov3.cfg'
# vehicle_netcfg = b'darknet/cfg/yolov3.cfg'
vehicle_dataset = b'data/vehicle-detector/coco.data'
vehicle_net = dn.load_net(vehicle_netcfg, vehicle_weights, 0)
vehicle_meta = dn.load_meta(vehicle_dataset)
imgs_paths = image_files_from_folder(input_dir)
imgs_paths.sort()
if not isdir(output_dir):
makedirs(output_dir)
print('Searching for vehicles using YOLO...')
for i, img_path in enumerate(imgs_paths):
print('\tScanning %s' % img_path)
bname = basename(splitext(img_path)[0])
def objectDetection(root, in_file, output_dir, classes):
try:
out_file = in_file.split('/')[-1].split('.')[0] + '_out'
vehicle_threshold = .5
vehicle_weights = 'object_detector/data/vehicle-detector/yolov2.weights'
vehicle_netcfg = 'object_detector/data/vehicle-detector/yolov2.cfg'
vehicle_dataset = 'object_detector/data/vehicle-detector/coco.data'
file_ptr = open(output_dir + '/' + out_file + '_obj.txt', 'w+')
vehicle_net = dn.load_net(vehicle_netcfg, vehicle_weights, 0)
vehicle_meta = dn.load_meta(vehicle_dataset)
if 'person' in classes:
gen_model = import_gender_model('./gender_model.h5')
writer = None
cap = cv2.VideoCapture(in_file)
cnt = 0
while cap.isOpened():
ret, frame = cap.read()
if not ret:
root.statusStrVar.set(
'Done...Video saved at {}'.format(output_dir + '/' +
out_file + '.mp4'))
break
print 'Processing frame {}..'.format(cnt)
root.statusStrVar.set('Processing frame {}..'.format(cnt))
img = nparray_to_image(frame)
R, _ = detect(vehicle_net, vehicle_meta, img, vehicle_threshold)
R = [r for r in R if r[0] in classes]
print '\t%d objects found' % len(R)
txt = 'Frame No : {}\n'.format(cnt)
if len(R):
WH = np.array(frame.shape[1::-1], dtype=float)
for i, r in enumerate(R):
name = r[0]
confidence = r[1] * 100
cx, cy, w, h = (np.array(r[2]) / np.concatenate(
(WH, WH))).tolist()
tl = (int((cx - w / 2.) * WH[0]), int(
(cy - h / 2.) * WH[1]))
br = (int((cx + w / 2.) * WH[0]), int(
(cy + h / 2.) * WH[1]))
object_img = frame[tl[1]:tl[1] + int(w * WH[1]),
tl[0]:tl[0] + int(h * WH[0])]
if name == 'car':
color = ''
if object_img.shape[0] > 0 and object_img.shape[1] > 0:
color = process_image(Image.fromarray(object_img))
# color = get_color(object_img)
txt += '\t Object Name, Coords : {}, ({}, {})\n'.format(
name, tl, br)
txt += '\t Object confidence : ' + '{:.2f}'.format(
confidence) + '\n'
txt += '\t Vehicle Color : {}\n\n'.format(color)
cv2.putText(
frame,
'{} ('.format(name) + '{:.2f}'.format(confidence) +
', {})'.format(color), tl,
cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2,
cv2.LINE_AA)
if name == 'person':
gender = ''
if object_img.shape[0] > 0 and object_img.shape[1] > 0:
gender = gender_detect(object_img, gen_model)
txt += '\t Object Name, Coords : {}, ({}, {})\n'.format(
name, tl, br)
txt += '\t Object confidence : ' + '{:.2f}'.format(
confidence) + '\n'
txt += '\t Person gender : {}\n\n'.format(gender)
cv2.putText(
frame,
'{} ('.format(name) + '{:.2f}'.format(confidence) +
', {})'.format(gender), tl,
cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 1,
cv2.LINE_AA)
if name != 'car' and name != 'person':
txt += '\t Object Name, Coords : {}, ({}, {})\n'.format(
name, tl, br)
txt += '\t Object confidence : ' + '{:.2f}'.format(
confidence) + '\n\n'
cv2.putText(
frame, '{} ('.format(name) +
'{:.2f}'.format(confidence) + ')', tl,
cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2,
cv2.LINE_AA)
print '\t\t{}th Coodrs : ({}, {})'.format(i, tl, br)
cv2.rectangle(frame, tl, br, (255, 0, 0),
2) #crop_region(Iorig,label)
file_ptr.write(txt)
print '\n'
if writer is None:
fourcc = cv2.VideoWriter_fourcc(*'DIVX')
writer = cv2.VideoWriter(output_dir + '/' + out_file + '.mp4',
fourcc, 30,
(frame.shape[1], frame.shape[0]),
True)
writer.write(frame)
del frame
cnt += 1
except:
traceback.print_exc()
sys.exit(1)
file_ptr.close()
writer.release()
cap.release()
sys.exit(0)