def run(self, video_path):
self.imageCount = 0
self.capture = cv2.VideoCapture(video_path)
while True:
ret, frame = self.capture.read()
if not ret:
print("视频解析")
break
im = darknet.nparray_to_image(frame)
boxes = darknet.detect(self.net, self.meta, im)
for i in range(len(boxes)):
score = boxes[i][1]
label = boxes[i][0]
xmin = boxes[i][2][0] - boxes[i][2][2] / 2
ymin = boxes[i][2][1] - boxes[i][2][3] / 2
xmax = boxes[i][2][0] + boxes[i][2][2] / 2
ymax = boxes[i][2][1] + boxes[i][2][3] / 2
cv2.rectangle(frame, (int(xmin), int(ymin)),
(int(xmax), int(ymax)), (0, 255, 0), 2)
cv2.putText(frame,
str(label) + str(round(score, 3)),
(int(xmin), int(ymin)),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.8,
color=(0, 255, 255),
thickness=2)
frame = cv2.resize(frame, (1024, 900),
interpolation=cv2.INTER_CUBIC)
cv2.imwrite("./temps/temp" + str(self.imageCount) + ".jpg", frame)
self.imageCount += 1
print(self.imageCount)
print("finish")
def run(self):
self.capture = cv2.VideoCapture(0)
self.imageCount = 0
while True:
ret, frame = self.capture.read()
if not ret:
break
im = darknet.nparray_to_image(frame)
boxes = darknet.detect(self.net, self.meta, im)
for i in range(len(boxes)):
score = boxes[i][1]
label = boxes[i][0]
xmin = boxes[i][2][0] - boxes[i][2][2] / 2
ymin = boxes[i][2][1] - boxes[i][2][3] / 2
xmax = boxes[i][2][0] + boxes[i][2][2] / 2
ymax = boxes[i][2][1] + boxes[i][2][3] / 2
cv2.rectangle(frame, (int(xmin), int(ymin)),
(int(xmax), int(ymax)), (0, 255, 0), 2)
cv2.putText(frame,
str(label)[2:-1] + str(round(score, 3)),
(int(xmin), int(ymin)),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.8,
color=(0, 255, 255),
thickness=2)
frame = cv2.resize(frame, (1024, 900),
interpolation=cv2.INTER_CUBIC)
cv2.imwrite("./temps/temp" + str(self.imageCount) + ".jpg",
frame) #检测后的图片保存
self.capture_image.emit("./temps/temp" + str(self.imageCount) +
".jpg") #发送完成图片保存的信号
time.sleep(0.05)
self.imageCount += 1
if cv2.waitKey(1) & 0xFF == ord('q'): # 按键盘q就停止拍照
break
def run(self):
self.imageCount = 0
self.capture = cv2.VideoCapture(self.video_path)
while True:
ret, frame = self.capture.read()
if not ret:
break
im = darknet.nparray_to_image(frame)
boxes = darknet.detect(self.net, self.meta, im)
for i in range(len(boxes)):
score = boxes[i][1]
label = boxes[i][0]
xmin = boxes[i][2][0] - boxes[i][2][2] / 2
ymin = boxes[i][2][1] - boxes[i][2][3] / 2
xmax = boxes[i][2][0] + boxes[i][2][2] / 2
ymax = boxes[i][2][1] + boxes[i][2][3] / 2
cv2.rectangle(frame, (int(xmin), int(ymin)),
(int(xmax), int(ymax)), (0, 255, 0), 2)
cv2.putText(frame,
str(label), (int(xmin), int(ymin)),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.8,
color=(0, 255, 255),
thickness=2)
cv2.imwrite("./temps/temp" + str(self.imageCount) + ".jpg", frame)
self.add_image.emit("./temps/temp" + str(self.imageCount) + ".jpg")
time.sleep(0.05)
self.imageCount += 1
def videoDetection(self):
answer = QMessageBox.warning(
self, "注意", "如果您上次使用了摄像头检测或视频检测后没有导出，当前操作会导致数据丢失，请确认是否要继续！",
QMessageBox.No | QMessageBox.Yes, QMessageBox.Yes)
if answer == QMessageBox.No:
return
if os._exists("./temps"):
shutil.rmtree("./temps")
os.mkdir("./temps")
fname, _ = QFileDialog.getOpenFileName(self, '请选择图片文件', ".", "(*.mp4)")
fnameSplit = fname.split('.')
index = len(fnameSplit) - 1
typeSuffix = fnameSplit[index] #文件名后缀
if typeSuffix != "mp4":
QMessageBox.critical(self, "文件类型错误", "您未指定视频或指定的文件不是mp4文件，请确认！",
QMessageBox.Yes, QMessageBox.Yes)
else: #正确的图片路径
cap = cv2.VideoCapture(fname)
while True:
ret, frame = cap.read()
if ret:
im = darknet.nparray_to_image(frame)
r = darknet.detect(self.net, self.meta, im)
else:
break
cap.release()
cv2.destroyAllWindows()
def classify(self, img):
self.log("Preparing image...")
# Convert to c image
c_img = darknet.nparray_to_image(img)
self.log("Done, classifying...")
result = darknet.detect(self.network, self.metadata, c_img)
self.log("Done, drawing boxes...")
new_image = self.draw_boxes(img, result)
self.log("Done")
return new_image
def imageDetection(self):
answer = QMessageBox.warning(
self, "注意", "如果您上次使用了摄像头检测或视频检测后没有导出，当前操作会导致数据丢失，请确认是否要继续！",
QMessageBox.No | QMessageBox.Yes, QMessageBox.Yes)
if answer == QMessageBox.No:
return
if os.path.exists("./temps"):
shutil.rmtree("./temps")
os.mkdir("./temps")
detection_image, _ = QFileDialog.getOpenFileName(
self, '请选择图片文件', ".", "(*.jpg)")
fnameSplit = detection_image.split('.')
index = len(fnameSplit) - 1
typeSuffix = fnameSplit[index] # 文件名后缀
if typeSuffix != "jpg":
QMessageBox.critical(self, "文件类型错误", "您未指定图片或指定的文件不是jpg文件，请确认！",
QMessageBox.Yes, QMessageBox.Yes)
else: # 正确的图片路径
frame = cv2.imread(detection_image)
im = darknet.nparray_to_image(frame)
boxes = darknet.detect(self.net, self.meta, im)
for i in range(len(boxes)):
score = boxes[i][1]
label = boxes[i][0]
xmin = max(5, boxes[i][2][0] -
boxes[i][2][2] / 2) #可以根据坐标关系调整文字位置
ymin = max(5, boxes[i][2][1] - boxes[i][2][3] / 2)
xmax = boxes[i][2][0] + boxes[i][2][2] / 2
ymax = boxes[i][2][1] + boxes[i][2][3] / 2
cv2.rectangle(frame, (int(xmin), int(ymin)),
(int(xmax), int(ymax)), (0, 255, 0), 2)
cv2.putText(frame,
str(label)[2:-1] + str(round(score, 3)),
(int(xmin), int(ymin)),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.8,
color=(255, 0, 0),
thickness=2)
frame = cv2.resize(frame, (1024, 900),
interpolation=cv2.INTER_CUBIC)
cv2.imwrite("./temps/temp" + ".jpg", frame) #检测后的图片保存
self.showImage("./temps/temp" + ".jpg")
thrs = [float(el) for el in args.threshold.split(',')]
if len(thrs) == 1: thrs = thrs * len(classnames)
assert (
len(thrs) == len(classnames)
), "Provide thresholds amount the same as classnames amount or single"
thrs = {cls: thrs[i] for i, cls in enumerate(classnames)}
print(cfg, met, wgh)
dn.set_gpu(0)
net = dn.load_net(cfg.encode('utf-8'), wgh.encode('utf-8'), 0)
meta = dn.load_names(met.encode('utf-8'))
for i, sample in enumerate(dataset):
print("{} ({} from {})".format(sample['imgFn'], i + 1, len(dataset)))
img = cv2.imread(sample['imgFn'])
dets = dn.detect(net, meta, dn.nparray_to_image(img))
dets = convertWithThreshold(dets, thrs)
if verbose:
print('\t{}'.format(sample['objs']))
print('\t{}'.format(dets))
estimateDets(dets, sample['objs'], classnames, generalEstimation)
for classname, estimation in perClassEstimations.items():
estimateDets(dets, sample['objs'], (classname, ), estimation)
estimationTable.analyseDetections(dets, sample['objs'])
print("General estimation results:")
generalEstimation.printStats()
print("Per class estimation results:")
for classname, estimation in perClassEstimations.items():
print("----- " + classname + " -----")
estimation.printStats()
origin_img = "data/dog.jpg"
out_img = "data/dog_test.jpg"
def showPicResult(image):
img = cv2.imread(image)
cv2.imwrite(out_img, img)
for i in range(len(r)):
x1 = r[i][2][0] - r[i][2][2] / 2
y1 = r[i][2][1] - r[i][2][3] / 2
x2 = r[i][2][0] + r[i][2][2] / 2
y2 = r[i][2][1] + r[i][2][3] / 2
im = cv2.imread(out_img)
cv2.rectangle(im, (int(x1), int(y1)), (int(x2), int(y2)), (0, 255, 0),
3)
#This is a method that works well.
cv2.imwrite(out_img, im)
cv2.imshow('yolo_image_detector', cv2.imread(out_img))
cv2.waitKey(0)
cv2.destroyAllWindows()
net = dn.load_net(str.encode("cfg/yolov3.cfg"), str.encode("yolov3.weights"),
0)
meta = dn.load_meta(str.encode("cfg/coco.data"))
image = dn.nparray_to_image(cv2.imread(origin_img))
r = dn.detect(net, meta, image)
#r = dn.detect(net, meta, "data/dog.jpg")
print r
showPicResult(origin_img)
cv2.imshow('yolo_image_detector', image)
#cv2.waitKey(0)
#cv2.destroyAllWindows()
net = dn.load_net(str.encode("cfg/yolov3.cfg"), str.encode("yolov3.weights"),
0)
meta = dn.load_meta(str.encode("cfg/coco.data"))
#print rq
cap = cv2.VideoCapture(video_path)
while (cap.isOpened()):
ret, frame = cap.read()
#cv2.imshow("lalala",frame)
start = time.time()
image = dn.nparray_to_image(frame)
r = dn.detect(net, meta, image)
showPicResult(frame)
end = time.time()
print r
#fps = cap.get(cv2.CAP_PROP_FPS)
seconds = end - start
fps = 1 / seconds
print "fps: ", fps
for i in range(len(r)):
if 'person' in r[i]:
print "ALERT!"
if cv2.waitKey(1) & 0xFF == ord('q'):
break
#cv2.release()
def __call__(self, img, threshold):
im = dn.nparray_to_image(img)
print("Created np image")
dets = dn.detect(self.net, self.meta, im, threshold)
return self._fromYoloFormat(dets)
meta = darknet.load_meta(
b"C:/Users/user/anaconda3/Lib/site-packages/darknet/data/1117/obj.data")
cap = cv2.VideoCapture(
"C:/Users/user/anaconda3/Lib/site-packages/darknet/data/1117/suwon_test.mp4"
)
print(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
print(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
while (cap.isOpened()):
ret, image = cap.read()
image = cv2.resize(image, dsize=(480, 640), interpolation=cv2.INTER_AREA)
print(image.shape)
if not ret:
break
frame = darknet.nparray_to_image(image)
r = darknet.detect_image(net, meta, frame)
boxes = []
for k in range(len(r)):
width = r[k][2][2]
height = r[k][2][3]
center_x = r[k][2][0]
center_y = r[k][2][1]
bottomLeft_x = center_x - (width / 2)
bottomLeft_y = center_y - (height / 2)
x, y, w, h = bottomLeft_x, bottomLeft_y, width, height
boxes.append((x, y, w, h))
for k in range(len(boxes)):