正确循环迭代，但对于视频帧，它仅适用于前四个索引

Question

count = 1
for i in real_videonames_index:
  videofile = filed[i]
  success = True
  vidcap = cv2.VideoCapture(videofile)
  while success:
    if (count%one_frame_each == 0):
      success,image = vidcap.read()
      image_gray = rgb2gray(image)
      if image.shape[1]>640:
        tmp = resize(image_gray,(math.floor(640 / image_gray.shape[1] * image_gray.shape[0]), 640),mode='constant')
      name = '/content/drive/MyDrive/Training/REAL/' + str(count) + '.jpg'
      print('Creating ....' + name)
      cv2.imwrite(name, image)
      print ('*', end="")
    else:
      success,image = vidcap.read()                                 
    count += 1  
  print('/n/n/n/n{} video completed successfully/n/n/n'.format(i))
  i += 1

在循环的简单迭代中工作正常，但在提取帧时它仅适用于前三个索引？

下面是第一个视频帧捕获后的错误。我正在使用 google colab。提前致谢。

Answer 1

实际上错误是在最后一帧 cv2 中得到了 None 值，所以它可以通过再次获取帧来解决。正确的代码是。

count = 0
face_cascade = cv2.CascadeClassifier('/content/drive/MyDrive/Training/haarcascade_frontalface_default.xml')
for i in real_videonames_index[51:76]:
  videofile = videonames_list[i]
  vidcap = cv2.VideoCapture(videofile)
  success,image = vidcap.read()
  while success:
    if (count%one_frame_each == 0):
      name = '/content/drive/MyDrive/Training/Validation Data/REAL/'+names[i]+ "_" + str(count) + '.jpg'
      image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
      facesBase = face_cascade.detectMultiScale(image, scaleFactor=1.1, minNeighbors=5)
      for f in facesBase:
        x, y, w, h = [ v for v in f ]
        cv2.rectangle(image, (x,y), (x+w, y+h), (255,0,0), 3)
        face_crop = image[y:y+h, x:x+w]
        cv2.imwrite(name, face_crop)
      success,image = vidcap.read()
      print('Read a new frame: {} and with name {}'.format(success,count))
    else:      
      success,image = vidcap.read()
      print('Read a new frame: {} and with name {}'.format(success,count))
    count += 1
  print('{} Real Video Extracted Successfully'.format(i))
  i += 1

正确循环迭代，但对于视频帧，它仅适用于前四个索引

Loop Iterate correctly but with video frames it work only for first four indexes

opencv

machine-learning

google-colaboratory

video-capture