OpenCV(4.5.5) :-1: 错误: (-5:Bad argument) 在函数“cvtColor”中

Question

你们能帮我吗？我正在尝试使用我自己的数据（路径=数据）通过在我的视频上应用媒体管道来创建数据集。处理后的 vid (.ny) 将位于我之前声明的输出文件夹 (path = O_Video) 中。 seq 数量为 30，因为我有 30 个视频，start_folder = 0 且 start_video = 0。

videos = cv2.VideoCapture(IMPORT_DATA)
videos_input = cv2.cvtColor(videos, cv2.COLOR_BGR2RGB)

# Set mediapipe model 
with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:
    
    # NEW LOOP
    # Loop through actions
    for action in actions:
    # Loop through sequences aka videos
        for sequence in range(start_folder, start_video+no_sequences):

                # get results
            results = mp_face.process(videos_input)

            for detection in results.detections: # iterate over each detection and draw on image
                  mp_drawing.draw_detection(videos, detection)
                
                
                # NEW Export keypoints
            keypoints = extract_keypoints(results)
            npy_path = os.path.join(DATA_PATH, action, str(sequence))
            np.save(npy_path, keypoints)

                # Break gracefully
            if cv2.waitKey(10) & 0xFF == ord('q'):
                break
                    
    cap.release()
    cv2.destroyAllWindows()

我收到的错误如下：

Error                                     Traceback (most recent call last)
Input In [15], in <module>
      1 videos = cv2.VideoCapture(IMPORT_DATA)
----> 2 videos_input = cv2.cvtColor(videos, cv2.COLOR_BGR2RGB)
      4 # Set mediapipe model 
      5 with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:
      6     
      7     # NEW LOOP
      8     # Loop through actions

error: OpenCV(4.5.5) :-1: error: (-5:Bad argument) in function 'cvtColor'
> Overload resolution failed:
>  - src is not a numpy array, neither a scalar
>  - Expected Ptr<cv::UMat> for argument 'src'

Answer 1

您无法将 VideoCapture 对象传递给 cvtColor。

您必须单独传递每个帧（numpy 数组）。

vid = cv.VideoCapture(...)
assert vid.isOpened()

while True:
    (valid, frame) = vid.read()
    if not valid:
        break

    converted = cv.cvtColor(frame, ...)
    ...

Answer 2

您需要捕获帧数据，如下所示。

videos = cv2.VideoCapture(IMPORT_DATA)
# Capture the video by frame
ret, frame = videos.read()
# check ret for success and then do this
videos_input = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)