python - Python + OpenCV + Base64:将框架转换为base64的问题

Question

我正在尝试将视频转换为帧，并将那些帧转换为base64字符串。我无法这样做，并且遇到一些异常(exception)。下面是我的代码:

import cv2
import base64


def footage_to_frame(video):
    vidcap = cv2.VideoCapture(video)
    success, frames = vidcap.read()
    if success:
        return frames


def frame_to_base64(frames):
    with frames as frame:
        frame_b64 = base64.b64encode(frame.read())
    return frame_b64

该方法的函数调用为:

frames = converter.footage_to_frame("/Users/myname/Desktop/video.mp4")
converter.frame_to_base64(frames)

以下是我在控制台中遇到的错误:

File "/Users/myname/Desktop/Test/src/service/converter.py", line 13, in frame_to_base64
    with frames as frame:
AttributeError: __enter__

Answer 1

在函数frame_to_base64(frames)中，帧已经是单个图像，因为VideoCapture.read返回单个图像。这也是一个opencv镜像(numpy数组)，您不能在其中使用“with”。

def frame_to_base64(frame):
    return base64.b64encode(frame)

如果要阅读视频的所有帧，则应执行以下操作:

import cv2
import base64


def footage_to_frame(video):
    vidcap = cv2.VideoCapture(video)
    frames = []

    #  read until no more frames exist in the video
    while True:
        success, frame = vidcap.read()
        if (success):
            frames.append(frame)
        else:
            #  unable to read a frame
            break
 
    return frames


def frames_to_base64(frames):
    frames_b64 = []
    #  iterate frames and convert each of them to base64
    for frame in frames:
        frames_b64.append(base64.b64encode(frame))
    return frames_b64

尽管取决于视频时长，您可能会遇到内存问题。