python - python 3.7 有没有办法手动清除内存？

Question

我正在尝试制作一个程序来删除视频中的背景，以便仅显示前景。它适用于短视频，但在较长的视频上，速度会呈指数级下降，并开始使用越来越多的内存，从而导致程序崩溃。

我尝试删除 for 循环中使用的变量，如下面的代码所示。

def video_generator(name_of_video, video, cap, number_of_frames):
    check, frame = cap.read()
    A = np.array([])
    a = np.array([])

    if cap.isOpened():
        width = int(cap.get(3))
        height = int(cap.get(4))

    dims = (height, width)

    while cap.isOpened():
        check, frame = cap.read()

        if check:
            gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY).flatten()
            A = np.append(A, gray)
            a = np.append(a, 'row')
            image = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) 
        else:
            break   

    B = np.resize(A, (number_of_frames, width * height))

    U, S, V = decomposition.randomized_svd(B, 2)
    background = U @ np.diag(S) @ V

    foreground = ([])

    def generate_frames(B, background):
        for i in range(number_of_frames):
            foreground = np.reshape(B[i] - background[i], dims)
            frame = plt.imshow(foreground, cmap = 'gray')
            plt.savefig('D:/Python_Programs/BR_SVD/Frames/' + name_of_video + '/file_' + str(i + x * number_of_frames).zfill(2) + '.png')   

            del frame
            del foreground
            gc.collect()

    generate_frames(B, background)  

if number_of_frames_initial <= 200:
    cap = cv2.VideoCapture(video)
    video_generator(name_of_video, video_initial, cap_initial, number_of_frames_initial)

else:
    number_of_videos = int(number_of_frames_initial/200 + 1)
    time_of_videos = duration/number_of_videos

    for x in range (number_of_videos):

        subprocess.call ([
            'ffmpeg', '-i', video_initial, '-ss', str(time_of_videos * x), '-t', str(time_of_videos), 'D:/Python_Programs/BR_SVD/Temp_Videos/video_%02d.mp4' % x])

        video = 'D:/Python_Programs/BR_SVD/Temp_Videos/video_%02d.mp4' % x
        cap = cv2.VideoCapture(video)
        number_of_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
        video_generator(name_of_video, video, cap, number_of_frames)

        del B
        del background
        gc.collect()

Answer 1

• http://www.digi.com/wiki/developer/index.php/Python_Garbage_Collection
• http://docs.python.org/library/gc.html
• http://arctrix.com/nas/python/gc/

没有关于 Python 如何进行垃圾回收的明确资源(除了源代码本身)，但这 3 个链接应该会给您一个很好的主意。

更新

该来源实际上非常有帮助。你能从中得到多少取决于你对 C 语言的理解程度，但这些评论实际上非常有帮助。跳到collect()函数，注释很好地解释了这个过程(尽管是非常技术性的术语)。