时间戳

首页 > 解决方案 > 使用并发期货通过多处理运行多个视频

python - 使用并发期货通过多处理运行多个视频

问题描述

我正在尝试使用 concurrent.futres.ProcessPoolExecuter 运行代码，但在运行视频时遇到下面提到的while true错误class get_frames

concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.

下面提到的是我的实际代码：

import cv2
import numpy as np
import imutils
import os
import threading
import concurrent.futures
import time

start = time.perf_counter()

class VideoCapture:

    def __init__(self, name):
        self.cap = cv2.VideoCapture(name)
        self.q = queue.Queue()
        t = threading.Thread(target=self._reader)
        # t.get_ident()
        t.start()
        t.daemon = True

    # read frames as soon as they are available, keeping only most recent one
    def _reader(self):
        while True:
            ret, frame = self.cap.read()
            if not ret:
                break
            if not self.q.empty():
                try:
                    self.q.get_nowait()  # discard previous (unprocessed) frame
                except Queue.Empty:
                    pass
            self.q.put(frame)

    @classmethod
    def read(cls):
        return cls(self.q.get())


names = ['test.mp4', 'test1.mp4']

frames = [None] * len(names);
#frame_in = [None] * len(names);
ret = [None] * len(names);

cap_list = [cv2.VideoCapture(k) for k in names]


# print(cap_list)

class get_frames:

    def __init__(self, cap, name):
        self.cap = cap
        self.name = name
        # print(self.cap)

    def display_frames(self):
        while True:
            ret, frames = self.cap.read()
            if ret is True:
                count += 1
                print(self.name)


with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:
    for i, cap in enumerate(cap_list):
        g_f = get_frames(cap, i)
    results = executor.map(g_f.display_frames, cap_list)
    print(results)

对此的任何建议都会有很大帮助

标签： pythonpython-3.xopencvmultiprocessingprocess-pool

解决方案

您可以坚持使用ProcessPoolExecutor Example。

示例如下： 

def main():
    with concurrent.futures.ProcessPoolExecutor() as executor:
        for number, prime in zip(PRIMES, executor.map(is_prime, PRIMES)):
            print('%d is prime: %s' % (number, prime))

if __name__ == '__main__':
    main()
  • 您必须使用if __name__ == '__main__':语法。按照方法
    放置准备代码。 没有它，整个脚本会被执行多次，并且会卡住。 cap_list = [cv2.VideoCapture(k) for k in names]main()
  • 该示例使用方法，而不是对象。
    我想你可以映射对象方法，但我不能让它工作。
  • 使用 Python 3.6 时，出现错误："TypeError: can't pickle cv2.VideoCapture objects".
    我将文件名传递给display_frames，并在方法cv2.VideoCapture的开头执行display_frames。
  • 您发布的代码未使用class VideoCapture.

用于检测：

  • 我的示例生成合成视频文件（为了使示例“自包含”）。
  • 该示例显示了用于测试的视频帧。

这是代码： 

import cv2
import numpy as np
import concurrent.futures

def display_frames(name):
    cap = cv2.VideoCapture(name)

    while True:
        ret, frames = cap.read()

        if ret is False:
            break

        # Show frames for testing:
        cv2.imshow(str(cap), frames)
        cv2.waitKey(100)

    cap.release()

    return name


def main():
    names = ['test2.mp4', 'test3.mp4']

    # Generate two synthetic video files to be used as input:
    ###############################################################################
    width, height, n_frames = 640, 480, 30  # 30 frames, resolution 640x480

    intput_filename1 = names[0]
    intput_filename2 = names[1]

    # Use MPEG4 codec (for testing)
    synthetic_out = cv2.VideoWriter(intput_filename1, cv2.VideoWriter_fourcc(*'mp4v'), 25, (width, height))

    for i in range(n_frames):
        img = np.full((height, width, 3), 60, np.uint8)
        cv2.putText(img, str(i+1), (width//2-100*len(str(i+1)), height//2+100), cv2.FONT_HERSHEY_DUPLEX, 10, (30, 255, 30), 20)  # Green number
        synthetic_out.write(img)

    synthetic_out.release()

    width, height, n_frames = 320, 240, 20 # 20 frames, resolution 320x240
    synthetic_out = cv2.VideoWriter(intput_filename2, cv2.VideoWriter_fourcc(*'mp4v'), 25, (width, height))

    for i in range(n_frames):
        img = np.full((height, width, 3), 60, np.uint8)
        cv2.putText(img, str(i+1), (width//2-50*len(str(i+1)), height//2+50), cv2.FONT_HERSHEY_DUPLEX, 5, (255, 30, 30), 10)  # Blue number
        synthetic_out.write(img)

    synthetic_out.release()
    ###############################################################################


    with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:
        for name in executor.map(display_frames, names):
            print(name)

    cv2.destroyAllWindows() # For testing


# Using Python 3.6 there is an error: "TypeError: can't pickle cv2.VideoCapture objects"
if __name__ == '__main__':
    main()

我真的不知道它是否回答了您的问题，因为您的原始代码所剩无几。
我希望它有帮助...

推荐阅读