Python：如何使用ffmpeg读取视频（视频文件或rtsp）并使用OpenCV显示它们？

Question

I know that if we want to read video file or USB camera or rtsp stream, we can use cv2.VideoCapture .
But for some reason, I want to use ffmpeg to do the thing and process obtained images using opencv. So the pseudocode would be:

import cv2
import ffmpeg

videoFile = 'XXX.mp4'

video = ffmpeg.XXX(videoFile)

while True:
    frame = video.read() # this read() function is made up by me
    frameCV = ffmpeg2cv(frame) # this ffmpeg2cv() function is made up by me

    cv2.imshow('frame', frameCV)

I have searched a something like https://github.com/kkroening/ffmpeg-python . But I still not know how to do this. I think ffmpeg is too hard and there are few online docs unlike opencv. So please help.

Answer 1

This is what I found on the topic: https://www.bugcodemaster.com/article/extract-images-frame-frame-video-file-using-ffmpeg They propose a call extracting one image each second called: fps=1 Below they suggest how to extract each frame like: -vf select=eq(pict_type\\,I)

So I would suggest using subprocess and try something like this:

import subprocess

time = 200  # of your movie
frame_rate = 50
pathToFfmpeg = "SOME\\PATH"
pathToInput = "ANOTHER\\PATH"
pathToOuput = "YET\\ANOTHER\\PATH"

proc = subprocess.run([pathToFfmeg + "\\ffmpeg.exe",
                       "-framerate", str(frame_rate),
                       "-loglevel", "panic",  # Suppress Log
                       "-i", pathToInput + "\\SomeName.avi",

                      # Now copying their code
                      "-vf", "select=eq(pict_type\,I)", 
                      "-vsync", "vfr", pathToOuput + "\\Name.jpg",  
                      "-hide_banner"])  # Specify Output

And then you can read the outout images with opencv.

Note: You need all the "" for subprocess

Choose an Output name that contains a counter/iterator.