When deep learning makes data sets, it specifies how many frames to extract an image from the long video to the specified file path

Doing deep learning collection Data sets when , Record video first , Later, the video frames are extracted as data sets , It is a good way to prepare data sets efficiently , The code for saving a specified number of frames in a video frame to the specified file path is shown below :

#  Export the video as several frames of pictures 
DATA_DIR = "Video/1.avi"  #  Video storage path 
SAVE_DIR = "Images/photo"  #  Save path of frame pictures 
GAP = 30  #  Export a picture every how many frames 

import cv2  # OpenCV library 
import os

def getphoto(video_in, video_save):
    number = 0
    cap = cv2.VideoCapture(video_in)  #  Open video file 
    n_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))  #  The number of frames in the video 
    fps = cap.get(cv2.CAP_PROP_FPS)  #  The frame rate of the video 
    dur = n_frames / fps  #  Time of video 
    num_frame = 0
    judge = cap.isOpened()
    while judge:
        flag, frame = cap.read()  # flag Is the read state ,frame The next frame 
        if cv2.waitKey(0) == 27:
            break
        if flag:
            num_frame += 1
            if num_frame % GAP == 0:
                print(" Saving the %d A picture " % number)
                cv2.imwrite(video_save + '/' + str(number) + '.jpg', frame)  # cv2.imwrite(‘ route ’ + ‘ name ’ + ‘ suffix ’,  Frames to save )
                number += 1
        else:
            break

    print(" Video duration : %d  second " % dur)
    print(" Total frames of video : %d  The number of saved frames is : %d" % (n_frames, number))
    print(" frame per second (FPS): %.1lf" % fps)


def main_1(path):
    video_in = path
    video_save = SAVE_DIR
    getphoto(video_in, video_save)

if __name__ == '__main__':
    paht = DATA_DIR  #  Video path 
    main_1(paht)

The above is how to quickly extract the number of frames between videos and save a frame to the specified file path , I hope I can help you , thank you ！