Opencv3.4.0实现视频中的帧保存为图片功能

Z时代
2024-01-10
分类：IT

现在有一个收集人脸的需求，每一张照片只有一个人。我的解决办法是用收集录制一段视频，然后上传到PC上。在PC上使用OpenCV将图片中的每一帧保存为JPG图片。

以下是代码：


import time
import cv2
if __name__ == '__main__':
 # 填写视频的绝对路径
 vidcap = cv2.VideoCapture('/home/shushi/video/猎场.mp4')
 success, image = vidcap.read()
 start_time = time.time()
 print(start_time)
 while success:
  end_time = time.time()
  file_name = str(end_time).replace('.', '')
  # 每隔三秒截屏
  if 3 == int(end_time - start_time):
   start_time = end_time
   # 保存JGP 的绝对路径
   cv2.imwrite('/home/shushi/video/' + file_name + ".jpg", image) # save frame as JPEG file
  success, image = vidcap.read()
  if cv2.waitKey(10) == 27: # exit if Escape is hit
   break

将视频放到与Python文件同级目录下，然后运行程序，你就会得到视频帧对应的JPG图片了：

小编再为大家分享一段代码：OpenCV按视频帧进行截取，作者是Haku_yyf


#include <opencv2/highgui/highgui.hpp>
#include<opencv2/imgproc/imgproc.hpp>
#include <iostream>
int main(int argc, char **argv)
{
 //打开视频文件
 cv::VideoCapture cap("E:\\New folder\\DSC_3543.MOV");
 if (!cap.isOpened())
 {
 std::cout << "不能打开视频文件" << std::endl;
 return -1;
 }
 //从3000ms开始播放视频 
 //cap.set(CV_CAP_PROP_POS_MSEC, 3000);
 //获取视频的帧速率
 double fps = cap.get(cv::CAP_PROP_FPS);
 double width = cap.get(cv::CAP_PROP_FRAME_WIDTH);
 double height = cap.get(cv::CAP_PROP_FRAME_HEIGHT);
 //std::cout << fps << std::endl;
 //std::cout << width << std::endl;
 //std::cout << height << std::endl;
 cv::VideoWriter wrt("C:\\Users\\Administrator\\Desktop\\1.avi", CV_FOURCC('M', 'J', 'P', 'G'), fps, cv::Size(width, height));
 int cout = 0;
 while (true)
 {
 cv::Mat frame;
 //从视频中读取一个帧
 bool bSuccess = cap.read(frame);
 ++cout;
 if (!bSuccess)
 {
 std::cout << "不能从视频文件读取帧" << std::endl;
 break;
 }
 //在MyVideo窗口上显示当前帧
 if (cout>=fps*1 && cout<=fps*56)//取视频1-56秒内容
 {
 std::cout << "正在截取..."<<(int)(((cout-fps)/(fps*(56-1)))*100)<<"%" << std::endl;
 wrt << frame;
 }
 if (cout>fps*57)
 {
 std::cout << "ENd!!!" << std::endl;
 break;
 }
 }
 system("pause");
 return 0;
}