Best practice how to display/receive IP network camera stream

Question

I have a IP network camera that can stream MJPEG, H.264 and others using TCP, UDP, RTSP and so on. Within my client application I need to access this stream to get still images (capture) or the full video stream itself.

Because of network load and latency (to get the most current image), I would prefer RTSP. So I tried the MediaElement from WPF but even with the help of many posts here on Stackoverflow I did not manage to get it running.

Any help how to achieve that or which other protocol should be used?

Answer 1

I found the solution in another thread/post here. You will see that I'm only displaying every 10th frame. This is because the frame are received so fast, that imshow(...) will display destroyed/warped images. If you supply only every 10th frame to imshow(...), the result will be okay. In case the IP camera delivers 1080p on 30fps, I had to show only every 30th or 40th frame.

#include "cv.h"
#include "highgui.h"
#include <iostream>

int main(int, char**) {
    cv::VideoCapture vcap;
    cv::Mat image;

    const std::string videoStreamAddress = "rtsp://cam_address:554/live.sdp"; 
    /* it may be an address of an mjpeg stream, 
    e.g. "http://user:pass@cam_address:8081/cgi/mjpg/mjpg.cgi?.mjpg" */

    //open the video stream and make sure it's opened
    if(!vcap.open(videoStreamAddress)) {
        std::cout << "Error opening video stream or file" << std::endl;
        return -1;
    }

    int counter = 0;
    for(;;) {
        counter++;

        if(!vcap.read(image)) {
            std::cout << "No frame" << std::endl;
            cv::waitKey();
        }

        // if the picture is too large, imshow will display warped images, so show only every 10th frame
        if (counter % 10 != 0)
            continue;

        cv::imshow("Output Window", image);
        if(cv::waitKey(1) >= 0) break;
    }   
}