How to enable Multiple Windows in Qt using OpenCV?

Question

I'm developing an application which requires inspection of certain industrial process using a live video feed from a single camera. I want to know if it's possible to create multiple windows using only one camera. Also, each window will have a different Region Of Interest . I'm running this code in QThread
Here is the code I've been trying but it crashes as soon as I run it.

mythread.cpp

#include "mythread.h"
#include "mainwindow.h"
#include <opencv2/opencv.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/imgproc/imgproc.hpp>

using namespace cv;
using namespace std;

char key;

MyThread::MyThread(QObject *parent, bool b) : QThread(parent), Stop(b)
{
}

// run() will be called when a thread starts
void MyThread::run()
{
    cvNamedWindow("Camera_Output", 1);    //Create window
    cvNamedWindow("Camera_Output1", 1);

        CvCapture* capture = cvCaptureFromCAM(0);  //Capture using camera 0 connected to system
        cvSetCaptureProperty( capture, CV_CAP_PROP_FRAME_WIDTH, 640 );
        cvSetCaptureProperty( capture, CV_CAP_PROP_FRAME_HEIGHT, 480 );

        CvCapture* capture1 = cvCaptureFromCAM(0);  //Capture using camera 0 connected to system
        cvSetCaptureProperty( capture1, CV_CAP_PROP_FRAME_WIDTH, 640 );
        cvSetCaptureProperty( capture1, CV_CAP_PROP_FRAME_HEIGHT, 480 );

     //Create loop for live streaming
     while(1){  

            IplImage* framein = cvQueryFrame(capture); //Create image frames from capture
            IplImage* framein1 = cvQueryFrame(capture1);

            /* sets the Region of Interest  - rectangle area has to be __INSIDE__ the image */
            cvSetImageROI(framein, cvRect(200, 200, 320, 240));
            cvSetImageROI(framein1, cvRect(500, 500, 320, 240));


            /* create destination image  - cvGetSize will return the width and the height of ROI */
            IplImage *frameout = cvCreateImage(cvGetSize(framein),  framein->depth, framein->nChannels);
            IplImage *frameout1 = cvCreateImage(cvGetSize(framein1),  framein1->depth, framein1->nChannels);

            /* copy subimage */
            cvCopy(framein, frameout, NULL);
            cvCopy(framein1, frameout1, NULL);

            /* always reset the Region of Interest */
            cvResetImageROI(framein);
            cvResetImageROI(framein1);

            cvShowImage("Camera_Output", frameout);   //Show image frames on created window
            cvShowImage("Camera_Output1", frameout1);

            key = cvWaitKey(10);     //Capture Keyboard stroke
            if (char(key) == 27){
                break;      //ESC key loop will break.
            }
        }
   }

Can anyone tell me where am I going wrong?

Answer 1

You can absolutely do that. Simply create multiple view and link the output of the camera (via some update-signal) to a display functionality in the window you want it to be (via some update-and-render-slot). So, conceptually:

• Create multiple views
• Set region of interests in each view
• Connect camera output to each view
• If a new image arrives, notify the view. Each view will extract the region-of-interest and render it.

Each view runs its own event loop then, so that should be fine. A general note: AFAIK it is considered bad practice to subclass QThread or do similar things. I would prefer the worker approach here.