Question

I spend the whole morning today searching for a way to save an array of images in IplImage type in openCV and failed.

This is what I am trying to do :

IplImage* GetThresholdedImage(IplImage* img) {

IplImage* imageTest[2];

IplImage* imgHSV = cvCreateImage(cvGetSize(img), 8, 3); // hold the resulted HSV image

cvCvtColor(img, imgHSV, CV_BGR2HSV); // convert the coming image from the camera from RGB format to HSV (Hue, Saturation, Value)

imageTest[0] = cvCreateImage(cvGetSize(img), 8, 1); //hold the thresholded image of the yellow color

imageTest[1] = cvCreateImage(cvGetSize(img), 8, 1); //hold the thresholded image of the red color

cvSmooth(imgHSV, imgHSV, CV_GAUSSIAN, 11, 11); //smooth the image to remove the noise from the image

cvInRangeS(imgHSV, cvScalar(24, 100, 150), cvScalar(34, 255, 255),
        imageTest[0]); //this function filter out the colors in this range (This is a yellow color)

cvInRangeS(imgHSV, cvScalar(172, 100, 150), cvScalar(179, 255, 255),
        imageTest[1]); //this function filter out the colors in this range (This is a red color)


cvReleaseImage(&imgHSV);

return *imageTest;

}

Now when I try to return the array in the main in order to process it –>

IplImage *thresholdedImage;// = cvCreateImage(cvGetSize(frame), 8, 1); // to store the thresholded image

IplImage *yellow = cvCreateImage(cvGetSize(frame), 8, 1);
IplImage *red = cvCreateImage(cvGetSize(frame), 8, 1);


//===========================================

// start creating three windows to show the video after being thresholded, after it passes the contour function and the final video

cvNamedWindow("display", CV_WINDOW_AUTOSIZE);
cvNamedWindow("Threshold", CV_WINDOW_AUTOSIZE);
cvNamedWindow("contoured", CV_WINDOW_AUTOSIZE);

while (key != 'q') { // grab the video unless the user press q button

    frame = cvQueryFrame(capture);

    if (!frame) {
        break;
    }

    //start the actual video processing on real-time frames

    //first output of the threshold method
    thresholdedImage = GetThresholdedImage(frame);

    yellow = *thresholdedImage;

    red = *thresholdedImage++;

    //insert the resulted frame from the above function into the find contour function
    cvFindContours(yellow, storage, &contours, sizeof(CvContour),
            CV_RETR_EXTERNAL, CV_CHAIN_APPROX_SIMPLE, cvPoint(0, 0));

    cvFindContours(red, storage, &contours, sizeof(CvContour),
            CV_RETR_EXTERNAL, CV_CHAIN_APPROX_SIMPLE, cvPoint(0, 0));

however it gives me error !!!

any help is appreciated , thank you

Answer 1

An array of pointers to IplImage would be IplImage**. Also note, that you have to provide the buffer from the outside, as you can’t return a pointer to local (non-static) data.

void GetImages(IplImage** images, unsigned int num) {
    // assign pointers here
    images[0] = ...
    images[1] = ...
}

IplImage *images[2];
GetImage(images, 2);

As an alternative, you could as well create the array within your function using new or malloc and return the pointer to it. Just ensure you delete/free it later on.

IplImage **CreateImages(unsigned int num) {
    IplImage **images = new IplImage*[num];
    // assign again
    images[0] = ...
    return images;
}

Also ensure you free the images, once you’re done with them. So you shouldn’t use the approach you used above (increasing the pointer). Instead simply use array syntax to access n-th element.

Your code fails due to returning only a single pointer to the first image instead of the actual array.