Edge detection in openCv gives runtime error

I have used cvCanny function to detect Edges.

cvCanny( img_b, out, lowThresh*N开发者_开发知识库*N, highThresh*N*N, aperature_size ); 

But in run time it gives runtime error. The error msg not clear at all. It refers some memory location. Please help me..!!

code:

void switch_callback_h( int position ){
 highInt = position;
}
void switch_callback_l( int position ){
 lowInt = position;
}

int _tmain(int argc, _TCHAR* argv[])
{

 const char* name = "Edge Detection Window";
 // Kernel size
 int N = 7;
CvCapture* capture = cvCaptureFromCAM(1);
IplImage* frame;

while(1) {
frame = cvQueryFrame( capture );

// Add convolution boarders
 CvPoint offset = cvPoint((N-1)/2,(N-1)/2);
 cvCopyMakeBorder(frame, img_b, offset, IPL_BORDER_REPLICATE, cvScalarAll(0));

 // Make window
 cvNamedWindow( name, 1 );

 // Edge Detection Variables
 int aperature_size = N;
 double lowThresh = 20;
 double highThresh = 40;

 // Create trackbars
 cvCreateTrackbar( "High", name, &high_switch_value, 4, switch_callback_h );
 cvCreateTrackbar( "Low", name, &low_switch_value, 4, switch_callback_l );
 highThresh = 800;
        lowThresh = 100;

     cvCanny( img_b, out, lowThresh*N*N, highThresh*N*N, aperature_size );  

        cvShowImage(name, out);
 cvReleaseImage( &frame );
 cvReleaseImage( &img_b );
 cvReleaseImage( &out );
 cvDestroyWindow( name );

   if( cvWaitKey( 15 ) == 27 ) 
 break;

  return 0;
}

You were very close to making this work. Basically, here are your problems:

• You're creating a window inside the loop. Since the window has the same name every time, you only to create it once.
• You're destroying your images before you're able to show them. Your images will be shown not when you called cvShowImage, but when you call cvWaitKey. By that time you've already freed the image, so nothing will be shown.
• You're freeing frames loaded from CvCapture. The documentation explicitly says not to do this:

The function cvQueryFrame grabs a frame from a camera or video file, decompresses it and returns it. This function is just a combination of GrabFrame and RetrieveFrame , but in one call. The returned image should not be released or modified by the user. In the event of an error, the return value may be NULL.

• You're not initializing image_b or out anywhere. They're not even declared in the code you posted. I don't know how your code even compiled, let alone ran.
• You're specifying image_b as the source to the canny edge detection, when it should really be frame
• You're not freeing the video capture struct

Like I said, a number of tiny points. Here's code that works:

#include <stdlib.h>
#include <cv.h>
#include <highgui.h>
//
// Comment this out to use the webcam.
//
#define LOAD_IMAGE "/home/misha/Desktop/Lenna.png"
int main(int argc, char **argv)
{
    const char *name = "Edge Detection Window";
    int N = 7;
#ifndef LOAD_IMAGE
    CvCapture *capture = cvCaptureFromCAM(1);
#endif
    IplImage *frame = NULL;
    IplImage *out = NULL;

    cvNamedWindow(name, 1);

    while (1) 
    {
#ifdef LOAD_IMAGE
        frame = cvLoadImage(LOAD_IMAGE, 0);
#else
        frame = cvQueryFrame(capture);
#endif
        out = cvCreateImage(cvGetSize(frame), frame->depth, frame->nChannels);

        int aperature_size = N;
        double lowThresh = 20;
        double highThresh = 40;

        highThresh = 800;
        lowThresh = 100;

        cvCanny(frame, out, lowThresh * N * N, highThresh * N * N,
                aperature_size);

        cvShowImage(name, out);
#ifndef LOAD_IMAGE
        if (cvWaitKey(15) == 27)
            break;
#else
        cvWaitKey(0);
        break;
#endif
        }
    cvReleaseImage(&out);
    cvDestroyWindow(name);
#ifdef LOAD_IMAGE
    cvReleaseImage(&frame);
#else
    cvReleaseCapture(&capture);
#endif
    return 0;
}

Compiles with:

gcc -ggdb -Wall -o devan.out devan.c `pkg-config --cflags --libs opencv`

Standard image and output: