"Access violation" when assigning matrix values

[karnavor]

Hello,

I am writing a program to estimate the fundamental matrix between two images, using OpenCV. This is a computer vision technique, but the error I am getting is something to do with memory in the program itself. I have copied the code exactly from this reliable source: http://books.google.co.uk/books?id=s...xample&f=false

Here is the code below:

int main()
{
	int point_count = 100;
	CvMat* points1;
	CvMat* points2;
	CvMat* status;
	CvMat* fundamental_matrix;

	points1 = cvCreateMat(1,point_count,CV_32FC2);
	points2 = cvCreateMat(1,point_count,CV_32FC2);
	status = cvCreateMat(1,point_count,CV_8UC1);

	/* Fill the points here … */
	for(int i = 0; i < point_count; i++ )
	{
	    points1->data.db[i*2] = 2;
	    points1->data.db[i*2+1] = 4;
	    points2->data.db[i*2] = 5;
	    points2->data.db[i*2+1] = 3;
	}

	fundamental_matrix = cvCreateMat(3,3,CV_32FC1);
	int fm_count = cvFindFundamentalMat( 
	points1,points2,fundamental_matrix,
	CV_FM_RANSAC,1.0,0.99,status );

        return 0;
}

However, when this is run, I get the following error:

"Unhandled exception at 0x0040122b in TestProgram.exe: 0xC0000005: Access violation writing location 0x0037d000."

This occurs when i = 54, at the line "points2->data.db[i*2] = 5;"

It seems to be suggesting that I don't have enough memory to make this assignment, but I have reserved enough. My "points2" matrix contains 100 elements, each of type CV_32FC2 (which represent two floating points), and thus the matrix should be able to hold 200 floating points.

Furthermore, this code is copied exactly from the source, which is a reliable text book. So, the code should run, and I am very confused as to why I am getting this error.

Thanks for any help

[Lindley]

You are using the .db member of the union, which assumes that the underlying array is of doubles. However, CV_32FC2 assumes that each element is a 32-bit float----not a 64-bit double.

[karnavor]

Hi,

Thanks - I managed to fix it by using cvmSet(), which seems to automatically convert to a float.

However, I am now having an issue with the "status" matrix. I want to print out the elements of the matrix, and so I use "cout << cvmGet(status, 0, i);", but it gives me the following runtime error:

"Debug Error! Program: ........MyProgram.exe The application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information."

Here is the code:

int main()
{
	CvMat* points1;
	CvMat* points2;
	CvMat* status;
	CvMat* fundamental_matrix;

	points1 = cvCreateMat(2, 8, CV_32FC1);
	points2 = cvCreateMat(2, 8, CV_32FC1);
	status = cvCreateMat(1, 8, CV_8UC1);

	for (int i = 0; i < 8; i ++)
	{
		cvmSet(points1, 0, i, i);
		cvmSet(points1, 1, i, i);
		cvmSet(points2, 0, i, i);
		cvmSet(points2, 1, i, i);
	}

	fundamental_matrix = cvCreateMat(3, 3, CV_32FC1);
	int fm_count = cvFindFundamentalMat(points1, points2, fundamental_matrix, CV_FM_RANSAC, 1.0, 0.99, status);

	cout << endl << endl << endl;
	for (int i = 0; i < 8; i ++)
	{
		cout << cvmGet(status, 0, i);
	}

       return 0;
}

Any ideas??

Thanks

[Paul McKenzie]

Hi,

Thanks - I managed to fix it by using cvmSet(), which seems to automatically convert to a float.

However, I am now having an issue with the "status" matrix. I want to print out the elements of the matrix, and so I use "cout << cvmGet(status, 0, i);", but it gives me the following runtime error:

"Debug Error! Program: ........MyProgram.exe The application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information."

Use divide and conquer.

Remove these lines:

fundamental_matrix = cvCreateMat(3, 3, CV_32FC1);
	int fm_count = cvFindFundamentalMat(points1, points2, fundamental_matrix, CV_FM_RANSAC, 1.0, 0.99, status);

Now what happens? If the error still occurs, then its something very basic, otherwise it's that function that was removed is the cause of the problem.

Regards,

Paul McKenzie