Use OpenCV to calculate cross-correlation of two arrays

Question

I am using OpenCV to calculate the cross-correlation of two rows taken from some matrix X (16x128 dimension). This is what I have done:

cv::Mat channel1(1,128,CV_32F, X.row(1).clone());     //take the first row of matrix X
cv::Mat channel2(1,128,CV_32F, X.row(10).clone());    //take the 10th row of matrix X
cv::Mat result;
cvMatchTemplate(channel1,channel2,result, CV_TM_CCORR);

However, I am getting the following error: "Error: no instance of constructor "cv::Mat::Mat" matches the argument list". Could anyone please help me fiz this issue. Thank you in advance.

Answer 1

There are a couple of problems with the code you posted:

1) There is no instance of a cv::Mat constructor that matches the way you are initializing the "channel1" and "channel2" matrices.

Solution:

• Declare your matrices in this way:

  cv::Mat channel1(X.row(1).clone());     //take the first row of matrix X
  cv::Mat channel2(X.row(10).clone());    //take the 10th row of matrix X
  

2) cvMatchTemplate is the old-style C-api, and (afaik) is not compatible with the C++ oriented cv::Mat.

Solution:

• (as berak suggested) use cv::matchTemplate instead.

3) Better yet, you could use matchTemplate directly:

    cvMatchTemplate(X.row(1).clone(),X.row(10).clone(),result, CV_TM_CCORR);

Hope it helps.