I am writing a C++ Mex source file to perform matchingFeatures operation after extracting corner feature points. Please help me to convert index of matching features of 2 images (in vector<DMatch>) to mxArray .

Here is my code-

    #include "opencvmex.hpp"
    using namespace cv;
    void checkforInputs(int nrhs,const mxArray *prhs[])
    {
        if (nrhs != 2)
        {
            mexErrMsgTxt("Incorrect number of inputs. Function expects 2 inputs.");
        }

        if (!(mxIsUint8(prhs[0])&&mxIsUint8(prhs[0])))
        {       
            mexErrMsgTxt("Input image must be uint8.");
        }
    }
    void mexFunction(int nlhs, mxArray *plhs[], int nrhs, const mxArray *prhs[])
    {   
        checkforInputs(nrhs, prhs);// check wheteher inputs are correct.

        cv::Ptr<cv::Mat> description1 = ocvMxArrayToImage_uint8(prhs[0], true);
        cv::Ptr<cv::Mat> description2 = ocvMxArrayToImage_uint8(prhs[1], true);

        Ptr<DescriptorMatcher> matcher= cv::DescriptorMatcher::create("BruteForce-Hamming");// pointer to DescriptorMatcher object 

        std::vector<DMatch> matches;

        matcher->match(*description2, *description1, matches); 
// how to get indexes of matches in 2 images get converted to mxArray 

}

[James Tursa]

I am not familiar with OpenCV DMatch. Can you get at the field data directly? I.e., can you get at the distance, imgIdx, queryIdx, and trainIdx data directly?