Btukfyl

I'm new opencv c++ developer. I applied the SLIC superpixel so I have Mat:labels contain the ID of each pixel, I applied SPLIT on the image to get the Green channel and then I'm looking to change the value of specific pixels who have the ID=150 ( each pixel =! of 150 must is converted to black pixel). after debug I got this bug and this window:

Exception thrown at 0x00007FFB63A0A388 : Microsoft C++ exception: cv::Exception at memory location 0x0000001496AFE290.
Unhandled exception at 0x00007FFB63A0A388 : Microsoft C++ exception: cv::Exception at memory location 0x0000001496AFE290.

Here's my code

   int main()

{



    Mat labels;

    Mat RGB[3], HSV[3],S_HSV,image3;

    vector<double> STD, generale_mean;

    Mat image, Red, Green, Blue, image2,mask,Gray;

    int region_number;

    vector<vector<Point> > contours;

    vector<Vec4i> hierarchy;



    image = imread("D:\mémoire\eye.png", CV_LOAD_IMAGE_COLOR);



    //applying the superpixel 



    Ptr<cv::ximgproc::SuperpixelSLIC> slic = cv::ximgproc::createSuperpixelSLIC(image, 100,50, 0.01);

    slic->iterate(10);

    if (true)

        slic->enforceLabelConnectivity(50);



    slic->getLabelContourMask(mask, true);

    image.setTo((0, 0, 255), mask);

    slic->getLabels(labels);

    region_number = slic->getNumberOfSuperpixels();



    cv::cvtColor(image, image, CV_BGR2RGB);



    cv::cvtColor(image, Gray,CV_RGB2GRAY);



    split(image, RGB);



    Red.push_back(RGB[0]);

    Green.push_back(RGB[1]);

    Blue.push_back(RGB[2]);



        int f = 150;

        Mat Green1 = Green.clone();

        int Y =0;



        for (int j = 0; j < Green1.rows; j++) {

            for (int i = 0; i < Green1.cols; i++) {



                double N = labels.at<double>(Point(i, j));

                if (N =! f) 



                    Green1.at<double>(Point(i, j)) = 00000;



            }

        }



        cv::imshow("green", Green1);        



    waitKey(0);         // Wait for a keystroke in the window

    return 0;

}

You have problems with the types of the image. When you load an image in OpenCV using:

image = imread(filename, CV_LOAD_IMAGE_COLOR);

it will be an image of type CV8UC3. This means, that it will have 3 channels (the C3 part) and that each pixel value for each channel is of type unsigned int of 8 bits. In C++ you can use uchar for it. Then when you use the at function, you need to use a type that can hold the information correctly... that is in this case uchar. In you code you have:

Green1.at<double>(Point(i, j)) = 00000;

which should be:

Green1.at<uchar>(j,i) = 0;

Note that also Point(i,j) is not needed and can be replaced by j,i.

The .at<type>(row, column) function is more or less equivalent to do:

Mat test;

type* data = reinterpret_cast<type*>(test.data);

data[row][column];

You also have a problem in:

double N = labels.at<double>(Point(i, j));

but labels it is assign by

slic->getLabels(labels);

which gives a mat of type CV_32SC1 -> that means, 1 Channel and 32 bits signed integer, or in C++, int.

Then this should be :

int N = labels.at<int>(j,i);

Finally:

this

Red.push_back(RGB[0]);

Green.push_back(RGB[1]);

Blue.push_back(RGB[2]);

should be:

Red = RGB[0];

Green = RGB[1];

Blue = RGB[2];