CODEWITHSUNDEEP
opencv
BBB
BBB

Reputation: 59

OpenCV split and merge program shows error

I have written the following program to split the image1 in three channels and then add the image2 to the blue channel of image1. I am using code blocks compiler and it's not showing any error but when I execute it the command prompt shows a runtime error and force closes my program. Can someone tell me what's wrong with my program? here is a link to the screenshot of the error I get https://dl.dropboxusercontent.com/u/13916799/Capture4.JPG

using namespace std;
using namespace cv;

void addImages(Mat &image1,Mat &image2,Mat &result)
{
    result.create(image1.size(),image1.type());

    vector<Mat> planes;

    split(image1,planes);

    planes[0] += image2;

    merge(planes,result);
}


int main()
{
    Mat image1 = imread("C:\\castle.jpg",CV_LOAD_IMAGE_UNCHANGED);
    Mat image2 = imread("C:\\rain.jpg",CV_LOAD_IMAGE_UNCHANGED);
    Mat result;

    addImages(image1,image2,result);

    namedWindow("vOut",CV_WINDOW_AUTOSIZE);
    imshow("vOut",result);

    waitKey(0);
    destroyAllWindows();
}

Upvotes: 0

Views: 560

Answers (2)

flankechen
flankechen

Reputation: 1255

you can't add a one channel image to a three channel image,

planes[0] += image2;

you could find your error message in arithm.cpp

 else if( !checkScalar(src2, src1.type(), kind2, kind1) )
 CV_Error( CV_StsUnmatchedSizes,
                 "The operation is neither 'array op array' (where arrays have the same size and the same number of channels), "
                 "nor 'array op scalar', nor 'scalar op array'" );

add the image2 to the blue channel of image1? I don't quite get it.

split image2 and add them with the same channel, type, size and so on...

Upvotes: 2

Mailerdaimon
Mailerdaimon

Reputation: 6090

The error states: Size of input argument do not match

To resolve this error check the following:

  • Size (cols and rows) of image1 and image2 have to be the same
  • image2 must be NxMx1 (1 Channel)

I would guess that the error lies in your imread call for image2. You read the image with the CV_LOAD_IMAGE_UNCHANGED this could load an image in rgb, bgr, rgba [...] format, all with more than one channel.

To read an image with as Grayscale (one channel, try: CV_LOAD_IMAGE_GRAYSCALE) Check that all your images (and planes) have the right dimension and your Code should work.

Upvotes: 1

Related Questions