Get floodfill outputs (connected pixels) copied into a new mat

Question

Is there a convenient way to create a new Mat from the outputs of a floodfill operation? I want to get a mat of only the pixels which are detected as connected to the seed pixel and are flood filled technically.

Guess that I executed floodFill method to a certain seed point and only 1/4 of the total pixels were filled as they were connected. I want to copy those pixels only to a new image, which represent only those 1/4 number of pixels and most probably smaller than the original input image.

I did this anyway via a very long,higher time+cpu consuming approach. In brief my approach was giving different colors for different floodfill calls and keep records of pixels of same color in a separate data structure, and etc.

I want to know if there is a direct and easier approach using the Mask created by floodfill or using any other approach.

Answer 1

It's not completely clear what you need exactly. Please take a look at this code, and check if croppedResult is what you want.

#include <opencv2\opencv.hpp>
using namespace cv;

int main()
{
    // Create a test image
    Mat1b img(300, 200, uchar(0));
    circle(img, Point(150, 200), 30, Scalar(255));
    rectangle(img, Rect(30, 50, 40, 20), Scalar(255));
    rectangle(img, Rect(100, 80, 30, 40), Scalar(255));

    // Seed inside the circle
    Point seed(160, 220);

    // Setting up a mask with correct dimensions
    Mat1b mask;
    copyMakeBorder(img, mask, 1, 1, 1, 1, BORDER_CONSTANT, Scalar(0));

    Rect roi;
    uchar seedColor = 200;
    floodFill(img, mask,
        seed + Point(1,1),  // Since the mask is larger than the filled image, a pixel (x,y) in image corresponds to the pixel (x+1,y+1) in the mask
        Scalar(0),          // If FLOODFILL_MASK_ONLY is set, the function does not change the image ( newVal is ignored),
        &roi,               // Minimum bounding rectangle of the repainted domain.
        Scalar(5),          // loDiff
        Scalar(5),          // upDiff 
        4 | (int(seedColor) << 8) | FLOODFILL_MASK_ONLY);
        // 4-connected | with defined seedColor | use only the mask 

    // B/W image, where white pixels are the one set to seedColor by floodFill
    Mat1b result = (mask == seedColor);

    // Cropped image
    roi += Point(1,1);
    Mat1b croppedResult = result(roi);

    return 0;
}

Test image img:

Mask mask after floodFill:

Mask result with only seedColor pixels:

Cropped mask croppedResult:

---

UPDATE

    // B/W image, where white pixels are the one set to seedColor by floodFill
    Mat1b resultMask = (mask == seedColor);
    Mat1b resultMaskWithoutBorder = resultMask(Rect(1,1,img.cols,img.rows));

    Mat3b originalImage;
    cvtColor(img, originalImage, COLOR_GRAY2BGR); // Probably your original image is already 3 channel

    Mat3b imgMasked(img.size(), Vec3b(0,0,0));
    originalImage.copyTo(imgMasked, resultMaskWithoutBorder);

    Mat3b croppedResult = imgMasked(roi);
    return 0;