How to copy grayscale image to RGB images red channel in OpenCV?

Question

I have two input images, which are greyscale and I am creating one more image, which is RGB and should contain in the red channel one of the grey image and in the green the other.

Mat img, img2;
img =imread("above.jpg", CV_LOAD_IMAGE_GRAYSCALE);
img2 =imread("left.jpg", CV_LOAD_IMAGE_GRAYSCALE);

Mat * aboveLeft = new Mat(img.rows, img.cols, CV_LOAD_IMAGE_COLOR);

int from_to [] = {0,1};
cv::mixChannels(&img, 1, aboveLeft, 3, from_to, 1);

I've tried to use the mixChannels command, to write img input into aboveLeft's first channel, but I couldn't figure out the right syntax and get assertion errors.

Is this the right way to fill in a channel?

Answer 1

I assume img and img2 has same size and type.

cv::Mat img = cv::imread("above.jpg", CV_LOAD_IMAGE_GRAYSCALE);
cv::Mat img2 = cv::imread("left.jpg", CV_LOAD_IMAGE_GRAYSCALE);

std::vector<cv::Mat> images(3);
Mat black = Mat::zeros(img.rows, img.cols, img.type());
images.at(0) = black; //for blue channel
images.at(1) = img;   //for green channel
images.at(2) = img2;  //for red channel

cv::Mat colorImage;
cv::merge(images, colorImage);

Answer 2

Assuming that you have 3 grayscale matrices, you can use cv::merge():

cv::Mat blue, green, red;
std::vector<cv::Mat> images(3);

// OpenCV works natively with BGR ordering
images.at(0) = blue;
images.at(1) = green;
images.at(2) = red;

cv::Mat color;
cv::merge(images, color);

As pointed out in the comments, you need to find a way to figure out what you want for the blue channel!