How to correctly publish and receive cv::Mat using cv_bridge? (C++)

edit

Hey everyone,

I am having troubles converting a cv::Mat to a sensor_msgs/Image, publishing it and doing the reverse operation to regain the original cv::Mat. The imshow() function shows me the correct image(white-gray-black image with numbers ranging from 0 to 1) before sending, but shows a seemingly random and black-white(binary) image after receiving it. Normalizing the matrix with cv::normalize() before displaying it does not do the trick either.

So I would be happy if you could check out my code below or give me a working sample code from one of your projects :). There is the fear that the mistake lies outside the shown code, but I don't think you want to read all my nodes.

Extract publisher code:

cv::Mat out_mat;
// fill mat....
std::cout << "out_mat.type() = " << out_mat.type() << "\n"; // prints out_mat.type() = 5
cv::namedWindow("out_mat", CV_WINDOW_NORMAL);
cv::resizeWindow("out_mat", 500, 500);
cv::imshow("out_mat", out_mat); // shows correct image
cv::waitKey(1);
cv_bridge::CvImage cvi_mat;
cvi_mat.encoding = sensor_msgs::image_encodings::MONO8;
cvi_mat.image = out_mat;
cvi_mat.toImageMsg(container); // container of type sensor_msgs/Image
pub_peaks_.publish(container);

Extract subscriber code:

// convert image msg to cv mat
cv_bridge::CvImagePtr img;
img = cv_bridge::toCvCopy(container, sensor_msgs::image_encodings::MONO8); // container of type sensor_msgs/Image
cv::Mat mat_received = img->image;
cv::Mat mat;
mat_received.convertTo(mat,5);
std::cout << "mat type: " << mat.type() << "\n"; // prints mat_c type: 5
cv::namedWindow("Grid", CV_WINDOW_NORMAL);
cv::resizeWindow("Grid", 500, 500);
cv::imshow("Grid", mat); // shows false/random (binary black-white) image
cv::waitKey(1);

Every help is appreciated.