stereoMatching's profile - activity

cv::Mat A(3, 3, CV_8UC1);
cv::Mat B;
B.create(A.size(), CV_32FC1);
std::cout<<A.type()<<", "<<B.type()<<std::endl;
A.copyTo(B);
std::cout<<A.type()<<", "<<B.type()<<std::endl;

I want to copy A to B without changing the type of B.

Any clean solution to do this without writing explicit loop?Thanks

ps : I don't want to convert B to CV_32FC1 after copied

I can use other openCV(imgproc, core)modules with Qt on android, but not the camera module of openCV.(Qt5.2.1 would allow us to capture the image buffer from QCamera).

As the graph show, the most easiest solution is return a bounding rect surrounded the contour, split the bounding rect to small rectangles, but it is much harder for irregular shape.What kind of algorithm would you suggest to split the contour?

I would like to implement an algorithm from "Super-Resolution from a Single Image" with the help of openCV.But I don't know I should pay the author patent fee or not, I remember that you have to pay money for some algorithms like z scan.

The website of the paper http://www.wisdom.weizmann.ac.il/~vision/SingleImageSR.html