Hi there, I have a question concerning the matching of ORB features.
Since the ORB features are located on different levels of a pyramid, it’s possible that some feature point on the m-th level of the first image pyramid is matched with a point on the n-th level of the second image pyramid. Since different level have different image size, those two points have different coordinates origin. Then it doesn’t make sense to call the function cv::drawMatch() to visualize the matching, does it? Nor does it make sense to call cv::drawKeyPoints() to draw all the feature points in different levels on the raw(0th) image?