Hi dear developers. I found a way to get all camera parameters and rectangle transform and translate from one camera and just one frame with one click. No matrix equations, no quaternion or eulers. It is simple. It is fast. My theoreme: If on the projection of a rectangle we can see the corners and its area is greater than zero, then this rectangle has a representation in space and, moreover, only one. https://www.youtube.com/watch?v=3rX_MQt0CGI Sorry openCV developers. As it turned out, everything is much simpler.
This forum is disabled, please visit https://forum.opencv.org