Actually, the fix as introduced due to the 'bug' with ID 2620065 should be considered a bug - the previous implementation was correct. The new cvInitUndistortMap implies a change of the intrinsic camera matrix which cannot be desired. Neither in the bug report of 2620065 nor in the referenced forum discussion is given a proof of the claim that "after undistortion the principle point should be in...
07:51PM UTC on Apr 04 2009 in Open Computer Vision Library
