The function cv::UndistortPoints(), applies reverse lens distortion to a set of observed point coordinates. The models for lens distortion available in openCV are not invertible, which means an approximation can be made and indeed, from the documentation:
...undistort() is an approximate iterative algorithm that estimates the normalized original point coordinates out of the normalized distorted point coordinates (“normalized” means that the coordinates do not depend on the camera matrix).
My question is therefore: where can I find information on the approximation method used in undistortPoints? What are it's characteristics? How was it derived? Under what conditions is it likely to succeed/fail?
Handling lens distortion well is integral to many 3D Reconstruction applications so it would really be helpful with some clarity here.