Using rigid motion constraints for the registration of free-form surfaces.

Abstract

We present a new method for registration of freeform surfaces based on the iterative closest point (ICP) method and on geometric properties of reflected correspondence vectors. The method is based on computing relative gaps between reflected correspondences and between the projections of reflected correspondences along the rotation axis and using these to eliminate false matches. Experimental results on synthetic data and on real range images demonstrate that the method is robust and accurate for image registration with small motions.