Towards automatic quality assessment of range images

In this paper we present our approach to automatic assessment of the quality of acquired range images. To this end, a robust method for image registration is presented based on motion constraints defined by geometric properties of reflected correspondence vectors. By combining the iterative closest point (ICP) method with rigid motion constraints it is demonstrated that false matches can be effectively eliminated and that the method leads to accurate and robust image registration whose error analysis indicates the quality of the acquired images.