Enhanced, robust genetic algorithms for multiview range image registration

Abstract

We present a new method for precise registration of multiple range images with low overlap based on genetic algorithms (GAs). The proposed method minimizes the alignment error within the common overlap area among a set of views, which is computed by a novel robust evaluation metric, called the surface interpenetration measure. Because they search in a space of transformations, GAs are capable of registering surfaces without need for prealignment, as opposed to methods based on the iterative closest point (ICP) algorithm, the most popular to date. The experimental results confirm that the new method ensures more precise alignments than combined sequential pairwise alignments for multiview registration, providing accurate global alignment among overlapping views.