Registration of 3D range images using particle swarm optimization

Abstract

Improvements to the application of a relatively new meta-heuristic strategy, called Particle Swarm Optimization (PSO), to the registration of 3D range images are presented. The proposed improvements include: implementation of the search intensification procedure, the von-Neumann particle neighborhood topology and the differential evolution operator. The 3D data registration results show significant improvements compared to the currently used PSO-based registration techniques, as well as techniques involving other meta-heuristic optimization strategies, such as genetic algorithms and adaptive simulated annealing. © Springer-Verlag Berlin Heidelberg 2004.