An improved artificial bee colony algorithm for range image registration

Abstract

Range image registration is an attractive topic in image processing field. It aims at finding an optimal transformation or correspondence between images captured from different views. Iterative closest point (ICP) algorithm is the most well-known method for registration. However, it needs a pre-alignment typically provided by the user. To overcome this drawback of ICP algorithm, many intelligent algorithms have been introduced to solve the registration problem. In this paper, we present an improved artificial bee colony (ABC) algorithm for range image registration. Inspired by particle swarm optimization algorithm, a new solution-updating strategy is proposed and introduced into the ABC. Consequently, the search ability of ABC algorithm is obviously enhanced while more precise results are obtained. Dozens of experiments are conducted to compare the performance of the proposed algorithm and other registration methods. The results demonstrate that the improved algorithm outperforms other algorithms.