Improved particle swarm optimization for fast block matching with application to motion estimation in micro/nano systems

Abstract

The function of motion estimation (ME) is crucial in micro/nano systems in which the use of RGB/gray cameras attracts an increasing attention in recent years. In this paper, to improve the computational efficiency of existing PSO-based motion estimation method using RGB/gray cameras, we present a novel fast algorithm for the motion estimation problem with application to nano/micro manipulation systems. Particularly, in the proposed method, we improve the PSO with introducing the narrow solution space (NSS), to develop an improved version of PSO which was experimentally demonstrated significant improvements over the standard PSO on the computation speed meanwhile with a comparable estimation accuracy. Moreover, the method can also potentially allow to be integrated with a sub-pixel interpolation technique for the purpose of allowing for more advanced manipulation tasks.