An image threshold segmentation method based on multi-behaviour global artificial fish swarm algorithm

Jing Zeng

School of Physics and Electronic Information, China West Normal University, Nanchong 637002, Sichuan, China

Firstly, this paper describes how the histogram analysis method pre-processes the images to be segmented. Then is makes a detailed analysis of the working principles and behaviour pattern of basic artificial fish swarm algorithm (AFSA); dissects the defects of AFSA in principle and proposes an improved AFSA with global convergence. Finally, it presents the main steps of image threshold segmentation method based on AFSA; compares the performances of AFSA with those of other intelligent algorithms and proves that this improved AFSA makes all-around improvements on image segmentation.