Abstract: Aiming at image extraction problem, optimal threshold selection is key for image segmentation results. In processing different kinds of image region,because of particle swarm optimization algorithm (PSO)’s premature phenomenon, it is difficult to accurately calculate the optimal segmentation threshold image segmentation, and accuracy rate is low. In order to improve the segmentation accuracy and accurate extraction of the image target, this paper proposed a image threshold segmentation methods based on chaos particle swarm optimization algorithm(CPS()).Benefit from chaotic operation of ergodicity, sensitivity to initial conditions and other advantages, CPSO improves particle swarm premature aggregation and can notbe traped in a local optimum problem,accelerates the overall optimal solution search ability. The CPSO image segmentation performance is best by simulation experiment, and the experimental results show that, compared with other image segmentation algorithm, CPSO not only accelerates the speed of operation, improves the efficiency of image segmentation, but also improves the segmentation accuracy, and it is very suitable for real-time image segmentation.

Key words: Image segmentation, PSO algorithm,Threshold segmentation