OTSU Multi-Threshold Image Segmentation Based on Improved Particle Swarm Algorithm

In view of the slow convergence speed traditional particle swarm optimization algorithms, which makes it easy to fall into local optimum, this paper proposes an OTSU multi-threshold image segmentation based on improved algorithm. After completes iterative update and position, method calculating contribution degree is used obtain approximate position direction, reduces scope search. At same time, asynchronous monotone increasing social learning factor decreasing individual are balance global Finally, chaos introduced increase diversity population achieve (IPSO). Twelve benchmark functions selected test performance algorithm compared with meta-heuristic The results show robustness superiority standard dataset images for experiments, some algorithms compare calculation efficiency, peak signal noise ratio (PSNR), structural similarity (SSIM), feature (FSIM), fitness value (FITNESS). that running time 30% faster than other in general, accuracy also better algorithms. Experiments proposed can higher efficiency.