Acceleration Strategy Improving the Particle Swarm Optimization for Multilevel Thresholding Problems

The acceleration strategy (AS) is proposed for improving the searching speed of particle swarm optimization (PSO) algorithm on solving multilevel thresholding problems. By adopting the dynamically changing range of every dimension during the evolutionary process, the acceleration strategy can make the searching work more efficiently on dealing with the constraint of relationship among the thresholds. The PSO with acceleration strategy is validated based on the Otsu´s method, and the experimental results show that the new PSO algorithm has a higher searching speed than the original PSO algorithm.