免疫克隆粒子群优化算法在水面舰船图像分割中的应用

摘要

介绍了一种免疫克隆粒子群优化(IC PSO)算法来进行函数优化,目的在于克服基本粒子群优化(PSO)算法容易陷入局部极值的不足,从而实现全局搜索.通过免疫克隆原理的应用,根据亲和度的高低进行粒子克隆选择、淘汰和高频变异,提高了种群的多样性,增强了算法全局搜索的能力,提高了收敛速度和精度.实验结果表明,该算法完成全局搜索所需的迭代次数明显少于PSO算法,在水面舰船图像分割中具有优良的性能.%An immunity clone with particle swarm optimization ( IC-PSO ) algorithm is proposed and applied to optimize functions, it overcomes the shortcoming of converging to local optimum for PSO algorithm. By introducing the IC algorithm, the diversity of particle swarms and the converging rate and accuracy are improved, and the capability of global searching is enhanced. From the experiment results, it is shown that the hybrid algorithm has perfect property in surface ships segmentation.