论文部分内容阅读
针对传统粒子算法容易陷入局部极值而产生过早收敛的问题,提出一种基于遗传粒子优化算法。利用遗传算法在全局搜索方面的优势,对粒子算法进行改进,从而得到一种遗传粒子优化算法。在图像的分类中,传统K-Means聚类算法聚类中心的选择影响较大的问题,引入遗传粒子优化算法,对聚类中心进行优化,从而避免聚类中心的随机选择给图像分类精度带来很大的影响。最后,通过系统仿真比较,验证了该算法的优势。
Aiming at the problem of premature convergence caused by the traditional particle algorithm falling into local extremum, a genetic algorithm based on particle optimization is proposed. By using the advantage of genetic algorithm in global search, the particle algorithm is improved, and a genetic algorithm for particle optimization is obtained. In the classification of images, the traditional K-Means clustering algorithm has a big influence on the choice of clustering centers. The genetic particle optimization algorithm is introduced to optimize the clustering centers so as to avoid the random selection of clustering centers to give the image classification accuracy band Come a big impact. Finally, through the system simulation comparison, the advantage of this algorithm is verified.