基于QPSO聚类算法的图像分割方法  

Image Segmentation Method Based on QPSO Clustering Algorithm

作　　者：王丹[1] 周锦程[1] 

机构地区：[1]黔南民族师范学院数学与统计学院,贵州都匀558000

出　　处：《科技视界》2016年第12期51-52,94,共3页Science & Technology Vision

基　　金：贵州省科技厅联合基金(LKQS201313;LKQS201314);黔南民族师范学院校级科研项目(QNSY2011QN10;2014ZCSX13;2014ZCSX18)

摘　　要：量子粒子群优化算法克服了传统粒子群优化算法中无法保证全局收敛、容易陷入局部最优的缺点,是近年来优化技术领域的一个研究热点。本文结合当前图像分割中常用的K-均值聚类算法中的相关技术,设计了基于QPSO的聚类算法并将其用于图像分割处理问题中。实验结果表明:在图像分割处理中,相对于K-均值聚类算法,QPSO聚类算法不仅不依赖于初始聚类中心的选择,而且还能得到相对于K-均值聚类算法精度更高的聚类中心,其在图像分割中的效果优于通常的K-均值聚类算法。The quantum particle swarm optimization（QPSO） algorithm overcomes the shortcomings of the traditional particle swarm optimization algorithm which can not guarantee the global convergence, easy to fall into the local optimal. QPSO algorithm has become a research hotspot in the optimization technology filed in recent years. In this paper, combined with the relevant technology of the current commonly K-Means clustering algorithm in the image segmentation problem, we propose a new QPSO based clustering algorithm and we use it to the image segmentation filed.Experimental results show that, with respect to the K-means clustering algorithm, QPSO clustering algorithm does not rely on the choice of the initial cluster center, but compared with the K-means clustering algorithm, it can get higher precision clustering center. Thus, it is better than the usually K-means clustering algorithm in the image segmentation problem.

关 键 词：PSO算法 QPSO算法 K-MEANS聚类 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]