粒子群优化的自适应区间二型模糊聚类算法  

PSO⁃based adaptive interval type-2 fuzzy clustering algorithm

作　　者：陈艺蕾 毕杨 王丽 CHEN Yilei;BI Yang;WANG Li(School of Electronic Engineering,Xi’an Aeronautical University,Xi’an 710077,China;School of Communication and Information Engineering,Xi’an University of Posts and Telecommunications,Xi’an 710121,China)

机构地区：[1]西安航空学院电子工程学院,陕西西安710077 [2]西安邮电大学通信与信息工程学院,陕西西安710121

出　　处：《电子设计工程》2023年第11期157-162,共6页Electronic Design Engineering

基　　金：国家自然科学基金(12004293,61901350)。

摘　　要：针对区间二型模糊聚类算法对聚类中心和模糊指数敏感的问题提出粒子群优化的自适应区间二型模糊聚类算法(PSO-AIT2FCM)。利用高斯滤波函数对获得的统计直方图做平滑处理,同时设计波峰与聚类中心的对应规则精准地确定聚类中心。构造新颖的融合直觉模糊信息的目标函数,通过类内紧致和类间分离两个方面的信息共同判断聚类效果。利用粒子群优化算法优秀的全局搜索能力在可行域内寻找合适的模糊指数组合。将该文算法应用于Berkeley图像分割中,实验结果表明该算法能够有效地确定聚类中心和模糊指数的组合并得到良好的分割结果。Interval type-2 fuzzy clustering algorithm is senstive to the selection of clustering centers and fuzzifiers.To solve these two problems,Particle Swarm Optimization based Adaptive Interval Type-2 Fuzzy Clustering Algorithm(PSO-AIT2FCM)is proposed in this paper.A Gaussian filtering function is used to smooth the statistical histograms of images and a corresponding rule of peaks and clustering centers is designed to determine the clustering centers accurately.A novel objective function with intuitionistic fuzzy information is constructed which contains intra⁃class compactness and inter⁃class separation information to judge the effect of clustering.A fuzzifier search method based on particle swarm optimization is proposed to search the suitable fuzzifiers in feasible region.Image segmentation experiments on Berkeley show that the proposed algorithm can adaptively determine the clustering centers and fuzzifiers and achieve good segmentation results.

关 键 词：区间二型模糊聚类 统计直方图 直觉模糊 粒子群优化 图像分割 

分 类 号：TN91[电子电信—通信与信息系统]