基于遗传算法的模糊模式识别及其应用  

Fuzzy pattern recognition based on genetic algorithm and its application

作　　者：刘素华[1] 韩萍[1] 

机构地区：[1]郑州工程学院计算机科学系,河南郑州450052

出　　处：《计算机工程与设计》2005年第4期932-934,共3页Computer Engineering and Design

基　　金：河南省科技攻关基金项目(0224010011)

摘　　要：针对谷物害虫图像识别的特点,提出了基于模糊理论的害虫图像识别方法。对模糊C-均值聚类做了简要分析和评论,在此基础上将遗传算法引入模糊聚类,利用其搜索的随机和并行性,克服了模糊C-均值聚类的局部性和对初始聚类中心的敏感性;采用了基于贴近度和择近原则的模糊识别方法,分析了格贴近度的不足之处,并对之进行了改进。实验结果表明了上述方法是有效的,可行的,扩大了遗传算法和模糊理论的应用范围。Aiming at the characteristic of recognizing grain pest, a method is proposed with fuzzy theory. Fuzzy C-means clustering is introduced and remarked firstly. On the basis of systematic analysis of current algorithms, genetic algorithm is inducted into fuzzy clus- tering to solve the locality and the sensitiveness of the initial condition of Fuzzy C-means clustering by searching randomly and in paral- lelism. Then, the fuzzy discern method is proposed based on approximation value and the principle of selecting the near. The deficiency of dose-approximation value is proposed and improved. The example shows that the method is feasible and efficient, and it widens the application of genetic algorithm and fuzzy theory.

关 键 词：模式识别 模糊C-均值聚类 遗传算法 贴近度 图像识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]