多阈值模糊互信息图像分割方法  

Image multi-threshold method based on fuzzy mutual information

作　　者：刘立新[1] 

机构地区：[1]西安邮电学院电子与信息工程系,西安710121

出　　处：《计算机工程与应用》2009年第25期166-168,197,共4页Computer Engineering and Applications

基　　金：国家自然科学基金No.60872058~~

摘　　要：提出了多阈值模糊互信息图像分割新方法。该方法首先将模糊隶属度函数引入到传统互信息量中并定义模糊信息量概念;其次将模糊互信息量用于图像分割时给出具体隶属度函数的构造;最后采用混沌优化法来获得多阈值模糊互信息分割方法的最佳阈值。实验结果表明,提出的多阈值模糊互信息图像分割方法是有效的。The new image multi-threshold method based on fuzzy mutual information is proposed.First,fuzzy membership function is used in classical mutual information and defines the new concept of fuzzy mutual information.Second,the method of constructing fuzzy membership function is proposed in image segmentation based on fuzzy mutual information.Finally,the optimal thresholds of image segmentation based on fuzzy mutual information is obtained by means of chaos optimization method. Experimental results show that multi-threshold image segmentation method based on fuzzy mutual information is feasible.

关 键 词：图像分割 阈值化 互信息 模糊隶属度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]