基于严格等价函数的区间二型模糊熵及其图像阈值分割方法  

Entropy of Interval Type-2 Fuzzy Sets and its Thresholding Method Based on Restricted Equivalence Functions

作　　者：姚兰[1] 肖建[2] 

机构地区：[1]成都信息工程学院控制工程学院,成都610225 [2]西南交通大学电气工程学院,成都610031

出　　处：《计算机辅助设计与图形学学报》2015年第6期1074-1081,共8页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(51177137)

摘　　要：文中利用严格等价函数提出一种基于区间二型模糊熵的图像阈值分割方法.首先基于公理化定义,利用严格定价函数提出一种区间二型模糊熵的构建方法,由此可以得到多个不同的模糊熵计算表达式;然后通过理论分析给出了利用最小化模糊熵准则选取最优阈值的方法.实验结果表明,与现有的其他模糊阈值分割法和改进的2维Otsu法等相比,该方法的分割更加准确,运行时间更少,具有更广泛的适应性.By using restricted equivalence functions, an image thresholding method based on entropy ot In- terval type-2 fuzzy sets is proposed. Firstly, on the basis of the axiom definition, a construction method of entropy measure for interval type-2 fuzzy sets is presented, so that some different formulas are derived to calculate this kind of entropy. Then, through theory justification the optimal threshold can be determined by minimum fuzzy entropy criterion. Finally, experiments demonstrate that, compared with other fuzzy thresh- olding methods and improved 2-D Otsu＇s methods, the method can achieve accurate segmentation results and has excellent adaptive capability.

关 键 词：图像处理 区间二型模糊集 最小化模糊熵 阈值分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]