基于改进蚁群算法和指数熵算子的三类阈值图像分割  

A THREE-LEVAL THRESHOLDING METHOD FOR IMAGE SEGMENTATION BASED ON IMPROVED ANT COLONY ALGORITHMS AND EXPONENTIAL ENTROPY OPERATOR

作　　者：安海涛[1] 

机构地区：[1]内蒙古电子信息职业技术学院实习实训中心,呼和浩特010070

出　　处：《内蒙古农业大学学报（自然科学版）》2012年第2期182-184,共3页Journal of Inner Mongolia Agricultural University(Natural Science Edition)

摘　　要：针对多目标图像分割问题,采用了1种基于二维灰度直方图的三类阈值分割方法,将图像划分为暗、灰和亮3种不同的区域,分别给出了其模糊隶属度函数,引入概率分析,定义了基于指数熵算子的最大模糊熵准则,通过改进的蚁群算法,搜索最优模糊参数组合,以此确定图像的分割阈值。实验结果表明,该算法能快速、有效的分割图像。For multi - target image segmentation, a three - level thresholding method is presented based on 2D gray histogram, the image is partion to three parts fuzzily through defining three fuzzy membership functions, including dark, gray and bright part, the max- imum fuzzy entropy is defined based on exponential entropy operator and probability analysis. According to the improved ant colony al- gorithms, the optimal combination of the fuzzy parameters is searched, and the optimal thresholds is determined. Tile experimental re- suits show that the proposed method can segment the image effectively and fast.

关 键 词：图像分割 最大模糊熵 指数熵 蚁群算法 

分 类 号：O236[理学—运筹学与控制论]