基于遗传算法的彩色图像二维熵多阈值自适应分割  

Self-adaptive threshold segmentation for color image using two-dimensional entropy method based on genetic algorithm

作　　者：常发亮[1] 刘静[1] 乔谊正[1] 

机构地区：[1]山东大学控制科学与工程学院,山东济南250061

出　　处：《控制与决策》2005年第6期674-678,共5页Control and Decision

基　　金：国家自然科学基金项目(60104009);山东省自然科学基金项目(Y2001G06).

摘　　要：提出一种基于遗传算法的二维熵多阈值自适应图像分割方法.在分析研究二维熵阈值分割原理的基础上,将可变码长的遗传算法应用于多阈值分割处理过程,采用基于多阈值的整数编码方式,将图像分割的类别数即染色体的码长融合到适应度函数中,从而实现了在对阈值寻优的同时得以优化分割类别数,最终实现图像的多阈值自适应分割处理.实验分析结果表明,该方法具有实现阈值寻优速度快,最优解对应图像分割效果好的特点.A method of color image multi-threshold segmentation using two-dimensional entropy method based on genetic algorithm (GA) is presented. Through analyzing the theory of entropy threshold segmentation, a GA of changeable code-length is applied. Every threshold is coded as integer data, and the fitness function includes classes of segmentation, which is the length of one chromosome. Then it realizes optimizing classes of segmentation as soon as optimizing every threshold. The experiments show that the applied segmentation method shows highly effective and speediness.

关 键 词：遗传算法 二维熵 多阚值 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]