混沌粒子群优化指数交叉熵的阈值分割  

Exponential cross entropy thresholding based on chaotic particle swarm optimization

作　　者：万水龙 罗国成[1] 余彪[1] 

机构地区：[1]南京航空航天大学电子信息工程学院,江苏南京210016

出　　处：《微型机与应用》2014年第7期71-73,共3页Microcomputer & Its Applications

基　　金：文件检验鉴定公安部重点实验室开放课题(10KFKT005)

摘　　要：为了提高指数交叉熵的阈值选取效率,提出了一种混沌粒子群优化指数交叉熵的阈值分割方法。首先导出指数交叉熵阈值选取方法,然后利用混沌粒子群算法对其进行优化。实验结果表明,相对于最大熵法和指数熵法,混沌粒子群优化指数交叉熵的阈值分割方法不仅分割结果精确,而且运行时间也相应缩短。In order to improve exponential cross entropy threshold selection efficiency, exponential cross entropy thresholding based on chaotic particle swarm optimization is proposed. Firstly, exponential cross entropy threshold selection is derived, then chaotic particle swarm optimization is used to search for the best thresholds. A large number of experimental results show that exponential cross entropy thresholding based on chaotic particle swarm optimization can achieve superior segmented results and greatly reduce the running time, in contrast with the maximum entropy method and the exponential entropy method.

关 键 词：阈值分割 指数交叉熵 混沌粒子群 

分 类 号：TN911.73[电子电信—通信与信息系统]