基于QPSO的二维模糊最大熵图像阈值分割方法  

2D Fuzzy Maximum Entropy Image Threshold Segmentation Method Based on QPSO

作　　者：田杰[1] 曾建潮[1] 

机构地区：[1]太原科技大学系统仿真与计算机应用研究所,太原030024

出　　处：《计算机工程》2009年第3期230-232,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60674104)

摘　　要：针对运用图像分割方法求取阈值时存在的计算复杂、时间长、实用性差等问题,提出一种新的二维最大熵图像分割方法,该方法利用基于量子行为的微粒群算法对图像的二维阈值空间进行全局搜索,并将搜索到的二维熵最大值所对应的点灰度-区域灰度均值作为阈值进行图像分割。实验结果表明,该方法具有一定优越性,在执行时间与收敛性方面均得到较理想的分割效果。Aiming at the problems such as complex calculation, long executive time, and worse practicability when using image segmentation method to seek threshold, a novel 2D maximum entropy image segmentation method is proposed, which uses Quantum-behaved Particle Swam Optimization（QPSO） algorithm to conduct global search of 2D image threshold space, and takes the gray scale value of pixel and the gray scale mean value of region corresponding to 2D maximum entropy value as the threshold for image segmentation. Experimental results show this method has some advantages in aspects of executive time and astringency.

关 键 词：图像分割 二维模糊最大熵 量子行为的微粒群优化算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]