文献类型：期刊文献

中文题名：基于二维熵的人工鱼群材料图像分割方法

英文题名：Segmentation for sem image based on two-dimensional entropy and artificial fish swarm algorithm

作者：丁贤云[1];朱煜[1]

机构：[1]华东理工大学信息科学与工程学院,上海200237

年份：2010

卷号：40

期号：2

起止页码：210

中文期刊名：激光与红外

外文期刊名：Laser & Infrared

收录：CSTPCD;;Scopus;北大核心:【北大核心2008】；CSCD:【CSCD2011_2012】；

基金：国家自然科学基金(No.50803016);华东理工大学青年骨干教师项目(No.0156101)资助

语种：中文

中文关键词：SEM材料;图像分割;二维熵;人工鱼群

外文关键词：SEM material image ; image segmentation ; two-dimensional entropy ; artificial fish swarm algorithm

摘要：阈值法是图像分割中的重要方法,并在图像处理中得到了广泛的应用。针对电子扫描显微镜(SEM)摄取的纤维材料图像的自身特性,在预处理的基础上,提出了一种基于二维灰度直方图的人工鱼群图像分割方法。二维直方图的阈值的选取,是一个求全局最优的优化问题,本文将人工鱼群的算法应用于图像分割中,利用人工鱼群算法寻求二维熵的最优值,在实验中,人工鱼群算法收敛速度快,结果稳定,取得了理想的效果。
Thresholding is an important method of image segmentation and is extensively used in image processing for many applications. The fibrous material image captured from scanning electronic microscopy （SEM）shows the porosity surface structure characters of the material. To some extent, quantity analysis of material image is necessary for material study. In the paper, a method of image segmentation by artificial fish swarm algorithm based on two-dimensional gray histogram is presented after pretreatment. According to the maximum 2D entropy principle,the optimal combination of the parameters is searched, and the optimal threshold is determined by artificial fish swarm algorithm. Experimental results show that the proposed method gives better performance, and has a quick convergence rate.