基于顶帽变换和最大类间方差法的图像分割方法研究  

Research on Image Segmentation Method Based on Top-hat Transformation and Otsu

作　　者：殷苏民[1] 朱锦萍[1] 王祖声 江煜[1] 陆文俊[1] 

机构地区：[1]江苏大学机械工程学院,镇江212013

出　　处：《科学技术与工程》2014年第7期60-64,共5页Science Technology and Engineering

基　　金：国家自然科学基金(50975126;51179076);江苏省自然科学基金(BK2010042);江苏省"333工程"科研项目(BRA2011180)资助

摘　　要：为了克服直接使用最大类间方差法(Otsu)的图像分割方法不能对具有亮度不均匀背景的图像进行准确分割的不足,提出了一种基于数学形态学中的顶帽变换和最大类间方差法(Otsu)的图像分割方法。首先从数学形态学出发,详细阐述了灰度形态学的基本运算(腐蚀、膨胀、开、闭运算),引申出可以消除图像不均匀背景的顶帽变换法。然后介绍Otsu图像分割算法的具体步骤。最后以MATLAB7.0为实验平台,将顶帽变换和Otsu两种算法结合起来对图像进行分割算法仿真实验。理论分析和仿真实验结果表明,提出的方法可以实现快速、准确地分割。It can＇ t segment the imagines with backgrounds of uneven brightness accurately if the Otsu method is used directly. In order to overcome the disadvantage, a method of image segmentation that based on top-hat trans- formation of mathematical morphology and Otsu are presented. Firstly starting from the mathematical morphology, the basic operations of grayscale morphology, such as erosion, dilation, opening and closing operations are intro- duced. And then top-hat transformation is presented. Secondly the specific steps of Otsu algorithm for image segmentation are introduced. Finally take MATLABT. 0 as experimental platform to make image segmentation algorithm simulation combining top-hat transformation and Otsu. The oretical analysis and simulation results show that the method can take quickly and accurately segmentation in the MATLAB environment.

关 键 词：图像分割 顶帽变换 数学形态学 最大类间方差法(Otsu) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]