边缘和区域多阶段结合的图像分割  

Image Segmentation Based on Multistage Combination of Edge and Region Information

作　　者：许向阳[1,2] 宋恩民[1,2] 金良海[1,2] 文耀光[1,2] 

机构地区：[1]华中科技大学计算机科学与技术学院,湖北武汉430074 [2]图像信息处理与智能控制教育部重点实验室,湖北武汉430074

出　　处：《小型微型计算机系统》2011年第5期943-946,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60972098)资助;湖北省自然科学基金项目(2009CDB075)资助

摘　　要：基于局部灰度的边缘检测和基于全局灰度统计的阈值分割是两种最要的图像分割方法.针对这两种方法单独使用时得不到理想分割结果的问题,提出一种边缘和区域多阶段结合的策略,利用边缘像素的灰度确定子区域分割阈值,将边缘和区域边界距离评估作为子区域递归分割的终止条件,最后叠加边缘图修订区域分割结果.实验表明,本文方法对于目标灰度不均的细胞图像能获得理想的分割效果.Edge detection and thresholding-based partition,which are based on local gray information and global statistics separately,are two important image segmentation methods.But when they are independently used to process images,the segmentation results are usually unsatisfactory.To address this problem,this paper proposes a new method which combines edge and region information in multistage.The proposed method computes the threshold of a sub-region by the gray levels of edge pixels,and evaluates recursive local region segmentations by the distance between edges and region boundaries,and finally combines local segmentation results and edge map to obtain final segmentation results.The experimental results show that good segmentation results are obtained even when the gray levels of objects are fluctuate in images.

关 键 词：图像分割 阈值选取 边缘检测 边缘区域结合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]