基于SUSAN算子与分水岭算法的图像分割方法  

Image Segment Method Using SUSAN Operator and Watershed Algorithm

作　　者：苏畅[1] 陈宇拓[1] 张潇云[1] 

机构地区：[1]中南林业科技大学计算机科学学院,湖南长沙410004

出　　处：《电脑与信息技术》2007年第6期47-49,53,共4页Computer and Information Technology

摘　　要：针对分水岭图像分割算法中的过分割问题,提出了一种结合SUSAN算子和分水岭算法进行图像分割的方法。该方法首先用SUSAN算子对原始图像进行划分,检测出图像中不包含边界的区域,然后将检测结果作为标记符在梯度图像上进行标记,最后用分水岭算法对带标记的梯度图像进行分割。试验结果表明,该方法具有较好的分割效果。To solve the over-segmentation problem produced by watershed segmentation algorithm,an efficient method based on SUSAN operator and watershed algorithm is proposed in this paper.This method use SUSAN operator to detect the regions which not contain edge.Then the regions are extracted as markers in gradient image.According to the markers, the labeled gradient image is segmented by the watershed labelling algorithm.The experimental results show that this method can efficiendy solve the over-segmentation problem.

关 键 词：SUSAN算子 分水岭算法 过度分割 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]