结合形态学分水岭的模糊聚类图像分割方法  

Segmentation for Fuzzy Clustering Image with Morphology Watershed Algorithm

作　　者：王建[1] 宋豫晓[1] 居本祥[2] 

机构地区：[1]华南农业大学工程学院,广东广州510642 [2]重庆大学光电工程学院,重庆400044

出　　处：《现代电子技术》2010年第6期119-121,共3页Modern Electronics Technique

基　　金：广东省工业攻关计划基金资助项目(2004B10101032);华南农业大学博士启动基金项目(B2004003)

摘　　要：图像分割是将图像分成各具特性的区域,并将感兴趣的目标提取出来的技术,是图像分析和计算机视觉中非常重要的研究内容。根据图像单一的属性标准对图像进行分割会产生过分割的现象,不容易提取到图像中有用的信息。结合数学形态学分水岭算法先对图像进行分割,得到了对图像的初步处理结果,再利用模糊聚类方法对分水岭分割产生的图像进行后处理,使得灰度信息相同或相近的点聚合在一起,得到了较好的图形轮廓。Image segmentation is a technique which can divide image into small areas with different characteristics and get the targets of interests. It is also a very important issue in the image processing and computer vision,but the image segmentation depending on the only attribute of the image, in addition, it can make excessive segmentation easily and can not get useful information of the image. Dividing image with the help of mathematical morphology watershed algorithm, and the primary result are gained,dealing with segmentation of watershed image under the method of fuzzy clustering. Because of this method, the same or similar points with grey information get together. The satisfactory image contours are given.

关 键 词：图像分割 分水岭算法 模糊聚类 计算机视觉 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]