基于分水岭分割的图像处理算法研究  

Research on Image Processing Algorithms Based on Watershed Segmentation

作　　者：夏梦琴 周建[1] 王静[1] 史道玲 曹齐松 XIA Meng-qin;ZHOU Jian;WANG Jing;SHI Dao-ling;CAO Qi-song(Xinhua university of anhui, Hefei Anhui 230088, China)

机构地区：[1]安徽新华学院

出　　处：《科技视界》2019年第17期71-72,共2页Science & Technology Vision

基　　金：安徽省省级教育教研项目(2017jyxm0531,2018jyxm1077);安徽新华学院校级质量工程(2018jy005,2014jqkcx10)

摘　　要：本文通过使用开闭运算和腐蚀膨胀运算等相关形态学算法,结合距离算法、标记算法和梯度图算法,提取梯度幅值图像的局部极小点,得到改进的梯度幅值图像,在此梯度图上使用分水岭分割,有效解决了传统分水岭分割存在的“过分割”问题,具有较高的分割精度。In this paper, some morphological algorithms, such as open-close operation and corrosion expansion operation, are used to extract the local minima of gradient magnitude image by combining distance algorithm, labeling algorithm and gradient graph algorithm, and an improved gradient magnitude image is obtained. In this gradient image, watershed segmentation is used to effectively solve the problem of “over-segmentation” existing in traditional watershed segmentation, which has high segmentation precision. degree.

关 键 词：分水岭 过分割 形态学 梯度幅值 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]