基于最大树法的模糊图像分割方法  

Fuzzy Image Segmentation Based on Maximum Tree Method

作　　者：杨梦宁[1] 杨丹[1] 张强劲[2] 

机构地区：[1]重庆大学数理学院,重庆400044 [2]云南农业大学基础与信息工程学院,昆明650201

出　　处：《计算机科学》2005年第8期190-191,共2页Computer Science

摘　　要：本文提出一种新的模糊图像分割方法。根据任一像素点与相邻像素点的关系,计算局部均值均方差向量,并建立像素间的模糊相似关系,采用最大树方法进行模糊聚类分析,在模糊意义下实现图像分割。模糊逻辑下的图像分割,具有刻画视觉系统不精确度量特性的优点。由于待处理图像像素点数目巨大,采用最大树方法分类,能够大幅度减少工作量。因此,用最大树方法进行模糊聚类,实现图像分割,是一种既简便又高效的智能化的处理方法。A novel fuzzy-based maxinum tree image segmentation method is presented in this work. According to the relation between a pixel and the 3× 3 neighborhood of pixels about tbis center pixel, calculate the local mean and variance vectors, construct the fuzzy relationship, cluster by maximum tree method, and perform fuzzy image segmentation. Thanks to fuzzy logic, it preserves the propriety of measure of imprecision. Because of the great number of the pixels in processing image, we cluster by the method of maximum tree to reduce the works largely. So, it is a simple and efficient intellectualized method that clusters the pixels with maximum tree in a fuzzy way and performs image segmentation.

关 键 词：最大树 模糊逻辑 图像分割 图像分割方法 最大树法 模糊聚类分析 像素点数 模糊相似关系 视觉系统 均方差 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP641[自动化与计算机技术—计算机科学与技术]