一种基于模糊聚类的快速二值化方法  

A FUZZY-CLUSTERING-BASED QUICK BIVALUED APPROACH

作　　者：丁震[1] 胡钟山[1] 杨静宇[1] 唐振民[1] 

机构地区：[1]南京理工大学计算机科学系,南京210094

出　　处：《计算机学报》1998年第S1期331-334,共4页Chinese Journal of Computers

摘　　要：通过确定阈值实现图像的二值化分割是一种重要且实用的图像分割技术．本文提出了一种基于模糊聚类的二值化方法．这种方法将模糊C-均值算法加以推广（GFCM）后，应用于图像的二值化分割．通过与Otsu阈值法的分割结果比较后表明，该方法的分割效果好，分割耗时少且适用性强．It is an important and practical technique for image segmentation to select a threshold forbinarization of images. In the paper, an approach of fuzzy-clustering-based binarization is proposed,which is a generalization of fuzzy C-means(FCM) algorithm and is used in bivalued segmentation ofimages. The comparison of segmentation results with Otsu's thresholding method show that theapproach has better segmentation effect, less segmentation time-consuming and more widespreadadaptability.

关 键 词：图像分割 图像二值化 模糊聚类 模糊C-均值算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]