基于模糊聚类分析的边缘检测算法  

Image edge-detection algorithm based on fuzzy cluster analysis

作　　者：陈波[1] 杨阳[2] 沈田双 

机构地区：[1]信息工程大学测绘学院,郑州450052 [2]桂林空军学院,桂林541003

出　　处：《仪器仪表学报》2006年第z2期1603-1604,共2页Chinese Journal of Scientific Instrument

基　　金：国家863计划(2002AA731213);空间目标图像特性提取和识别的串行实时处理技术资助项目

摘　　要：将模糊c-均值聚类算法(FCM)应用到图像的边缘检测中。首先,将灰度图像中的每一个像素点看成是一个数据样本,将该点的灰度值经过Robert算子、Sobel算子和Prewitt算子处理构成它的特性向量,形成具有三维特征的数据集,然后对这个数据集应用模糊聚类算法进行分类,自适应地检测出图像的边缘点,达到提取边缘的目的。实验结果表明,这种混合算法能得到很好的边缘检测效果,并且得到的结果无需再细化处理,提高了边缘定位的精度。In this paper, an fuzzy c-means clustering method (FCM) is presented to solve edge detection problems. Firstly, a pixel point in a gray image is regarded as a data sample, and its gray values which is processed by Robert operator, Sobel operator and Prewitt operator makes up of the feature vectors of this data sample. In this way a data set with three-dimensional features can be obtained. Then the FCM is used for this data set, it can detect out the edge points adaptively. The experimental results show that this method can detect out the edge of an image correctly, its results need not to be adjusted, it can enhance the precision of edge orientation.

关 键 词：模糊聚类 边缘检测 ROBERT算子 SOBEL算子 PREWITT算子 

分 类 号：TH7-55[机械工程—仪器科学与技术]