摘要： 模糊聚类算法（FCM）应用于数字图像的边缘检测已取得了较好的效果。通常情况下要进行检测的图像是完整的，但也存在边缘不完整或模糊的情况，这时原来的FCM算法就难以直接应用。该文提出了不完整数据的模糊聚类算法，对原FCM的算法公式进行了修正，并将其应用到局部域上有缺损的图像边缘检测中。实验结果表明，修正后的算法能较精确地恢复边缘缺损部分的信息，并实现其边缘检测。

关键词: 边缘检测, 模糊聚类算法(FCM), 缺损图像, 最优化原理

Abstract: Fuzzy c-clustering algorithm(FCM) is a useful tool in edge detection of digital image. Usually, the detected image is complete, but it may be incomplete or unclear in edge sometimes, and FCM is not directly applicable. In this paper, a changed FCM algorithm(CFCM) for incomplete data clustering is given, and it is also used to do edge-detecting of incomplete image. The experimental results of edge-detection of two incomplete images through CFCM show: the incomplete data of the images can be recovered quite exactly, resulting in a good edge-detection of the images.

中图分类号: 

• TP391.41