摘要： 传统聚类图像分割算法在对含有噪声的图片进行分割时图片中的噪声往往引起较大的偏差。针对含有一定噪声的图片提出一种算法，分别在传统模糊聚类算法中的距离公式和隶属度值公式中增加修正量的方法，以减少噪声对聚类结果的影响。实验结果证明该方法对一些噪声图片具有抗噪性能。

关键词: 图像分割, 聚类, 噪声, 矫正, 隶属度

Abstract: Traditional image segmentation algorithms by clustering usually cause bad result in images contain noises. Aiming at these images, this paper proposes a method which adds rectifying variables in the equations of distance and membership function in traditional clustering algorithm in order to reduce the influence of noises. Experimental result proves it is good for some noisy images.

Key words: image segmentation, clustering, noise, rectifying, membership

中图分类号: 

• TP391