中文摘要：

结合粗糙集理论和K-均值聚类算法,提出了一种图像的粗糙聚类分割方法,试验结果表明,其比随机选取聚类的中心点和个数减少了运算量,提高了分类精度和准确性.

英文摘要：

This paper presents an image segmentation method based on rough set theory and u _ j K-means clustering. By using equivalence relations of attributes of image, the rough set theory offers the number and the centroids of the clusters, which initialize the K-means clustering. Then the image is segmented by K-means clustering algorithm.