摘要： 提出的改进的模糊c-均值聚类方法采用基于标准协方差矩阵的Mahalanobis距离,即椭球体聚类方法,这种聚类算法更接近遥感数据散点图的实际情况,从而可以显著提高聚类效果。对北京卫星ASTER数据的聚类分析实验表明,改进的模糊c-均值聚类方法的聚类效果要优于K-均值聚类方法和常规的模糊c-均值聚类方法。

关键词: 遥感数据, K-均值聚类, 模糊c均值聚类, Mahalanobis距离

Abstract: This article presents an improved fuzzy c-mean classifier in which Mahalanobis distance is adopted using standard connivance matrix, shown as ellipse spheroid cluster algorithm. This methodis more likely close to remote sensing data scatter map then that of other cluster algorithm so that the classification results are better either. Satellite ASTER Beijing data is used for testing the results proved that the improved the Mahalanobis distance classifier is more precedence than k-mean classifier and fuzzy c-mean classifier.

中图分类号: 

• TP75