基于SSKM算法的遥感图像半监督聚类

半监督聚类是近几年提出的一种新的聚类方法,具有良好的聚类性能,但是,它们绝大多数都需要有完整的先验信息,即对于所有的样本类别,都需要有至少一个标签数据。本文提出了一种基于不完整信息的遥感图像半监督聚类方法——SSKM聚类算法,算法利用部分样本类别的先验信息,辅助遥感图像聚类。实验表明,相比于传统的K均值聚类,该算法能够有效地改善遥感图像的聚类效果。

Semi-supervised Clustering of Remote Sensing Images Based on SSKM Algorithm

Semi-supervised Clustering,proposed in recent years,is a new clustering method which can get satisfactory result.However,most of these methods depend heavily on the completeness of the prior knowledge,which means that all classes in the dataset need at least one labeled object.In this paper a new semi-supervised clustering algorithm of remote sensing images using incomplete prior knowledge--SSKM clustering algorithm is proposed to solve this problem.It can make use of the partial prior knowledge to lead the process of clustering.An empirical study shows that this method can efficiently improve the quality of clustering comparing to traditional K Means Clustering.