A SPA-based K-means clustering algorithm for the remote sensing information extraction

Set Pair Analysis (SPA) is a new methodology to describe and process uncertainty system, which has been applied in many fields recently. In this paper, a new approach to remote sensing information extraction, the SPA-based k-means clustering algorithm (SPAKM), has been proposed based on the principle of SPA. The basic ideals and steps of SPAKM are discussed. The proposed algorithm can overcome the limitation of K-means clustering algorithm to certain extent. Finally, cluster analysis experiments of LANDSAT TM image have been made. The results show that the improved K-means clustering algorithm is superior to K-means in classification accuracy of land cover classes of mixed pixels.