改进的RPCCL聚类方法及在遥感影像分割中的应用

An Improved RPCCL Approach and Its Application to Segmentations of Remotely Sensed Imagery

RPCCL(rival penalized controlle dcompetitive learning)方法被认为是一种性能良好的聚类方法,但是将RPCCL聚类方法应用于高分辨率遥感影像分割任务中存在着聚类性能不稳定的局限。基于此,论文提出了基于簇初始化的RPCCL方法。研究表明,该方法能在保证聚类精度的情况下,改善RPCCL方法的聚类性能。

RPCCL is considered as a good clustering approach.However,some drawback like the instability of clustering performance exists when applied to the segmentation of high-resolution remote sensing images.Therefore,an improved RPCCL approach based on clusters initialization is proposed.Experiments results indicate that the proposed approach can improve clustering performance while keeping the clustering accuracy.