遥感影像分类中的模糊聚类有效性研究
Fuzzy Cluster Validation for Remote Sensing Image Classification
-
摘要: 通过不同遥感影像对常用的模糊聚类有效性指数进行测试,检验这些指数能否适用于遥感影像分类。实验结果证明,Xie-Beni系列指数能够有效地获得遥感影像最优分类数,相对于其他模糊聚类有效性指数更适用于遥感影像分类。Abstract: Fuzzy cluster validation plays a very important role in fuzzy clustering of remote sensing because it can find the optimal number of clusters without apriori assumptions on the number of clusters.This paper presents a review of fuzzy cluster validity indices and tests these indices whether they are suitable for remote sensing image classification using different remote sensing images.The results of our experiment illustrate that Xie-Beni index is more suitable for remote sensing image classification than other indices as it can effectively obtain the optimal number of clusters.