HYPERSPECTRAL IMAGE CLASSIFICATION USING ITERATIVE AUTO-WEIGHTED DIMENSION REDUCTION

Abstract

In hyperspectral image classification task, achieving suitable dimension reduction is important to obtain desired classification performance. There are dozens of approaches to achieve this process. In this paper, a supervised autoweighted dimension reduction method is applied on hyperspectral images for classification purposes. The proposed method examines auto-weighted condition with a view to analyzing the effects on hyperspectral images. Comparative experimental studies are realized in order to demonstrate the advantage and disadvantage of the used method.