Hyper Spectral Image Classification Using Spectral and Spatial Dimension Reduction

Due to the high dimensional data produced by the several spectral bands used for the same spatial area, hyperspectral image classification is a challenging task. This issue is addressed by spectral and spatial dimension reduction methods. In this paper we propose a new method based on principal component analysis for spectral dimensions and filters for spatial dimension reduction. Spectral dimensions are reduced by the SubXPCA and spatial dimensions are reduced by extended morphological profiles (EMP), perform the classification to classify the various hyperspectral objects in a scene. The proposed method was compared with various classifiers and tested on three hyperspectral datasets. The evaluation results indicate that the proposed method improved accuracy.