Feature Reduction Based on the Fusion of Spectral and Spatial Transformation for Hyperspectral Image Classification

In recent years, the classification of Hyper Spectral Image (HSI) has posed a big challenge for the analysis of multidimensional property of the image. So it is of utmost importance to reduce the dimension of HSIs. There are several ways to reduce the dimension of hyperspectral images such as Principle Component Analysis (PCA), Kernel Principal Component Analysis (KPCA), Kernel Entropy Component Analysis (KECA) and so on. Through this article, We proposed an updated variant of KPCA using multiple kernels such as Linear, RBF, Cosine, Sigmoid, etc. We fused their spectral and special properties by classifying the HSIs using Hybrid Spectral Net Model (HybridSN) which is a recently trending modified deep neural network algorithm using Convolutional Neural Network (CNN). This paper presents empirical outcomes of the effects of using different kernels of KPCA algorithm and their performances regarding the classification of well-known hyperspectral data sets.