Hyperspectral Image Classification Using SVM and Neural Network

Abstract

A Satellite image classification is a significant method used in remote sensing for the automated analysis and pattern recognition of satellite data, which facilitate the automated understanding of a large amount of information. These days, there exist many types of classification algorithms, such as parallelepiped and minimum distance classifiers, but it is still essential to get better performance in terms of correctness rate. Alternatively, over the last few years, cellular automata have been utilized in remote sensing to implement procedure related to simulation. While there is little preceding research of cellular automata related to satellite image classification, they offer much reward that can improve the results of classical categorization algorithms. In this research work, we firstly segment the pixels presented in the satellite image by using K - means to divide the image into clusters. Features are extracted with the help of PCA technique. After extracting the features, optimization is done by using PSO technique. Classification will be done by using Support vector machine (SVM) and Neural network (NN). The performance parameters like PSNR, MSE, and index error will be measured after simulating the proposed work.