Determining an Efficient Supervised Classification Method for Hyperspectral Image

This paper proposes a research work done in search of best-supervised learning algorithm and the best kernel for Hyperspectral Image classification. In this work, we find that SVM outperforms other supervised algorithms. Many kernels are utilized in support vector machines for classification. Among them Linear, Polynomial and RBF kernels are analysed and the kernel that best suits for the application is determined. Cuprite (Nevada, USA) is the Hyperspectral image used in this paper.