Choose Best SVM Kernels for Hyperspectral Satellite Image

Abstract

The use of  a remote sensing technologie was  a gained more  an attention due to a increasing need to   a collect data for  an environmental changes. Satellite images classification is an relatively a recent type of the remote sensing that a uses satellite an imagery to indicate several virtual environment a characteristics. The support a vector machine (SVM) method with various kernel functions (i.e.,  RBF, a sigmoid and a polynomial) were employed to the recognize and classify the hyperspectral satellite image. The SVM with RBF kernel function achieves the best classification accuracy of 87.3% based on the overall Classification.