The use of hyper-spectral pictures is expanding rapidly with the advancement of remote sensing technologies. The precise categorization of ground features using hyper-spectral pictures is a crucial research topic that has garnered considerable interest. The classification of hyper-spectral pictures has shown positive classification results using a variety of techniques. The three steps of this technique include pre-processing, feature extraction, and classification for the hyperspectral pictures. Each nondiagonal element in the matrix describes how various spectral bands correlate with one another. These matrices are then supplied to a support vector machine for classification as spatial-spectral characteristics. The suggested technique provides a fresh way to describe the spatial-spectral data in the HSI before categorization.Three publicly accessible hyperspectral data sets were employed in investigations, because the results demonstrate that the suggested method can perform superior than several cutting-edge algorithms, particularly when there are few training examples available.

Hyper Spectral Image Classification, Support Vector Machine, Hyperspectral Data Sets, Feature Extraction.