Hyperspectral Images Classification Based on KLT

摘要

We propose an approach, based on KLT—Karhunen-Loève Transform, in order to automatically classify hyperspectral images. We used an AVIRIS image with 224 bands taken from the surroundings of Alto Paraiso City, on Goias State, in the middle of Brazil. Our approach performs dimensionality reduction, using only the eigenvectors with the highest eigenvalues, generating an eigenspace of low dimension. We also con-sider the spectral signatures of each class we work with. The classification is done finding the shortest Euclidean distance among the primitives of the new images and the primitives of the classes. We built a thematic map with 4 different classes.