Block principal component analysis for extraction of informative features for classification of hyperspectral images

This paper proposes a method to reduce the dimensionality of feature space for recognition of hyperspectral images. The method consists of dividing the spectral channels into blocks with high in-block correlation and the subsequent application of principal component analysis. It is shown that the proposed method allows to reduce the number of channels used in the classification by an order of magnitude with no signifi cant degradation of recognition quality.