Feature extraction by linear spectral unmixing

Abstract

Linear Spectral Unmixing (LSU) has been proposed for the analysis of hyperspectral images, to compute the fractional contribution of the detected endmembers to each pixel in the image. In this paper we propose that the fractional abundance coefficients to be used as features for the supervised classification of the pixels. Thus we compare them with two well-known linear feature extraction algorithms: Principal Component Analysis (ICA) and Independent Component Analysis (ICA). A specific problem of LSU is the determination of the endmembers, to this end we employ two approaches, the Convex Cone Analysis and another one based on the detection of morphological independence. © Springer-Verlag 2001.