About the Applications of Unmixing-Based Denoising for Hyperspectral Data

Abstract

Unmixing-based Denoising is a recently defined method which exploits spectral unmixing to recover bands characterized by a low Signal-to-Noise Ratio in a hyperspectral scene. The output of the unmixing process, which aims at decomposing each image element in signals typically related to pure materials, is inferred into the pixelwise reconstruction of a given band, ignoring the residual vector which is mainly characterized by undesired atmospheric influences and sensor-induced noise. The reconstructed images exhibit both high visual quality and reduced spectral distortions. This paper analyses the main problems that must be taken into account when applying this technique to real data. Special attention is given to the reference spectra used in the linear mixing model, which should be selected in order to keep the informational content of a given band unaltered in the reconstruction step.