The Effect of Impulse Denoising on Geometric Based Hyperspectral Unmixing

Abstract

Hyperspectral unmixing is a process to find number of spectral component (called endmember), estimation of endmember signatures and their abundance fractions in each pixel on the scene. Geometric based algorithms are developed for hyperspectral unmixing problem in the literature. The distribution of spectra (points in n-dimensional scatterplot) can be used to estimate endmember signatures geometrically. Impulse denoising before unmixing process can help getting better results for endmember extraction. For this reason, General Prior Algorithm (GAP) is used before unmixing process. Experiments using real data demonstrate that this preprocessing step provided better results for endmember estimation.