Quantitative detection of settle dust over green canopy using sparse unmixing of airborne hyperspectral data

The main task of environmental and geosciences applications are efficient and accurate quantitative classification of earth surfaces and spatial phenomena. Recently, the ground-truth and laboratory measured spectral signatures promoted by advanced algorithms are proposed as a new path toward solving the unmixing problem of hyperspectral remote sensing (HRS) imagery in semi-supervised fashion. In this paper, the sensitivity of sparse non-linear unmixing techniques to extract and identify a small amount of settle dust over green vegetation canopy using HRS airborne imagery data is proposed. Among the available techniques, this study present results of two selected algorithms: 1) L_1/2 sparsity-constrained nonnegative matrix factorization (L_1/2-NMF) and 2) orthogonal matching pursuit (OMP). The performance is evaluated on real HRS imagery data via detailed experimental assessment. The first dataset including a conducted study area in Hadera, Israel and the second dataset is APEX Open Science Data Set (OSDS) in Baden, Switzerland. The results compared with performances of selected conventional unmixing techniques.