Augmented Gaussian Linear Mixture Model For Spectral Variability In Hyperspectral Unmixing

In this paper, we propose a novel hyperspectral unmixing through the perturbed linear mixture model to take into account the spectral variability offset of the linear mixture model. In our proposed approach, we reformulate the LMM by adding a term to account for the spectral variations of endmember spectra of the dictionary. We use a white Additive Gaussian distribution for the perturbations in the LMM and employ the maximum likelihood estimation. Our proposed Augmented Gaussian LMM (AGLMM) employs the multiplicative updating rules to accelerate the convergence and exploit the sparsity of the unknown parameters. We evaluate our proposed unmixing approach on different datasets. Our results show the superior performance of the proposed AGLMM method over the state-of-the-art methods.