Estimation of parameters of a linear spectral mixture for hyperspectral images with atmospheric distortions

In this paper, we propose a novel method for estimating parameters of a linear spectral mixture for hyperspectral images. This method allows omitting a preliminary atmospheric correction of the input image. In order to derive a solution of the mixture problem different models of radiation transmission in atmosphere are considered. An evaluation of the effects of noise, the number of input pixels, and the number of signatures on the accuracy of the linear mixture coefficient restoration and the input pixel representation error is made.