Spectral Reweighting and Spectral Similarity Weighting for Sparse Hyperspectral Unmixing

Sparse unmixing separates the pixel of hyperspectral images (HSIs) into a collection of pure spectral signatures and the associated fractional coefficients with a complete spectral library as a priori, avoiding the drawback of inaccurate extraction of endmember information from the original HSI. As a state-of-the-art sparse unmixing method, fast multiscale spatial regularization unmixing algorithm (MUA) consists of two procedures, concerning on the approximation image domain and the original domain, respectively. However, it ignores the intersuperpixel correlation of the original domain that each superpixel only involves a small number of spectral signatures, and ignores the spectral variability of the approximate image domain. We address these two issues by introducing two different weighting factors to enhance the unmixing result. The effectiveness of our proposed algorithm is demonstrated by the experimental results on both synthetic and real hyperspectral data. The code and datasets of this letter can be found at https://github.com/wangtaowei11/Unmixing-Algorithm.

Automated summary

Sparse unmixing separates the pixels of hyperspectral images into pure spectral signatures and coefficients, avoiding the inaccuracy of endmember extraction. The fast multiscale spatial regularization unmixing algorithm ignores pixel correlation and spectral variability, which we address by introducing weighting factors to improve the unmixing result.