New Hyperspectral Endmember Extraction Algorithms using Convex Geometry

Abstract

The decomposition of the mixed pixels into individual pure material (endmember) along with its newlineproposition is called spectral unmixing for hyperspectral images. Spectral unmixing is considered newlinea three-stage problem for the hyperspectral image. The first is the subspace dimension which finds newlinethe number of pure materials in the image. The second one is endmember extraction which extracts newlinethe pure material spectra from the image and the third one is abundance estimation which estimates newlinethe proportions of each material in mixing. The endmember extraction is a very challenging stage newlinein spectral unmixing as abundance mapping greatly depends on extracted endmembers. In the newlineliterature, endmember extraction is addressed using a geometrical, statistical, sparse regression, newlineand deep learning approach. Due to simplicity and easy understanding, many researchers use the newlinegeometrical approach. In our research work, we focus on the geometrical endmember extraction newlineapproach which majorly uses the concept of convex geometry. In our work, we have developed newlinenew eight algorithms that improve the endmember extraction accuracy and abundance estimation newlineaccuracy. The first new algorithm explores entropy-based spatial information with convex set newlineoptimization-based spectral information. The second algorithm uses K-medoids clustering with newlineconvex geometry. The K-medoids clustering is used for removing redundant points that make our newlinesecond algorithm as noise-robust The third algorithm uses the area maximization approach instead newlineof conventional volume maximization. Surveour s formula is used for finding the area of a convex newlinepolygon in this third algorithm. The fourth algorithm uses the Rank correlation coefficient to find newlineonly effective bands for applying convex geometry. This fourth algorithm used Pearson s correlation newlinecoefficient. The fifth algorithm combines the geometrical features with statistical features. The newlineconvex geometry is used as a geometrical feature and covariance of the band is used as a statistical newlinefeature. The sixth a