Endmember Extraction in Hyperspectral Images Using 1-1Minimization and Linear Complementary Programming

摘要

Endmember extraction in Hyperspectral Images (HSI) is a critical step for target detection and abundance estimation. Inthis paper, we propose a new approach to endmember extraction, which takes advantage of the sparsity property of thelinear representation of HSI's spectral vector. Sparsity is measured by the I_0norm of the abundance vector. It is also wellknown that I_1 norm well resembles I_0in boosting sparsity while keeping the minimization problem convex and tractable.By adding the I_1 norm term to the objective function, we result in a constrained quadratic programming which can besolved effectively using the Linear Complementary Programming (LCP). Unlike existing methods which requireexpensive computations in each iteration, LCP only requires pivoting steps, which are extremely simple and efficient forthe un-mixing problem, since the number of signatures in the reconstructing basis is reasonably small. Preliminaryexperiments of the proposed methods for both supervised and unsupervised abundance decomposition showedcompetitive results as compared to LS-based method like Fully Constrained Least Square (FCLS). Furthermore,combination of our unsupervised decomposition with anomaly detection makes a decent target detection algorithm ascompared to methods which require prior information of target and background signatures.