Hyperspectral Unmixing Using Robust Deep Nonnegative Matrix Factorization

Nonnegative matrix factorization (NMF) and its numerous variants have been extensively studied used in hyperspectral unmixing (HU). With the aid of designed deep structure, NMF-based methods demonstrate advantages exploring hierarchical features complex data. However, a noise corruption problem commonly exists data severely degrades performance when applied to HU. In this study, we propose an ℓ2,1 norm-based robust nonnegative (ℓ2,1-RDNMF) for HU, which incorporates norm into two stages structure achieve robustness. The multiplicative updating rules ℓ2,1-RDNMF are efficiently learned provided. efficiency presented method is verified experiments using both synthetic genuine