Hyperspectral and Mutlispectral Image Fusion Based on Spectral Library and Sparse Unmixing to Address Spectral Variability

In this paper, a novel method for fusing hyperspectral and multispectral images to generate a high-spatial-resolution hyperspectral image, that has both a high spatial and a high spectral resolution, is presented. The introduced method employs an extended version of the Linear Mixing Model (LMM) addressing spectral variability (SV). The proposed method consists of four main stages: (1) constructing a spectral library through the generalized linear mixing model (GLMM), (2) applying a spectral library pruning strategy, (3) extracting spatial information from the multispectral image using a sparse regression technique and (4) generating the fused hyperspectral product. The method is evaluated on synthetic datasets and the results clearly demonstrate its effectiveness, yielding fusion products with superior spectral and spatial fidelity.