Augmented GBM Nonlinear Model to Address Spectral Variability for Hyperspectral Unmixing

Spectral unmixing (SU) is a significant preprocessing task for handling hyperspectral images (HSI), but its process affected by nonlinearity and spectral variability (SV). Currently, SV considered within the framework of linear mixing models (LMM), which ignores nonlinear effects in scene. To address that issue, we consider on SU while investigating images. Furthermore, an augmented generalized bilinear model proposed to (abbreviated AGBM-SV). First, AGBM-SV adopts (GBM) as basic caused second-order scattering. Secondly, scaling factors dictionaries are introduced issues illumination conditions, material intrinsic variability, other environmental factors. Then, data-driven learning strategy employed set sparse orthogonal bases abundance according distribution characteristics real materials. Finally, alternating direction method multipliers (ADMM) optimization used split solve objective function, enabling algorithm estimate learn dictionary more effectively. The experimental results demonstrate comparative superiority both qualitative quantitative perspectives, can effectively problem scenes improve accuracy.