Hyperspectral Band Selection Based on Generalized 3DTV and Tensor CUR Decomposition

Hyperspectral Imaging (HSI) serves as an important technique in remote sensing. However, high dimensionality and data volume typically pose significant computational challenges. Band selection is essential for reducing spectral redundancy in hyperspectral imagery while retaining intrinsic critical information. In this work, we propose a novel hyperspectral band selection model by decomposing the data into a low-rank, smooth component and a sparse component. In particular, we develop a generalized 3D total variation (G3DTV) that employs the ℓ₁^p-norm on derivatives to preserve spatial-spectral smoothness. By applying the alternating direction method of multipliers (ADMM), we derive an efficient algorithm, where the tensor low-rankness is achieved through the tensor CUR decomposition. We demonstrate the effectiveness of the proposed approach through comparisons with various state-of-the-art band selection techniques on two benchmark real-world datasets. In addition, we offer practical guidelines for parameter selection in both noise-free and noisy scenarios. Demo codes for our algorithm are available at https://github.com/khenneberger/THBSCUR.