An approximate message passing approach for compressive hyperspectral imaging using a simultaneous low-rank and joint-sparsity prior

TL;DR

This paper introduces a novel compressive sensing method for hyperspectral imaging that leverages a simultaneous low-rank and joint-sparsity prior, improving reconstruction accuracy and compression ratio.

Contribution

It proposes a new algorithm based on loopy belief propagation that jointly exploits structured sparsity and amplitude correlations in hyperspectral data.

Findings

Outperforms existing CS-based methods in reconstruction accuracy

Achieves higher compression ratios in hyperspectral data acquisition

Reduces reconstruction error significantly

Abstract

This paper considers a compressive sensing (CS) approach for hyperspectral data acquisition, which results in a practical compression ratio substantially higher than the state-of-the-art. Applying simultaneous low-rank and joint-sparse (L&S) model to the hyperspectral data, we propose a novel algorithm to joint reconstruction of hyperspectral data based on loopy belief propagation that enables the exploitation of both structured sparsity and amplitude correlations in the data. Experimental results with real hyperspectral datasets demonstrate that the proposed algorithm outperforms the state-of-the-art CS-based solutions with substantial reductions in reconstruction error.