A Comparative Study of Compressive Sensing Algorithms for Hyperspectral Imaging Reconstruction
Jon Alvarez Justo, Daniela Lupu, Milica Orlandic, Ion Necoara, Tor Arne Johansen
TL;DR
This paper addresses hyperspectral image reconstruction from compressed measurements using a $2.5\times$ compression factor within a compressive sensing framework. It compares convex solvers (FISTA, ADMM) against greedy pursuits (gOMP, BIHT, CoSaMP) on three datasets (Salinas, Jasper Ridge, China) and reports PSNR and convergence metrics. Results show that gOMP delivers the best accuracy and fastest recovery across datasets, though its performance depends on the unknown sparsity level $\kappa$; CoSaMP also achieves very high PSNR in several cases, while the convex methods guarantee convergence with longer runtimes. These findings guide algorithm selection for practical HSI CS reconstruction and point to future improvements in speed-accuracy trade-offs and scalability.
Abstract
Hyperspectral Imaging comprises excessive data consequently leading to significant challenges for data processing, storage and transmission. Compressive Sensing has been used in the field of Hyperspectral Imaging as a technique to compress the large amount of data. This work addresses the recovery of hyperspectral images 2.5x compressed. A comparative study in terms of the accuracy and the performance of the convex FISTA/ADMM in addition to the greedy gOMP/BIHT/CoSaMP recovery algorithms is presented. The results indicate that the algorithms recover successfully the compressed data, yet the gOMP algorithm achieves superior accuracy and faster recovery in comparison to the other algorithms at the expense of high dependence on unknown sparsity level of the data to recover.