Reproduce “Beyond the Visible: Jointly Attending to Spectral and Spatial Dimensions with HSI-Diffusion for the FINCH Spacecraft” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ian Vyse, Rishit Dagli, Dav Vrat Chadha, John P. Ma, Hector Chen, Isha Ruparelia, Prithvi Seran, Matthew Xie and 21 more
The University of Toronto Aerospace Team developed HSI-Diffusion, a 3D diffusion model for denoising and destriping hyperspectral satellite imagery, tailored for the FINCH CubeSat mission. This approach demonstrates superior performance in mitigating striping noise on synthetic satellite data, achieving PSNRs above 36 and meeting the FINCH mission's SNR requirements for scientific data retrieval.
Key insights
- A 3D diffusion model, by simultaneously considering spatial and spectral information, can effectively remove striping noise from satellite hyperspectral images while preserving spectral fidelity.
- Training a denoising model on synthetically generated stripe patterns applied to publicly available HSI datasets can prepare it robustly for real satellite data that is not yet operational.
- Achieving a PSNR of 36 or higher on remote sensing data is sufficient to meet the FINCH mission's SNR requirement of 30 for reliable scientific data retrieval.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.