🤖 AI Summary
This study addresses the long-standing challenge of atmospheric compensation in long-range passive long-wave infrared (LWIR) hyperspectral imaging, which is severely degraded by atmospheric absorption, emission, and path radiance, yet has been largely overlooked due to modeling complexity. To recover true target spectra efficiently, this work proposes a lightweight ensemble deep learning framework that introduces, for the first time, an ensemble Transformer architecture to jointly estimate atmospheric transmittance, path radiance, and a shared downwelling radiance spectrum. A sparse autoencoder embedded within the framework discovers latent activation patterns aligned with geographic regions, even without explicit geolocation supervision. Evaluated on a MODTRAN-simulated dataset, the method achieves consistently low spectral distortion across all estimated atmospheric components. The code and dataset are publicly released to facilitate further research.
📝 Abstract
Passive long-wave infrared (LWIR) hyperspectral imaging under a standoff geometry depends on atmospheric absorption and emission, as well as reflected radiance, thus making atmospheric compensation essential to get knowledge of a target of interest. Despite its importance, this compensation has been largely overlooked due to its practical and modeling difficulty. In this paper, we present a lightweight set-based deep learning framework that takes multiple radiance measurements, collected at different standoff ranges, as input and jointly estimates transmittance, atmospheric path radiance, and a shared downwelling spectrum. We analyze the learned representation with a sparse autoencoder and observe that several latent features do activate on geographically coherent subsets of the test data despite the absence of location supervision. Experiments on a MODTRAN generated standoff LWIR dataset demonstrate low spectral distortion across all estimated products. The dataset and code is publicly available at: https://factral.co/SAE-LWIR/