Data-Driven Invertible Neural Surrogates of Atmospheric Transmission
This work addresses atmospheric modeling for remote sensing applications, presenting an incremental improvement through the integration of differentiable programming into existing simulation-based methods.
The paper tackles the problem of inferring atmospheric transmission profiles from spectral scenes by introducing a framework that uses a physics-based simulator tuned with autodifferentiation to create invertible neural surrogates, enabling tasks like atmospheric correction and modality recasting.
We present a framework for inferring an atmospheric transmission profile from a spectral scene. This framework leverages a lightweight, physics-based simulator that is automatically tuned - by virtue of autodifferentiation and differentiable programming - to construct a surrogate atmospheric profile to model the observed data. We demonstrate utility of the methodology by (i) performing atmospheric correction, (ii) recasting spectral data between various modalities (e.g. radiance and reflectance at the surface and at the sensor), and (iii) inferring atmospheric transmission profiles, such as absorbing bands and their relative magnitudes.