Abstract
A multimodal generative modeling approach combined with permutation-invariant set attention is investigated in this article to support long-wave infrared (LWIR) in-scene atmospheric compensation. The generative model can produce realistic atmospheric state vectors (T, H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O, O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> ) and their corresponding transmittance, upwelling radiance, and downwelling radiance (TUD) vectors by sampling a low-dimensional space. Variational loss, LWIR radiative transfer loss, and atmospheric state loss constrain the low-dimensional space, resulting in lower reconstruction error compared to standard mean-squared error approaches. A permutation-invariant network predicts the generative model low-dimensional components from in-scene data, allowing for simultaneous estimates of the atmospheric state and TUD vector. Forward modeling the predicted atmospheric state vector results in a second atmospheric compensation estimate. Results are reported for collected LWIR data and compared against fast line-of-sight atmospheric analysis of hypercubes-infrared (FLAASHIR), demonstrating commensurate performance when applied to a target detection scenario. Additionally, an approximate eight times reduction in detection time is realized using this neural network-based algorithm compared to FLAASH-IR. Accelerating the target detection pipeline while providing multiple atmospheric estimates is necessary for many real world, time sensitive tasks.
Highlights
L ONG wave infrared hyperspectral sensors collect data between 8–14 μm across hundreds of contiguous bands, Manuscript received July 27, 2020; revised September 17, 2020 and October 5, 2020; accepted October 23, 2020
The MMAE is used as part of the overall MDAC algorithm to perform in-scene atmospheric compensation and atmospheric state estimation
This study has presented a new long-wave infrared (LWIR) in-scene atmospheric compensation approach, producing both an atmospheric state vector and TUD vector from in-scene data only
Summary
L ONG wave infrared hyperspectral sensors collect data between 8–14 μm across hundreds of contiguous bands, Manuscript received July 27, 2020; revised September 17, 2020 and October 5, 2020; accepted October 23, 2020. Leveraging thermal hyperspectral data for these applications requires precise atmospheric compensation algorithms for accurate material characterization These compensation methods should be efficient and require minimal user input to operate on the large volumes of data collected by modern sensors. This article extends previous research in efficient LWIR atmospheric compensation [4], investigating new architectures to form a joint representation of atmospheric measurements, and their corresponding radiometric quantities. By focusing or attending to the most salient data aspects for a particular task, model performance can be improved while increasing interpretability [7] These advantages are achieved through a weighted average where the weights are attention scores that highlight feature importance
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
