Ill-posed Surface Emissivity Retrieval from Multi-Geometry Hyperspectral Images using a Hybrid Deep Neural Network

Abstract

Atmospheric correction is a fundamental task in remote sensing because observations are taken either of the atmosphere or looking through the atmosphere. Atmospheric correction errors can significantly alter the spectral signature of the observations, and lead to invalid classifications or target detection. This is even more crucial when working with hyperspectral data, where a precise measurement of spectral properties is required. State-of-the-art physicsbased atmospheric correction approaches require extensive prior knowledge about sensor characteristics, collection geometry, and environmental characteristics of the scene being collected. These approaches are computationally expensive, prone to inaccuracy due to lack of sufficient environmental and collection information, and often impossible for real-time applications.In this paper, a geometry-dependent hybrid neural network is proposed for automatic atmospheric correction using multi-scan hyperspectral data collected from different geometries. Atmospheric correction requires solving the downwelling, upwelling and transmission components of the atmosphere contemporaneously. However, these are dependent on each other. This leads to ill-posed mathematical problem solutions. The proposed network is able to maintain the physical relationship among different radiative components from the observed at-sensor total radiance.It is equipped with a novel network structure and a loss function, which are trained using two different longwave (7.5 -13.5µm) hyperspectral dataset, one simulated using MODTRAN, and the second observed using the Blue Heron instrument in a dedicated field study. The proposed network can characterize the atmosphere without any additional meteorological data.A grid-search method is also proposed to solve the temperature emissivity separation problem. Results show that the proposed network has the capacity to accurately characterize the atmosphere and estimate target emissivity spectra with a Mean Absolute Error (MAE) under 0.02 for 29 different materials. This solution can lead to accurate atmospheric correction to improve target detection for real time applications.

Physics-guided neural networks for hyperspectral target identification

“…Other authors have also explored neural networks that retrieve physical quantities similar to our PGNN. Leveraging multi-geometry HSI, Xu et al retrieve atmospheric components with a neural network, 37 in contrast to our PGNN that operates on a single spectral signature without multi-geometry information. Neural networks have also been applied to LWIR for atmospheric compensation, a task closely related to our PGNN prediction of atmospheric quantities, by simultaneously processing multiple pixels.…”