Atmospheric correction of airborne infrared hyperspectral images by direct estimation of the radiative terms
François Lemaître, Laurent Poutier, Yannick Boucher
Abstract
François Lemaître, Laurent Poutier, Yannick Boucher
Abstract
This paper deals with an autonomous method to retrieve the spectral emissivity and surface temperature from a hyperspectral infrared airborne sensor, covering both the MWIR and the LWIR bands. Atmospheric spectral radiative terms are assessed from the acquired sensor radiance by using a parametric model of the atmosphere and an optimization of likelihood criteria based on the whole image. Decomposition of the spectral emissivity on eigenvectors (obtained by PCA of an emissivity data base) allows the resolution of the temperature-emissivity separation (TES) by a linear constrained system (in radiance unit). This method gives also pixel temperature estimations and insures realistic results in narrow absorption bands. The approach is exposed, and illustrated with simulation results.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper deals with an autonomous method to retrieve the spectral emissivity and surface temperature from a hyperspectral infrared airborne sensor, covering both the MWIR and the LWIR bands. Atmospheric spectral radiative terms are assessed from the acquired sensor radiance by using a parametric model of the atmosphere and an optimization of likelihood criteria based on the whole image. Decomposition of the spectral emissivity on eigenvectors (obtained by PCA of an emissivity data base) allows the resolution of the temperature-emissivity separation (TES) by a linear constrained system (in radiance unit). This method gives also pixel temperature estimations and insures realistic results in narrow absorption bands. The approach is exposed, and illustrated with simulation results.
Key concepts: Radiance, Emissivity, Hyperspectral imaging, Remote sensing, Radiative transfer, MODTRAN, Full spectral imaging, Infrared