Wavelength Calibration Correction Technique for Improved Emissivity Retrieval

Abstract

Accurate retrieval of surface emissivity from long-wave infrared hyperspectral imaging data is necessary for many scientific and defense applications. Emissivity retrieval requires an atmospheric model for the scene and consists of two interwoven steps: atmospheric compensation (AC) and temperature-emissivity separation (TES). AC converts the at-aperture radiance to ground radiance, and TES uses the ground radiance to produce a temperature and emissivity estimate. TES assumes that emissivity spectra for solids are smooth, compared to atmospheric features. Model-based techniques find an atmospheric model, which produces the smoothest emissivity estimates. The high-resolution model must be band averaged to the sensor's spectral response function (SRF), which is difficult to characterize and maintain. Any errors in the SRF cause errors in the atmospheric spectra and roughness in the emissivity estimates. We propose a technique that improves the quality of the retrieved emissivity by correcting band-averaging errors of the model from the SRF. An in-scene AC (ISAC) technique is used to find pixels containing a high emissivity material. Atmospheric model bands far from absorption features and a material library are then used to identify the material and estimate the pixel temperatures. At-aperture radiance spectra are generated for the pixels, instead of the ground radiance spectra generated by ISAC. The linear relationship between the generated and measured at-aperture radiances is used to determine model correction factors. Using simulated data, we demonstrate the capability of this technique to substantially reduce calibration error effects in the corrected atmospheric spectra and retrieved emissivities.

Keywords

• Hyperspectral imaging
• remote sensing
• calibration