Temperature and Emissivity Retrievals From Hyperspectral Thermal Infrared Data Using Linear Spectral Emissivity Constraint

Owing to the ill-posed problem of radiometric equations, the separation of land surface temperature (LST) and land surface emissivity (LSE) from observed data has always been a troublesome problem. On the basis of the assumption that the LSE spectrum can be described by a piecewise linear function, a new method has been proposed to retrieve LST and LSE from atmospherically corrected hyperspectral thermal infrared data using linear spectral emissivity constraint. Comparisons with the existing methods found in literature show that our proposed method is more noise immune than the existing methods. Even with a NEΔT of 0.5 K, the rmse of LST is observed to be only 0.16 K, and that of LSE is 0.006. In addition, our proposed method is simple and efficient and does not encounter the problem of singular values unlike the existing methods. As for the impact of the atmosphere, the results show that our proposed method performs well with the uncertainty of the atmospheric downwelling radiance but suffers from the inaccuracy of the atmospheric upwelling radiance and atmospheric transmittance, which implies that an accurate atmospheric correction is still needed to convert the radiance measured at the satellite level to the at-ground radiance. To validate the proposed method, a field experiment was conducted, and the results show that 80% of the samples have an accuracy of LST within 1 K and that the mean values of LSE are accurate to 0.01.