Temperature and Emissivity Separation from Thermal Airborne Hyperspectral Imager (TASI) Data

The Thermal Airborne Hyperspectral Imager (TASI) acquires 32 bands to provide continuous spectral coverage in the wavelength range of 8 to 11.5 mu m. The instrument was used during a field campaign in the City of Shiliazhuang, Hebei Province, China, in 2010. Land surface temperature and emissivity were measured near simultaneous with the airborne campaign for calibration and validation of the instrument. Radiance calibration was performed band-by-band using calibration coefficients, and atmospheric correction was performed using data from in situ measurements and the MODTRAN model. Surface temperature and emissivity separation were determined using the ASTER temperature-emissivity separation (ASTER_TES) and iterative spectral smooth temperature and emissivity separation (ISSTES) methods. The ASTER_TES method resulted in satisfactory agreement with ground data, with root mean square error (RMSE) values of 2.2 K for temperature and 0.0460 for broad-emissivity. The ISSTES method provided better ground validation results, with a RMSE for temperature of 1.8 K and a RMSE for broad-emissivity of 0.0394. The emissivity shapes acquired by the two methods were very similar. The results have relevance to studies of global climate change, environmental monitoring, classification, feature mining, and target recognition.