LSC-TES: Local spectral curvature-based accurate temperature-emissivity separation in indoor environments

Temperature-emissivity separation (TES) is a very interesting technique that estimates both the surface temperature and spectral emissivity from hyperspectral data such as Fourier transform infrared spectroscopy . The TES method is necessary for measuring the temperature of an object to which a contact thermometer cannot be attached. The iterative spectral smoothness TES (ISS-TES) method works well in outdoor environments with strong absorption in downwelling . In indoor environments , ISS-TES cannot work properly because of two factors: almost no absorption in downwelling and thermal equilibrium between an object and its surroundings. This study analyzes the causes of incorrect TES of ISS-TES through mathematical derivations and simulations. Based on these analyses, a novel local spectral curvature-TES (LSC-TES) method is proposed for stable temperature and spectral emissivity estimation in indoor environments. The application of local spectral smoothness instead of global spectral smoothness in ISS-TES on a heated object can mitigate the limitations of ISS-TES. The experimental results obtained using simulations and TELOPS Hyper-CAM LWIR validate the feasibility of the proposed LSC-TES in indoor environments. • This study analyzes the causes of incorrect TES of ISS-TES through mathematical derivations. • A novel LSC-TES method is proposed for indoor temperature and spectral emissivity estimation. • The application of local spectral smoothness can mitigate the limitations of ISS-TES. • The experimental results with TELOPS Hyper-CAM LWIR validate the feasibility of the LSC-TES.