In several applications, it is important to determine if a target can be easily detected, recognized or identified. It is common practice to measure its thermal signature, represented by the thermal contrast, defined as the target background differencial temperature. In order to obtain the thermal contrast, it is necessary to establish a relationship between the radiance detected by the sensor and the target temperature. Energy emitted by the target will depend not only on its temperature but also on its emissivity. On the other hand, energy received on the sensor will depend, among other factors, on the atmospheric conditions which will affect the path radiance. Hence, target temperature, emissivity and atmospheric conditions are essential parameters in order to give accurate measurements of the thermal contrast. In this work, we propose a comparison between two methods to split temperature and emissivity from radiance measurements, "The Grey Body Emissivity (GBE)" algorithm and the "Bayesian Inference (BI)" method. This
comparison has been done with spectral radiance measured with a FTIR spectroradiometer in the MWIR (3-5μm) and LWIR (8-12μm) bands. Measurements were done over a blackbody at different distances (ranging from 6cm to 15m) and temperatures (ranging from 0°C to 140°C), and over metallic plates of different materials and finishing at a fixed temperature of 55°C. Atmospheric conditions were modeled using MODTRAN 4.0 v3r1 computer code.