Parameter Estimation and Performance Bounds for Radiometric Model-based Non-contact Temperature Measurements

This paper describes a heuristic technique for contactless temperature measurement. To measure the temperature contactless of metal surfaces, the emissivity must be known. This is one of the problems at radiation temperature measurements. Different emissivity models are shown and compared. The emissivity is depending on temperature, wavelength, surface quality and oxide layer thickness. The evaluation of the measurement is based on the least squares estimator. The algorithm fits the measurement data to the Planck's radiation law, thereby the unknown parameter temperature T and emissivity \textgreek{e} will be varied. Furthermore the lower bound of the variance of this estimator is calculated by the Cramèr-Rao lower Bound (CRLB) and will be checked by the Monte Carlo simulation.