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

This paper treats parameter estimation and performance bounds for multi-spectral model-based non-contact temperature measurement techniques. Based on Planck's law and different models for the unknown emissivity, different signal models for the interrogation results of a hot object with a multi-spectral camera are discussed. Based on these models, estimators and Cramer-Rao lower bounds for the unknown temperature and emissivity are discussed. To the best of our knowledge, the presented Cramer-Rao lower bound for achievable temperature estimation accuracy has not been published in the available literature so far. The resulting algorithms and performance bounds are validated both with simulated and real-world measurement data.