Determination of emissivity dependence on wavelength using a hybrid least squares method for non-contact temperature measurements

The paper is devoted to non-contact measurements for temperature and emissivity. There are many parameters controlling emissivity: temperature, wavelength, surface condition, etc. There are cases for insufficiently detailed approximation for a body with emissivity depends on the wavelength as an issue of processing the remote measurements of emissivity. A hybrid method aimed to reducing the multidimensional minimization to one-dimensional case for calculating the emissivity decomposition coefficients is presented. For a gray body, linear and quadratic emissivity models are considered. An increasing number of solutions with increasing number coefficients of polynomial decomposition for emissivity is observed. The approach of choosing a proper solution is developed. The limitations on the model parameters numbers are presented: this is related to a growth of relative uncertainties in each parameter and with using the Akaike criterion.