A data processing method for multispectral emissivity and temperature with the introduction of new objective function and nonlinear constraints

Abstract

The underdetermined equation in multispectral pyrometer temperature measurement involves simultaneous unknowns of emissivity and temperature, posing a challenging obstacle to achieving accurate temperature inversion. In recent years, constrained optimization algorithms have been increasingly employed to address this issue. However, these algorithms need to set the appropriate initial emissivity values in particular and the imposition of manual constraints on the search range for emissivity. In this paper, a new data processing method that does not require these artificial Settings is proposed. Our method incorporates new objective functions and nonlinear constraints into the inversion of multispectral emissivity and temperature, while employing the Barrier Function Interior Point Method as an optimization tool. In addition, it has to be mentioned that in the blackbody temperature setting of the reference temperature model, the temperature of the blackbody is set very close to the target temperature by the constrained optimization algorithm, which obviously does not meet the needs of large-scale temperature measurement. The data processing method proposed in this paper addresses situations where there is a significant difference between the blackbody set temperature and the target temperature, ensuring both accuracy and speed over a wide range. Experiments demonstrate that our proposed method achieves a relative error of less than 0.42% in emissivity inversion, less than 0.57% in temperature inversion, and a calculation time of under 0.2 s. Our method can be applied to some high-precision and fast temperature measurement occasions that require short processing time and small relative error.