单目标极小值优化法的多波长真温反演研究

在多波长温度测量系统中，光谱发射率假设为波长或温度模型来实现真温的求解。由于模型假设存在一定的不确定性，光谱发射率模型与实际光谱发射率的变化规律可能不符，会造成较大的真温反演误差。另外，光谱发射率与波长或温度之间的函数关系通用性较差，尤其是待测辐射体发生改变时，这种关系自然也就不复存在了。为了获得正确的真温，首次将优化思想引入到方法中，建立了单目标极小值优化法（Single Objective Minimization Optimization Method，SMO）完成真温反演。新方法无需建立光谱发射率和波长或温度之间的模型，降低了真温求解方法的复杂性和技术难度。与原有的光谱发射率与温度之间的模型（也称二次测量法，Second Measurement Method，SMM法）相比，在相同的初始条件下，新方法与二次测量法相比，反演速度提高了95%以上。

Multi wavelength true temperature inversion of single objective minimization method

In the multi-wavelenth temperature measurement system, the spectral emissivity is assumed to be a wavelength or temperature model to solve the true temperature. Because of the uncertainty of the model assumption, spectral emissivity model is inconsistent with the variation rule of the actual spectral emissivity, which will cause larger inversion error. In addition, the poor universality of the mathematical relationship between the emissivity and wavelength or temperature, especially the measured radiator is changed, the relationship will cease to exist. In order to the obtain the true temperature, the optimization idea is introduced into the method for the first time, and a single objective minimization optimization method(SMO) was established to complete the inversion of true temperature. The new method does not need to establish a model between spectral emissivity and wavelength or temperature, and has a certain universality. Compared with the second measurement method(SMM), the inversion speed of the proposed method was increased over 95% in the same initial conditions.