多光谱辐射测温法是一种能够反演出辐射体真实温度(真温)的非接触式的温度测量方法, 这种方法通过采集不同波长下被测目标的亮度温度信息, 利用相关的算法反演被测目标的真温, 多光谱高温计就是利用这种方法反演被测目标真温的最重要的测量工具之一。 经过近半个世纪的不懈努力与探索, 国内外众多学者在此领域取得了长足的进步。 由于光谱发射率都小于1, 因此采用辐射高温计不能直接测量得到被测目标的真温, 只有通过处理多个光谱通道的波长和亮度温度, 利用多光谱辐射测温数据的处理方法才可以获得被测目标的真温。 在真温的反演的过程中, 一般都需要找到光谱发射率与波长或温度等变量之间的函数关系, 用含有波长或温度等变量的表达式代替光谱发射率, 这类方法模型选取缺乏足够的理论支持, 对于非专业人员, 选择合适的光谱发射率模型存在一定的难度。 由于光谱发射率具有瞬时多变性, 假设的光谱发射率模型与实际光谱发射率的变化之间往往存在一定的差异, 有可能导致真温反演产生较大的误差。 另外, 光谱发射率与波长或温度等变量之间的数学模型是需要通过大量的实验和经验才能确定, 这种数学模型通用性较差, 尤其是当待测辐射体发生改变时, 这种数学模型也就失去了意义。 因此, 找到一种无需假定光谱发射率与波长或温度之间数学模型而且又具有一定通用性的多光谱真温反演方法成为一种迫切的需要。 对于多波长高温计的N个光谱通道, 每一个光谱通道的测量数据满足一个数学方程, 对于N个光谱通道可以构成一个欠定方程组。 为了求解这个方程组, 将优化的思想引入多光谱求解过程中, 提出了一种基于参考温度的多目标极小值优化原理的多光谱真温反演方法, 将多光谱真温的求解问题转化为多目标极值优化问题, 实现了无需假设光谱发射率模型的真温和光谱发射率的反演。 与传统的二次测量法相比, 新方法在反演精度上与二次测量法大体相同, 但在反演速度上得到了较大幅度的提高。 借助于以往学者测量的真实数据, 利用基于参考温度的多目标极小值优化原理的多光谱真温反演方法实现了真温和光谱发射率的反演。 新方法在反演速度上得到了较大幅度的提高, 借助于以往的固体火箭发动机羽焰温度真实的测量数据, 利用基于参考温度的多目标极小值优化方法实现了真温的反演。

Multi-spectral radiation thermometry is a non-contact temperature measurement method which can retrieve the true temperature of the radiator. The method collects the brightness and temperature information of the target under different wavelengths and retrieves the true temperature of the target using related algorithms. Multi-spectral pyrometer is one of the most important measuring tools to retrieve the true temperature of the target by this method. After nearly half a century of unremitting efforts and exploration, many scholars at home and abroad have made considerable progress. Because the spectral emissivity is less than 1, the true temperature of the target can not be measured directly by using a radiation pyrometer. Only by processing the wavelength and brightness temperatures of multiple spectral channels and using the processing technology of multi-spectral radiation temperature measurement data can the true temperature of the target be obtained. In the process of true temperature inversion, it is generally necessary to find the functional relationship between spectral emissivity and wavelength or temperature variables, and replace spectral emissivity with expressions containing wavelength or temperature variables. This method lacks sufficient theoretical support for model selection. For non-professionals, it is difficult to select a suitable spectral emissivity model. The solution to the equation is realized. Because of the instantaneous variability of spectral emissivity, there are always some differences between the assumed spectral emissivity model and the actual spectral emissivity, which may lead to large errors in a true temperature inversion. In addition, the mathematical model between spectral emissivity and wavelength or temperature variables needs a lot of experiments and experience to determine. This mathematical model has poor generality, especially when the radiator to be measured changes; this mathematical model loses its significance. Therefore, it is an urgent need to find a universal multi-spectral true temperature inversion method without assuming a mathematical model between spectral emissivity and wavelength or temperature. For each spectral channel of a multi-wavelength pyrometer, the measured data of each spectral channel satisfies a mathematical equation, and for each spectral channel, an undetermined system of equations can be formed. In order to solve this system of equations, the idea of optimization is introduced into the process of multi-spectral solution. A multi-spectral true temperature inversion method based on multi-objective minimum optimization principle of reference temperature is proposed. The problem of solving multi-spectral true temperature is transformed into a multi-objective extreme value optimization problem, and the inversion of true temperature and spectral emissivity without assuming the spectral emissivity model is realized. Compared with the traditional quadratic measurement method, the new method has the same inversion accuracy as the quadratic measurement method, but inversion speed has been greatly improved. With the help of the true measurement data measured by previous scholars, the inversion of true temperature and spectral emissivity is realized by using the multi-spectral true temperature inversion method based on the multi-objective minimum optimization principle of reference temperature. The new method has greatly improved the inversion speed. With the help of the true measurement data of plume temperature of solid rocket motor in the past, the inversion of true temperature is realized by using a multi-objective minimum optimization method based on reference temperature.