Multispectral radiation temperature measurement inversion calculation method based on generalized inverse-neural network, computer and storage medium

Abstract

The invention provides a multispectral radiation temperature measurement inversion calculation method based on a generalized inverse-neural network, a computer and a storage medium, and belongs to the technical field of multispectral radiation temperature measurement inversion calculation. The method comprises the following steps: firstly, calculating a voltage value for each spectrum channel by a simulation multispectral radiation thermodetector; secondly, constructing an equation for each spectrum channel, forming an equation set, and obtaining temperature and emissivity data with similar rules in an emissivity model through generalized inverse matrix calculation; comparing the data with an emissivity change rule in an emissivity model, and performing classification; thridly, defining a group of hyper-parameters, training the neural network, modifying the parameters by using a simulated annealing algorithm in the training process, and training the optimal parameters; and finally, inputting the test set verification sample set into the neural network, and outputting multispectral radiation temperature measurement data. The problem that in the prior art, an original data processing method cannot be universally suitable for materials with different emissivity, and inversion cannot be rapidly carried out is solved.

Description

technical field [0001] This application relates to a multispectral radiation temperature measurement inversion calculation method, in particular to a multispectral radiation temperature measurement inversion calculation method based on a generalized inverse-neural network, a computer and a storage medium, which belong to the multispectral radiation temperature measurement inversion calculation method technology field. Background technique [0002] In terms of data processing and inversion algorithms in the field of multispectral temperature measurement, the difficulty of multispectral radiation temperature measurement data processing methods lies in the data obtained through n spectral channels. According to the Planck formula, n radiation equations are established, but there are n+1 Unknowns (n ​​unknown spectral emissivity and 1 unknown true temperature). [0003] At present, the data processing methods of multispectral pyrometers can be roughly divided into two categorie...

AI Technical Summary

Problems solved by technology

[0004] The characteristic of the data processing method based on the variable emissivity model is that the emissivity assumption model can be changed within a certain range according to the different measured objects. This kind of method can already solve the real temperature of a certain type or several types of measured objects. and emissivity measurement problems, but still not suitable for all materials

[0005] At present, multi-wavelength radiation temperature measurement technology is still limited to single-point temperature measurement in terms of instrument development, and multi-point measurement needs to completely replicate the single-point optical and circuit systems. The device is complicated, the cost is high, and only a limited number of temperature points can be achieved. Measurement

The reason is that the existing multi-wavelength pyrometers are difficult to meet the problems of accurate collection of a large amount of radiation information and rapid inversion of real temperature based on a large amount of radiation information in the process of line temperature measurement from the inversion theory.

Therefore, in order to realize the online measurement of the multi-wavelength pyrometer line temperature, it is also necessary to solve the problem of the influence of the unknown spectral emissivity in the data processing process, which affects the accuracy of the real temperature inversion, and the multi-point, multi-wavelength and large data volume obtained by the online temperature measurement. In this situation, it is difficult to realize the problem of fast inversion of multi-point real temperature

Images