Infrared Target Emissivity Model Identification And Surface Temperature Measurement

Spectral emissivity is one of the most important parameters for infrared radiation characteristics of objects,which can fully reflect the law of infrared radiation for one objects.Thus the spectral emissivity of one object can be applied in many fields,such as medicine inducting and national defense.In this paper,spectral radiance of object was collected by spectral radiometer.Based on the related knowledge theory of infrared physics,a neural network model was established through the relationship between the wavelength of the object and the spectral radiant brightness.The established neural network model was used to calculate the infrared radiation brightness of the object between 3~5?m and 8~12?m.The temperature was calculated by using the Wien's displacement law and Planck's formula,and the spectral emissivity of the object can be obtained when the ratio of infrared radiance to the standard blackbody was obtained,and in the same temperature the spectral emissivity of the object can be obtained by comparing to the infrared radiance of the standard blackbody with the same condition.The main research contents of this paper are included as follows.Firstly,combined with the relevant knowledge of infrared physics,the calibrated radiometer was used to measure the spectral radiance interference data for one object.Then the data of the spectral radiance of the object was calculated.Secondly,RBF neural network was used to model spectral radiant brightness of 3~5?m and 8~12?m bands.The spectral radiance of the object was calculated between 3~5?m and8~12?m bands.The surface temperature of target object was calculated by the basic law of infrared radiation characteristics.Then,the spectral emissivity of the object was calculated according to the energy method.Finally,the extreme learning machine neural network model was built by using wavelength and spectral radiance data as neural network training data.The data of spectral emission brightness which is obtained from the neural network model was used to calculate the spectral emissivity and surface temperature between 3~5?m and 8~12?m bands.The two algorithms were compared through the same amounts of sample data.The extreme learning machine method was faster than the RBF algorithm under the condition of high accuracy.Consequently,this method was applied to calculate the infrared radiation characteristics of aircraft skin.