The Research Of Infrared Image Inversion Based On Infrared Illumination Mode

During the development of new imaging systems,a large number of infrared images are needed to test and evaluate them,and some infrared target detection algorithm based on machine learning also require a large number of infrared images for neural network training.In the acquisition of infrared images,it is difficult and restrictive to take the measured infrared date,and the simulation images have low realism.So the researchers proposed an infrared image/scene simulation method based on real data,using the measured data to obtain the real radiation distribution of the target,and applying the real radiation distribution of the target to the infrared scene simulation to improve the simulation authenticity.This method of converting measured data into the true radiation distribution of the target based on infrared imaging principles is called temperature inversion.Intrinsic image acquisition technology based on temperature inversion has been widely used in infrared scene simulation.Although the traditional inversion method considers the effects of atmospheric and sensor effects on the imaging results during the radiative transmission of the target,it does not take into account the effects of ambient radiation in the scene on the inversion results.The result of the traditional inversion algorithm approximates the target's intrinsic radiation to the target's spontaneous radiation.In the process of inversion,this paper analyzes the components of the target's intrinsic radiation by establishing an infrared illumination model,and removes the reflected part of the target radiation to extract the accurate target spontaneous radiation energy.First of all,this paper researches and analyzes the principle of infrared imaging and the model of infrared imaging process,and introduces the inversion principle of traditional inversion algorithms and the effects of each model.Traditional inversion processing was performed on infrared target images under different illumination.By comparing and analyzing the inversion results of targets under different illuminations,the disadvantages of traditional inversion algorithms are pointed out,which also leads to the research content of this paper.Then,the reflection principle and bidirectional reflection distribution function of light waves are introduced.In combination with the spontaneous emission of targets in the infrared band,an infrared radiation model is established.After analyzing the components of the illumination model,it is proposed to extract the radiation information in the infrared scene by establishing a calibration ball method,and use the extracted scene radiation information in combination with the target material and other information to obtain the reflected energy portion of the target intrinsic radiation.The extraction principle and extraction process are described.Finally,the framework of the inversion algorithm is introduced.Based on the traditionalinversion algorithm,the target's intrinsic radiation information is further processed to extract the spontaneous radiation information from the target's intrinsic radiation.According to the inversion steps of this paper,the infrared image is processed,and the results of the algorithm in this paper are compared with those of the traditional inversion algorithm.