Research On Target Recognition Techniques Based On Spectrometry Analysis And Polarization Imaging

Polarization imaging technology has unique advantages for detecting and identifying camouflage and false targets based on differences in target polarization characteristics.However,polarization imaging technology also has limitations.For some cases where the difference between the polarization characteristics of artificial targets and natural background is small,polarization imaging cannot highlight the target from a complex background.Spectral detection technology detects different spectra by using the difference of the target surface of the ground object and the difference of internal constituent elements.Some natural targets such as vegetation,rocks,snow,etc.will exhibit typical spectral curves,which are highly recognizable and show differences with the spectral curves of artificial targets.However,the spectrum detection technology is also susceptible to changes in the atmosphere and the environment,and the problem of "foreign matter is the same spectrum" appears.In view of the above problems,this paper studies the polarization imaging technology and spectral detection technology in the recognition target method.The typical ground object is selected for polarization imaging experiment and spectral detection experiment.In this paper,using rotating polaroid method for polarization imaging experiments to obtain the polarization information of the target.Spectral detection experiments were carried out using a spectrometer,and spectral information such as spectral polarization of the target was obtained by constructing a pre-polarization spectrum modulation module.A comprehensive analysis of the target's degree of polarization,polarization mean and variance,polarization contrast and spectral characteristic parameters is aimed at increasing the amount of information to identify the target and improving the recognition ability,so as to achieve the purpose of distinguishing and identifying the target.The experimental results show that the proposed method has a significantly higher recognition ability than the traditional single recognition technology when identifying targets with similar polarization or spectral properties.They are a useful complement to the detection and identification of targets.Through the comprehensive analysis of the target polarization characteristics and spectral features,the problem of low contrast in polarization imaging technology and the problem of "foreign matter in the same spectrum" in spectral detection technology can be effectively solved,and the recognition rate of the target is obviously improved.