Optimization Of Spatial Demodulation Algorithm For Spatially Modulated Fully Polarized Imaging

Polarization imaging is an advanced optical detection technology,which can not only obtain the intensity image,but also obtain the polarization information of the target.It has important application in the field of target detection and recognition.Spatially modulated fully polarized imaging is a new type of simultaneous fully polarized imaging technology.Based on the Stokes theory of polarization,the image of the target with all Stokes parameters（S₀、₁S、S₂、S₃）is spatially modulated and coupled into a single interferogram.The scheme can realize snapshot operation in principle,the corresponding instrument is compact,stable performance,and has a good application prospect.The main contents of this paper are as follows:1)For spatial modulation type all the basic principle of polarization imaging solutions,from the basic theory of the polarization and the basic principle of interference,comprehensively using Fourier optics,geometrical optics and wave optics and related concepts in the field of linear system,the principle and method of radiation is established step by step link of double amplitude transformation relations,analyzed the physical mechanism of spatial modulation,The mathematical model of spatial modulation is established.2)Based on the modulation mechanism in frequency domain and spatial domain of polarization interference of imaging information and the mathematical characteristics of the model,the basic principle of spatial reconstruction of fully polarized information is analyzed.By solving the overdetermined linear equations in the template,the implementation method of spatial demodulation algorithm is given.3)Using neural network model to test the simulation of the input image,get the spatial demodulation refactoring as a result,the structural similarity（SSIM）parameters quantitatively evaluate the spatial and frequency domain demodulation algorithm of demodulation algorithm of demodulation effect,and then error model is established,using the measurement error model analysis to understand the matrix condition number and measurement results,the relationship between The results show that the carrier matrix condition number K is correlated with the measurement error,and the method to improve the measurement accuracy is explored.4)Finally,the simulation parameters are set to obtain the system selection of the minimum number of conditions,and then the developed algorithm is modified by using the selection.SSIM is used to evaluate the modified demodulation quality in the algorithm test analysis of simulated polarization state image and measured image with given polarization state.For the simulated polarization images,the detection error of Stokes parameters corresponding to the modified algorithm is less than 5%,and the SSIM values of measured data before and after modification are above 0.9680 and 0.9718,respectively.The simulation results show that the SSIM of demodulation images obtained by the restriction condition optimization algorithm is higher than the SSIM of demodulation images obtained by the algorithm before optimization.The feasibility of the algorithm is verified.