Research On Spectral Imaging Technology Based On Compressed Sensing

Spectral imaging technology has great significance because it can capture the spectral dimension information of the two-dimensional object field and obtain material properties that cannot be detected by ordinary imaging.How to obtain the spectral image information of the two-dimensional object field is the core problem in the field of spectral imaging.The traditional method obtains the spectral information of the one-dimensional object field through the slit,and perfects the spectral image of the two-dimensional object field by means of spatial scanning.Alternatively,a single-band image of the two-dimensional object field can be obtained at one time through narrow-band filtering,and then all wavelengths are complemented by spectral scanning,so as to achieve the purpose of obtaining a two-dimensional object field spectral image.However,neither of these two schemes can complete the acquisition of images with a single exposure.Aiming at the problem that the traditional method cannot solve the problem of obtaining a complete spectral image of a two-dimensional object field with a single exposure,this paper adopts the coded aperture technology based on the compressed sensing theory in the computational imaging theory to solve the time resolution limitation in the spectral acquisition.The theory of compressed sensing requires that the measurement matrices are all random matrices,and the encoded images with different wavelengths keep the same spatial position on the image plane,so it puts forward extremely high requirements for the system chromatic aberration correction.In this paper,the coded aperture technology based on compressed sensing is used to design a high-efficiency and low-distortion achromatic relay lens group suitable for the compressed sensing theory,so as to realize the high-efficiency spatial consistency of the continuous waveband of the two-dimensional object field on the image plane.This paper builds a spectral microscopy imaging platform based on coded aperture technology based on compressed sensing theory,and comprehensively describes the built spectral microscopy imaging system from the aspects of early performance index design,key attribute analysis,actual optical path construction,imaging quality analysis,etc.At the same time,a detailed data analysis is carried out for the key technical parameters,and the focus is on the correction effect of image surface chromatic aberration,the paraxial optical path optimization scheme based on linear space invariant system,and the accuracy of spatial resolution and spectral resolution of reconstructed spectral data.Specific research contents include:(1)In conjunction with the theoretical physical imaging model of compressed sensing,a set of high-efficiency and low-distortion achromatic relay lens sets based on grating beam splitting elements are designed to achieve the same spatial position on the image plane of the two-dimensional object field images of different wavelengths encoded by the mask.(2)An optimization scheme is designed to improve the spectral imaging data compression efficiency and imaging signal-to-noise ratio,and improve the spectral accuracy of reconstructed data.The scheme includes optimization of mask position,optimization of numerical aperture of illumination light source and optimization of grating dispersion correction.The effects of the mask position and the numerical aperture of the illumination light source on the imaging system are studied through simulation experiments and data analysis,so as to improve the spatial resolution and spectral resolution accuracy of the reconstructed spectral data.(3)A spectral microscopy imaging system is built,which overcomes the situation that traditional imaging equipment requires fixed samples and can only process signals for specific static scenes,and can realize dynamic data acquisition of samples in the field of view without sacrificing time resolution,a 3D spectral data block with 2D spatial resolution and 1D spectral resolution of the sample is obtained with a single exposure.