摘要
针对菲涅尔波带片多级衍射特性,提出等效点扩散函数的方法,建立基于等效点扩散函数的衍射成像模型,采用数值仿真和实验相结合的方法,通过盲优化算法实现衍射成像模型的迭代寻优,进而利用成像模型实现衍射图像复原。并就算法收敛缓慢或容易陷入局部最优的问题,提出了测量衍射效率初始值的方法,建立了菲涅尔波带片衍射效率测量系统,得到了其衍射效率的初始值,加快了求解成像模型的速度。实验结果表明,衍射效率初始值接近于最优值,优化后的衍射图像在对比度、梯度等方面均有明显提升,验证了基于成像模型和盲优化算法复原衍射图像的可行性。
Membrane diffraction imaging has become an important technology for high-quality imaging of large aperture space telescopes due to the advantages of Diffractive Optical Elements(DOE)such as light weight,loose surface tolerance,small size,foldable and unfoldable,and low cost.The main DOEs include Fresnel Zone Plate(FZP),Photon Sieve(PS),Composite Photon Sieve(CPS)and Stitched Zone Plate(SZP).Among them,the FZP has the advantages of ultra-lightness and foldability,and is widely used in optical microscopy,X-ray imaging,astronomical observation,optical communication,optical ranging and aerospace technology.However,due to the multi-level diffraction characteristics of FZP,its non-imaging diffraction order light becomes a strong background noise of the imaging order light,resulting in low contrast of the imaging image;on the other hand,the inevitable processing errors and structural defects in the FZP manufacturing process lead to low diffraction efficiency of FZP,and introduce wavefront distortion of the imaging beam,resulting in a sharp drop in the resolution of the imaging image.In addition,multi-order diffraction is coupled with the beam wavefront aberration,which leads to further degradation of FZP imaging quality.Many factors limit the performance of FZP diffraction imaging,resulting in the fact that its imaging contrast and resolution cannot meet the high requirements of space imaging applications.In this paper,FZP is used as the primary imaging mirror,and the FZP diffraction imaging system is built to study its image degradation process.Starting from the multi-order diffraction characteristics of FZP,the establishment,solution and optimization of its imaging model are carried out.In view of the problem of FZP multi-order diffraction imaging,the equivalent point spread function method is proposed,the numerical solution of the equivalent point spread function is obtained based on the imaging model,and the image restoration study is carried out.In view of the problem of slow algorithm convergence in the process of solving the FZP imaging model,a FZP diffraction efficiency initial value measurement system and measurement method are designed,a measurement experimental platform is built,the initial value of the FZP diffraction efficiency is obtained,and the algorithm convergence speed is accelerated.Specifically,the FZP diffraction efficiency measurement study is carried out.According to the measurement reciprocity principle,synchronous trigger imaging technology is adopted,and micron-level optical path adjustment is achieved by worm and stepper motor.The FZP diffraction efficiency measurement device is established,and the initial value of FZP diffraction efficiency is obtained by averaging multi-frame image data.The parameter solution and image restoration research based on the imaging model are carried out.The equivalent point spread function method is proposed to characterize the interference of many non-imaging diffraction lights of FZP,and the mathematical relationship between the point spread function,equivalent point spread function and diffraction efficiency of the primary and secondary light of FZP is analyzed,and the FZP imaging model is established;then,the equivalent point spread function is solved based on the FZP imaging model and SPGD algorithm by combining numerical simulation and experiment.To avoid the SPGD algorithm from falling into the local optimum during the iterative solution process,the SA algorithm is used for verification.The similar solution results of the two algorithms prove the feasibility of the method;finally,the experimental research of image restoration is carried out based on experimental data.The experimental results show that the restored diffraction image has improvements in contrast,gradient,etc.,verifying the feasibility of the parameter solution method of the FZP imaging model proposed in this paper and the model-based diffraction image restoration research.
作者
李嘉勋
李灿
杜正聪
程翔
李庆
文良华
LI Jiaxun;LI Can;DU Zhengcong;CHENG Xaing;LI Qing;WEN Lianghua(School of Electronic Information and Engineering,Yibin University,Yibin 644000,China;College of Aviation Engineering,Civil Aviation Flight University of China,Guanghan 618307,China)
出处
《光子学报》
北大核心
2025年第2期143-155,共13页
Acta Photonica Sinica
基金
四川省科技厅计划项目(No.2023YFQ0093)
国家自然科学基金(No.61975171)。
关键词
衍射成像
多级衍射
成像模型
图像复原
盲优化算法
Diffraction imaging
Multi-order diffraction
Imaging model
Image restoration
Blind optimization algorithm
