SPGD算法高精度静态像差校正方法  

High-Precision Static Aberration Correction Method of SPGD Algorithm

作　　者：陈梅蕊 姜律 毛红敏[1] 孙会娟[3] 彭建涛 徐国定 胡立发[2] 陆焕钧 曹召良 Chen Meirui;Jiang Lü;Mao Hongmin;Sun Huijuan;Peng Jiantao;Xu Guoding;Hu Lifa;Lu Huanjun;Cao Zhaoliang(Jiangsu Key Laboratory of Micro and Nano Heat Fluid Flow Technology and Energy Application,School of Physical Science and Technology,Suzhou University of Science and Technology,Suzhou 215009,Jiangsu,China;School of Science,Jiangnan University,Wuxi 214122,Jiangsu,China;Institute of Mathematics and Physics,Beijing Union University,Beijing 100101,China;Shanghai Institute of Satellite Engineering,China Aerospace Science and Technology Corporation,Shanghai 201109,China)

机构地区：[1]苏州科技大学物理科学与技术学院,江苏省微纳热流技术与能源应用重点实验室,江苏苏州215009 [2]江南大学理学院,江苏无锡214122 [3]北京联合大学数理部,北京100101 [4]中国航天科技集团公司上海卫星工程研究所,上海201109

出　　处：《光学学报》2023年第5期97-109,共13页Acta Optica Sinica

基　　金：中国航天科技集团公司第八研究院产学研合作基金(SAST2020-025);“十四五”江苏省重点学科资助(2021135);江苏省自然科学基金青年基金(BK20220640);江苏省基础科学(自然科学)研究面上项目(22KJB150011);北京联合大学科研项目(ZK70202007)。

摘　　要：随机并行梯度下降(SPGD)算法广泛应用于光学系统静态像差校正,其性能指标对校正效果影响较大。由于传统的环围能量(EE)校正精度低、平均半径(MR)稳定性差,提出性能指标组合法,以实现静态像差的高精度、稳定校正。所提方法将EE与MR性能指标相结合进行像差校正,首先以EE作为性能指标对畸变图像进行校正,待能量集中于环围区域后,利用MR性能指标继续进行校正,直至能量分布均匀,校正终止。首先进行了校正仿真,结果显示:相比于EE和MR方法,性能指标组合法对静态像差的校正精度高、稳定性好。搭建实验光路,验证了所提方法的有效性。模拟和实验结果均表明,采用性能指标组合法可以获得高的校正精度且校正稳定。该方法可以应用于光学系统静态像差的校正和消除,实现其接近衍射极限的光学性能。Objective Due to the influence of the processing error of optical elements,system assembly error,and ambient temperature,the optical system usually has static aberrations,which reduce the imaging quality and significantly affects the optical performance.The stochastic parallel gradient descent(SPGD)algorithm is widely used to correct static aberrations.For the SPGD algorithm,the performance metrics,usually including the Strehl ratio(SR),encircled energy(EE),and mean radius(MR),have a great influence on the correction accuracy.In practical applications,SR is seldom chosen as it is difficult to achieve.In terms of EE,the convergence speed and correction accuracy depend closely on the selected encircled area,and hence,EE can only represent the total energy distribution of the surrounding area but cannot effectively reflect the entire intensity distribution.In contrast,the performance metric MR takes the whole intensity distribution into consideration,and higher correction accuracy is thus obtained.However,it is sensitive to small disturbance voltage,which makes the correction unstable.Therefore,to achieve high-precision and stable correction of static aberrations,we propose a combination method of performance metrics,which can not only concentrate the spot energy but also make the energy distribution uniform.Methods The proposed method combines the performance metrics EE and MR to correct the static aberrations.EE is first chosen as the performance metric and is computed with the acquired image.The control voltage is calculated and applied on the deformable mirror to correct the distorted wavefront.When most of the energy is concentrated in the encircled area,the performance metric is switched from EE to MR.Afterward,the energy distribution is further unified,which can also reduce the MR of the light spot.Thus,with the combination method,the energy can be better concentrated,and the intensity distribution can be more uniform.Meanwhile,the static aberrations can be corrected with higher accuracy.Results and Discussio

关 键 词：成像系统 随机并行梯度下降算法 静态像差 性能指标 校正精度 稳定性 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]