基于物理增强神经网络的连续太赫兹波同轴数字全息成像  

Continuous-Wave Terahertz In-Line Digital Holography Based on Physics-Enhanced Deep Neural Network

作　　者：赵洁[1,2] 金晓宇 王大勇 戎路[1,2] 王云新 林述锋[1] Zhao Jie;Jin Xiaoyu;Wang Dayong;Rong Lu;Wang Yunxin;Lin Shufeng(Faculty of Science,Beijing University of Technology,Beijing 100124,China;Beijing Engineering Research Center of Precision Measurement Technology and Instruments,Beijing 100124,China)

机构地区：[1]北京工业大学理学部,北京100124 [2]北京市精密测控技术与仪器工程技术研究中心,北京100124

出　　处：《激光与光电子学进展》2023年第18期20-28,共9页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(62220106005,62175004,62075001);北京市教委重点项目(KZ202010005008);北京市自然科学基金项目(4222061,4222063)。

摘　　要：太赫兹同轴数字全息是一种全场、无透镜、定量相衬成像方法,具有简单且稳健的光路结构,适合太赫兹波应用,然而其固有的孪生像问题会严重降低再现像的质量.提出一种将物理模型和卷积神经网络相结合的迭代相位复原方法,在无需施加约束以及准备预训练的标记数据集情况下,可从单幅同轴数字全息图中高保真度地恢复出样品的复振幅分布,并充分抑制孪生像干扰.仿真和实验结果表明了该方法的可行性,再现像质量优于目前主流方法,即基于物理增强神经网络的方法可以进一步拓展太赫兹数字全息成像的应用范围.Terahertz(THz)in-line digital holography is a promising full-field,lens-free,and quantitative phase-contrast imaging method with an extremely compact and stable optical configuration.Hence,it is suitable for the application of THz waves.However,the inherent twin-image problem can impair the quality of its reconstructions.In this study,a novel learning-based iterative phase retrieval algorithm,termed as physics-enhanced deep neural network(PhysenNet),is introduced.This method combines a physical model with a convolutional neural network to mitigate the twin-image issue in THz waves.Notably,PhysenNet can reconstruct the complex fields of a sample with high fidelity from just a single inline digital hologram,without the need for constraints or a pre-training labeled dataset.Based on simulations and experimental results,it is evident that PhysenNet surpasses existing phase retrieval algorithms in imaging quality,further enhancing the application range of THz in-line digital holography.

关 键 词：连续太赫兹波 同轴数字全息 神经网络 相位复原 

分 类 号：O438.1[机械工程—光学工程]