[1]闫恪涛,张洽铭,佘世刚,等.端到端的神经网络相位解包裹方法[J].常州大学学报(自然科学版),2025,37(01):85-92.[doi:10.3969/j.issn.2095-0411.2025.01.010]
　YAN Ketao,ZHANG Qiaming,SHE Shigang,et al.End-to-end neural network phase unwrapping method[J].Journal of Changzhou University(Natural Science Edition),2025,37(01):85-92.[doi:10.3969/j.issn.2095-0411.2025.01.010]
点击复制

端到端的神经网络相位解包裹方法()
分享到：

常州大学学报(自然科学版)[ISSN:2095-0411/CN:32-1822/N]

卷:
第37卷
期数:
2025年01期
页码:
85-92
栏目:
机械与动力工程
出版日期:
2025-01-22

文章信息/Info

Title:
End-to-end neural network phase unwrapping method
文章编号:
2095-0411(2025)01-0085-08
作者:
闫恪涛1; 张洽铭1; 佘世刚1; 高书苑1; 余文君2; 于瀛洁2
1.常州大学 机械与轨道交通学院, 江苏 常州 213164; 2.上海大学 机电工程与自动化学院, 上海 200444
Author(s):
YAN Ketao1;  ZHANG Qiaming1;  SHE Shigang1;  GAO Shuyuan1;  YU Wenjun2;  YU Yingjie2
1.School of Mechanical Engineering and Rail Transit, Changzhou University, Changzhou 213164, China; 2.School of Mechanical Engineering and Automation, Shanghai University, Shanghai 200444, China
关键词:
干涉测量;  相位解包裹;  神经网络;  UNet;  相位去噪
Keywords:
interferometry;  phase unwrapping;  neural network;  UNet;  phase denoising
分类号:
TK 8
DOI:
10.3969/j.issn.2095-0411.2025.01.010
文献标志码:
A
摘要:
相位解包裹为重要的信号处理过程,其目的是将被包裹相位恢复到原始相位。文章探讨了基于U型神经网络结构(UNet)对测量区域的包裹相位进行相位展开的方法,该方法可以有效地处理包裹相位区域,能够从噪声包裹相位中直接估计平滑的展开相位。以圆域和花瓣状域的包裹相位数据为例来建立训练数据库,通过模拟数据验证该方法的解包裹性能。分析表明,在噪声情况下UNet方法的均方误差小于传统解包裹算法,通过实测数据验证了UNet方法的性能。
Abstract:
Phase unwrapping is an important signal processing process, which aims to restore the wrapped phase to the original phase. This paper investigates the method of phase unwrapping for the wrapped phase of the measurement region using a U-shaped neural network structure(UNet). This method can effectively deal with the wrapped phase area, and can directly estimate the smooth unwrapped phase from the noisy wrapped phase. In this paper, the wrapped phase data of the circular domain and the petal domain were used as examples to establish the training database, and the unwrapped performance of the method was verified by simulated data. Analysis shows that the mean square error of the UNet method is smaller than that of the traditional unwrapping algorithm in the case of noise. The performance of the UNet method is verified by the measured data.

参考文献/References:

[1] SUN Y, BIAN Y X, SHEN H, et al. High-accuracy simultaneous phase extraction and unwrapping method for single interferogram based on convolutional neural network[J]. Optics and Lasers in Engineering, 2022, 151: 106941.
[2] ZHANG J K, LI Q G.EESANet: edge-enhanced self-attention network for two-dimensional phase unwrapping[J]. Optics Express, 2022, 30(7): 10470-10490.
[3] LIAN S Z, YANG H Q, KUDO H. Simple phase unwrapping method with continuous convex minimization[J]. Optics Express, 2022, 30(18): 33395-33411.
[4] ZHENG D L, DA F P. A novel algorithm for branch cut phase unwrapping[J]. Optics and Lasers in Engineering, 2011, 49(5): 609-617.
[5] ZHAO M, HUANG L, ZHANG Q C, et al. Quality-guided phase unwrapping technique: comparison of quality maps and guiding strategies[J]. Applied Optics, 2011, 50(33): 6214-6224.
[6] GAO D P, YIN F L. Mask cut optimization in two-dimensional phase unwrapping[J]. IEEE Geoscience and Remote Sensing Letters, 2012, 9(3): 338-342.
[7] GHIGLIA D C, ROMERO L A. Robust two-dimensional weighted and unweighted phase unwrapping that uses fast transforms and iterative methods[J]. Journal of the Optical Society of America A, 1994, 11(1): 107-117.
[8] KAUFMANN G H, GALIZZI G E, RUIZ P D. Evaluation of a preconditioned conjugate-gradient algorithm for weighted least-squares unwrapping of digital speckle-pattern interferometry phase maps[J]. Applied Optics, 1998, 37(14): 3076-3084.
[9] SPOORTHI G E, GORTHI S, GORTHI R K S S. PhaseNet: a deep convolutional neural network for two-dimensional phase unwrapping[J]. IEEE Signal Processing Letters, 2019, 26(1): 54-58.
[10] YU H T, CHEN X Y, HUANG R B, et al. Untrained deep learning-based phase retrieval for fringe projection profilometry[J]. Optics and Lasers in Engineering, 2023, 164: 107483.
[11] WANG C X, PANG Q. The elimination of errors caused by shadow in fringe projection profilometry based on deep learning[J]. Optics and Lasers in Engineering, 2022, 159: 107203.
[12] RAVI V, GORTHI R K. CF3DNet: a learning-based approach for single-shot 3D reconstruction from circular fringes[J]. Optics and Lasers in Engineering, 2023, 167: 107597.
[13] QIAO G, HUANG Y Y, SONG Y P, et al. A single-shot phase retrieval method for phase measuring deflectometry based on deep learning[J]. Optics Communications, 2020, 476: 126303.
[14] 王帅, 陈波, 程朋飞, 等. 卷积神经网络相位解包裹方法[J]. 现代电子技术, 2021, 44(21): 70-73.
[15] 王硕, 王华英, 王学, 等. 融合注意力机制的相位解包裹方法[J]. 光学技术, 2022, 48(4): 385-390.
[16] 邸江磊, 唐雎, 吴计, 等. 卷积神经网络在光学信息处理中的应用研究进展[J]. 激光与光电子学进展, 2021, 58(16): 9-35.
[17] SPOORTHI G E, SAI SUBRAHMANYAM GORTHI R K, GORTHI S. PhaseNet 2.0: phase unwrapping of noisy data based on deep learning approach[J]. IEEE Transactions on Image Processing, 2020, 29: 4862-4872.
[18] ZHANG J C, TIAN X B, SHAO J B, et al. Phase unwrapping in optical metrology viadenoised and convolutional segmentation networks[J]. Optics Express, 2019, 27(10): 14903-14912.
[19] SUMANTH K, RAVI V, GORTHI R K. A multi-task learning for 2D phase unwrapping in fringe projection[J]. IEEE Signal Processing Letters, 2022, 29: 797-801.
[20] WANG K Q, LI Y, KEMAO Q, et al. One-step robust deep learning phase unwrapping[J]. Optics Express, 2019, 27(10): 15100-15115.
[21] BRIÑEZ-DE LEÓN J C, RICO-GARCÍA M, RESTREPO-MARTÍNEZ A.PhotoelastNet: a deep convolutional neural network for evaluating the stress field by using a single color photoelasticity image[J]. Applied Optics, 2022, 61(7): D50-D62.
[22] LI H, ZHONG H P, NING M Q, et al. Using neural networks to create a reliable phase quality map for phase unwrapping[J]. Applied Optics, 2023, 62(5): 1206-1213.
[23] PARK S, KIM Y, MOON I. Automated phase unwrapping in digital holography with deep learning[J]. Biomedical Optics Express, 2021, 12(11): 7064-7081.
[24] WANG K Q, QIAN K M, DI J L, et al. Deep learning spatial phase unwrapping: a comparative review[J]. Advanced Photonics Nexus, 2022, 1(1): 014001.
[25] 李国栋, 韦春龙, 于瀛洁, 等. 圆形域干涉图中的相位解包裹[J]. 光学精密工程, 2000, 8(5): 473-477.
[26] NAIR V, HINTON G E. Rectified linear units improve Restricted Boltzmann machines[C]//ICML 2010-Proceedings, 27th International Conference on Machine Learning. Haifa: ICML, 2010: 807-814.
[27] IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift[C]//32nd International Conference on Machine Learning. Lille: ICML, 2015: 448-456.
[28] ITOH K. Analysis of the phase unwrapping algorithm[J]. Applied Optics, 1982, 21(14): 2470.
[29] HERRÁEZ M A, BURTON D R, LALOR M J, et al. Fast two-dimensional phase-unwrapping algorithm based on sorting by reliability following anoncontinuous path[J]. Applied Optics, 2002, 41(35): 7437-7444.

备注/Memo

备注/Memo:
收稿日期: 2024-08-14。
基金项目: 国家自然科学基金资助项目(52205552, 52075314)。
作者简介: 闫恪涛(1991—), 男, 山东泰安人, 博士, 讲师。通信联系人: 于瀛洁(1969—), E-mail: yingjieyu@staff.shu.edu.cn
更新日期/Last Update: 1900-01-01