GPU加速三维面形测量

赵亚龙; 刘守起; 张启灿

doi:10.3788/IRLA201847.0317003

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

GPU加速三维面形测量

赵亚龙, 刘守起, 张启灿

文章导航 > 红外与激光工程 > 2018 > 47(3): 317003-0317003(7)

赵亚龙, 刘守起, 张启灿. GPU加速三维面形测量[J]. 红外与激光工程, 2018, 47(3): 317003-0317003(7). doi: 10.3788/IRLA201847.0317003

引用本文:

赵亚龙, 刘守起, 张启灿. GPU加速三维面形测量[J]. 红外与激光工程, 2018, 47(3): 317003-0317003(7). doi: 10.3788/IRLA201847.0317003

Zhao Yalong, Liu Shouqi, Zhang Qican. 3D shape measurement accelerated by GPU[J]. Infrared and Laser Engineering, 2018, 47(3): 317003-0317003(7). doi: 10.3788/IRLA201847.0317003

Citation:

Zhao Yalong, Liu Shouqi, Zhang Qican. 3D shape measurement accelerated by GPU[J]. Infrared and Laser Engineering, 2018, 47(3): 317003-0317003(7). doi: 10.3788/IRLA201847.0317003

GPU加速三维面形测量

doi: 10.3788/IRLA201847.0317003

1.
四川大学电子信息学院,四川成都 610064

基金项目:

国家自然科学基金（61675141）;国家重大科学仪器设备开发专项（2013YQ490879）

详细信息

作者简介:
赵亚龙(1990-),男,硕士生,主要从事三维面形测量方面的研究。Email:616475529@qq.com

中图分类号: TN247

3D shape measurement accelerated by GPU

1.
School of Electronics and Information Engineering,Sichuan University,Chengdu 610064,China

摘要: 随着通用计算和图形显示需求的不断增加，图形处理器（Graphics Processing Unit，GPU）在医学、科学计算、图像处理等领域得到了广泛的应用。但它在三维测量领域的应用还只是一个开始。文中基于傅里叶变换轮廓术（Fourier Transform Profilometry，FTP）和三频外差法设计了两套三维测量系统，并利用计算统一设备架构（Compute Unified Device Architecture，CUDA）方法，加速了静态或动态物体的三维重建。在三频外差测量系统中，需要利用高速数字投影模块和相机，同步触发采集小视场表面的12个变形条纹图，然后对图像数据进行处理。实验结果表明：对12幅1 360 pixel1 024 pixel大小的图像进行相位展开运算，GPU方法比CPU方法的效率提高了2 089倍。在基于FTP方法的测量系统中，摄像机只需记录一幅变形条纹图，然后拷贝到显存中，并用CUDA编程的算法进行处理，进而重建出物体的三维面形。基于GPU的FTP方法对一幅1 024 pixel1 280 pixel大小的图像进行计算，其计算时间比CPU方法缩短了27倍。
- 三维面形测量 /
- GPU /
- 条纹投影 /
- 傅里叶变换轮廓术 /
- 三频外差
Abstract: Driven by the increasing demands of the general purpose in computation and image display, Graphics Processing Unit(GPU) has been developed and used in many fields, such as medical field, scientific calculations, image processing etc.. But, its application in 3D shape measurement is still a beginning. In this paper, two 3D shape measurement systems based on Fourier Transform Profilometry(FTP) and tri-frequency heterodyne method were implemented with Compute Unified Device Architecture(CUDA) technology to speed up their 3D shape construction of a measured static or dynamic object. In the first 3D shape measuring system based on tri-frequency heterodyne method, a high-speed digital projection module and a synchronously triggered camera were used to record 12 deformed fringe images on the surface of a small object. The experimental result demonstrates that the efficiency of the unwrapping phase calculation by GPU is improved 2 089 times than that of CPU for doing same task on 12 images with 1 360 pixel1 024 pixel each. In the second system based on FTP, only one deformed fringe image was recorded by a camera, then transferred into GPU and processed by the programmed CUDA algorithm to restore the corresponding 3D shape. Compared with the traditional processing method by CPU, the time consumption of FTP method completed by GPU is shortened 27 times for a 1 024 pixel1 280 pixel image.
- 3D shape measurement /
- GPU /
- fringe projection /
- Fourier transform profilometry /
- tri-frequency heterodyne

[1]	Tang Wei, Ye Dong. 3D computer vision measurement systems[J]. Infrared and Laser Engineering, 2008, 37(S1):328-332. (in Chinese)
[2]	Tian Qingguo, Ge Baozhen, Du Pu, et al. Measurement of human figure size based on laser 3D scanning[J]. Optics and Precision Engineering, 2007, 15(1):84-88. (in Chinese)
[3]	Ye Haijia, Chen Gang, Xing Yuan. Stereo matching in 3D measurement system using double CCD structured light[J]. Optics and Precision Engineering, 2004, 12(1):71-75. (in Chinese)
[4]	Zhang S, Royer D, Yau S T. GPU-assisted high-resolution, real-time 3-D shape measurement[J]. Optics Express, 2006, 14(20):9120-9129.
[5]	Nikolaus Karpinsky, Morgan Hoke, Vincent Chen, et al. High-resolution, real-time three-dimensional shape measurement on graphics processing unit[J]. Optical Engineering, 2014, 53(2):024105-1-8.
[6]	Lu Jin. The surface morphology measurement system based on GPU[D]. Hangzhou:Zhejiang University, 2011. (in Chinese)
[7]	Zhang Shu, Chu Yanli, Zhao Kaiyong, et al. GPU Performance Computing of CUDA[M]. Beijing:China Waterpower Press, 2009:11-13.
[8]	Chen Qian, Qiu Yuehong, Yi Hongwei. Star image registration algorithm based on GPU parallel program design[J]. Infrared and Laser Engineering, 2014, 43(11):3756-3761. (in Chinese)
[9]	Wang Xinhua, Wang Xiaokun. Real time image mosaic of the transient gigapixel imaging system[J]. Chinese Optics, 2015, 8(5):785-793. (in Chinese)
[10]	Huntley J M, Salder H O. Shape measurement by temporal phase unwrapping:comparison of unwrapping algorithms[J]. Measurement Science Technology, 1997, 8(9):986-992.
[11]	Carsten Reich, Reinhold Ritter, Jan Thesing. White light heterodyne principle for 3D-measurement[C]//SPIE, 1997, 3100:236-344.
[12]	Takeda M, Mutoh K. Fourier transform profilometry for the auto measurement of 3-D object shapes[J]. Applied Optics, 1983, 22(24):3977-3982.
[13]	Lei Gundong, Lu Yinhuan, Wang Ruli. Adaptive main frequency bandpass filters used in Fourier transform profilometry[J]. Chinese Journal of Optics and Applied Optics, 2010, 3(3):245-251. (in Chinese)

[1]	张庆辉, 李浩, 吕磊, 卢盛林, 潘威. 基于相移轮廓术的双采样运动物体三维重构 . 红外与激光工程, 2023, 52(12): 20220891-1-20220891-6. doi: 10.3788/IRLA20220891
[2]	张宗华, 李雁玲, 高峰, 高楠, 孟召宗, 蒋向前. 面向结构光三维测量的相位展开技术综述（特邀） . 红外与激光工程, 2023, 52(8): 20230126-1-20230126-23. doi: 10.3788/IRLA20230126
[3]	朱坡, 张宗华, 高楠, 高峰, 王张颖. 彩色高反光物体表面三维形貌测量技术 . 红外与激光工程, 2023, 52(7): 20220761-1-20220761-7. doi: 10.3788/IRLA20220761
[4]	杨静雯, 张宗华, 付莉娜, 李雁玲, 高楠, 高峰. 利用抖动算法扩展深度范围的三维形貌测量术 . 红外与激光工程, 2023, 52(8): 20230059-1-20230059-10. doi: 10.3788/IRLA20230059
[5]	王张颖, 张宁宁, 高楠, 李奎, 孟召宗, 张宗华. 基于单色条纹投影的高动态范围物体表面形貌三维测量 . 红外与激光工程, 2023, 52(8): 20230327-1-20230327-9. doi: 10.3788/IRLA20230327
[6]	刘飞, 罗惠方, 江翰立, 张茵楠, 严谨. 改进的三频三步相移结构光三维重建方法 . 红外与激光工程, 2022, 51(4): 20210179-1-20210179-9. doi: 10.3788/IRLA20210179
[7]	朱其幸, 王道档, 卢毅伟, 孔明, 许新科. 双光纤点衍射干涉投影系统误差校正及优化 . 红外与激光工程, 2022, 51(3): 20210140-1-20210140-8. doi: 10.3788/IRLA20210140
[8]	吴荣, 赵世丽, 赵洋, 谢锋云. 条纹投影用于不同景深物体的三维测量 . 红外与激光工程, 2022, 51(11): 20220088-1-20220088-10. doi: 10.3788/IRLA20220088
[9]	殷永凯, 张宗华, 刘晓利, 彭翔. 条纹投影轮廓术系统模型与标定综述 . 红外与激光工程, 2020, 49(3): 0303008-0303008-18. doi: 10.3788/IRLA202049.0303008
[10]	王玉伟, 陈向成, 王亚军. 改进的双频几何约束条纹投影三维测量方法 . 红外与激光工程, 2020, 49(6): 20200049-1-20200049-7. doi: 10.3788/IRLA20200049
[11]	张钊, 韩博文, 于浩天, 张毅, 郑东亮, 韩静. 多阶段深度学习单帧条纹投影三维测量方法 . 红外与激光工程, 2020, 49(6): 20200023-1-20200023-8. doi: 10.3788/IRLA20200023
[12]	李勇, 张广汇, 马利红, 应晓霖, 姚建云. 条纹投影动态三维表面成像技术综述 . 红外与激光工程, 2020, 49(3): 0303005-0303005-13. doi: 10.3788/IRLA202049.0303005
[13]	张宗华, 于瑾, 高楠, 孟召宗. 高反光表面三维形貌测量技术 . 红外与激光工程, 2020, 49(3): 0303006-0303006-13. doi: 10.3788/IRLA202049.0303006
[14]	党源源, 王昕. CPU-GPU异构系统在光学遥感影像处理中的应用 . 红外与激光工程, 2020, 49(S1): 20200092-20200092. doi: 10.3788/IRLA20200092
[15]	张启灿, 吴周杰. 基于格雷码图案投影的结构光三维成像技术 . 红外与激光工程, 2020, 49(3): 0303004-0303004-13. doi: 10.3788/IRLA202049.0303004
[16]	冯世杰, 左超, 尹维, 陈钱. 深度学习技术在条纹投影三维成像中的应用 . 红外与激光工程, 2020, 49(3): 0303018-0303018-17. doi: 10.3788/IRLA202049.0303018
[17]	陈超, 於燕琴, 黄淑君, 刘晓利, 徐静, 张宗华. 三维小视场成像系统 . 红外与激光工程, 2016, 45(8): 824002-0824002(6). doi: 10.3788/IRLA201645.0824002
[18]	陈茜, 邱跃洪, 易红伟. 基于GPU的星图配准算法并行程序设计 . 红外与激光工程, 2014, 43(11): 3756-3761.
[19]	王茂芝, 郭科, 徐文皙. 基于集群和GPU的高光谱遥感影像并行处理 . 红外与激光工程, 2013, 42(11): 3070-3075.
[20]	石坤, 郝颖明, 王明明, 付双飞. 海面背景红外实时仿真 . 红外与激光工程, 2012, 41(1): 25-29.

点击查看大图

计量

文章访问数: 614
HTML全文浏览量: 116
PDF下载量: 92
被引次数: 0

GPU加速三维面形测量

doi: 10.3788/IRLA201847.0317003

1. 四川大学电子信息学院,四川成都 610064

作者简介:
赵亚龙(1990-),男,硕士生,主要从事三维面形测量方面的研究。Email:616475529@qq.com

收稿日期: 2017-10-05
修回日期: 2017-11-15
刊出日期: 2018-03-25

基金项目:

国家自然科学基金（61675141）;国家重大科学仪器设备开发专项（2013YQ490879）

中图分类号: TN247

关键词:

摘要: 随着通用计算和图形显示需求的不断增加，图形处理器（Graphics Processing Unit，GPU）在医学、科学计算、图像处理等领域得到了广泛的应用。但它在三维测量领域的应用还只是一个开始。文中基于傅里叶变换轮廓术（Fourier Transform Profilometry，FTP）和三频外差法设计了两套三维测量系统，并利用计算统一设备架构（Compute Unified Device Architecture，CUDA）方法，加速了静态或动态物体的三维重建。在三频外差测量系统中，需要利用高速数字投影模块和相机，同步触发采集小视场表面的12个变形条纹图，然后对图像数据进行处理。实验结果表明：对12幅1 360 pixel1 024 pixel大小的图像进行相位展开运算，GPU方法比CPU方法的效率提高了2 089倍。在基于FTP方法的测量系统中，摄像机只需记录一幅变形条纹图，然后拷贝到显存中，并用CUDA编程的算法进行处理，进而重建出物体的三维面形。基于GPU的FTP方法对一幅1 024 pixel1 280 pixel大小的图像进行计算，其计算时间比CPU方法缩短了27倍。

English Abstract

3D shape measurement accelerated by GPU

1. School of Electronics and Information Engineering,Sichuan University,Chengdu 610064,China

Received Date: 2017-10-05
Rev Recd Date: 2017-11-15
Publish Date: 2018-03-25

Keywords:

Abstract: Driven by the increasing demands of the general purpose in computation and image display, Graphics Processing Unit(GPU) has been developed and used in many fields, such as medical field, scientific calculations, image processing etc.. But, its application in 3D shape measurement is still a beginning. In this paper, two 3D shape measurement systems based on Fourier Transform Profilometry(FTP) and tri-frequency heterodyne method were implemented with Compute Unified Device Architecture(CUDA) technology to speed up their 3D shape construction of a measured static or dynamic object. In the first 3D shape measuring system based on tri-frequency heterodyne method, a high-speed digital projection module and a synchronously triggered camera were used to record 12 deformed fringe images on the surface of a small object. The experimental result demonstrates that the efficiency of the unwrapping phase calculation by GPU is improved 2 089 times than that of CPU for doing same task on 12 images with 1 360 pixel1 024 pixel each. In the second system based on FTP, only one deformed fringe image was recorded by a camera, then transferred into GPU and processed by the programmed CUDA algorithm to restore the corresponding 3D shape. Compared with the traditional processing method by CPU, the time consumption of FTP method completed by GPU is shortened 27 times for a 1 024 pixel1 280 pixel image.