GPU-assisted real-time 3-D measurement using speckle-embedded fringe

282

发表时间：2019-07-10 00:00

Shijie Feng, Qian Chen and Chao Zuo

Jiangsu Key Laboratory of Spectral Imaging & Intelligent Sense,

Nanjing University of Science and Technology, Nanjing, Jiangsu Province 210094, China

Abstract

This paper presents a novel two-frame fringe projection technique for real-time, accurate and unambiguous three-dimensional (3-D) measurement. One of the used frames is a digital speckle pattern and the other one is a composite image which is generated by fusing that speckle image with sinusoidal fringes. The involved sinusoidal component is used to obtain a wrapped phase map by Fourier transform profilometry (FTP) and the speckle image helps determine the fringe order for the solved phase map. Compared with traditional methods, the proposed pattern scheme enables measurements of discontinuous surfaces with only two frames, greatly reducing the number of required patterns and thus lowering the sensitivity to movements. This merit makes the method very suitable to inspect dynamic scenes. Moreover, it shows close performance in measurement accuracy compared with the phase-shifting method from our experiments. To process data in real time, a CUDA-enabled Graphics Processing Unit (GPU) is adopted to accelerate some time-consuming computations. With our system, measurements can be performed at 21 frames per second with a resolution of 307K points per frame.

Files:

PDF Media 1 Media 2

Citations

Shijie Feng, Qian Chen, and Chao Zuo, "Graphics processing unit–assisted real-time three-dimensional measurement using speckle-embedded fringe," Appl. Opt. 54, 6865-6873 (2015).

Results

Fig. 1. (a) Captured spekcle image; (b) Captured speckle-embedded sinusoidal pattern; (c) Resultant sinusoidal fringe after the subtruction; (d) Fourier frequncy spectrum; (e) Wrapped phase map solved by FTP.

(a) (b) (c)

Fig. 2. (a) show the obtained fringe order; (b) is the 3-D reconstruction and (c) a plot of the 350th row of the 3-D model.

Fig. 3. Snapshots of real-time measurements of dynamic scenes. (a) Measurement result of a shaking paper (see Media 1); (b) Measurement result of a static box and a moving hand (see Media 2).

Methods

Fig. 1. Typical arrangement of a fringe projection system.

A set of fringe patterns are projected on a measured object and the images of the patterns distorted by the object surface are captured by a camera from different angle.

Fig. 2. The flowchart of our method.

A novel two-frame speckle-embedded fringe projection method is presented. One of the patterns is the speckle image and the other is a composite fringe pattern which is composed of that speckle image and sinusoidal fringes. Through a subtraction of these two images, we can obtain the contained sinusoidal component which is then used to retrieve the phase through FTP. Next to unwrap the solved phase unambiguously, the idea of digital speckle temporal sequence correlation is introduced. Finally, an accurate unwrapped phase map is obtained which is then converted to high distribution.

Fig. 3. The data processing procedures in the host PC and the GPU.

Modern GPU is very efficient at manipulating computer graphics and image processing, and its highly parallel structure makes it more effective than general-purpose CPU for algorithms where processing of large blocks of data is done in parallel. In this work, our programming is based on CUDA which is a parallel computing platform and programming model created by NVIDIA and implemented by GPU.