Journal of Experiental Mechanicals

基于运动散斑的物体三维形貌测量方法

On the measurement method of three-dimensional topography of object based on moving speckle

Received:June 20, 2017 Revised:August 28, 2017

英文关键词:stereo vision stereo match spacetime stereo moving speckles

基金项目:

Author Name	Affiliation
SUN Jun-feng	College of Areo Science and Engineering, National University of Defense Technology, Changsha 410073, Hunan, China
XU Yu-hua*	College of Areo Science and Engineering, National University of Defense Technology, Changsha 410073, Hunan, China
ZHAO Xiao-feng	Xi'an Research Institute of Hi-Tech, Xi'an 710025, Shannxi, China
XIE Liang	College of Areo Science and Engineering, National University of Defense Technology, Changsha 410073, Hunan, China
ZHANG Xiao-hu	College of Areo Science and Engineering, National University of Defense Technology, Changsha 410073, Hunan, China

Hits: 548

Download times: 390

中文摘要:

设计了一种基于近红外运动散斑的低成本双目立体视觉系统。该系统将一个低成本散斑投影器安装在一个摇摆运动机构上，向目标表面投射运动的散斑，利用空间-时间立体视觉原理，在立体匹配的时候，不同于传统的立体视觉方法(采用二维的滑动窗口进行对应点查找)，它不仅利用图像像素之间的空间信息，还引入了图像的时间信息，可以采用一个三维的滑动窗口进行立体匹配。与传统的立体视觉方法相比，该方法具有更高的重建精度。此外，本文采用金字塔搜索策略、积分图像和GPU对立体匹配算法进行加速，使得立体匹配的计算速度提高了800多倍。

英文摘要:

A low-cost binocular stereoscopic vision system based on near-infrared (NIR)moving speckle is designed and presented in this paper. In this system, a low cost NIR projector is mounted on a wobbler mechanism to project moving speckles on the surface of the target. The space-time stereo method is used to compute the disparity map by combing the spatial and temporal information. The advantage is that a 3D sliding matching window can be used to establish the pixel correspondences between the images of the left and right cameras. Experimental results show that the method presented outperforms the traditional stereo methods in accuracy. Additionally, the stereo matching algorithm is speed up by more than 800 times by combing the pyramid searching strategy, integral image technique, and the graphics processing unit (GPU).

View Full Text Download reader