• CN:11-2187/TH
  • ISSN:0577-6686

›› 2010, Vol. 46 ›› Issue (18): 7-12.

• 论文 • 上一篇    下一篇

基于一维靶标的结构光视觉传感器标定

周富强;蔡斐华   

  1. 北京航空航天大学精密光机电一体化技术教育部重点实验室
  • 发布日期:2010-09-20

Calibrating Structured-light Vision Sensor with One-dimensional Target

ZHOU Fuqiang;CAI Feihua   

  1. Key Laboratory of Precision Opto-mechatronics Technology of Ministry of Education, Beihang University
  • Published:2010-09-20

摘要: 针对结构光视觉三维测量模型参数的现场标定,提出一种基于自由移动的一维靶标(至少包含3个共线特征点,取其中之一为特征线的原点)的结构光视觉传感器标定方法。根据一维射影变换获取靶标上特征线的消隐点,并与摄像机投影中心确定特征线在摄像机坐标系下的方向矢量;基于特征点的长度约束及方向约束计算特征点的摄像机坐标,得到特征线的直线方程;利用射影变换和特征线的方程获得多个非共线的光条上控制点的摄像机坐标,将控制点拟合成光平面。试验表明,在一般试验条件下,一维靶标结构光参数标定方法可以达到平面靶标法的标定精度。相对于二维靶标,高精度大尺寸一维靶标加工制造容易,维护简单,更重要的是可以用于大尺寸结构光视觉测量的现场在线标定。

关键词: 标定, 结构光视觉, 射影变换, 消隐点, 一维靶标

Abstract: In order to calibrate the three-dimensional measurement model parameters of structured-light vision in situ, a novel approach to easily calibrating a structured-light vision sensor with free-moving one-dimensional target (at least 3 feature-points aligned on a feature-line and one of them chosen as the origin) is proposed. In the camera coordinate frame, the vanishing point of the feature-line can be obtained on the basis of 1D projective transformation, with at least three collinear feature-points on the feature-line, so the direction vector of the feature-line can be determined by joining the projective center and the vanishing point. The camera coordinates of the origin can be computed in accordance with the direction of the feature-line and the known distance between the origin and the other feature-point. The camera coordinates of a control point on the light-stripe can be calculated by the feature-line equation and its distance to the origin which is calculated with its projective coordinates based on 1D projective transformation. The light plane is fitted with at least three nonlinear control points in the camera coordinate frame. The experimental results prove that the proposed approach can achieve the same calibration accuracy as the planar-target-calibration approach under general laboratory conditions. Compared to planar target, 1D target with high accuracy and large dimensions is manufactured more easily and maintained more simply. Moreover, the proposed technique is also important in practice especially when calibrating structured-light vision sensor with large-scale measuring range.

Key words: Calibration, One-dimensional target, Projective transformation, Structured-light vision, Vanishing point

中图分类号: