高级检索
张然. 一种简单高精度的线结构光传感器标定方法[J]. 安徽工业大学学报(自然科学版),2024,41(3):314-320. doi: 10.12415/j.issn.1671-7872.23095
引用本文: 张然. 一种简单高精度的线结构光传感器标定方法[J]. 安徽工业大学学报(自然科学版),2024,41(3):314-320. doi: 10.12415/j.issn.1671-7872.23095
ZHANG Ran. A Simple and High Precision Calibration Method for Line Structured Light Sensor[J]. Journal of Anhui University of Technology(Natural Science), 2024, 41(3): 314-320. DOI: 10.12415/j.issn.1671-7872.23095
Citation: ZHANG Ran. A Simple and High Precision Calibration Method for Line Structured Light Sensor[J]. Journal of Anhui University of Technology(Natural Science), 2024, 41(3): 314-320. DOI: 10.12415/j.issn.1671-7872.23095

一种简单高精度的线结构光传感器标定方法

A Simple and High Precision Calibration Method for Line Structured Light Sensor

  • 摘要: 线结构光三维视觉测量是获取物体表面三维数据的主要方法之一,传统的线结构光标定结果完全依赖提取光条纹特征点的数量,实际操作中无法保证提取足够多的特征点来拟合平面。针对这一问题,设计1种新的线结构光标定方法,将相机中心与结构光光条纹中心的任一特征点连成直线,不需求得光条纹中心线所有点的坐标,只需找到光条纹与棋盘格格边的交点;将其作为光条纹的特征点,控制特征点的数量,基于一次拟合特征点在像素坐标系下的直线方程,再联立直线方程与棋盘格平面方程,得到相机坐标系下的特征点坐标,依次求得3条及以上光条纹的中心坐标;利用多条不重合的直线拟合平面,采用二次拟合的方法提高标定精度。通过实验室搭建的线结构光单目相机对本文方法进行标定实验,结果表明:特征点到光平面的平均距离为0.052 mm,该距离在标定误差范围内,本文方法可有效解决因特征点较少而导致标定误差大的问题。

     

    Abstract: Line structured light 3D vision measurement is one of the main methods for obtaining 3D data on the surface of an object, the traditional line structured light calibration results rely entirely on the number of feature points on the extracted light streak, and in practice operations, it cannot guarantee the extraction of enough feature points to fit the plane. To address this problem, a new line structured light calibration method was proposed. The center of the camera was connected to any feature point in the center of the structured light stripe as a straight line, and there was no need to obtain the coordinates of all the points in the center line of the light stripe, but only to find the intersection of the light stripe with the checkerboard grid edge, which was used as the feature point of the light stripe.The number of such points was controlled, and a single linear regression was performed to obtain the equation of these characteristic points in pixel coordinates. Subsequently, this linear equation was combined with the equation of the chessboard plane to derive the corresponding feature point coordinates in the camera coordinate system. This process was repeated for three or more non-overlapping light stripes, enabling the fitting of a plane using multiple lines and enhancing calibration accuracy through a second-order regression. An experimental calibration was conducted using a line-structured light monocular camera setup in a laboratory environment. The results show an average distance of 0.052 mm between the characteristic points and the light plane, which falls within the acceptable calibration error range. This demonstrates that the proposed method can effectively solve the problem of large calibration errors caused by a limited number of feature points.

     

/

返回文章
返回