一种高精度线结构光视觉传感器现场标定方法

An Accurate Calibration Method for Line Structured Light Vision Sensor

针对现有线结构光视觉传感器标定方法存在的局限性,提出一种不需要求解光平面标定点的标定方法。根据光条图像求解平面靶标上光条在摄像机坐标系下的Plücker矩阵。在视觉传感器前合适位置将平面靶标摆放多次,联立所有光条空间直线的Plücker矩阵,求解光平面在摄像机坐标系的平面方程。最后通过非线性优化方法得到光平面方程在最大似然准则下的最优解。在标定过程中,所有光条点都参与光平面参数的计算过程,因此该方法标定结果精度高、稳健性强。实验证明,与现有方法相比该方法标定精度提高30%左右。

A calibration method for line structured light vision sensor that does not require extraction of control points in the light plane is proposed to solve the calibration problem of line structured light vision sensor. The planar target is placed in front of the vision sensor several times. The P1/icker matrix of light strip on the target plane is computed. The equation of the light plane in the camera coordinate frame can be solved by combining the Plficker matrixs of the light stripe on the target plane at different positions. The maximum likelihood estimation of the light plane equation is obtained through nonlinear optimization. The proposed method is robust and accurate, because all of the light stripe image points are used to determine the equation of the light plane. Experimental results show that, compared with that of existing calibration methods, the calibration accuracy of the proposed method improves about 30 % .