基于改进ArUco标签的位姿估计方法

Pose Estimation Method Based on Improved ArUco Tags

针对使用ArUco标签在远距离测量时精度不高的问题,提出改进ArUco标签的位姿估计方法。该方法在ArUco标签的基础上设计了阵列特征圆环,提取圆环的中心点并对其进行排序,根据交比不变特性判断出圆心检测的准确性;利用加权最小二乘法建立ArUco标签角点与阵列圆环圆心之间的直线拟合关系,实现角点修正;采用EPnP算法解算出改进ArUco标签的位姿参数,并利用LM算法对位姿进行优化,提高位姿测量精度。设计了机械臂视觉定位系统验证所提出方法的可行性和有效性。实验结果表明,此改进方法在2.5~4 m范围内测量的最大平移误差为8.72 mm,最大旋转误差为1.78°。所提出的方法在4 m处纵向距离测量的相对误差降低0.69%,提高了测量精度。

Aiming at the problem of low accuracy when using ArUco tags for long-distance measurement,an improved pose estimation method is proposed.The method designs array feature rings based on ArUco tags,the center points of the rings are extractd and sortd,and the accuracy of circle center detection are judged based on the cross-ratio invariant characteristics.The weighted least squares method is used to establish a straight line fitting relationship between the corner points of the ArUco tag and the center of the array rings to achieve corner point correction.The EPnP algorithm is used to calculate the pose parameters of the improved ArUco tag,and the LM algorithm is used to optimize the pose and improve the pose measurement accuracy.A robotic arm visual positioning system is designed to verify the feasibility and effectiveness of the proposed method.Experimental results show that the maximum translation error measured by this method in the range of 2.5~4 m is 8.72 mm,and the maximum rotation error is 1.78°,and the maximum rotation error is 4.65°.The proposed method reduces the relative error of longitudinal distance measurement by 0.69% at 4 m,improving the measurement accuracy.