视觉测量中光学测头姿态估计方法

基于光学测头特征点成像的单机视觉坐标测量系统具有测量空间大、测量精度高及可进行现场测量等优点,是替代传统的大型坐标测量机的最有发展前途的技术方案之一.为实现该测量方法,必须先解决光学测头姿态估计问题,为此,提取一种利用SVD与矢量观测值的姿态确定方法.首先,通过基于SVD的5点优化算法确定特征点的空间坐标;然后,利用特征点的测头坐标和空间坐标建立关于测头坐标矢量、像机坐标矢量的矩阵方程,优化求解矩阵方程确定光学测头(测头坐标系)相对于像机坐标系的姿态.该算法利用所有数据冗余信息,提高了测量系统的精度和稳定性.实验证明,该算法切实可行,系统的单点重复性及测量不确定度<0.1 mm.

Optical probe pose estimation in the vision measurement

Single camera vision coordinates measurement based on feature points imaging of optical probe has the characteristics such as large measurement space,high measurement precision,measurement on-line and so on,so it is one of good foreground technology schemes that may replace traditional large-scale coordinates measurement machine.But in order to realize the measurement method,the problem of optical probe pose estimation must be resolved firstly.So attitude determination method using SVD and vector observation is presented.Firstly,space coordinates of optical feature points are acquired by 5 points optimized algorithm using SVD.Then,the attitude of optical probe relative to camera coordinates system is determined using the vector observations which are defined using the coordinates of feature points in the camera coordinates system and the probe coordinates system.The method makes use of all the redundancy data information,measurement precision and stability of measurement system are improved.Experiments have proved that the method is feasible,the deviation of repeatability and measurement uncertainty is less than 0.1 mm.