提出了一种快速高精度的摄像机主动视觉标定方法,建立了摄像机模型并详细分析了其各项参数的求解算法。标定时,令摄像机作一组二维的平移运动,采集圆孔靶标件的图像并计算圆心的像点坐标,同时记录摄像机的移动距离,得到标定所需的特征点。利用这些特征点计算摄像机标定参数,标定精度可达到0.005 mm。利用该方法定制的标定模块实现了摄像机的自动标定。该方法对摄像机运动的限制条件较少,并基本实现了摄像机模型参数的线性求解,为主动视觉系统的摄像机标定提供了一种有效的解决方案。

A rapid and high-precision camera calibration technique based on active vision is proposed. A camera model and detailed analysis on the solution algorithm of its parameter are established. By in-plane translational motion of the camera,images of a hollow target plate are collected to calculate locations of the circular center. Meanwhile,the exact displacements of the camera are recorded;therefore the needed coordinates of the feature points in the object plane and in the correlative image plane are gained. Using these feature points,the camera calibration parameters can be calculated with a precision of 0.005 mm. Finally,automatic calibration is realized by use of the custom calibration module. The proposed method with less limiting conditions of camera motion basically realizes linear solution of the parameters of camera model,and provides an effective solution of camera calibration for active vision systems.