一种实用的平板高精度摄像机标定方法

在分析现有平板标定法所存在问题及完整摄像机模型的基础上,提出了一种实用的平板标定方法;首先,针对方格在进行亚像素角点提取时可能出现奇异点的问题对所有角点进行SVD分解;在未考虑畸变的情况下通过最小二乘法求取相机的初始内外参数;然后使用求取出来的初始内外参数及世界坐标求出理想图像坐标,根据畸变模型建立理想图像坐标和实际图像坐标之间的方程,解出畸变系数;最后采用LM算法作全局非线性优化,并根据反投影误差来控制角点校正及迭代运算,得到视觉系统更精确的内外参数;实验表明,该方法在对图像中心领域附近空间距离测量精度能到达0.0526mm,能满足高精度视觉测量及其它方面应用,迭代一次的情况下精度优于张正友方法。

Practical High Precision Camera Calibration Method from a Plane

On the basis of analyzing the problem of present flat panel method and the complete camera model,a practical flat panel calibration method is presented.First,considering the problem of singular points when using the sub-pixel method to locate the corner,SVD is used to solve it.The initial camera parameters are computed using Root Mean Square and close-form algorithm on a distortion-free camera model.Then,estimate the distortion coefficient according to the relationship between the real image coordinate and ...