基于相位映射的三维主动视觉标定

Calibration Method for 3-D Vision System Using Phase Mapping

通过在结构光发射单元和成像单元间施加几何约束条件,建立统一的坐标系。利用条纹投影和相位映射技术,构造了将编码相位图映射为物体的三维空间坐标的数学模型,进而获得深度图像空间坐标的计算值,然后将其与物体空间的三维标定数据基准进行比较,建立目标函数为误差平方和最小的非线性优化方程。两步法迭代求解这个优化方程,最终获得三维系统的结构参数。实验结果证明了本文提出的三维成像系统标定理论和方法的有效性。

A calibration method for three-dimensional sensing scheme based on phase mapping is described. The calibration procedure is based on the definition of a unique coordinate system in the structure system,and thus,a constraint is introduced between two devices. Fringe projection and phase mapping are used to convert the phase map to world coordinates of the object surface by a matbematieal model. Using the model,the 3-D coordinates of range image can be obtained with moderate accuracy,which will be compared with the prescribed benchmark in order to build up a square-error based objective function, leading to a non-linear equation. A two-step procedure is developed to solve the equation, and all structural parameters involved in 3-D sensing system can be detemained with appropriate accuracies. Experimental results show that these proposed approaches are accurate and robust.