一种新的非线性相机模型标定方法

A New Method of Calibration for Non-Linear Camera Model

提出一种新的非线性相机模型标定方法.该方法考虑各种成像误差,用二元函数的多项式建立成像误差模型,用一个中间变量代替非线性畸变因素.在已知至少6个不共面点的空间坐标及对应的图像坐标,利用线性算法求解出非线性畸变模型的投影矩阵,用修正后的投影矩阵进行三维重建.实验结果表明,整个修正过程避免非线性方程的求解,10个修正参数由线性方程组一并解出,修正模型三维重建的精度整体得到提高.

A new method is proposed for the non-linear model camera calibration. In this method, all kinds of imaging errors are considered. The imaging error module is established in the form of a binary function polynomial, where nonlinear distortion is replaced by a variable parameter and projection matrix is answered with linear algorithms on the condition that at least six non-coplanar points＇ world coordinates and the corresponding image coordinates are given. This calibrated projection matrix is utilized in 3-dimensional reconstruction. The experimental results show that the precision of the callbrated model can be improved greatly because ten calibrated parameters can be solved by a set of linear equations instead of solving nonlinear equations during the calibration process.