基于DLT模型的摄像机标定简化方法

Simplified method to calibrate cameras based on DLT model

  • 摘要: 在忽略摄像机镜头畸变误差的情况下,通过建立像坐标系与物坐标系,利用DLT模型,结合最优化算法的标定方法和两步法,由于在物坐标系中平面物体的任意一点在Z轴上的分量为0,减少变量数至5个,从而推导出一种简化的摄像机标定方法。借助实例,在文中建立的2个命题的基础上,通过对圆的切线、切线交点及所成像的几何关系确定的的坐标进行参数求解,得到确定模型的二维像和三维像。此外,考虑到更为一般的情况,引入畸变误差(径向误差),对模型先忽略误差项,得到初始参数值,然后利用迭代方法求解。

     

    Abstract: In the case of neglecting the distortion error of the camera lens, the simplified method to calibrate cameras was deduced by establishing the proper coordinate systems, utilizing DLT mode, combining the optimization algorithm calibration method and two-step method, and decreasing the variable number to five. By the aid of a example, the parameters were solved with the tangent line of a circle, the intersection point of tangent and thecoordinates defined by the geometrical relationship of the image based on two propositions proposed in the paper. The numerical example shows that the method has less computation and high accuracy.

     

/

返回文章
返回