用三线法标定车载摄像机的外部参数

Calibration of external parameters of vehicle-mounted camera with trilinear method

根据摄像机透视投影原理,三条互不重合的平行线在图像平面上具有相同的消失点和不同的斜率。按照针孔成像模型,经过数学推导可以建立摄像机的外部参数与消失点和成像斜率之间的解析表达式,无需经过迭代和优化计算。实现过程只需要地面上有三条平行线,然后通过手工或自动的方式确定出这三条直线成像的交点和斜率,就可以标定出摄像机的外部参数。实验表明,该方法可以在不同环境下进行车载摄像机标定,方位角和俯仰角的标定误差小于百分之一弧度。

In accordance with the perspective projection principle of camera, three mutually non-coincident parallel lines have same vanishing point and different slopes on imaging plane. Without iteration and optimization calculation, an analytical expression of the external parameters of a camera, vanishing point and imaging slopes can be established through mathematical derivation according to pinhole imaging model. The calibration of external parameters of a camera can be performed only needing to draw three parallel lines on the ground and determine the crossover points and slopes through manual or automatic schemes. The experiments show that the method can be applied to calibration for vehicle-mounted camera, the calibrating error for azimuth and pitch angle is less than 1/100 radian.