摄像机标定中靶标圆心像点坐标的精确计算

Calculation exact center coordinate of a target circle in camera calibration

摄像机标定是机器视觉中的一个关键问题。当靶标特征像点在世界坐标系中的坐标值确定后,便可以计算得到内参和外参,进而完成摄像机标定。为了简化靶标特征像点的计算方法并提高其计算精度,提出了基于透视不变性的计算方法来得到虚圆心像点坐标。该方法利用计算机搜索算法找出3个圆心不共线的圆形靶标的椭圆像的公切线,每个椭圆上两对切点连线的交点即为对应靶标虚圆心的像点。模拟实验验证了该方法的实用性。结果表明:算法不存在模型误差,计算过程简洁高效,提高了靶标特征像点坐标的计算精度,使摄像机的标定更加精确。

Camera calibration is a key problem for machine vision.When the target characteristic points coordinates in world coordinates system are confirmed,the internal and external parameters are calculated and the camera calibration is implemented.In order to simplify the algorithms and improve the accuracy of the calculation,an algorithm based on the invariants of projection for calculating coordinates of suppositional centers was presented.This algorithm is to use computer visual search to get all common tangents of three ellipses images of corresponding circular target whose centers were not on one beeline.The intersection of two beelines determined by two pairs of tangency points on one ellipse was just the center of the corresponding circle on the target image.The practicability of this algorithm is illustrated by the simulation experiment.The result shows that this algorithm has no model error,and the numerical procedure is simplified.The calculating precision of the target characteristic points coordinates has thus been improved,and the camera calibration will be more accurate.