基于两相同圆的自标定方法

Self-calibration using two same circles

摄像机标定在机器视觉中有着十分重要的作用。本文提出了一种基于两相同圆的自标定方法。首先,用摄像机从不同角度拍摄三幅包含两相同圆的图像,根据圆环点的不变性求取摄像机内部参数。然后根据透视投影变换的相切不变性,利用两圆的内外公切线求得摄像机的外部参数。该方法不需要任何已知的摄像机参数和烦琐的标志点匹配过程,简单方便地求解摄像机的内部参数和外部参数。大量的实验验证了该方法的精确性,鲁棒性和适用的广泛性。

Camera calibration plays a very important role in machine vision.A novel self-calibration method using two same circles is proposed.First,three photos containing two same circles are captured by the camera from different views.Then the intrinsic parameters of the camera are obtained from the invariance of the circular points.According to the invariance of tangency in projective projection,the extrinsic parameters of the camera are solved from the common tangents of the two circles.The proposed method could get intrinsic parameters as well as exterior parameters without any known camera parameters and complicated matching processes.Extensive experiment results validate the accuracy,robustness and widespread applicability of this method.