基于椭圆拟合的圆的摄像机标定方法

摘要

This paper proposes a approach for camera calibration based on conic and ellipse fitting. The approach only requires camera to take a planar template which contains a circle and its any 3 groups orthogonal diameters is used and taken 3 or more figures from different orientations. Through the nonlinear least squares of ellipse fitting gets accurate elliptic boundary, thus locate the feature points in the pattern precisely, further improve the calibration precision of the camera. Computer simulation and real experiment prove the feasibility and the robustness of this approach.%本文提出了基于椭圆拟合的圆的摄像机标定方法。该方法的平面模板要求圆内有3组两两正交的直径，从3个或多个不同的方位摄取平面圆模板的图像。通过非线性的最小二乘的椭圆的拟合，得到精确的椭圆边界，从而精确定位模板中的特征点，进一步提高摄像机的标定精度。真实图像实验表明，该方法易于实现，精度较高，稳定性较好。