采用同心圆图像的摄像机标定方法

Camera Calibration from Concentric Circles Images

为研究计算机视觉问题,介绍了一种新的采用同心圆图像进行摄像机内参数标定的算法,在标定中,采用了针孔摄像机模型。与已有的方法不同,算法以3D圆的投影方程为基础,包含了摄像机的内部参数矩阵。根据其在代数上的特殊形式,可以得到关于摄像机内部参数的一个线性方程组,通过创建足够的方程组,可以很容易求得所有的摄像机内部参数。方法是完全线性的算法,无需迭代优化,无需求取虚圆点,仅需要三幅两同心圆的图像即可实现全部摄像机内部参数的标定。最后通过仿真验证了内参数标定算法的准确性和鲁棒性。

A new algorithm for camera calibration using two concentric circles was proposed. For the calibration, the pinhole camera model was used. Different from previous methods, it takes the projective equations of 3 - D circles, which include the intrinsic parameter matrix of the camera, as the basis of our calibration approach. According to the special structure of the projective equations in algebra, a linear equation system about the intrinsic parameters can be obtained. After constructing enough equations, the calibration can be easily finished. This method is a linear approach and only needs three images of the two concentric circles to recover all the five intrinsic parameters. Experiments using computer simulated data demonstrate the robustness and accuracy of this method.