基于两个正交一维物体的单幅图像相机标定

Camera Calibration with Single Image Based on Two Orthogonal One-Dimensional Objects

提出了一种利用两个正交一维物体构成"T"型靶标进行摄像机标定的新方法。该方法只需对"T"型靶标上已知坐标的5点投影一幅图像,然后根据柔性靶标原理计算出由虚点和标记点组成的共直线的4点,由射影变换同素性、接合性以及交比不变性标定出镜头的一阶径向畸变参数。利用已知畸变参数对图像进行畸变校正,然后由基于两个正交一维物体坐标变换的方法即可标定出相机的内外参数。该方法线性求解镜头畸变参数,避免了传统方法非线性迭代优化过程中产生的参数耦合现象。实验表明,不进行镜头畸变校正则相机标定精度随着图像噪声的增加呈不稳定状态;进行畸变校正后对简单标定计算的初始值进行优化得到稳定的高精度标定结果。整个实验设备简单,操作方便,只需一幅图像即可实现镜头畸变和相机内外参数的标定,可以达到实时的效果。

A new method for camera calibration is proposed by utilizing a T-shaped target of two orthogonal one- dimensional （1D） objects. Through this method, only one projected image of 5 points with known coordinates on the T-shaped target is needed. Then according to the flexible target principle, 4 collinear points composed by the virtual and marked points can be determined, and finally the first-order radial distortion parameters of the lens are calibrated through the primality, associativity and cross-ratio invariability of the projective transformation. Using the calibrated distortion parameters to correct the image distortions, it is possible to calibrate the intrinsic and extrinsic parameters of the camera on the basis of coordinate transformation of the two orthogonal 1D objects. This method, using the linear solution to gain the distortion parameters of the lens, is able to prevent the parameters coupling in the nonlinear iterative optimization process in traditional methods. Experiments prove that if the distortion correction of the camera lens is not conducted, the camera calibration result will be unstable with the increase of image noises; while after conducting the distortion correction, a stable calibration result with high precision will be ensured when the initial values of a simple calibration are optimized. The calibration experiment is simple and easy to operate. Only one image is needed in realizing the calibration of the camera lens distortion and of the intrinsic and extrinsic parameters, thus the real-time operation can be realized.