采用棋盘格靶标的全自动相机标定方法

为了去除相机标定过程中的人为干预，提出了一种采用改进的棋盘格靶标的全自动相机标定方法。识别出每幅标定图像中的四个标志圆，利用四个标志圆圆心的图像坐标和物理坐标计算二维射影变换矩阵。依据该射影变换矩阵计算出棋盘格角点的初始图像位置，接着提取亚像素精度的角点位置。迭代求解需要标定的相机参数。由实验可知，该全自动相机标定方法的棋盘格角点识别能力和相机标定精度，与Bouguet的相机标定工具箱相当，且可以显著地减少标定时间和工作量。利用20幅分辨率为640×480的靶标图像标定相机仅需16 s。

Fully automatic method for camera calibration based on checkerboard target

In order to get rid of the human intervention in the camera calibration process, this paper proposes a fully automatic method taking advantage of an improved planar checkerboard target. The four circular makers are recognized. Thus a 2D projective transformation matrix can be uniquely determined by the image coordinates and correspondingly 2D physical coordinates of the four circular makers. The initial coordinates of all checkerboard corners are calculated using the projective matrix, followed by a refinement step to obtain the sub-pixel corner coordinates. The camera parameters are solved by successive iterations. According to the experiments, the proposed automatic calibration method not only has a capability that can identify all the checkerboard corners in images, and a good calibration accuracy as well as Bouguet’s camera calibration toolbox does, but also significantly reduces the time and effort spent in the calibration process. As shown in the experiments, it takes only 16 seconds for the proposed method to calibrate a camera using 20 images with a resolution of 640×480.