【摘要】 相机内参及相机与激光雷达外参标定是两者融合感知的前提，为解决平面标定方法存在的过拟合问题，并提高相机与低线束激光雷达的外参标定精度，提出一种基于三维标定板的相机内参及相机/激光雷达外参的联合标定方法。首先通过三维标定板校准相机内参以解决内参标定过程中的过拟合问题；然后设计了一种基于线曲率的平面提取方法来提高单帧点云的平面提取精度；最后通过图像与激光点云中对应的平面特征计算两个传感器之间的外参。实验结果表明，所提方法相比平面标定方法能得到更精确相机内参，与velo2cam＿calibration算法相比，外参标定的平均平移误差减少了39.4%，平均旋转误差减少了37.1%。更多还原

【Abstract】 Camera intrinsic and camera/LiDAR extrinsic calibration is a prerequisite for the fusion of the two senses. In order to solve the overfitting problem of planar calibration method and improve the accuracy of camera and low line-channel Li DAR calibration, a joint calibration method using 3D calibration plate to realize camera intrinsic and camera/LiDAR extrinsic is proposed. Firstly, the camera intrinsic is calibrated by the 3D calibration plate to solve the overfitting problem in the intrinsic calibration process. Then a line curvature based plane extraction method is designed to improve the plane extraction accuracy in a single frame point cloud. Finally, the extrinsic between the two sensors is calculated by the corresponding planar features in the image and LiDAR. It is demonstrated that the proposed method can obtain more accurate camera intrinsic than the planar calibration method. Compared with the velo2cam＿calibration algorithm, the average translation error of the extrinsic calibration is reduced by 39.4% and the average rotation error is reduced by 37.1%.更多还原