基于共面约束的多线激光雷达相对位置关系检校

Relative position calibration of multibeam LIDAR based on coplanar constraints

多线激光雷达具有成本低、体积小、能直接获取场景地物表面的三维点云数据等优点,已被广泛应用在无人驾驶、移动测量、机器人等领域。为减少遮挡,提高点云密度,两个或多个激光雷达常被集成在一起,互为补充。不同激光雷达的安装位置和姿态不同,要融合激光雷达的点云数据,关键在于对激光雷达之间相对位置关系的检校。为检校激光雷达之间的相互位置关系,提出了基于共面约束的检校算法。算法要求不同的激光雷达同时扫到相同的平面,利用平面在不同坐标系下的对应关系求解激光雷达之间的相互位置关系,并结合Levenberg-Marquardt (L-M)优化算法,提高检校精度。该算法操作简单、通用性强、检校精度高。

Multibeam LIDAR has been widely applied in the fields of unmanned ground vehicle,mobile measurement and robots because of its low cost,small size and capability of acquiring 3-dimensional distance of objects in the scene.To reduce occlusion and improve the density and coverage of point cloud,two or more LIDAR devices are integrated together to complement each other.As the installation position and attitude of the LIDAR are different,relative position calibration is a key step before fusing different laser data.In order to calibrate the relative spatial position relationship,a method based on coplanar constraints was proposed.Different LIDARs captured the range data of the same plane simultaneously.Although the range data had different coordinate systems,they represented the same plane.Relative positions between LIADR were initially obtained by fitting the common plane of multiple range data of different coordinate systems,and then optimized by L-M algorithm to enhance the calibration accuracy.This method is simple,accurate,and suitable for most LIDAR configurations in practical applications.