平面特征的多站地面激光雷达点云配准

Plane-based registration of terrestrial LiDAR point clouds acquired by multiple scans

为了进一步研究建筑物密集区域多站地面激光雷达(LiDAR)点云数据的配准问题,该文提出一种基于平面特征的地面LiDAR数据配准方法:对点云数据进行分割获取平面信息;人工选择典型的平面,对相应的点云数据进行平面拟合,得到相应法向量;利用罗德里格矩阵的性质,建立三维激光扫描数据配准模型。实验结果表明,该方法的配准精度较高、计算速度快,可以取得较好的点云配准效果。

In order to further study on registration of terrestrial LiDAR point clouds acquired by mul- tiple scans in urban areas, this paper presented a plane-based method： first, the point clouds was segmented to get the plane information; second, the plane segments were manually selected, and the normal vectors of the planes fitted from the segmented point cloud data were obtained; finally, a 3D laser scanning data registration model was established by using the character of the Rodriguez Matrix. Experimental result showed that the proposed method could get good registration results with high precision and fast computation speed.