融合位置信息的智能驾驶高精地图三维重建

3D reconstruction of intelligent driving high-precision maps with location information convergence

针对因高精地图三维重建需计算大量点云数据而带来的重建效率低的问题,提出了一种融合车辆位置信息和道路结构特性的高精地图三维重建算法。该算法将车辆位置信息融入道路主干点云提取模型,将道路点云划分为道路主干与非道路主干区域,并提取道路主干点云中的道路元素点云用于高精地图三维重建时的ICP配准,减少了点云计算数量。实验结果表明,该算法可缩短高精地图三维重建时间,提高地图重建效率。

The high-precision maps is the basis of intelligent driving.However,a large amount of point clouds needs to be computed in the process of 3D reconstruction of high-precision map generating,which makes the reconstruction efficiency lowly.Therefore,a high-precision map 3D reconstruction algorithm that combines vehicle location information and road structure characteristics is proposed in this paper.The algorithm integrates vehicle location information into the road trunk point cloud extraction model,divides the road point cloud into road trunk and non-road trunk areas and extracts road element point clouds from the road trunk point cloud for ICP registration during the process of high-precision map 3D reconstruction.The experimental results show that the algorithm can shorten the time of high-precision map 3D reconstruction and improve the reconstruction efficiency.