APR：通过点云重构进行在线远程点云配准（英文版）

英文标题：APR: Online Distant Point Cloud Registration Through Aggregated Point Cloud Reconstruction中文摘要：本文提出一种名为 APR 的新型特征提取框架，该框架利用自编码器设计重建了一个带有几个帧的密集聚合点云，从而提取具有丰富局部几何信息的特征进行在线的远程点云配准。实验结果表明，APR 较 KITTI 和 nuScenes 数据集上的其他特征提取器表现更优秀。英文摘要：For many driving safety applications, it is of great importance to accuratelyregister LiDAR point clouds generated on distant moving vehicles. However, suchpoint clouds have extremely different point density and sensor perspective onthe same object, making registra