PilotSCAN-Orchard-Summer-2024-ReconstructionDataset

This dataset provides reconstructed 3D point clouds and tree-level structural data derived from a multi-sensor LiDAR acquisition in orchard environments. It includes point clouds generated using EKF-based trajectory estimation and refined through ICP registration and pose graph optimisation, together with output files related to the reconstruction pipeline. The dataset is organised as follows: Processed data Global point cloud (cloudCompare_*.bin)Fully reconstructed and aligned 3D point cloud generated from the complete dataset using the reference reconstruction pipeline (Software available on https://www.cloudcompare.org/ ). clouds/ directoryContains individual point clouds exported in .txt format: EKF/: point clouds projected using the EKF-based trajectory only ICP_PG/: point clouds after refinement using ICP registration and pose graph optimisation These outputs enable direct comparison between initial and refined reconstructions. tf/ directoryContains the transformation matrices associated with the reconstruction process: EKF: poses obtained from EKF-based sensor fusion ICP + Pose Graph: refined poses after scan alignment and global optimisation Tree-level data trees/ directoryContains manually segmented and registered individual trees extracted from the reconstructed point clouds.This subset is intended for structural analysis and multi-temporal comparison at tree level.本数据集提供了果园环境下多传感器激光雷达（LiDAR）采集得到的重建三维点云与单木级结构数据。数据集包含基于扩展卡尔曼滤波（EKF）的轨迹估计生成的点云，以及经迭代最近点（ICP）配准与位姿图优化精修后的点云，同时附带与重建流程相关的各类输出文件。 数据集结构如下： 一、 处理后数据 全局点云（cloudCompare_*.bin）：通过参考重建流程对完整数据集处理得到的完整对齐三维点云，软件可从https://www.cloudcompare.org/ 获取。 clouds/ 目录：包含以.txt格式导出的单帧点云： EKF/：仅基于EKF轨迹估计投影得到的点云 ICP_PG/：经过ICP配准与位姿图优化精修后的点云 上述输出可用于直接对比初始重建结果与精修后重建结果的差异。 tf/ 目录：包含与重建流程相关的变换矩阵： EKF：基于EKF传感器融合得到的位姿信息 ICP + 位姿图：经过点云配准与全局优化后得到的精修位姿 二、 单木级数据 trees/ 目录：包含从重建点云中手动分割并配准得到的单棵树木点云子集，可用于单木尺度的结构分析与多时序对比研究。