ConSLAM

该数据集名为ConSLAM，是一项开创性的工作，它提供了首个开放获取的室内杂乱建筑工地数据集，其中包含了RGB和激光雷达（LiDAR）数据序列，以及地面激光扫描（TLS）点云。该数据集用于通过SLAM2REF（BIM-SLAM的增强版）和OGM2PGBM计算地面真实姿态，并采用多种指标来量化轨迹质量。该任务旨在室内环境中进行姿态优化和地图构建。This dataset, named ConSLAM, is a pioneering work that provides the first open-access dataset for cluttered indoor construction sites, which contains RGB and LiDAR data sequences as well as terrestrial laser scanning (TLS) point clouds. It is utilized to calculate ground-truth poses via SLAM2REF (an enhanced version of BIM-SLAM) and OGM2PGBM, and adopts multiple metrics to quantify trajectory quality. This task aims at pose optimization and map construction in indoor environments.

搜集汇总

背景与挑战

背景概述

ConSLAM是一个专为SLAM和建筑进度监控设计的真实世界建筑工地数据集，包含多传感器（RGB相机、激光雷达、IMU）数据，并提供了轨迹评估工具和相关研究论文。