M2DGR
收藏资源简介:
M2DGR是由上海交通大学上海导航与位置服务重点实验室创建的一个大规模多传感器多场景SLAM数据集,专为地面机器人设计。该数据集包含36个序列,总数据量约1TB,涵盖室内外多种环境。数据集通过配备多种传感器(包括鱼眼相机、红外相机、事件相机、视觉惯性传感器等)的地面机器人收集,确保数据的高质量和多样性。创建过程中,通过精确校准和同步所有传感器,以及使用运动捕捉设备和激光3D跟踪器获取地面真实轨迹,确保数据的准确性。M2DGR数据集旨在推动SLAM技术在复杂环境中的应用,特别是在物流、安全、仓库管理等领域,通过提供丰富的实际场景数据,帮助研究人员评估和改进SLAM算法。
M2DGR is a large-scale multi-sensor, multi-scenario SLAM dataset created by the Shanghai Key Laboratory of Navigation and Location-Based Services at Shanghai Jiao Tong University, purpose-built for ground robots. It contains 36 sequences with a total data volume of approximately 1 TB, covering a wide range of indoor and outdoor environments. The dataset is collected using ground robots equipped with an array of sensors, including fisheye cameras, infrared cameras, event cameras, visual-inertial sensors and others, to guarantee high data quality and diversity. During dataset construction, all onboard sensors were precisely calibrated and synchronized, and motion capture systems and laser 3D trackers were utilized to acquire ground-truth trajectories, thus ensuring the overall accuracy of the dataset. The M2DGR dataset aims to advance the application of SLAM technologies in complex environments, particularly in domains such as logistics, security and warehouse management. By providing abundant real-world scene data, it helps researchers evaluate and refine SLAM algorithms.



