OpenLORIS-Scene
收藏资源简介:
OpenLORIS-Scene数据集由清华大学电子工程系与清华伯克利深圳学院创建,旨在为服务机器人的终身同步定位与地图构建(SLAM)研究提供真实世界的数据支持。该数据集包含22个序列,覆盖了办公室、走廊、家庭、咖啡馆和市场等多样化室内场景,记录了多次不同时间的数据,以模拟真实生活中的场景变化。数据采集使用了包括RGBD、立体鱼眼、惯性测量单元(IMUs)、轮式里程计和激光雷达等多种传感器,确保数据的丰富性和多样性。数据集的创建过程涉及精确的传感器校准和同步,以及使用高精度激光SLAM方法生成地面实况轨迹。该数据集主要应用于服务机器人领域,旨在解决机器人在长期运行中面临的定位和地图维护挑战,特别是在环境变化频繁的场景中。
The OpenLORIS-Scene dataset was developed by the Department of Electronic Engineering of Tsinghua University and Tsinghua-Berkeley Shenzhen Institute, aiming to provide real-world data support for lifelong simultaneous localization and mapping (SLAM) research of service robots. This dataset includes 22 sequences covering diverse indoor scenarios such as offices, corridors, homes, cafes and markets, and records data collected at different times to simulate scene changes in real-life environments. Multiple sensors including RGBD devices, stereo fisheye cameras, inertial measurement units (IMUs), wheel odometry and LiDAR were used for data collection, ensuring the richness and diversity of the dataset. The creation of this dataset involves precise sensor calibration and synchronization, as well as the generation of ground-truth trajectories via high-precision laser SLAM methods. This dataset is primarily applied in the service robotics field, targeting the resolution of localization and map maintenance challenges faced by robots during long-term operations, especially in scenarios with frequent environmental changes.

- 1Are We Ready for Service Robots? The OpenLORIS-Scene Datasets for Lifelong SLAM清华大学电子工程系与清华伯克利深圳学院 · 2020年



