1. 📥 Dataset

We evaluate ScaRF-SLAM on a real-world dataset collected at the Oxford Robotics Institute (ORI) with accurate ground-truth camera trajectories and LiDAR point clouds for quantitative evaluation (download link).

The dataset is recorded using the front fisheye camera and IMU of an Insta360 ONE RS 1-Inch, rigidly mounted to a LiDAR–inertial mapping system. Ground-truth poses are obtained by registering the undistorted LiDAR scans to a high-precision terrestrial laser scanner map. It contains five sequences, each organized according to the folder structure below (using R01 as an example):

r01
├── r01_bag
│   ├── metadata.yaml
│   └── r01_bag_0.mcap
└── r01_gt
    ├── cloud_gt_fov
    │   ├── <sec>_<nsec>.pcd
    │   ├── <sec>_<nsec>.pcd
    │   └── ...
    ├── cloud_gt.pcd
    ├── poses_gt.csv
    └── poses_gt.txt

• r01_bag: ROS 2 data bag containing fisheye images, IMU measurements, and ground-truth poses.
• cloud_gt_fov: sparse undistorted LiDAR point clouds at each timestamp in the local camera coordinate frame, with points outside the camera field of view removed. Used for recall evaluation.
• cloud_gt.pcd: dense registered and undistorted LiDAR point cloud. Used for precision and reconstruction error evaluation.
• poses_gt.csv: ground-truth camera trajectory in CSV format.
• poses_gt.txt: ground-truth camera trajectory in TUM format.