WHU-Helmet Dataset

Main web page of our group : http://3s.whu.edu.cn/ybs/index.htm

If you use the dataset, please cite our paper:

@article{li2023whu,
    title={WHU-Helmet: A helmet-based multi-sensor SLAM dataset for the evaluation of real-time 3D mapping in large-scale GNSS-denied environments},
    author={Li, Jianping and Wu, Weitong and Yang, Bisheng and Zou, Xianghong and Yang, Yandi and Zhao, Xin and Dong, Zhen},
    journal={IEEE Transactions on Geoscience and Remote Sensing},
    year={2023},
    publisher={IEEE}
}

Paper Link: WHU-Helmet: A helmet-based multi-sensor SLAM dataset for the evaluation of real-time 3D mapping in large-scale GNSS-denied environments

1. Hardware configuration

Mechanical structure design of the WHU-Helmet. (a) Head-mounted system. (b) Head-mounted system with a reference FOG (Fiber Optic Gyro) IMU.

Coordinate systems involved in the Helmet system.

2. Dataset links and overview

If you need GT files, please send us a E-mail with your organization information to (kafeiyin00@gmail.com).

2.1 Forest

2.2 Mountain

2.3 Underground

2.4 Infrastructure

2.5 Heritage

2.6 Road side

2.7 Calibration

we provide the extrinsic parameters for LiDAR-IMU，LiDAR-camera, and camera intrinsic parameters in the calibration folder.

The extrinsic parameters of LiDAR-IMU are calculated using the AFLI-Calib

Weitong Wu, Jianping Li, Chi Chen, Bisheng Yang,et al.,2023. AFLI-Calib: Robust LiDAR-IMU extrinsic self-calibration based on adaptive frame length LiDAR odometry. ISPRS Jouranl of Photogrammetry and Remote Sensing.199,157-181.https://doi.org/10.1016/j.isprsjprs.2023.04.004.

2.8 Run Dataset

Data preprocess (Important)

We provide the rosbag (link) for running. Before you play the bag, use the rosbag_edit.cpp to preprocess the data.

rosrun your_package_name rosbag_edit_node $input_rosbag_path $output_rosbag_path

Example for running [fast-lio2] (https://github.com/hku-mars/FAST_LIO)

In the config_file folder, We provide the config file for running fast-lio2. You can easily use it for running.

2.9 Evaluation of SOTA SLAM

2.9.1 Evaluation Tool and Criteria

We use evo to evaluate results with four metrics:

• APE_Trans(m): Absolute pose error about translation.
• APE_Angle(deg): Absolute pose error about angle.
• RPE_Trans(m/frame): Relative pose error about translation.
• RPE_Angle(deg/frame): Relative pose error about angle.

  2.9.2 Results of SOTA Algorithm

  We test our dataset using Fast_lio, LIO_Livox , LOAM_Livox and MULLS. The following are results: