/WHU-HelmetDataset

Werable Mapping Dataset

Primary LanguageC++

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

image

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

image

Coordinate systems involved in the Helmet system.

2. Dataset links and overview

Scenes Bag Links Scanner Type
1_forest 1_forest_2021-12-06-11-10-14.7z.001 ; 1_forest_2021-12-06-11-10-14.7z.002 Livox AVIA
2_mountain 2_mountain_2021-11-12-12-31-45.7z.001 ; 2_mountain_2021-11-12-12-31-45.7z.002 Livox Mid70
3.1_underground_tunnel 3.1_underground_tunnel_2021-11-14-10-21-01.7z.001 Livox Mid70
3.2_underground_park 3.2_underground_park_2021-12-29-15-05-33.7z.001 ; 3.3_underground_subway_2021-12-29-16-19-51.7z.001 Livox AVIA
3.3_underground_subway 3.3_underground_subway_2021-12-29-16-19-51.7z.001 ; 3.3_underground_subway_2021-12-29-16-19-51.7z.002 Livox AVIA
4_Infrustracture 4_Infrustracture_2021-11-14-09-18-30.7z.001 Livox Mid70
5.1_Heritage_library 5.1_Heritage_library_2021-11-10-16-03-20.bag Livox Mid70
5.2_Heritage_residence 5.2_Heritage_residence_2021-12-30-16-37-49.7z.001 Livox AVIA
6.1_roadside_campus 6.1_roadside_campus_2021-11-12-10-10-57.7z.001; 6.1_roadside_campus_2021-11-12-10-10-57.7z.002;6.1_roadside_campus_2021-11-12-10-10-57.7z.003 Livox Mid70
6.2_roadside_street 6.2_roadside_street_2021-12-29-11-12-17.7z.001; 6.2_roadside_street_2021-12-29-11-12-17.7z.002 Livox AVIA
6.3_roadside_mall 6.3_roadside_mall_2021-12-29-10-32-35.7z.001;6.3_roadside_mall_2021-12-29-10-32-35.7z.002 Livox AVIA

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

2.1 Forest

image

2.2 Mountain

image

2.3 Underground

image

image

image

2.4 Infrastructure

image

2.5 Heritage

image

image

2.6 Road side

image

image

image

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: image