FusionPortable_dataset_tools

News

• (20240408) Initial development tools have been released
• (20240407) Data of FusionPortable are also stored in Google Drive. Please check the webpage to download.

Download Dataset

1. Please visit FusionPortable dataset and FusionPortableV2_dataset to check and download data.
2. Download the compressed rosbag.
3. When finished, use the similar command 7z l 20220216_garden_day.7z to extract data.

Installation

Clone the Repo

git clone https://github.com/fusionportable/fusionportable_dataset_tools.git --recursive
cd fusionportable_dataset_tools

Setup the Python environment (tested on Python-3.9) or using Anaconda directly and run

conda create -n fp_dataset python=3.9
pip install -r requirements

Usage

Usage of Data-Loader-Related Functions

1. Convert the raw rosbag into individual files: write_bag_to_data.ipynb

2. Convert the algorithms' results (i.e., R3LIVE, FAST-LIO2) stored as the rosbag into individual files: write_alg_bag_to_data.ipynb

3. Convert the raw files into the KITTI-360 format (including synchronized sensor data, callibration files, odometry): write_data_to_kitti360.ipynb

4. Generate depth map with respect to the frame_left camera for the depth evaluation purpose: write_depthmap_to_kitti360.ipynb

Usage of Tool Functions

1. Project undistorted point cloud onto image to verify the error in extrinsics: visualize_depthmap.ipynb

Usage of Evaluation Tools

• Trajectory Evaluation

• Mapping Evaluation

Usage of Applications

Some Issues with Dependencies

1. Something wrong with the ros_numpy

File ~/anaconda3/envs/fp_dataset/lib/python3.9/site-packages/ros_numpy/point_cloud2.py:224
    221             new_cloud_arr[field_name] = cloud_arr[field_name]
...
AttributeError: module 'numpy' has no attribute 'float'.

Solution: Please goto path_to_python/site-packages/ros_numpy/point_cloud2.py:224 and replace the original line with

def get_xyz_points(cloud_array, remove_nans=True, dtype=np.float64):

Contribution

Please refer to Contribution Guidance to make contributions to this project.

Inquiry

Please feel free to contact Dr.Jianhao Jiao (jiaojh1994 at gmail.com) or Mr.Hexiang Wei (hweiak at connect.ust.hk) if you have any questions.

Citation

@misc{wei2024fusionportablev2,
  title={FusionPortableV2: A Unified Multi-Sensor Dataset for Generalized SLAM Across Diverse Platforms and Scalable Environments}, 
  author={Hexiang Wei and Jianhao Jiao and Xiangcheng Hu and Jingwen Yu and Xupeng Xie and Jin Wu and Yilong Zhu and Yuxuan Liu and Lujia Wang and Ming Liu},
  year={2024},
  eprint={2404.08563},
  archivePrefix={arXiv},
  primaryClass={cs.RO}
}

@inproceedings{jiao2022fusionportable,
  title={Fusionportable: A multi-sensor campus-scene dataset for evaluation of localization and mapping accuracy on diverse platforms},
  author={Jiao, Jianhao and Wei, Hexiang and Hu, Tianshuai and Hu, Xiangcheng and Zhu, Yilong and He, Zhijian and Wu, Jin and Yu, Jingwen and Xie, Xupeng and Huang, Huaiyang and others},
  booktitle={2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  pages={3851--3856},
  year={2022},
  organization={IEEE}
}