kth-rpl / dufomap Goto Github PK

View Code? Open in Web Editor NEW

133.0 1.0 14.0 81.13 MB

[RA-L'24] DUFOMap: Efficient Dynamic Awareness Mapping

Home Page: https://kth-rpl.github.io/dufomap

License: BSD 3-Clause "New" or "Revised" License

C++ 99.58% CMake 0.16% Dockerfile 0.25%

dynamic-point-removal mapping robotics autonomous-driving ufomap

dufomap's Introduction

DUFOMap: Efficient Dynamic Awareness Mapping

[video coming soon] [poster coming soon]. Accepted by RA-L'24.

Quick Demo: Run with the same parameter setting without tuning for different sensor (e.g 16, 32, 64, and 128 channel LiDAR and Livox-series mid360), the following shows the data collected from:

Leica-RTC360	128-channel LiDAR	Livox-mid360

0. Setup

sudo apt update && sudo apt install gcc-10 g++-10
sudo apt install libtbb-dev liblz4-dev

Dockerfile will be soon available.

Clone quickly and init submodules:

git clone --recursive -b main --single-branch https://github.com/KTH-RPL/dufomap.git

1. Build & Run

Build:

cmake -B build -D CMAKE_CXX_COMPILER=g++-10 && cmake --build build

Prepare Data: Teaser data (KITTI 00: 384.4Mb) can be downloaded via follow commands, more data detail can be found in the dataset section or format your own dataset follow custom dataset section.

wget https://zenodo.org/records/8160051/files/00.zip -p data
unzip data/00.zip -d data

Run:

./build/dufomap_run data/00 assets/config.toml

2. Evaluation

Please reference to DynamicMap_Benchmark for the evaluation of DUFOMap and comparison with other dynamic removal methods.

Evaluation Section link

Acknowledgements

Thanks to HKUST Ramlab's members: Bowen Yang, Lu Gan, Mingkai Tang, and Yingbing Chen, who help collect additional datasets.

This work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation including the WASP NEST PerCorSo.

Feel free to explore below projects that use ufomap (attach code links as follows):

Citation

Please cite our works if you find these useful for your research.

@article{daniel2024dufomap,
  author={Duberg, Daniel and Zhang, Qingwen and Jia, MingKai and Jensfelt, Patric},
  journal={IEEE Robotics and Automation Letters}, 
  title={{DUFOMap}: Efficient Dynamic Awareness Mapping}, 
  year={2024},
  volume={9},
  number={6},
  pages={1-8},
  doi={10.1109/LRA.2024.3387658}
}
@article{duberg2020ufomap,
  author={Duberg, Daniel and Jensfelt, Patric},
  journal={IEEE Robotics and Automation Letters}, 
  title={{UFOMap}: An Efficient Probabilistic 3D Mapping Framework That Embraces the Unknown}, 
  year={2020},
  volume={5},
  number={4},
  pages={6411-6418},
  doi={10.1109/LRA.2020.3013861}
}
@inproceedings{zhang2023benchmark,
  author={Zhang, Qingwen and Duberg, Daniel and Geng, Ruoyu and Jia, Mingkai and Wang, Lujia and Jensfelt, Patric},
  booktitle={IEEE 26th International Conference on Intelligent Transportation Systems (ITSC)}, 
  title={A Dynamic Points Removal Benchmark in Point Cloud Maps}, 
  year={2023},
  pages={608-614},
  doi={10.1109/ITSC57777.2023.10422094}
}

dufomap's People

Contributors

Stargazers

Watchers

Forkers

kin-zhang rickkcir wwdcrobot shrijitsingh99 qian5683 rick-960123 p-hoanganh hkwww weixiaopassking gatsby23 chaoyanseu

dufomap's Issues

关于论文中的limizations

你好，我在拜读论文时，看到最后的局限性部分说，如果在点云地图边界上的体素，虽然是动态体素，但是由于没有fully observed，也无法被划分为动态体素。我在测试数据时可能遇到了类似的问题，

测试时动态点云下面一半被过滤掉了，但是路面上方仍然有一串漂浮的点云，不知道有没有什么解决方案呢

The dynamic and static maps obtained using the kitti dataset are not accurate.

Using the kitti dataset and the default configuration gave poor results.

This is a dynamic point cloud, which considers the ground to be dynamic.

Below is a static point cloud.

Recently, I have been researching SLAM based on Livox mid-360, and very interested in your research.
I wonder could you share the dataset
[Indoor-Floor] Our own dataset, Collected by Livox mid-360 in a quadruped robot.

Thank you very much for considering my request. I look forward to your prompt response.

Best regards,

Cino

dynamic removal result between octomap and dufomap

Hi Qing Wen
I test dufomap and octomap with my dataset.

The result above is from dufomap with the default config.toml. We can see that some dynamic objects(in red circle) are not trimmed.
The result below is from octomap(https://github.com/Kin-Zhang/octomap.git) with default config.yml. Even though the static objects(in green rectangle) are removed, the dynamic objects are almost deleted.

In your paper, the SA\DA of dufomap performs better than octomap, which is inconsistent with my experiment. Can you explain why?
If you want to reproduce, you can download the dataset from: https://drive.google.com/drive/folders/1kpevt1zkFlDNiw-h6ISCD13-W8EdX6By?usp=drive_link

你好，请问queryNearest函数怎么使用？

nothing different with the gt_point.pcd?

Hi, Kin
I run the code you pushed 4 days ago with configs in assets/. But the result dufomap_output.pcd seems no different with the gt_cloud.pcd. The dynamic points are still saved. Is it right?
Here is the screentshot:

static areas remove too much

Hi Dr.Zhang

First of all, thank you for sharing the data. I have run dufomap with twofloor and our own data and found that static areas such as floors and walls will also be remove, which with configs in assets/.

I did not find this phenomenon on Octomap, but the running time is very slow

Can this phenomenon be optimized by adjusting parameters?

Here is the screenshots, run with twofloor:

dufomap_output_dynamic.pcd

dufomap_output_static.pcd

octomapfg_output.pcd

Best regards,

Cino

Is there a version that can run on windows?

This works great, I want to run it on windows, but I have some problems? Is there a windows version?

哈喽哈喽，我最近想在一个ros2项目中使用dufomap，但是我一旦引入ros2相关头文件，就会引发一串编译错误：

          哈喽哈喽，我最近想在一个ros2项目中使用dufomap，但是我一旦引入ros2相关头文件，就会引发一串编译错误：

有没有碰到过类似的问题呢

Originally posted by @whw-create in #10 (comment)

Ask for Leica-RTC360 dataset

Hi Dr.Zhang
I have tested the Teaser data (KITTI 00: 384.4Mb) you provided and the results are very good.
Now I plan to test the Leica RTC360 data and see how this algorithm performs in dense crowds.
Can you provide me with this dataset you are using.
Thanks.