CoInfra: A Large-Scale Cooperative Infrastructure Perception System and Dataset in Adverse Weather

CoInfra is a real-world, large-scale cooperative perception system and dataset designed to support research in multi-agent perception, especially under adverse weather conditions (snow, rain, freezing rain). It features 14 synchronized infrastructure nodes, each with a LiDAR and dual cameras, deployed across a shared region for full-scene coverage.

Demo videos

Download Instructions

Use the huggingface_hub CLI or library

pip install huggingface_hub

Create a python script in "CoInfra" folder, and run it:

from huggingface_hub import snapshot_download

path = snapshot_download(
    repo_id="MinghaoNing/CoInfra",
    repo_type="dataset",
    local_dir=".",                # <-- download into current folder
    local_dir_use_symlinks=False  # <-- ensure actual files are copied
)

print("Downloaded to:", path)

This will download the full dataset repo (all files) to a local folder.

Extraction Instructions

The archives are compressed using zstd for fast decompression. To extract the dataset, you need to have zstd installed. You can install it via package managers like apt, brew.

To extract an archive, use the following command:

tar --use-compress-program=zstd -xf [archive_file.tar.zst]

To extract all archives in a directory, you can use:

for archive in *.tar.zst; do
    echo "Extracting $archive..."
    tar --use-compress-program=zstd -xf "$archive"
done

📖 Citation

If you use CoInfra, please cite:

@article
{ning2025coinfra,
title={CoInfra: A Large-Scale Cooperative Infrastructure Perception System and Dataset in Adverse Weather},
author={Ning, Minghao and Yang, Yufeng and Shu, Keqi and Huang, Shucheng and Zhong, Jiaming and Salehi, Maryam and Rahmani, Mahdi and Lu, Yukun and Sun, Chen and Saleh, Aladdin and Hashemi, Ehsan and Khajepour, Amir},
journal={arXiv preprint arXiv:2507.02245},
url={https://arxiv.org/abs/2507.02245},
year={2025}
}