Files | Download |
---|
A dataset designed to capture short-term environmental changes across both natural and semi-urban environments, with specific aim of analyzing how such dynamics impact localization and re-localization performance.
View the Project on GitHub RaoufDannaoui1/Natural_and_Semi-Urban_Dataset
Weekly Multi-Temporal Dataset for Short-Term Localization and Environmental Change Analysis, designed to study the impact of real-world environmental changes on localization.
The dataset is a short-term, high-resolution, multi-modal dataset focused on understanding how real-world changes, such as vegetation growth, trimming, and object displacement, affect 3D LiDAR-based localization in dynamic outdoor environments.
The data was collected weekly from February to April 2025, across two contrasting outdoor scenarios:
This dataset provides a unique opportunity to analyze and benchmark localization robustness over short time intervals, making it ideal for applications in re-localization, change detection, SLAM, and dynamic mapping.
All data is recorded using a Leica Pegasus TRK100 mounted on a Zoe electric vehicle, the Leica is equipped with:
weekXX_hhmm-DD-MM-YYYY/
├── assets/
│ └── track_trajectories/
├── images_360/
│ ├── SemiUrban_track1-2/
│ └── Natural_track3-4/
└── point_clouds/
├── SemiUrban_track1-2/
└── Natural_track3-4/
Loading dataset files...
If you use this dataset in your research, please cite our paper.
@inproceedings{ardannaoui_2025_icp_analysis,
title={When and Where Localization Fails: An Analysis of the Iterative Closest Point in Evolving Environments},
author={Dannaoui, Abdelraouf and Laconte, Johann and Debain, Christophe and Pomerleau, Fran{\c{c}}ois and Checchin, Paul},
booktitle={2025 European Conference on Mobile Robots (ECMR)},
year={2025},
organization={IEEE},
note={Accepted for publication}
}
Data will be released under a permissive academic license (TBD).
For questions or collaboration inquiries, please contact:
Abdel-Raouf Dannaoui
Ph.D. Candidate in Robotics -- INRAE
dannaoui.abdelraouf@gmail.com