Wild-Places: A Large-Scale Dataset for Lidar Place Recognition in Unstructured Natural Environments

This repository contains pre-trained checkpoints for the dataset introduced in the paper Wild-Places: A Large-Scale Dataset for Lidar Place Recognition in Unstructured Natural Environments, which has been published at ICRA2023.

If you find this dataset helpful for your research, please cite our paper using the following reference:

@inproceedings{2023wildplaces,
  title={Wild-places: A large-scale dataset for lidar place recognition in unstructured natural environments},
  author={Knights, Joshua and Vidanapathirana, Kavisha and Ramezani, Milad and Sridharan, Sridha and Fookes, Clinton and Moghadam, Peyman},
  booktitle={2023 IEEE international conference on robotics and automation (ICRA)},
  pages={11322--11328},
  year={2023},
  organization={IEEE}
}

Download Instructions

Our dataset can be downloaded through The CSIRO Data Access Portal. Detailed instructions for downloading the dataset can be found in the README file provided on the data access portal page.

Training and Benchmarking

Here we provide pre-trained checkpoints and results for benchmarking several state-of-the-art LPR methods on the Wild-Places dataset.

Update Nov. 2025: With the release of Wild-Places v3.0, we have also re-run training for two state-of-the-art methods (LoGG3D-Net, MinkLoc3Dv2) on the Wild-Places dataset using expanded batch sizes to provide new training checkpoints which better reflect the capabilities of recent state-of-the-art GPUs. We provide checkpoints and benchmarked results for both the recently trained models and the checkpoints released with the ICRA2023 paper.

Checkpoints

Release	Model	Checkpoint
ICRA2023	TransLoc3D	Link
ICRA2023	MinkLoc3DV2	Link
ICRA2023	LoGG3D-Net	Link
2025 Re-Training	MinkLoc3DV2	Link
2025 Re-Training	LoGG3D-Net	Link

For further instructions on training and evaluating these checkpoints on the Wild-Places dataset, please follow the instructions found at the Wild-Places GitHub