Welcome to The ROVER Dataset

Hello everyone! We’re excited to share The ROVER Dataset with you. This dataset is designed to support Visual SLAM research, especially for challenging outdoor environments. Here, you’ll find a resources, information, and tools to assist you in your projects.

Explore our dataset, discover the recording system used, and dive into the chapters below to get started!

Chapters

• Recording System
  Discover the sensors, platforms, and technical specifications behind the dataset collection.

• Dataset Overview
  Learn about the dataset structure, contents, and the various scenarios we’ve included to simulate real-world conditions.

• Benchmark
  Find detailed results and code of the benchmark.

• Download
  Access the dataset files directly, with details on file formats, storage requirements, and download options.

Collaborators

This is joint work of Institute for Intelligent Systems (Esslingen University), ANDREAS STIHL AG & Co. KG, and Robot Learning Lab (University of Freiburg).

by & & by

Citation

If you use our dataset, please cite our paper:

@ARTICLE{schmidt2025rover,
  author={Schmidt, Fabian and Daubermann, Julian and Mitschke, Marcel and Blessing, Constantin and Meyer, Stephan and Enzweiler, Markus and Valada, Abhinav},
  journal={IEEE Transactions on Robotics}, 
  title={ROVER: A Multi-Season Dataset for Visual SLAM}, 
  year={2025},
  volume={},
  number={},
  pages={1-19},
  keywords={Simultaneous localization and mapping;Visualization;Forestry;Robots;Lighting;Global navigation satellite system;Benchmark testing;Location awareness;Automobiles;Hands;SLAM;Visual SLAM;datasets;benchmark},
  doi={10.1109/TRO.2025.3577026}}

Thank you for visiting, and we hope The ROVER Dataset proves valuable in your work!