IROS 2017 – Voxblox Dataset

Year: 2017
DOI: 10.3929/ethz-b-000721636

This dataset accompanies the Voxblox 3D volumetric mapping system, released in conjunction with the IROS 2017 paper “Voxblox: Incremental 3D Euclidean Signed Distance Fields for On-board MAV Planning.”

It provides RGB-D and/or depth-only sequences suitable for TSDF/ESDF reconstruction, incremental mapping, and robotic planning evaluation.

Original dataset page: https://projects.asl.ethz.ch/datasets/doku.php?id=iros2017

Data Access

The dataset is now permanently hosted on the ETH Research Collection:

➡️ Download / Landing Page: https://doi.org/10.3929/ethz-b-000721636

Contents

The dataset includes sequences used for evaluating:

• voxel-based TSDF reconstruction
• incremental ESDF generation
• planning-aware mapping
• online MAV/local robot navigation

Typical data provided:

• RGB–D recordings
• depth images
• calibrated intrinsic/extrinsic parameters
• ground-truth or reference reconstructions (in some sequences)
• example output reconstructions from Voxblox

Paper

Voxblox: Incremental 3D Euclidean Signed Distance Fields for On-board MAV Planning Helen Oleynikova, Zachary Taylor, Markus Fehr, Juan Nieto, Roland Siegwart IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017.

GitHub: https://github.com/ethz-asl/voxblox

BibTeX

@inproceedings{oleynikova2017voxblox,
  title        = {Voxblox: Incremental {3D} Euclidean Signed Distance Fields for On-Board {MAV} Planning},
  author       = {Oleynikova, Helen and Taylor, Zachary and Fehr, Markus and Nieto, Juan and Siegwart, Roland},
  booktitle    = {2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  pages        = {1366--1373},
  year         = {2017},
  organization = {IEEE}
}

Tags: tsdf, esdf, mapping, reconstruction, planning, mav