Skip to content

Acknowledgements

The benchmark, the Datasets section and the Device wiki rest on recordings that other people made and chose to share. Thank you.

Please cite each source you use. Contributors & citation explains how, and the dataset card lists the paper for each set.

Research groups that publish calibration recordings openly:

ProviderDatasets we useUsed for
Computer Vision Group, TU MunichTUM-VI, TUM-VIE, TUM-Fisheye (Double Sphere), TUM RGB-D, TUM RSresults, examples, Why kb4, download recipes
CollaboraMonado SLAM datasets MIC, MGC and MOCresults, examples, the Valve Index, HP Reverb G2 and Samsung Odyssey+ pages
Robotics and Perception Group, University of ZurichUZH-FPVresults, examples
Hilti, Trimble and the University of OxfordHilti-Trimble SLAM Challenge 2026the Insta360 ONE RS 1-Inch 360 page
HiltiHilti 2022, Hilti 2023 handhelddownload recipes
Autonomous Systems Lab, ETH ZurichEuRoC MAVdownload recipes
GRASP Laboratory, University of PennsylvaniaPennCOSYVIOdownload recipes
Steffen UrbanUrban fisheye (ImprovedOcamCalib)download recipes

The Monado and TUM-VIE frame-camera recordings are CC BY 4.0, which lets us host converted copies under converted/. We link to every other set at its original host.

Thank you to the makers and partners who shared recordings of their devices. Each card links the device pages their recordings made possible.

Tujian logo

Tujian

Shenzhen Tujian Intelligent Technology

Four-camera inertial module recording

DIBULI logo

DIBULI

迪博凌(上海)科技有限公司, Shanghai

The camera + IMU calibration recordings they publish

Thanks to Li Xiangfu (李相辅)

Product photos on the device pages come from the makers’ product pages and are credited under each photo.

Your device could be next. To share a calibration recording, see Submit a dataset.