This is an old revision of the document!
Monocular Visual Odometry Dataset
We present a dataset for evaluating the tracking accuracy of monocular Visual Odometry (VO) and SLAM methods. It contains 50 real-world sequences comprising over 100 minutes of video, recorded across different environments – ranging from narrow indoor corridors to wide outdoor scenes.
All sequences contain mostly exploring camera motion, starting and ending at the same position: this allows to evaluate tracking accuracy via the accumulated drift from start to end, without requiring ground-truth for the full sequence.
In contrast to existing datasets, all sequences are photometrically calibrated: We provide the exposure times for each frame as reported by the sensor, the camera response function and the lens attenuation factors (vignetting). Further, we propose a simple approach to non-parametric vignette and camera response calibration, which require minimal set-up and are easy to reproduce.
Dataset Download
- Full dataset with all 50 sequences zip (43GB)
- Code for reading and undistorting the dataset sequences; performing photometric calibration with proposed approach https://github.com/tum-vision/mono_dataset_code
- More notes on the intrinsic calibration format: https://github.com/JakobEngel/dso#31-dataset-format
- Supplementary material with ORB-SLAM and DSO results zip (2.7GB) and Matlab scripts for running evaluation zip (30MB) (14.10.2016.: We have updated the supplementary material with the fixed real-time results for ORB-SLAM, corresponding to the revised version of the papers.)
- Individual sequences:
Calibration sequences
- All calibration sequences zip (13GB)
- Individual sequences:
Sequence name | Download | Preview Video |
narrowGamma_scene1 | zip (2.02GB) | play (5x) |
narrowGamma_scene2 | zip (1.20GB) | play (5x) |
narrowGamma_sweep1 | zip (0.37GB) | play (5x) |
narrowGamma_sweep2 | zip (0.76GB) | play (5x) |
narrowGamma_sweep3 | zip (0.29GB) | play (5x) |
narrowGamma_vignette | zip (0.35GB) | play (5x) |
narrow_checkerboard1 | zip (0.27GB) | play (5x) |
narrow_checkerboard2 | zip (0.05GB) | |
narrow_sweep1 | zip (0.51GB) | play (5x) |
narrow_sweep2 | zip (0.40GB) | play (5x) |
narrow_sweep3 | zip (0.20GB) | play (5x) |
narrow_vignette | zip (0.32GB) | play (5x) |
narrow_whitePaper | zip (0.04GB) | play (5x) |
wideGamma_scene1 | zip (1.56GB) | play (5x) |
wideGamma_sweep1 | zip (0.53GB) | play (5x) |
wideGamma_sweep2 | zip (0.70GB) | play (5x) |
wideGamma_vignette | zip (0.41GB) | play (5x) |
wideGamma_vignette2 | zip (0.42GB) | play (5x) |
wide_checkerboard1 | zip (0.24GB) | play (5x) |
wide_checkerboard2 | zip (0.04GB) | |
wide_sweep1 | zip (0.46GB) | play (5x) |
wide_sweep2 | zip (0.59GB) | play (5x) |
wide_vignette | zip (0.55GB) | play (5x) |
wide_vignette2 | zip (0.27GB) | play (5x) |
wide_whitePaper | zip (0.06GB) | play (5x) |
License
Unless stated otherwise, all data in the Monocular Visual Odometry Dataset is licensed under a Creative Commons 4.0 Attribution License (CC BY 4.0) and the accompanying source code is licensed under a BSD-2-Clause License.
Publications
Export as PDF, XML, TEX or BIB
Journal Articles
2022
[] DM-VIO: Delayed Marginalization Visual-Inertial Odometry , In IEEE Robotics and Automation Letters (RA-L) & International Conference on Robotics and Automation (ICRA), volume 7, 2022. ([arXiv][video][project page][supplementary][code])
2018
[] Omnidirectional DSO: Direct Sparse Odometry with Fisheye Cameras , In IEEE Robotics and Automation Letters & Int. Conference on Intelligent Robots and Systems (IROS), IEEE, 2018. ([arxiv])
[] Online Photometric Calibration of Auto Exposure Video for Realtime Visual Odometry and SLAM , In IEEE Robotics and Automation Letters (RA-L), volume 3, 2018. (This paper was also selected by ICRA'18 for presentation at the conference.[arxiv][video][code][project])
ICRA'18 Best Vision Paper Award - Finalist [] Direct Sparse Odometry , In IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018.
Conference and Workshop Papers
2021
[] Tight Integration of Feature-based Relocalization in Monocular Direct Visual Odometry , In Proc. of the IEEE International Conference on Robotics and Automation (ICRA), 2021. ([project page])
2020
[] D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry , In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
Oral Presentation
2019
[] Rolling-Shutter Modelling for Visual-Inertial Odometry , In International Conference on Intelligent Robots and Systems (IROS), 2019. ([arxiv])
2018
[] Direct Sparse Odometry With Rolling Shutter , In European Conference on Computer Vision (ECCV), 2018. ([supplementary][arxiv])
Oral Presentation [] Deep Virtual Stereo Odometry: Leveraging Deep Depth Prediction for Monocular Direct Sparse Odometry , In European Conference on Computer Vision (ECCV), 2018. ([arxiv],[supplementary],[project])
Oral Presentation [] LDSO: Direct Sparse Odometry with Loop Closure , In International Conference on Intelligent Robots and Systems (IROS), 2018. ([arxiv][video][code][project])
[] Direct Sparse Visual-Inertial Odometry using Dynamic Marginalization , In International Conference on Robotics and Automation (ICRA), 2018. ([supplementary][video][arxiv])
2017
[] Stereo DSO: Large-Scale Direct Sparse Visual Odometry with Stereo Cameras , In International Conference on Computer Vision (ICCV), 2017. ([supplementary][video][arxiv][project])
2016
[] Direct Sparse Odometry , In arXiv:1607.02565, 2016.
[] A Photometrically Calibrated Benchmark For Monocular Visual Odometry , In arXiv:1607.02555, 2016.