Automotive
A2D2
Multimodal driving perception, segmentation and vehicle-bus learning
Surround cameras and five LiDARs provide annotated perception frames, long unlabelled drives and timestamped vehicle-bus signals.
Access links
Primary places to inspect, download, read about, or reproduce the dataset.
Quick facts
- Vehicle type
- Automotive
- Environment
- Outdoor
- Data origin
- Real-world
- Sensor count
- 11
- Ground truth
- Not reported
- Calibration
- Parameters Only
- Annotations
- Available
Sensor overview
High-level sensor availability before the detailed sensor records below.
Ground truth
Reference scope, method, rate, coverage, and provenance.
Not reported
Calibration and synchronization
Reported calibration level, reproducibility signals, and supporting notes.
- Level
- Parameters Only
- Processed parameters
- Available
- Raw calibration data
- Not reported
- Calibration targets
- Not reported
Parameters only
Processed calibration parameters are reported, but raw calibration data is not reported.
Notes
Distributed sensor poses and camera intrinsics/distortion; ICP and depth/RGB edge alignment.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
The 41,277 annotated images, 12,497 front-box frames and 392,556 unlabelled sequential images describe different products. Three unlabelled drives are not a whole-corpus sequence count.
Camera model spelling and raw versus experiment image widths conflict in the source. Models, image dimensions and camera cadence remain unknown in this entry.
Registered depth derives from input LiDAR. No independent navigation reference or separately released raw GNSS/IMU observation stream is established from vehicle-bus signals.
Faces and plates are blurred; box coverage is front-facing. Timing accuracy and full release totals remain unverified.
Cameras
Camera 1
- Model
- Not reported
- Setup
- Multi Camera
- Count
- 6
- Effective cameras
- 6
- Modality
- Rgb
- Resolution
- Not reported
- Rate
- Not reported
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Front rectilinear plus five fisheye views; source model and resolution conflicts retained for review.
IMUs
Not reported
GNSS
Not reported
LiDAR
Velodyne VLP-16
- Model or name
- Velodyne VLP-16
- Count
- 5
- Dimensions
- 3
- Rate
- 10 Hz
- Channels
- 16
- Range
- Not reported
Radar
Not reported
Additional sensors
Not reported
Annotations
Reported annotation availability and task support.
- Availability
- Available
- Format
- Image semantics and traffic-participant instances; projected point labels; front-frustum 3D boxes.
- Class count
- 38
Semantic segmentation
Available
Instance segmentation
Available
Object detection
Available
Optical flow
Not reported
Depth ground truth
Not reported
Citation
Citation key and BibTeX kept at the end of the page for reference.
@misc{geyer2020a2d2,
author = {Geyer, Jakob and Kassahun, Yohannes and Mahmudi, Mentar and Ricou, Xavier and Durgesh, Rupesh and Chung, Andrew S. and Hauswald, Lorenz and Pham, Viet Hoang and Mühlegg, Maximilian and Dorn, Sebastian and Fernandez, Tiffany and Jänicke, Martin and Mirashi, Sudesh and Savani, Chiragkumar and Sturm, Martin and Vorobiov, Oleksandr and Oelker, Martin and Garreis, Sebastian and Schuberth, Peter},
title = {A2D2: Audi Autonomous Driving Dataset},
year = {2020},
eprint = {2004.06320},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2004.06320v2}
}Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy