Automotive, UAV, Quadruped
M3ED
Event-based robotic perception
Synchronized stereo event, grayscale, RGB, inertial, LiDAR and RTK recordings from cars, UAVs and Spot across urban, indoor and forest settings, with derived pose, depth and semantic labels.
Access links
Primary places to inspect, download, read about, or reproduce the dataset.
Quick facts
- Vehicle type
- Automotive, UAV, Quadruped
- Environment
- Indoor, Outdoor
- Data origin
- Real-world
- Sensor count
- 8
- Ground truth
- Available
- Calibration
- Partial Raw Data
- Annotations
- Available
Sensor overview
High-level sensor availability before the detailed sensor records below.
Ground truth
Reference scope, method, rate, coverage, and provenance.
| Scope | Method | Reference system | Count | Rate | Coverage |
Reference provenance
|
|---|---|---|---|---|---|---|
| Not reported | LiDAR SLAM | FasterLIO | Not reported | Not reported | Full |
Reference provenance
Derived
|
|
||||||
| Other | LiDAR SLAM | FasterLIO projection with hidden-point removal | Not reported | Not reported | Unknown |
Reference provenance
Derived
|
|
||||||
Calibration and synchronization
Reported calibration level, reproducibility signals, and supporting notes.
- Level
- Partial Raw Data
- Processed parameters
- Available
- Raw calibration data
- Available
- Calibration targets
- AprilTag
Reproducible calibration data
Raw calibration data is reported, so users can reproduce or inspect the calibration process.
Notes
Announced calibration sequences include camera-LiDAR, camera intrinsics/extrinsics and camera-IMU. Precomputed all-sensor extrinsics and camera intrinsics are supplied. LiDAR rotation is refined per sequence.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
Compact 12 cm baseline limits single-frame driving depth quality.
UAV camera and LiDAR overlap covers only the top half of the image.
Soft mounting can alter camera-LiDAR extrinsics slightly during a sequence.
Pose references are estimated and can drift; 2 m drift was observed over an 800 m interval.
Cameras
ovc-grayscale
- Model
- OVC 3b AR0144
- Setup
- Stereo
- Count
- 1
- Effective cameras
- 2
- Modality
- Grayscale
- Resolution
- 1280 x 800
- Rate
- Not reported
- Shutter
- Global
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- 12 cm stereo baseline.
ovc-rgb
- Model
- OVC 3b AR0144
- Setup
- Mono
- Count
- 1
- Effective cameras
- 1
- Modality
- Rgb
- Resolution
- 1280 x 800
- Rate
- Not reported
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
prophesee-event
- Model
- Prophesee EVKv4
- Setup
- Stereo
- Count
- 1
- Effective cameras
- 2
- Modality
- Event
- Resolution
- 1280 x 720
- Rate
- Not reported
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- IMX636 sensors; event-rate controller disabled and bias held constant across sequences. Observed forest throughput is not a camera frame rate.
IMUs
OVC 3b IMU
- Model
- VectorNav VN100T-SMD
- Count
- 1
- Accel / gyro
- Not reported
GNSS
UBlox ZED-F9P
- Model or name
- UBlox ZED-F9P
- Count
- 1
- Rate
- 5 Hz
- GNSS type
- RTK
- Position output
- Available
- Velocity output
- Not reported
- Heading output
- Not reported
LiDAR
Ouster OS1-64U
- Model or name
- Ouster OS1-64U
- Count
- 1
- Dimensions
- Not reported
- Rate
- Not reported
- Channels
- 64
- Range
- Not reported
Radar
Not reported
Pitot tube
Not reported
Additional sensors
Not reported
Annotations
Reported annotation availability and task support.
- Availability
- Available
- Format
- Not reported
- Class count
- 11
Semantic segmentation
Available
Instance segmentation
Not reported
Object detection
Not reported
Optical flow
Not reported
Depth ground truth
Available
Citation
Citation key and BibTeX kept at the end of the page for reference.
Not reported
Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy