Marine underwater
SubPipe
Pipeline inspection and visual-inertial localization
Real AUV pipeline inspection with RGB/grayscale imagery, sonar, navigation state and selective annotations.
Access links
Primary places to inspect, download, read about, or reproduce the dataset.
Quick facts
- Vehicle type
- Marine underwater
- Environment
- Outdoor
- Data origin
- Real-world
- Sensor count
- 2
- Ground truth
- Available
- 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.
| Scope | Method | Reference system | Count | Rate | Coverage |
Reference provenance
|
|---|---|---|---|---|---|---|
| 6DoF | Other | INS/DVL estimated navigation state | Not reported | Not reported | Unknown |
Reference provenance
Derived
|
|
||||||
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
Official release layout documents config.yaml. Complete raw calibration sessions and achieved parameter accuracy remain unverified.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
The collection is divided into five chunks of a pipeline inspection mission; these are not asserted to be five independent missions.
High-resolution video lacks exact synchronization and its reported 240 Hz remains unverified. Navigation is an INS/DVL estimate interpolated to image times, with unknown achieved errors and coverage.
Sonar acoustic carrier frequency is not a sample rate. COCO sonar detection and RGB segmentation have different annotation scopes; full raw calibration delivery remains unverified.
Cameras
Camera 1
- Model
- Not reported
- Setup
- Mono
- Count
- 1
- Effective cameras
- 1
- Modality
- Rgb
- Resolution
- 2704 x 1520
- Rate
- Not reported
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Reported 240 Hz requires confirmation against extracted-frame timing; unsynchronized high-resolution video.
Camera 2
- Model
- Not reported
- Setup
- Mono
- Count
- 1
- Effective cameras
- 1
- Modality
- Grayscale
- Resolution
- 1936 x 1216
- Rate
- 4 Hz
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
IMUs
Not reported
GNSS
Not reported
LiDAR
Not reported
Radar
Not reported
Additional sensors
Other sensor
- Model
- Klein 3500
- Name
- Side-scan sonar
- Count
- 1
- Notes
- 900/455 kHz acoustic carrier frequencies, not rates. Native waterfall images 5000×500/2500×500; benchmark 640×640.
Other sensor
- Name
- DVL and auxiliary navigation/environment measurements
- Count
- 1
- Notes
- Velocity, depth/altitude, pressure/temperature, RPM, forward echo-sounder distance and inertial values; distinguish measurements from estimated state.
Annotations
Reported annotation availability and task support.
- Availability
- Available
- Format
- COCO detection; LabelMe-derived segmentation masks
- Class count
- Not reported
Semantic segmentation
Available
Instance segmentation
Not reported
Object detection
Available
Optical flow
Not available
Depth ground truth
Not reported
Citation
Citation key and BibTeX kept at the end of the page for reference.
@misc{alvareztunon2024subpipe,
author = {Olaya Álvarez-Tuñón and Luiza Ribeiro Marnet and László Antal and Martin Aubard and Maria Costa and Yury Brodskiy},
title = {SubPipe: A Submarine Pipeline Inspection Dataset for Segmentation and Visual-inertial Localization},
year = {2024},
url = {https://arxiv.org/abs/2401.17907v2}
}Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy