Platform not reported

UWStereo

Underwater stereo matching

A synthetic underwater stereo collection with 29,568 image pairs at 1280 by 720 pixels across four virtual scene groups, with dense disparity annotations for the left view.

Metadata: Partial Metadata needs review

Quick facts

Vehicle type
Not reported
Environment
Not reported
Data origin
Simulation
Sensor count
2
Ground truth
Available
Calibration
Not reported
Annotations
Available

Sensor overview

High-level sensor availability before the detailed sensor records below.

Ground truth

Reference scope, method, rate, coverage, and provenance.

Ground-truth references for UWStereo
Scope Method Reference system Count Rate Coverage Reference provenance
External
Measured using equipment independent of the platform’s normal onboard sensor pipeline.
Derived
Computed primarily from the dataset’s own recorded sensors.
Hybrid
Combines independent reference equipment with onboard measurements.
Other Other Unreal Engine 5 Not reported Not reported Full Reference provenance
External
Measured using equipment independent of the platform’s normal onboard sensor pipeline.
Derived
Computed primarily from the dataset’s own recorded sensors.
Hybrid
Combines independent reference equipment with onboard measurements.
Unknown
Reference notes
Dense left-view disparity is provided for every stereo pair. The simulator synthesizes depth, then computes disparity as baseline times horizontal focal length divided by depth. This reference is pixel disparity, with no asserted released metric-depth stream, camera trajectory, measured accuracy bound or physical reference rig.

Calibration and synchronization

Reported calibration level, reproducibility signals, and supporting notes.

Level
Not reported
Processed parameters
Not reported
Raw calibration data
Not reported
Calibration targets
Not reported

Calibration data

Calibration reproducibility details are not reported.

Notes

Paper rendering parameters include a 32 by 18 mm film back, focus plane of 150 simulator units, fx=fy=1400, cx=640 and cy=360. Equation 1 prints an all-zero last intrinsic-matrix row; no corrected matrix or released calibration package is asserted here.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

Synthetic rendering retains a domain gap from real underwater imagery.

The final subset is sampled and filtered for its disparity distribution.

Cameras

virtual-stereo

Model
Not reported
Setup
Stereo
Count
1
Effective cameras
2
Modality
Rgb
Resolution
1280 x 720
Rate
Not reported
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Rendered binocular imagery from two virtual camera components, with translation-only separation and baselines of 6, 12, 18, 24 or 30 simulator units. Animations run at 15 fps with at least 1,500 frames per scene before sampling every fourth frame and further filtering. The final collection is not asserted to retain a continuous 15 Hz sequence or any physical camera-unit count.

IMUs

Not reported

GNSS

Not reported

LiDAR

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
Not reported
Semantic segmentation Not reported
Instance segmentation Not reported
Object detection Not reported
Optical flow Not reported
Depth ground truth Not reported

Citation

Citation key and BibTeX kept at the end of the page for reference.

Citation key
lv2024uwstereo

BibTeX

@misc{lv2024uwstereo,
  author = {Qingxuan Lv and Junyu Dong and Yuezun Li and Sheng Chen and Hui Yu and Shu Zhang and Wenhan Wang},
  title = {UWStereo: A Large Synthetic Dataset for Underwater Stereo Matching},
  year = {2024},
  eprint = {2409.01782},
  archivePrefix = {arXiv},
  primaryClass = {cs.CV},
  url = {https://arxiv.org/abs/2409.01782v1}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy