Automotive
ORFD
Off-road traversability and multimodal freespace segmentation
Thirty off-road runs yield 12,198 RGB/LiDAR pairs with three-way traversability masks and projected/interpolated depth.
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
- 2
- 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
Camera intrinsics and camera–LiDAR extrinsics are supplied; raw calibration sessions are not established.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
The 12,198 image/cloud pairs are split into 8,398 training, 1,245 validation and 2,555 test pairs. All 361 grassland pairs are in training, so held-out grassland performance is not measured by that split.
Sparse projected and interpolated dense depth are derived from input LiDAR and should not be treated as independent depth ground truth.
The released forward RGB stream is established; the four additional mounted monochrome cameras are not included in the catalog sensor inventory.
Camera and LiDAR acquisition rates, synchronization precision, raw calibration sessions and quantitative label uncertainty remain unverified. The 1280×704 network input is distinct from the 1280×720 released images.
Traversability masks depend on visible geometry and the recording vehicle; they do not guarantee safety or feasibility for another vehicle.
Cameras
Camera 1
- Model
- Hesai Pandora color camera
- Setup
- Mono
- Count
- 1
- Effective cameras
- 1
- Modality
- Rgb
- Resolution
- 1280 x 720
- Rate
- Not reported
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Only released RGB is established; four mounted monochrome cameras are omitted.
IMUs
Not reported
GNSS
Not reported
LiDAR
Hesai Pandora
- Model or name
- Hesai Pandora
- Count
- 1
- Dimensions
- 3
- Rate
- Not reported
- Channels
- 40
- Range
- Not reported
Radar
Not reported
Additional sensors
Not reported
Annotations
Reported annotation availability and task support.
- Availability
- Available
- Format
- Image-plane traversable/non-traversable/unreachable pixel masks
- Class count
- 3
Semantic segmentation
Available
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.
@misc{min2022orfd,
author = {Min, Chen and Jiang, Weizhong and Zhao, Dawei and Xu, Jiaolong and Xiao, Liang and Nie, Yiming and Dai, Bin},
title = {{ORFD}: A Dataset and Benchmark for Off-Road Freespace Detection},
year = {2022},
eprint = {2206.09907},
archivePrefix = {arXiv},
primaryClass = {cs.CV},
url = {https://arxiv.org/abs/2206.09907v2}
}Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy