List of Datasets

Automotive

K-Radar

4D radar tensors, adverse-weather detection and multimodal fusion

Fifty-eight driving sequences release dense 4D radar, two LiDARs, stereo images and five-class forward boxes.

Metadata: Complete Metadata source documented Released 2022

Primary places to inspect, download, read about, or reproduce the dataset.

Quick facts

Vehicle type
Automotive
Environment
Outdoor
Data origin
Real-world
Sensor count
15
Ground truth
Available
Calibration
Parameters Only
Annotations
Available

Sensor overview

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

Cameras

9 reported

IMUs

2 reported

GNSS

1 reported

LiDAR

2 reported

Radar

1 reported

Magnetometers

Not reported

Barometers

Not reported

UWB

Not reported

Ground truth

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

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.
6DoF LiDAR SLAM LeGO-LOAM / local navigation reference Not reported Not reported Partial · Label Associated Frames 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.
Derived
Reference notes
Official resources/odometry supplies KITTI poses, relative quaternions/translations and local GPS coordinates. Shared LiDAR/GPS and qualitative camera checks generate/assess the estimate; missing label-associated poses and difficult sequences preclude independent full-coverage truth.

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

Per-sequence info_calib and projection tools are documented. Complete raw calibration sessions and independently measured residual errors remain unknown.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

Fifty-eight sequences include adverse weather and different road domains. Full Doppler/range/azimuth/elevation tensors are very large; cloud or reduced-tensor mirrors do not contain the same information as the complete raw release.

Official download documentation distinguishes local-server/full-raw or shipment options from Google Drive camera/LiDAR/GPS and processed-radar products. No account login or data/shipping request was made during review.

Later odometry labels are derived from shared LeGO-LOAM/GPS/camera observations, with sequence-specific degradation and fewer poses than some raw sensor streams. Local coordinates and identity initial poses are intentional.

Native camera/IMU/GNSS configurations, receiver topology, full calibration and independent trajectory accuracy remain incomplete. Capture rates and device range specifications do not certify delivered cadence or valid measurement extent.

Cameras

Camera 1

Model
ZED2i
Setup
Stereo
Count
4
Effective cameras
8
Modality
Rgb
Resolution
1280 x 720
Rate
Not reported
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Four directional stereo systems are described at 1280x720. Paper capture setting is 30 Hz; complete native-rate image delivery is not established by the selected radar-frame release.

Camera 2

Model
Not reported
Setup
Mono
Count
1
Effective cameras
1
Modality
Rgb
Resolution
Not reported
Rate
Not reported
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Released MP4 dashcam supports annotation; model, dimensions and rate are unreported.

IMUs

Ouster OS2-64 integrated IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

Ouster OS1-128 integrated IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

GNSS

GNSS 1

Model or name
Not reported
Count
1
Rate
1 Hz
GNSS type
RTK
Position output
Available
Velocity output
Not reported
Heading output
Not reported

LiDAR

Ouster OS2-64

Model or name
Ouster OS2-64
Count
1
Dimensions
3
Rate
10 Hz
Channels
64
Range
Not reported

Ouster OS1-128

Model or name
Ouster OS1-128
Count
1
Dimensions
3
Rate
10 Hz
Channels
128
Range
Not reported

Radar

RETINA-4ST

Count
1
Released data
Spectra / dense tensors
Frame rate
10 Hz
Doppler velocity
Available
Elevation angle
Available
Configured range
118 m
Field of view
107° H × 37° V (per radar)
Notes
Dense Doppler/range/azimuth/elevation tensor; 118 m range and 107×37-degree field of view.

Additional sensors

Not reported

Annotations

Reported annotation availability and task support.

Availability
Available
Format
TXT 3D boxes with tracking IDs; forward-only coverage
Class count
5
Semantic segmentation Not reported
Instance segmentation Not reported
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.

Citation key
paek2022kradar

BibTeX

@misc{paek2022kradar,
  author = {{Dong-Hee Paek} and {Seung-Hyun Kong} and {Kevin Tirta Wijaya}},
  title = {K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions},
  year = {2022},
  url = {https://arxiv.org/abs/2206.08171v4}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy