List of Datasets

Automotive, Handheld

MCD

SLAM and semantic perception

Multi-campus day/night lidar, vision, inertial and UWB recordings with continuous-time survey-map reference poses.

Metadata: Complete Metadata source documented Released 2023

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

Quick facts

Vehicle type
Automotive, Handheld
Environment
Outdoor
Data origin
Real-world
Sensor count
19
Ground truth
Available
Calibration
Parameters Only
Annotations
Available

Sensor overview

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

Cameras

8 reported

IMUs

6 reported

GNSS

1 reported

LiDAR

2 reported

Radar

Not reported

Magnetometers

Not reported

Barometers

Not reported

UWB

2 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 Continuous-time survey-map registration Not reported 10 Hz Unknown 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.
Hybrid
Reference notes
pose_inW.csv is sampled from the survey-map-constrained continuous-time reference every 0.1 seconds. spline.csv permits arbitrary queries; knot spacing is not a second independent reference cadence. Numerical independent dynamic accuracy and exact valid intervals remain unknown.

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

Separate calibrated-parameter YAML files are explicitly supplied for ATV and handheld rigs. Complete raw sessions and measured residual calibration/timing error are not established.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

Eighteen public recordings span NTU, KTH and TUHH. The ATV and handheld rigs have different camera, IMU and Ouster configurations; flat catalog counts can sum mutually exclusive variants.

D455 provides two grayscale streams plus RGB without an established depth topic. Published IMU/GPS ROS frequencies do not prove native component/receiver sampling; optional modalities vary by download row.

References combine shared LiDAR/inertial observations with externally surveyed maps. Discrete 10 Hz poses and continuous spline queries describe one underlying trajectory; full valid-frame membership and independent dynamic accuracy remain unknown.

Ouster header timestamps mark the end of a sweep, while Livox headers mark its start. Respect each point-time convention; calibration parameters and semantic labels do not establish complete raw calibration or uniform annotation coverage.

Cameras

Camera 1

Model
Intel RealSense D435i
Setup
Stereo
Count
1
Effective cameras
2
Modality
Grayscale
Resolution
640 x 480
Rate
30 Hz
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
ATV/NTU only: two infrared grayscale streams; no RGB or depth topic is established for this module.

Camera 2

Model
Intel RealSense D455 bottom
Setup
Trinocular
Count
1
Effective cameras
3
Modality
Rgb
Resolution
640 x 480
Rate
30 Hz
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Both configurations: two grayscale infrared streams and one RGB stream, represented once as three imagers in one module. No depth topic is listed.

Camera 3

Model
Intel RealSense D455 top
Setup
Trinocular
Count
1
Effective cameras
3
Modality
Rgb
Resolution
640 x 480
Rate
30 Hz
Shutter
Not Reported
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Handheld KTH/TUHH only: two grayscale infrared streams plus one RGB stream. Alternative to ATV D435i; these are not all simultaneously present in one rig.

IMUs

Ouster OS1 integrated IMU

Model
IAM-20680HT
Count
1
Accel / gyro
Not reported

D435i integrated IMU

Model
Bosch BMI055
Count
1
Accel / gyro
Not reported

D455 bottom integrated IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

D455 top integrated IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

VectorNav VN100

Model
VectorNav VN100
Count
1
Accel / gyro
Not reported

VectorNav VN200

Model
VectorNav VN200
Count
1
Accel / gyro
Not reported

GNSS

VectorNav VN200

Model or name
VectorNav VN200
Count
1
Rate
Not reported
GNSS type
GPS_IMU
Position output
Available
Velocity output
Not reported
Heading output
Not reported

LiDAR

Ouster OS1

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

Livox Mid70

Model or name
Livox Mid70
Count
1
Dimensions
3
Rate
10 Hz
Channels
Not reported
Range
Not reported

Radar

Not reported

Additional sensors

UWB

Model
LinkTrack P
Name
On-platform ranging nodes
Count
2
Placement
platform
Rate
20 Hz
Role
tag
Notes
Manual lists tag0/tag1 ranging messages for both platforms. Reference-node deployment and exact per-sequence usable ranges vary; no extra environmental-node count is inferred.

Annotations

Reported annotation availability and task support.

Availability
Available
Format
Pointwise semantic labels for annotated non-repetitive LiDAR scans; the 29-class task has subset-specific membership, separate from raw sensor recordings.
Class count
29
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.

Citation key
nguyen2024mcd

BibTeX

@inproceedings{nguyen2024mcd,
  title        = {MCD: Diverse Large-Scale Multi-Campus Dataset for Robot Perception},
  author       = {Nguyen, Thien-Minh and Yuan, Shenghai and Nguyen, Thien Hoang and Yin, Pengyu and Cao, Haozhi and Xie, Lihua and Wozniak, Maciej and Jensfelt, Patric and Thiel, Marko and Ziegenbein, Justin and Blunder, Noel},
  booktitle    = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages        = {22304--22313},
  license      = {CC BY-NC-SA 4.0},
  url          = {https://mcdviral.github.io/},
  version      = {1.0},
  year         = {2024},
  month        = {6}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy