Scene #1
Canadian Adverse Driving Conditions Dataset
Open-source dataset for autonomous driving in wintry weather.
Overview
High quality data for adverse driving conditions
Data Collection
Complex Driving Scenarios in Adverse Conditions
Car Setup
Vehicle, Sensor and Camera Details
- 8
- 10 Hz capture frequency
- 1/1.8” CMOS sensor of 1280x1024 resolution
- Images are stored as PNG
Wide Angle Cameras - 1
- 10 Hz capture frequency
- 32 channels
- 200m range
- 360° horizontal FOV; 40° vertical FOV (-25° to +15°)
LiDAR - 1Post-processed GPS and IMU
More on Autonomoose: The University of Waterloo's self-driving research platform.
Sensor Calibration
Data alignment between sensors and cameras
LiDAR extrinsics
Camera extrinsics
Camera intrinsic calibration
IMU extrinsics
Data Annotation
Complex Label Taxonomy
The resulting accuracy is consistently higher than what a human or synthetic labeling approach can achieve independently as measured against seven rigorous quality areas for each annotation.
The CADC includes 3D Bounding boxes for X object classes and a rich set of class attributes related to X, Y. For detailed definitions of each class and example images, please see the annotation instructions.
Instances Per Label
28194
Cars
62851
Pedestrians
20441
Trucks
4867
Bus
4808
Garbage Containers on Wheels
3205
Traffic Guidance Objects
705
Bicycle
638
Pedestrian With Object
75
Horse and Buggy
26
Animals
Get Started with CADC Dataset
View our paper here and download the development kit.
If you use our dataset please cite our paper.
Download Dataset