250,000 images of Chinese license plates across categories like weather, tilt, and illumination for oriented bounding box detection. Each image includes annotations for plate localization and character recognition across various parking lot environments.
Use Cases
- Train license plate localization models using the four-point oriented bounding box coordinates
- Develop character recognition algorithms using the alphanumeric and provincial character labels
- Evaluate model performance under adverse conditions using the 'ccpd_weather' and 'ccpd_blur' image sets
Strengths
- 250,000+ images captured in real-world parking scenarios
- Oriented bounding box (OBB) annotations for precise four-corner localization
- Subsets categorized by environmental challenges including 'ccpd_weather', 'ccpd_tilt', and 'ccpd_blur'