Traffic-IT Dataset Statistics

Explore the detailed breakdown of Traffic-IT's structure, showcasing the diversity in traffic scenes, weather conditions, and annotation types that make this dataset a comprehensive resource for traffic scene understanding.

Overall Dataset Summary

Total Images: 30,000

Total Question-Answer Pairs: 230,000

Total Annotations: 220,950

Geographical Coverage: Includes locations from Beijing, Chengdu, Melbourne, and others

Categories Covered: Weather conditions, location types, and times of day

Category-Specific Statistics

Weather Conditions:

Sunny: 6,426 images (28%)

Rainy: 5,956 images (26%)

Foggy: 2,578 images (11%)

Snowy: 2,199 images (9%)

Location Types:

Parking Lot: 4,907 images (20%)

City Road: 1,831 images (7%)

Time of Day:

Daytime: 18,000 images (60%)

Nighttime: 12,000 images (40%)

Annotation Type Statistics

Question-Answer Pairs: 230,000 pairs

Scene Annotations: 220,950 annotations, covering various traffic scenarios

Expert Validation: 1,800+ hours dedicated to ensuring accuracy and relevance

Global Diversity: Images collected from multiple countries and cities, capturing diverse traffic environments