Road Surface Analysis Dataset for Pothole Detection

Overview:

This dataset is designed for the development and validation of computer vision models aimed at detecting potholes on various road surfaces. It includes a comprehensive collection of video footage taken from vehicle-mounted cameras, providing a real-world view of diverse road conditions.

Contents:

• Videos: Over 1,000 hours of high-resolution video covering a range of lighting and weather conditions.
• Annotations: Each video is annotated with bounding boxes and labels identifying potholes, with severity levels indicated where applicable.
• Metadata: Includes information on the location, time of day, and weather conditions during each recording.

Applications:

Ideal for researchers and engineers developing AI-driven diagnostics tools for road maintenance, enhancing safety measures, and improving vehicle navigation systems.