This research introduces a comprehensive online benchmark system for evaluating deep learning algorithms in road crack detection, advancing intelligent road inspection technology.

• Establishes a standardized evaluation framework for comparing state-of-the-art algorithms
• Enables more accurate and efficient road crack detection through deep learning models
• Supports the development of Urban Digital Twins (UDTs) for smarter infrastructure management
• Provides a path to replace subjective manual inspections with objective AI-based assessment

For engineering teams, this research offers critical tools to implement automated inspection systems that reduce maintenance costs and extend infrastructure lifespan through early detection of structural issues.

Vehicular Road Crack Detection with Deep Learning: A New Online Benchmark for Comprehensive Evaluation of Existing Algorithms