A Comprehensive Survey on Road Damage Detection: Datasets, Methods, and Platforms

Abstract

Road damage detection has emerged as a strategic cornerstone of modern infrastructure management, directly impacting transportation safety, economic sustainability, and urban resilience. This review provides a comprehensive overview of state-of-the-art methodologies integrating diverse sensors (including cameras, LiDAR, ground penetrating radar, and accelerometers) with cutting-edge deep neural networks (such as CNNs, Transformers, and YOLO models) for precise classification and localization of road defects, including cracks, potholes, and subsurface anomalies. We also examines publicly available datasets, such as road damage detection dataset-series and UAV-Benchmark-Dataset, and discuss emerging themes in synthetic data generation and lightweight architectures aimed to mitigate dataset scarcity and improve model robustness. By synthesizing technical breakthroughs with real-world applications, this review provides a foundational reference for researchers and practitioners aiming to develop intelligent, scalable solutions for road infrastructure monitoring, ultimately contributing to safer, more sustainable transportation ecosystems.