Improved automatic road damage detection for YOLOv8

作者：Yiying Zhang, Xuefeng Tang, Chunlin He, T. Yang, Zekuan Zhao, Lirong He, Linbo Li, Hao Liu, Xiaoning Hou · 年份：2025 · DOI：10.1117/12.3060708 · 被引用次数：4 · 研究领域：Advanced Neural Network Applications

Road construction serves as a crucial pillar of global development, hence the routine maintenance and inspection of roads hold significant importance. Asphalt pavements are often plagued by issues such as cracks and potholes, and relying solely on manual visual inspection is inefficient and prone to overlooking microscopic problems. Failure to promptly detect road defects may lead to serious traffic accidents, decreased traffic efficiency, as well as exacerbated environmental pollution and economic losses. To address this, this paper proposes an optimized automatic detection model, Road-YOLOv8, for real-time detection of road damage issues. The experimental results demonstrate that the optimized model exhibits higher accuracy and performance compared to other popular models in the same series, particularly excelling in the precise localization of small targets.