Real-time road damage detection plays a crucial role in ensuring efficient transportation infrastructure maintenance and enhancing road safety. This study focuses on the application of the YOLOv8 algorithm for real-time road damage detection, aiming to provide timely and accurate information for effective decision-making and prompt repairs. The YOLOv8 algorithm was fine-tuned to identify bounding boxes associated with various types of road damage, including cracks of different sizes and irregular shapes. The model’s performance was evaluated using precision, accuracy (R), F1-score, and mean average precision (mAP) indicators, demonstrating its effectiveness in detecting road damage with high accuracy. The model had an impressive accuracy of 80%. The successful implementation of real-time road damage detection systems has significant implications for proactive and targeted repairs, ultimately contributing to improved road safety and infrastructure longevity. By leveraging the power of YOLOv8 and utilizing a comprehensive data set, this study aims to provide an accurate, efficient, and deployable solution for real-time road damage detection. The findings of this study have the potential to advance transportation infrastructure maintenance practices, particularly in the United States.
Real-Time Road Damage Detection Using YOLOv8
International Conference on Transportation and Development 2025 ; 2025 ; Glendale, Arizona
2025-06-05
Conference paper
Electronic Resource
English