Real time road defect monitoring from UAV visual data sources

Iason Katsamenis, Nikolaos Bakalos, Eftychios Protopapadakis, Eleni Eirini Karolou, Georgios Kopsiaftis, Athanasios Voulodimos

Abstract: The use of UAVs and artificial intelligence has emerged as a promising approach for monitoring road defects. This paper highlights the importance of these technologies in improving road inspection, maintenance, and safety. Traditional methods for inspecting roads are often time-consuming, expensive, and can put human inspectors in dangerous situations. However, drones equipped with high-resolution cameras and sensors can capture pavement image data quickly and safely. Deep learning algorithms can then analyze this data to identify and localize areas in need of repair. By leveraging these technologies, engineers and road construction experts can more efficiently monitor and maintain roads, reducing the costs associated with repairs and maintenance, while in parallel improving safety. To this end, this work emphasizes the potential of UAVs in conjunction with deep learning techniques to provide a more comprehensive view of road conditions, allowing for targeted repairs and more effective maintenance strategies, such as prefabrication and robotic interventions. Experimental results using objective evaluation criteria, such as precision, recall, F1-score, and IoU are promising, which entails that this study advocates for the adoption of these technologies to enhance the monitoring and maintenance of road infrastructures.