Intelligent Pavement Monitoring with Real-Time Geolocated Defect Detection on Smartphones

Abstract: Efficient road pavement inspection is essential to ensure the safety and proper maintenance of transportation infrastructure. However, continuous monitoring of these roads presents challenges, such as the manual identification of defects and their accurate geolocation. This study aims to develop a mobile application capable of real-time detection of defects in asphalt pavements, utilizing embedded computer vision algorithms and machine learning. The proposed method involves the automatic capture of images and georeferenced data every 100 meters, combined with the assessment of longitudinal irregularities and the identification of defects, such as potholes, cracks, and patches. Local processing on the smartphone enables instantaneous analysis of road conditions, with results displayed in real-time through a graphical interface that integrates a detailed map of detected defects along the route. The solution was validated on real roads, demonstrating high accuracy in defect identification and highlighting its potential for practical applications in continuous pavement monitoring. This approach significantly contributes to the automation and efficiency of the road inspection process, allowing infrastructure managers to quickly and effectively collect and analyze data, even in remote areas. The application of this technology has the potential to optimize road maintenance planning, reduce operational costs, and facilitate strategic decision-making for infrastructure managers.