Designing an Intelligent Pavement Maintenance and Management System using Drone Imagery and Artificial Intelligence

Abstract

This paper presents the development of an innovative pavement maintenance and management system leveraging advanced drone imagery and Convolutional Neural Network (CNN) image classification. Our system is designed to perform 2D modelling of road surfaces using high-resolution images captured by drones. These images are then analysed by a CNN model specifically trained to detect and classify pavement damages in accordance with the IRC:82 'Code of Practice for Maintenance of Bituminous Surfaces of Highways'. The classification process identifies various types of road distresses such as cracks, potholes, and surface wear. Each identified distress is documented in a comprehensive report detailing the nature of the damage and recommending specific remedies as per IRC guidelines. Furthermore, the system categorizes the severity of the damages, facilitating the dispatch of these results to maintenance authorities for immediate action. This ensures that repair efforts are prioritized effectively, contributing to the maintenance of safer and higher quality roadways. By automating the detection and classification of road damages, this system not only accelerates the repair process but also plays a crucial role in reducing road accidents by maintaining better road conditions. This approach showcases the potential of integrating artificial intelligence and drone technology in the field of road maintenance, marking a significant step towards smarter and safer road infrastructure.