Deep Learning Based Road Damage Identification Using Unmanned Aerial Vehicle Images

The transport infrastructure, especially roads, is of great importance in making transport safe and efficient. Nevertheless, the conventional road damage monitoring procedure can be time-consuming and needs to be done manually. The paper will suggest a smart road damage detention system with the help of deep learning and computer vision methods. The suggested system can be designed to identify and tag various forms of road damages including longitudinal cracks, transverse cracks, alligator cracks, and potholes amongst other surface damages. The web-based inspection system is developed based on Django framework that enables a user to submit an image of the road or conduct a road inspection in real time with the help of a live camera. The result of the experimentation process indicates that the developed system can indeed measure damages on roads very efficiently and effectively and this is quite useful in road inspection in real-time in the application of smart city surveillance.