Road Damages Detection and Classification using Deep Learning and UAVs

The Road health management is particularly important, especially for big cities and countries. Problems that occur on roads like road cracks can be extremely dangerous to drivers' and passengers' lives. In this paper, a road monitoring system is proposed to detect and classify the occurring problems on the road that happened due to obstacles, such as excavations. This work will help in repairing critical road damages faster and save people from accidents that are caused by these damages. The proposed model will detect and classify the problem related to road damage into categories (cracks, potholes, and other damages). This proposed model will be built using a deep learning technique which is convolutional neural networks (CNN). It has been found that CNN is widely used in this area and images detection and classification because it shows high performance. Numerous works have been done in this field, but it is hoped that this proposed technique will achieve better results. The proposed model will be connected with a drone, and it is linked to a web application to demonstrate the results and manage the system. Also, an announcement to government agencies such as the Ministry of Transportation or police could be sent using the web application. Theoretically, the outcomes from of this work shall demonstrate extremely reasonable findings in detecting and classifying road damages from real-time recording.