PotholeVision: An Automated Pothole Detection and Reporting System using Computer Vision

The rapid growth of remote sensing technologies, superior computing power, and machine learning techniques can help local governments in making pothole detection and reporting more efficient. In this paper, we propose an automated pothole detection and reporting system that utilizes edge computing devices installed on garbage trucks to detect and report potholes automatically. The installed devices capture images of the road surface, and object detection techniques are used to detect potholes. If a pothole is detected, the edge computing device sends the road surface image, GPS latitude, and longitude to the server. The server then counts the number of potholes and prioritizes them based on severity. The proposed system also provides a user-friendly interface for visualizing the potholes' locations on a map. Thus, reducing the need for manual reporting, minimizing time and resources for road maintenance, and increasing road safety.