Pothole Detection Using YOLO Models & Machine Learning for Early-Warning Alert Systems
Pavan Kalyan T Kalyan T
Paper Contents
Abstract
In the ceaseless cadence of everyday life, the infrastructure we depend on, especially the roads we travel on, typically falls victim to neglect. This neglect comes in the form of degrading road surfaces, where potholes come as a ubiquitous menace, contributing immensely to road accidents globally. This paper offers an innovative solution towards addressing this problem through sophisticated pothole detection and analysis. Utilizing video recorded by dash cams or carefully positioned cameras, we investigate and contrast a number of cutting-edge approaches, such as LiDAR based methods and different iterations of You Only Look Once (YOLO) object detection algorithms, to detect and pinpoint potholes. After this preliminary detection, we utilize advanced depth analysis algorithms to determine the severity of these road defects. Through the convergence of these technologies, we are suggesting a complete system with real-time pothole detection and severity grading. Not only will this system provide timely warning to drivers, but it also produces in-depth reports for the local authorities to allow for swift and focused road maintenance interventions. At the end of the day, our intention is to promote road safety and mitigate the frequency of accidents triggered by potholes, and enable smoother and safer travel for everyone.
Copyright
Copyright © 2025 Pavan Kalyan T. This is an open access article distributed under the Creative Commons Attribution License.