Real-Time Pothole Detection Using YOLOv8 for Smart Road Monitoring Systems

Abstract:Potholes are a common form of road surface damage and can contribute to accidents, traffic delays, increased vehicle maintenance costs. Detecting them early is essential, but manual examination is slow and often unreliable. The new deep-learning model YOLOv8 was adopted in this study, an automatic detection of potholes from images and videos. The model has been trained on real road data, so is able to spot potholes in different light conditions and weather. The high processing speed of the YOLOv8model makes it can process in real time and so this technology is suitable for mobile phone cameras, surveillance, drone monitoring etc.the results indicates that the system effectively detects potholes and can help authorities to better maintain the smoothness and safety of roads. Keywords: YOLOv8, pothole identification, deep neural network, visual computing, real-time monitoring, object detection, road safety.