Enhanced Pothole Detection Using YOLOv8 Nano

Abstract

With over 50 million cars sold annually and over 1.3 million motor vehicle accident deaths worldwide each year, road safety is an urgent concern. It is critical to address driving behavior in countries like India, which contributes 11% of all traffic fatalities worldwide. This study uses YOLOv8, a state-of- the-art deep learning model, to improve pothole identification, hence improving road safety. Our technology achieves a detection accuracy of over 90% by utilizing YOLOv8, which considerably lowers the likelihood of accidents caused by potholes. Moreover, our method has exceptional scalability, processing more than thirty frames per second, which makes it appropriate for real- time implementation in a variety of road conditions. Through rigorous experimentation and analysis, we showcase the potential of our system to revolutionize road safety measures by proactively addressing pothole-related hazards. Index Terms—Pothole detection, YOLOv8, Deep learning, Object detection, Real-time detection, Road safety.