Abstract
The present scientific research focuses on the creation of a system for detecting potholes in asphalt to avoid traffic accidents. The motivation for choosing this topic of wide interest comes from the desire to apply the new modern technologies of artificial intelligence and deep learning in an extremely important field of road safety. The use of platforms and systems based on artificial intelligence in smart cities by correctly and quickly detecting potholes in asphalt is very important for autonomous vehicles. To achieve this goal, we used the YOLOv5 algorithm, one for real-time object detection, and the built-in Jetson Nano system, a powerful and efficient platform for AI applications. The final goal of this project is to develop a high-performance system that can recognize and classify potholes in asphalt from images and video streams in real time, with as much precision as possible and that can be used in smart cars.