A Real-Time AI Vision-Based Vehicle Counting Smart Sensor on the Jetson Orin Nano
Hamza Elouiaazzani, Nicolas Marilleau, Tri Nguyen-Huu, Ahmed Laatabi, Mohamed Ait Babram · Procedia Computer Science · 2026
Accurate vehicle counting is essential for urban mobility and environmental health applications, including pollution modeling, congestion analysis, and traffic simulation. This paper presents an empirical study conducted in Marrakech evaluating an AI vision-based smart sensor built on a low-cost Jetson Orin Nano. Our results show that a software pipeline combining a frame grabber, video strider, YOLO12n detector, OcSORT tracker, and a line-vicinity counter delivers reliable, real-time vehicle counting performance.