Speed Estimation and Vehicle Tracking using YoloV8

Vehicle tracking and speed detection play a vital role in enhancing road safety and managing traffic flow efficiently. Traditional systems like RADAR and LIDAR can be expensive and complex to deploy. In this work, authors investigate and leverage the YOLOv8 deep learning model, which is based on a convolutional neural network (CNN) architecture, combined with centroid tracking and FPS-based speed estimation, to introduce a real-time, cost- effective solution for accurately detecting and monitoring vehicles in video streams. YOLOv8 is a powerful object detection algorithm that offers fast and accurate detection of vehicles in video frames, making it ideal for real-time applications like traffic monitoring. Through thorough evaluation, our system delivers high accuracy in speed estimation. The proposed model is versatile, non-intrusive, and adaptable for various applications such as traffic law enforcement, smart city infrastructure, and autonomous vehicle systems.