Real-Time Traffic Density Estimation Using Yolov8

Abstract

This project uses the YOLOv8 Deep Learning algorithm to accurately identify and classify vehicles in video streams for live traffic density. The model can accurately identify multiple types of vehicles; cars, buses, trucks, and motorcycles, all while achieving high accuracy and running in real-time. Stream lit is integrated with the system to provide an interactive user interface displaying the vehicle count, classification statistics, and visual representations of detected vehicles on traffic footage. Advanced object-tracking methods have also been utilized to monitor vehicles moving through a specific area allowing for in-and-out vehicle counting, flow direction analysis, and estimating how congested an area is based on the real-time traffic patterns. By analyzing the density of vehicles and the distribution of the vehicle categories, the system will produce valuable information for urban traffic management, optimizing traffic signals, developing smart parking solutions, and creating intelligent transportation systems. The proposed solution is a cost-effective, scalable, and fully automated method for intelligent traffic monitoring and greatly reduces the need for manual input and improves data-driven decision-making processes for smart city initiatives.