YOLO-Based Vehicle Detection and Counting for Traffic Control on Highway

Abstract

The roads are becoming overcrowded with increasing vehicles, expressways, and highways. Therefore, vehicles' detection, classification, tracking, and counting are crucial for various civilian, military, and government applications, including highway monitoring, toll collection, traffic planning, and traffic flow management. Vehicle detection is a critical step in traffic management. Computer Vision (CV) based techniques are well-suited for this task as they allow installation without disrupting traffic and can be easily modified. This paper proposes a cost-effective, portable system based on CV techniques for detecting and counting vehicles moving over the road. The proposed system analyses images extracted from video sequences to detect moving vehicles by separating the background from the images. The background extracted is subsequently used for additional analysis to identify and categorize various vehicles, including light, heavy, and motorcycles.