<![CDATA[The significant increase in the number of vehicles entering the Universitas Jenderal Achmad Yani area—especially after the construction of the Faculty of Science and Informatics building—has caused congestion at several strategic points on campus, including the area in front of the campus mosque. This study aims to develop a real-time vehicle density monitoring system to support more efficient campus traffic management. The method used involves applying the YOLOv5 object detection algorithm to identify and count vehicles from video recordings in selected monitoring areas. The system is designed to deliver fast and accurate detection while providing real-time vehicle density information. Testing results show that the system achieved strong detection performance, with a maximum precision value of 1.00 at a confidence threshold of 0.983. The maximum recall value of 0.90 was obtained at a lower confidence threshold, reflecting the system’s ability to detect most objects present. These findings highlight the trade-off between model confidence in predictions and its ability to avoid missing relevant objects. The contribution of this study is the development of a prototype system capable of automatically and in real time monitoring vehicle density in campus areas. This system has the potential to become part of a smarter, data-driven campus traffic management solution to reduce congestion and improve the comfort and mobility of the academic community.]]>
Copyrights © 2025
