<![CDATA[Traffic congestion at urban intersections is largely caused by fixed-time traffic light systems that are unable to adjust to fluctuating vehicle volumes. Adaptive traffic light control systems address this problem by regulating signal duration based on actual vehicle density, while computer vision-based vehicle detection provides input data for the control process. This article analyzes the gap between the data required by the control system and the data produced by detection methods under real field conditions. This study is a conceptual literature-based review and does not involve direct system testing. The findings show that several detection methods, including blob detection, YOLOv3, YOLOv4, YOLOv8, HSV, and K-Means, experience performance degradation due to vehicle density, weather, lighting, video quality, camera angle, visibility, and overlapping vehicle positions. The main gap lies in the mismatch between the control system’s need for accurate data and the limitations of detection methods in real traffic conditions. The development of adaptive traffic light systems needs to prioritize vehicle detection accuracy so that traffic signal timing can operate effectively.]]>
Copyrights © 2026
