Optimizing real-time traffic flow detection:an integrated approach combining YOLO with advanced image preprocessing
SUN Jiahao
ZOU Ruibin
LI Hefu
GAO Yang
ZHANG Weilin
LIU Yuanxin
Abstract：To improve the efficiency of urban traffic management,the application of intelligent transporta-tion systems(ITS)has become a practical and effective approach.Real-time traffic flow monitoring tech-nology provides essential data support for such systems.This study focuses on enhancing the performance of terminal devices used for real-time traffic monitoring and proposes a method based on the YOLO object detection algorithm.By incorporating image preprocessing techniques,the system achieves improved de-tection accuracy while reducing computational resource consumption at the terminal.The research utilizes a subset of the COCO public dataset to construct a customized training dataset suitable for this task.YOLOv5s and YOLOv8s models were trained and comprehensively evaluated across various scenarios,in-cluding dynamic video and real-time video streams.Techniques such as background subtraction,Contrast Limited Adaptive Histogram Equalization(CLAHE),and median filtering were applied to enhance input image quality.Experimental results demonstrate that these preprocessing methods improve detection accu-racy by approximately 1.2%to 1.8%under different testing conditions and environmental complexities,while also reducing resource usage.This study systematically analyzes key components including model training,image processing,and performance evaluation.Through a series of video-based and real-world simulation experiments,the proposed approach is shown to have practical value for intelligent traffic appli-cations.
Keywords：intelligent transportationYOLOreal-time vehicle flow monitoringmodel trainingimage preprocessing
Publication Date：2026-04-25
Online Publishing Date：2026-08-28(First online date of this platform, not the publication date of the document)
Pages：11( 238-248 )