Search

An Intelligent Traffic Flow Monitoring Method Based on Deep Learning and Multi-Indicator Fusion

Author(s)

Yongyi Lin1,*, Xiaoxu Wei2

Affiliation(s)

1School of Intelligent Media Engineering, Communication University of China Nanjing, Nanjing, China 2Department of Information and Intelligent Engineering, Shanghai Publishing and Printing College, Shanghai, China *Corresponding Author

Abstract

In intelligent traffic video analysis, there is a semantic gap between low-level vehicle detection and tracking outputs and the macro-level traffic state. Besides, congestion assessment based on a single metric often fails to identify latent congestion. To address these problems, this paper formulates intersection traffic state assessment as a mapping from vehicle trajectories to discrete congestion levels and proposes a three-tier progressive monitoring method following the perception, measurement and analysis workflow. The perception layer combines a YOLO detector with a tracking-by-detection strategy to generate vehicle trajectories. The measurement layer performs deduplicated vehicle counting and computes three complementary indicators, namely vehicle density, speed loss ratio, and road space occupancy ratio. The analysis layer fuses these indicators into an interpretable weighted congestion index to determine the congestion level. Experiments on a self-built intersection dataset show that the proposed method achieves a congestion classification accuracy of 93.6% and a macro-averaged F1 score of 0.921, improving accuracy by 12.8 percentage points over the density-only baseline. Ablation experiments further confirm that the speed loss ratio and the road space occupancy ratio are key indicators for identifying latent congestion. The method supports multi-channel video monitoring on general-purpose computers without dedicated GPUs, achieving a good balance between classification accuracy and ease of engineering deployment.

Keywords

Deep Learning; Object Detection; Multi-Object Tracking; Multi-Indicator Fusion; Traffic Congestion Detection

References

[1] Amenah S. M. Thabit, Chaker Abdelaziz Kerrache, Carlos T. Calafate. A survey on monitoring and management techniques for road traffic congestion in vehicular networks. ICT Express, 2024, 10(6): 1186-1198. [2] Zhengxia Zou, Keyan Chen, Zhenwei Shi, et al. Object detection in 20 years: A survey. Proceedings of the IEEE, 2023, 111(3): 257-276. [3] Jahongir Azimjonov, Ahmet Özmen. A real-time vehicle detection and a novel vehicle tracking systems for estimating and monitoring traffic flow on highways. Advanced Engineering Informatics, 2021, 50: 101393. [4] Shankha N. Chattopadhyay, Arvind K. Gupta. Do multivariate measures anticipate traffic congestion? Royal Society Open Science, 2026, 13(5): 252185. [5] Ross Girshick, Jeff Donahue, Trevor Darrell, et al. Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Columbus, USA, 2014: 580-587. [6] Joseph Redmon, Santosh Divvala, Ross Girshick, et al. You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016: 779-788. [7] Wei Liu, Dragomir Anguelov, Dumitru Erhan, et al. SSD: Single shot multibox detector. Proceedings of the European Conference on Computer Vision (ECCV), Amsterdam, Netherlands, 2016: 21-37. [8] Chien-Yao Wang, I-Hau Yeh, Hong-Yuan Mark Liao. YOLOv9: Learning what you want to learn using programmable gradient information. Proceedings of the European Conference on Computer Vision (ECCV), Milan, Italy, 2024: 1-21. [9] Alex Bewley, Zongyuan Ge, Lionel Ott, et al. Simple online and realtime tracking. Proceedings of the IEEE International Conference on Image Processing (ICIP), Phoenix, USA, 2016: 3464-3468. [10]Nicolai Wojke, Alex Bewley, Dietrich Paulus. Simple online and realtime tracking with a deep association metric. Proceedings of the IEEE International Conference on Image Processing (ICIP), Beijing, China, 2017: 3645-3649. [11]Yifu Zhang, Peize Sun, Yi Jiang, et al. ByteTrack: Multi-object tracking by associating every detection box. Proceedings of the European Conference on Computer Vision (ECCV), Tel Aviv, Israel, 2022: 1-21. [12]Yifu Zhang, Chunyu Wang, Xinggang Wang, et al. FairMOT: On the fairness of detection and re-identification in multiple object tracking. International Journal of Computer Vision, 2021, 129(11): 3069-3087. [13]Yunhao Du, Zhicheng Zhao, Yang Song, et al. StrongSORT: Make DeepSORT great again. IEEE Transactions on Multimedia, 2023, 25: 8725-8737. [14]Jinkun Cao, Jiangmiao Pang, Xinshuo Weng, et al. Observation-centric SORT: Rethinking SORT for robust multi-object tracking. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, Canada, 2023: 9686-9696.