Article Details

Abstract

In order to solve the shortcomings of traditional traffic flow statistics algorithms in real-time, stability and accuracy, a target detection algorithm based on improved YOLOv10 algorithm is proposed. The improved YOLOv10 algorithm is combined with DeepSORT tracking technology to construct a new traffic flow statistical architecture. The traffic scene video of the actual road condition is collected to verify the accuracy of the framework. The experimental results show that compared with the original algorithm, the average accuracy of the improved new algorithm in vehicle detection is improved by 1%, and the statistical accuracy of video traffic flow is improved by 3.46%.

Keywords

computer vision; traffic flow statistic; YOLO; DeepSORT; intelligent transportation

References

1. [1] Gong Hua, Zhang Zhongyao, Hu Yutong, etc Infrared video multi-target tracking based on improved YOLOv7 DeepSort [J]. Journal of Shenyang University of Technology, 2024, 43 (6): 20-27
2. [2] Ma Jianmin, Zhu Lei, Luo Youzeng Research on Station Crowd Monitoring Based on Multi object Tracking Algorithm [J]. Computing Technology and Automation, 2024, 43 (3): 75-81
3. [3] Zhang Haichuan A car following assistance warning algorithm based on lightweight YOLOv5 and DeepSort [J]. Modern Information Technology, 2024, 8 (18): 59-65
4. [4] He Shuilong, Zhang Jingjia, Zhang Linjun, etc Vehicle tracking algorithm based on YOLOv5+DeepSORT improved by Transformer [J]. Automotive Technology, 2024 (7): 9-16
5. [5] Wang Jiaru, Lv Bin, Wu Jianqing, etc Queue length sensing method based on shock wave model and YOLOv5 DeepSORT unidirectional coupling [J]. Journal of Shandong University: Engineering Edition, 2024, 54 (5): 42-49
6. [6] Huang Hao Research and Implementation of Multi object Tracking Algorithm Based on DeepSORT [D]. Nanjing: Nanjing University of Information Science and Technology, 2024
7. [7] Wang Jinhuan, Li Baomin Research on Real time Traffic Flow Collection System Based on YOLO Model [J]. Computer Technology and Development, 2024, 34 (9): 209-214
8. [8] Zheng Yangzhang Research on Road Pedestrian Abnormal Behavior Detection Based on Deep Learning [D]. Zhengzhou: Henan Agricultural University, 2024
9. [9] Zhu Hao Research on Vehicle Detection and Tracking Methods Based on Deep Learning [D]. Huainan: Anhui University of Science and Technology, 2024
10. [10] Ma Qinglu, Wu Yuechuan, Zhang Zixuan, etc Improvement and Application of YOLOv7 Algorithm for Intelligent Detection of Traffic Flow [J]. Highway, 2024, 69 (1): 242-249

Funding

2023 Autonomous Region University Basic Research Business Fund Project - Key Technology Research on Environmental Perception in Complex Traffic Scenarios (RZ2300001569); 2023 Autonomous Region Science and Technology Plan - Key Technology Research on Traffic Safety Risk Monitoring System and Early Warning Model (RZ2300002822)

About the authors

Baiyun (1973-), male, Han ethnicity, from Ulanqab, Inner Mongolia, associate professor, master's degree, research direction: traffic information processing, vehicle network; Liu Dandan (2001-), female, Han ethnicity, from Dazhou, Sichuan, is currently pursuing a master's degree in deep learning and image processing.