Performance optimization of traffic efficiency diagnosis system based on asynchronous pipeline and sliding window

ZHAO Xinyu; ZHANG Shuojin; YE Yutong; WANG Ruijie

doi:10.11959/j.issn.2096-0271.BDR26249

Chinese

您当前的位置：

首页 >

文章列表页 >

Performance optimization of traffic efficiency diagnosis system based on asynchronous pipeline and sliding window

更新时间：2026-09-24

- Performance optimization of traffic efficiency diagnosis system based on asynchronous pipeline and sliding window
- Big Data Research (2026)
- 作者机构：
  
  1.北京航空航天大学软件学院，北京市100191
  2.北京航空航天大学计算机学院，北京市100191
- 作者简介：
- 基金信息：
- DOI：10.11959/j.issn.2096-0271.BDR26249
  CLC： TP391
- Received：11 June 2026，
  
  Revised：2026-07-17，
  
  Accepted：24 September 2026，
- 稿件说明：
移动端阅览
ZHAO Xinyu, ZHANG Shuojin, YE Yutong, et al. Performance optimization of traffic efficiency diagnosis system based on asynchronous pipeline and sliding window[J/OL]. Big Data Research, 2026.
DOI：

ZHAO Xinyu, ZHANG Shuojin, YE Yutong, et al. Performance optimization of traffic efficiency diagnosis system based on asynchronous pipeline and sliding window[J/OL]. Big Data Research, 2026. DOI： 10.11959/j.issn.2096-0271.BDR26249.

摘要

城市道路交叉口交通效率诊断需从监控视频中持续完成车辆检测、多目标跟踪、车道级特征提取和拥堵传播预测。针对现有系统前端视频处理串行等待与后端连续窗口重复构造等效率瓶颈，围绕交通效率诊断场景构建了异步流水线与滑动窗口协同优化方案。前端将视频读取、车辆检测、多目标跟踪和结果输出拆分为衔接阶段，通过线程安全队列实现并行流动；后端维护固定长度窗口缓冲并复用历史窗口状态，减少重复构造与调度开销。在约109

300帧丁字路口监控视频上的实验表明，系统总处理时间由281.90分钟缩短至27.60分钟，平均帧率（frames per second

FPS）由6.46 FPS提升至65.94 FPS，整体效率提升约10倍。一致性实验表明，滑动窗口机制不改变ST-GCN输出，异步流水线不改变检测与跟踪结果。

Abstract

Traffic efficiency diagnosis at urban intersections requires continuous processing of surveillance video streams for vehicle detection

multi-object tracking

lane-level feature extraction

and congestion propagation prediction

combining high-frequency video processing with continuous temporal reasoning. Existing systems can complete the full workflow

but their practical implementations often suffered from serial waiting in front-end video processing and

redundant window construction in back-end temporal inference

affecting system real-time

关键词

Keywords

references

Krizhevsky A , Sutskever I , Hinton G E . ImageNet classification with deep convolutional neural networks [J ] . Communications of the ACM , 2017 , 60 ( 6 ): 84 - 90 .

Simonyan K , Zisserman A . Very deep convolutional networks for large-scale image recognition [C ] // Proceedings of the 3rd International Conference on Learning Representations . San Diego : ICLR , 2015 : 1 - 14 .

董红召 , 林少轩 , 佘翊妮 . 交通目标YOLO检测技术的研究进展 [J ] . 浙江大学学报(工学版) , 2025 , 59 ( 2 ): 245 - 260 .

Dong H Z , Lin S X , She Y N . Research progress of YOLO detection technology for traffic targets [J ] . Journal of Zhejiang University (Engineering Science) , 2025 , 59 ( 2 ): 245 - 260 .

张浩晨 , 张竹林 , 史瑞岩 , 等 . 基于YOLO-NPDL的复杂交通场景检测方法 [J ] . 山东交通学院学报 , 2025 , 33 ( 2 ): 34 - 47 .

Zhang H C , Zhang Z L , Shi R Y , et al . Complex traffic scene detection method based on YOLO-NPDL [J ] . Journal of Shandong Jiaotong University , 2025 , 33 ( 2 ): 34 - 47 .

谭金菁 , 毛文逸 , 付俊强 , 等 . 基于YOLOv8的道路交通目标检测优化算法 [J ] . 上海第二工业大学学报 , 2025 , 44 ( 4 ): 78 - 89 .

Tan J J , Mao W Y , Fu J Q , et al . Optimization algorithm for road traffic target detection based on YOLOv8 [J ] . Journal of Shanghai Second Polytechnic University , 2025 , 44 ( 4 ): 78 - 89 .

Wojke N , Bewley A , Paulus D . Simple online and realtime tracking with a deep association metric [C ] // 2017 IEEE International Conference on Image Processing (ICIP) . Beijing : IEEE , 2017 : 3645 - 3649 .

金立生 , 华强 , 郭柏苍 , 等 . 基于优化DeepSORT的前方车辆多目标跟踪 [J ] . 浙江大学学报(工学版) , 2021 , 55 ( 6 ): 1056 - 1064 .

Jin L S , Hua Q , Guo B C , et al . Multi-object tracking of preceding vehicles based on optimized DeepSORT [J ] . Journal of Zhejiang University (Engineering Science) , 2021 , 55 ( 6 ): 1056 - 1064 .

徐福良 . 城市交通视域下基于YOLOv8和DeepSORT的检测跟踪算法研究 [D ] . 青岛 : 青岛大学 , 2025 .

Xu F L . Research on detection and tracking algorithm based on YOLOv8 and DeepSORT in urban traffic scenario [D ] . Qingdao : Qingdao University , 2025 .

Yu B , Yin H , Zhu Z . Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting [C ] // Proceedings of the 27th International Joint Conference on Artificial Intelligence . Stockholm : IJCAI , 2018 : 3634 - 3640 .

钟强 . 基于图卷积神经网络的交通流预测研究 [D ] . 长沙 : 中南大学 , 2023 .

Zhong Q . Research on traffic flow prediction based on graph convolutional neural network [D ] . Changsha : Central South University , 2023 .

徐红飞 , 李婧 . 基于时空卷积神经网络GL-GCN的交通流异常检测算法 [J ] . 现代信息科技 , 2021 , 5 ( 2 ): 70 - 75 .

Xu H F , Li J . Traffic flow anomaly detection algorithm based on spatio-temporal convolutional neural network GL-GCN [J ] . Modern Information Technology , 2021 , 5 ( 2 ): 70 - 75 .

Szegedy C , Liu W , Jia Y , et al . Going deeper with convolutions [C ] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Boston : IEEE , 2015 : 1 - 9 .

He K , Zhang X , Ren S , et al . Deep residual learning for image recognition [C ] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Las Vegas : IEEE , 2016 : 770 - 778 .

Huang G , Liu Z , Van Der Maaten L , et al . Densely connected convolutional networks [C ] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Honolulu : IEEE , 2017 : 4700 - 4708 .

Redmon J , Farhadi A . YOLOv3: An incremental improvement [J ] . arXiv preprint arXiv: 1804.02767 , 2018 .

Ultralytics . YOLOv5: An improved version of YOLO [EB/OL ] . 2021 .

Wang C Y , Bochkovskiy A , Liao H Y M . YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver : IEEE , 2023 : 7464 - 7475 .

钱睿珂 , 胡海霞 . 基于改进YOLOv8的交通场景目标检测方法 [J ] . 商洛学院学报 , 2025 , 39 ( 6 ): 52 - 61 .

Qian R K , Hu H X . Traffic scene object detection method based on improved YOLOv8 [J ] . Journal of Shangluo University , 2025 , 39 ( 6 ): 52 - 61 .

张雨 , 罗素云 . YOLOv8-DSM: 高精度的小目标交通检测算法 [J ] . 传感器与微系统 , 2025 , 44 ( 3 ): 145 - 152 .

Zhang Y , Luo S Y . YOLOv8-DSM: High-precision small target traffic detection algorithm [J ] . Transducer and Microsystem Technologies , 2025 , 44 ( 3 ): 145 - 152 .

王桂宝 , 李鹏飞 , 熊召新 . YOLOv8-NTS: 一种针对交通标志检测的目标识别方法 [J ] . 陕西理工大学学报(自然科学版) , 2025 , 39 ( 2 ): 45 - 51 .

Wang G B , Li P F , Xiong Z X . YOLOv8-NTS: A target recognition method for traffic sign detection [J ] . Journal of Shaanxi University of Technology (Natural Science Edition) , 2025 , 39 ( 2 ): 45 - 51 .

顾立鹏 , 孙韶媛 , 李想 , 等 . 无人车驾驶场景下的多目标车辆与行人跟踪算法 [J ] . 小型微型计算机系统 , 2021 , 42 ( 3 ): 542 - 549 .

Gu L P , Sun S Y , Li X , et al . Multi-object vehicle and pedestrian tracking algorithm in unmanned vehicle driving scenarios [J ] . Journal of Chinese Computer Systems , 2021 , 42 ( 3 ): 542 - 549 .

金沙沙 , 龙伟 , 胡灵犀 , 等 . 多目标检测与跟踪算法在智能交通监控系统中的研究进展 [J ] . 控制与决策 , 2023 , 38 ( 5 ): 1201 - 1210 .

Jin S S , Long W , Hu L X , et al . Research progress of multi-object detection and tracking algorithms in intelligent traffic surveillance systems [J ] . Control and Decision , 2023 , 38 ( 5 ): 1201 - 1210 .

袁凯 , 户佐安 , 邓锦程 , 等 . 图神经网络在交通预测中的应用综述 [J ] . 交通运输工程学报 , 2023 , 23 ( 5 ): 39 - 61 .

Yuan K , Hu Z A , Deng J C , et al . Review on application of graph neural networks in traffic prediction [J ] . Journal of Traffic and Transportation Engineering , 2023 , 23 ( 5 ): 39 - 61 .

汤泽慧 . 基于时空图卷积网络的道路行程车速预测模型 [D ] . 北京 : 中国人民公安大学 , 2025 .

Tang Z H . Road travel speed prediction model based on spatio-temporal graph convolutional network [D ] . Beijing : People's Public Security University of China , 2025 .

张祥玉 . 基于周期解耦的动态图卷积神经网络交通流预测 [D ] . 兰州 : 兰州大学 , 2024 .

Zhang X Y . Traffic flow prediction based on periodic decoupled dynamic graph convolutional neural network [D ] . Lanzhou : Lanzhou University , 2024 .

张鹏程 . 基于多尺度动态时空图卷积网络的交通流预测研究 [D ] . 兰州 : 兰州理工大学 , 2025 .

Zhang P C . Research on traffic flow prediction based on multi-scale dynamic spatio-temporal graph convolutional network [D ] . Lanzhou : Lanzhou University of Technology , 2025 .

王建龙 . 基于图过程神经网络的城市道路拥堵时空预测 [D ] . 武汉 : 武汉大学 , 2022 .

Wang J L . Spatio-temporal prediction of urban road congestion based on graph process neural network [D ] . Wuhan : Wuhan University , 2022 .

吕甜甜 . 深度学习流水线并行优化方法研究 [D ] . 济南 : 齐鲁工业大学 , 2024 .

Lyu T T . Research on pipeline parallelism optimization methods for deep learning [D ] . Jinan : Qilu University of Technology , 2024 .

朱昌宏 . 面向行人监测的多路实时视频流处理框架设计与应用 [D ] . 北京 : 北京邮电大学 , 2024 .

Zhu C H . Design and application of multi-channel real-time video stream processing framework for pedestrian monitoring [D ] . Beijing : Beijing University of Posts and Telecommunications , 2024 .

NVIDIA Corporation . NVIDIA TensorRT: A programmable inference accelerator [R ] . Santa Clara : NVIDIA , 2024 .

Ultralytics . Ultralytics YOLO: StreamLoader for real-time video inference [EB/OL ] . 2023 .

Microsoft Corporation . ONNX Runtime: Cross-platform ML inference and training accelerator [EB/OL ] . 2021 .

NVIDIA Corporation . NVIDIA DeepStream SDK developer guide [R ] . Santa Clara : NVIDIA , 2024 .

Demidovskij A , Tarkhanov Y , Gorbachev E , et al . OpenVINO deep learning workflow toolkit: Analysis of improvements [C ] // Proceedings of the 30th International Conference on Computer Graphics and Machine Vision . St . Petersburg : CEUR-WS , 2020 : 1 - 7 .

Taymans W , Baker S , Wingo A , et al . GStreamer application development manual [M ] . Boston : Free Software Foundation , 2023 .

傅浩 . 面向视频分析的CNN并行计算优化研究 [D ] . 天津 : 天津大学 , 2021 .

Fu H . Research on CNN parallel computing optimization for video analysis [D ] . Tianjin : Tianjin University , 2021 .

张霞 , 乔环宇 , 曹峰 . 基于改进YOLOv8的高分辨率遥感图像目标检测算法 [J ] . 大数据 , 2025 , 11 ( 5 ): 152 - 169 .

Zhang X , Qiao H Y , Cao F . High-resolution remote sensing image object detection algorithm based on improved YOLOv8 [J ] . Big Data Research , 2025 , 11 ( 5 ): 152 - 169 .

彭森 , 张珂珂 , 杨一群 , 等 . 基于改进YOLOv8的矿井复杂环境小目标检测算法 [J ] . 大数据 , 2025 , 11 ( 6 ): 166 - 181 .

Peng S , Zhang K K , Yang Y Q , et al . Small object detection algorithm for mine complex environment based on improved YOLOv8 [J ] . Big Data Research , 2025 , 11 ( 6 ): 166 - 181 .

王亮 . 基于边缘计算的高效实时视频分析系统研究 [D ] . 武汉 : 华中科技大学 , 2024 .

Wang L . Research on efficient real-time video analysis system based on edge computing [D ] . Wuhan : Huazhong University of Science and Technology , 2024 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Intensive paradigm innovation for digital government development under the provincial dual-tier platform architecture

Analysis of the current status of terminology conflicts in data standards

Constructing handwritten grammatical error correction dataset for CSL learners: methodology and application

A review of key technologies for medical data governance

Construction and leakage-controlled evaluation of an airborne multi-source microphysical dataset for weather modification

Related Author

No data

Related Institution

No data

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰