基于空地信息融合的智慧高速交通状态判别  

Traffic Status Discrimination for Smart Expressway based on Air-ground Information Fusion

作　　者：王浩楠 郭亚娟 张萌萌[1] 康传刚 李镇 刘亮 WANG Hao-nan;GUO Ya-juan;ZHANG Meng-meng;KANG Chuan-gang;LI Zhen;LIU Liang(School of Transportation and Logistics Engineering,Shandong Jiaotong University,Jinan 250357,China;Shandong High-speed Co.Ltd.,Jinan 250000,China)

机构地区：[1]山东交通学院交通与物流工程学院,济南市2250357 [2]山东高速股份有限公司,济南市250000

出　　处：《公路》2025年第4期292-299,共8页Highway

基　　金：山东省自然科学基金项目,项目编号ZR2021QF110、ZR2021MF019;国家自然科学基金,项目编号52102412;山东省工业和信息化厅,项目编号202350100829;山东省交通运输厅,项目编号2023B74;山东高速股份有限公司,项目编号SDGS-2023-0281;山东省科学技术厅(新-代信息技术),项目编号2022TSGC2096,2022-11至2024-11。

摘　　要：准确高效的交通态势感知能够为智能交通系统提供重要的交通流信息。为解决高速公路上单一传感器交通状态判别精度低、可靠性差等问题,提出一种空地信息融合的智慧高速交通状态判别方法。根据高速公路基本路段的有效时空覆盖距离和无人机续航时间,提出一种全时空覆盖的空地一体交通数据采集模型方法,并采用YOLOv8物体检测和Deep-Sort多目标跟踪相结合的视频处理方式提取时空交通流参数。最后,基于模糊理论建立隶属度函数,提出智慧高速交通状态综合评价方法。结果表明:空地一体融合判定模型较无人机等单一传感器模型相比具有较高的准确率,交通流状态判别精度显著提升,平均绝对百分比误差降低了64.29%,均方根误差降低了57.9%,平均绝对误差降低了63.16%。通过实例分析可知本文方法能够有效反映高速公路区间路段的交通态势变化情况,有利于提升交通流实时监测、预警和调控优化水平。Accurate and efficient traffic state sensing can provide important traffic flow information for intelligent transportation systems.In order to solve the problems of low precision and poor reliability of single sensor traffic state discrimination on expressways,in the paper an intelligent expressway traffic state discrimination method with air-ground information fusion is proposed.According to the effective spatiotemporal coverage distance of the basic section of the expressway and the endurance time of the UAV,an air-ground integrated traffic data acquisition model method with full spatio-temporal coverage is proposed,and the video processing method combining YOLOv8 object detection and Deep-Sort multi-target tracking is used to extract the spatio-temporal traffic flow parameters.Finally,based on the fuzzy theory,the affiliation function is constructed,and the intelligent high-speed traffic state comprehensive evaluation method is proposed.The results show that:the air-ground integrated fusion determination model has higher accuracy compared with single sensor model such as UAV,and the traffic flow state discrimination accuracy is significantly improved,the average absolute percentage error is reduced by 64.29%,the root mean square error is reduced by 57.9%,and the average absolute error is reduced by 63.16%.Through the example analysis,it can be seen that the method proposed in the paper can effectively respond to the traffic state changes of the expressway interval section,which is conducive to improving the real-time monitoring,early warning and regulation and optimization level of traffic flow.

关 键 词：智能交通 空地信息融合 智慧高速 交通状态判别 交通信息采集 

分 类 号：U491.112[交通运输工程—交通运输规划与管理]