Traffic light is the most significant component of road transportation. Traffic lights control the traffic movement at intersections by allowing the movement of some lanes and forbidding that of the others to ensure safety and efficiency. It is important for intelligent vehicles to understand the meaning of traffic lights. Conventional research works on traffic light detection and tracking consider each light independently, so the occlusion of some lights may mislead the decision of the vehicle. Since all traffic lights at an intersection work together, it is more essential to understand the traffic movement status controlled by the traffic lights rather than the perception of a single light. Therefore, this paper proposes a spatial-temporal-rule-based framework to interpret the current traffic movement status and predict the next status. The spatial information, including position and class, is obtained by traffic light detection. The temporal information is perceived by the proposed topologic-feature-based tracking method. The rule information is recognized by combination of spatial and temporal information. A dual-view camera system consisting of a telephoto camera and a wide-angle camera is adopted to ensure long perception distance and wide coverage simultaneously. The successful tracking of traffic lights from 200 m to 10 m away provides enough data to recognize the phase-change rule, which leads to correct interpretation and prediction. A customized dataset was collected, and experiments were conducted to illustrate the capability of the proposed method.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Spatial-Temporal-Rule-based Traffic Movement Status Interpretation at Intersections *


    Beteiligte:
    Zhuang, Hanyang (Autor:in) / Gao, Liyang (Autor:in) / Qian, Yeqiang (Autor:in) / Wang, Chunxiang (Autor:in) / Yang, Ming (Autor:in)


    Erscheinungsdatum :

    2023-07-17


    Format / Umfang :

    4968353 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Real-Time Movement-Based Traffic Volume Prediction at Signalized Intersections

    Li, Wan / Ban, Xuegang “Jeff” / Zheng, Jianfeng et al. | ASCE | 2020


    Automated Approach for Computer Vision-Based Vehicle Movement Classification at Traffic Intersections

    Udita Jana / Jyoti Prakash Das Karmakar / Pranamesh Chakraborty et al. | DOAJ | 2023

    Freier Zugriff


    Eliminating street traffic intersections

    Malcher, F. | Engineering Index Backfile | 1930


    Traffic flow at intersections

    TSFASMAN ARKADIY O / WERNER JOHN S / CATALANO PASQUALE A et al. | Europäisches Patentamt | 2021

    Freier Zugriff