The invention discloses an intersection scene safety risk quantification method based on street scene space multi-mode fusion, and the method comprises the steps: calculating a traffic accident occurrence intensity index of an intersection scene region based on historical traffic accident data and considering a distance attenuation effect, and taking the traffic accident occurrence intensity index as training label data; based on the streetscape panoramic image and the road network data, calculating visual features of the intersection scene, and calculating spatial structure features of the intersection scene; calculating semantic features of the intersection scene based on the interest points and the traffic flow data; combining the features to construct intersection scene safety risk feature vectors; and combining the strength index and the security risk feature vector, and constructing a security risk quantitative model of the intersection scene by using a random forest regression algorithm. According to the method, the traffic accident occurrence intensity index is designed, the visual, spatial structure and semantic dimension feature modeling of the intersection scene is comprehensively considered, the dependence on computing power and label data is low, the operability is high, and technical support can be provided for road safety risk early warning, safe city construction and the like.

    本发明公开一种街景空间多模态融合的路口场景安全风险量化方法,基于历史交通事故数据并考虑距离衰减效应,计算路口场景区域的交通事故发生强度指标并作为训练标签数据;基于街景全景影像和路网数据,计算路口场景的视觉特征,计算路口场景的空间结构特征;基于兴趣点和交通流量数据,计算路口场景的语义特征;组合上述特征构建路口场景安全风险特征向量;结合强度指标和安全风险特征向量,利用随机森林回归算法构建路口场景的安全风险量化模型。本发明通过设计交通事故发生强度指标,并综合考虑路口场景的视觉、空间结构和语义维度特征建模,对算力和标签数据依赖性低、操作性强,可为道路安全风险预警和安全城市建设等提供技术支持。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Streetscape space multi-mode fusion intersection scene safety risk quantification method


    Weitere Titelangaben:

    一种街景空间多模态融合的路口场景安全风险量化方法


    Beteiligte:
    GUAN FANGLI (Autor:in) / ZHANG JIANHUI (Autor:in)

    Erscheinungsdatum :

    2023-12-29


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES / G06V



    Urban Streetscape Design and Crash Severity

    Harvey, Chester / Aultman-Hall, Lisa | Transportation Research Record | 2019


    Unsupervised Deep Learning to Explore Streetscape Factors Associated with Urban Cyclist Safety

    Zhao, Haifeng / Wijnands, Jasper S. / Nice, Kerry A. et al. | Springer Verlag | 2019


    Automatic driving decision-making safety collision risk assessment method based on multi-scene fusion

    YANG SHICHUN / ZHANG MENGYUE / CAO YAOGUANG et al. | Europäisches Patentamt | 2024

    Freier Zugriff

    Sensor fusion based on intersection scene to determine vehicle collision potential

    JIANG SHENGBING / ZENG SHUQING / PERANANDAM PRAKASH MOHAN et al. | Europäisches Patentamt | 2022

    Freier Zugriff

    Use of Aerial LiDAR in Measuring Streetscape and Street Trees

    Golombek, Yaneev / Marshall, Wesley E. | Transportation Research Record | 2019