地理科学 ›› 2019, Vol. 39 ›› Issue (8): 1238-1245.doi: 10.13249/j.cnki.sgs.2019.08.005

• • 上一篇    下一篇

基于网络时空核密度的交通事故多发点鉴别方法

王颖志1,王立君2   

  1. 1. 浙江警察学院交通管理工程系, 浙江 杭州 310053
    2. 浙江大学地理信息科学研究所, 浙江 杭州310028
  • 收稿日期:2018-08-22 修回日期:2018-12-12 出版日期:2019-08-10 发布日期:2019-10-10
  • 作者简介:王颖志(1973-),男,浙江杭州人,硕士,讲师,主要从事道路交通管理研究。E-mail: wangyingzhi@zjjcxy.cn
  • 基金资助:
    国家自然科学基金项目资助(41471313)

An Identification Method of Traffic Accident Black Point Based on Street-Network Spatial-Temporal Kernel Density Estimation

Wang Yingzhi1,Wang Lijun2   

  1. 1. Department ofTraffic Management Engineering, Zhejiang Police College, Hangzhou 310053, Zhejiang, China
    2. Geographic Information Science Institute, Zhejiang University, Hangzhou 310028, Zhejiang, China
  • Received:2018-08-22 Revised:2018-12-12 Online:2019-08-10 Published:2019-10-10
  • Supported by:
    National Natural Science Foundation of China(41471313)

摘要:

交通事故多发点是道路交通安全管理的重要治理对象,如何利用空间统计方法对其进行高效鉴别是研究热点。以华东某地为研究区域,以2013~2015年该研究区域的道路交通事故数据为研究对象,以时空道路网络为视角,通过路网匹配和路网裁剪形成事故时空子路段,提出一种基于交通事故场景的网络时空核密度估计值作为鉴别指标,用累计频率法和零膨胀的负二项回归模型确定鉴别阈值的事故多发点鉴别方法。

关键词: 道路交通事故, 事故多发点鉴别, 网络时空核密度估计

Abstract:

With the rapid progress of society and increase of motorization level, traffic accident is also increasing. It has become the third leading cause of accidental death in China, seriously threatening human life and social development. How to identify the black point of traffic accident is one of the important issues of traffic safety management. This article took the view of spatio-temporal street-network, matched traffic accidents to the spatial-temporal subsection of street-network for calculating the network spatio-temporal kernel density estimate. A method of identifying accident black points, which using street-network spatio-temporal kernel density estimated value as identification index and cumulative frequency analysis to determine the identification threshold, was proposed. We took the traffic accident records of a county in East China from 2013 to 2015 as the research data, to identify the traffic accident black points. It can be seen that traffic accidents have significant clustering characteristics in spatial road network and temporal features. In addition, most of the spatio-temporal sub-sections of accident black appear at roads intersections, which proves that there is indeed geometric heterogeneity between intersections and ordinary sections. The accident black points is a closed area formed by the continuous sub-sections of spatio-temporal with high accident rate. Moreover, the accident incidence rate in the center of the area must be higher than that in the surrounding area. In this study, the accident black points are displayed in 3d scene with 2d plane and temporal dimension. Further, the road characteristics, surrounding location and temporal characteristics of 8 high-risk accident black points are analyzed. The experiments demonstrated that this method can more accurately identify the spatio-temporal boundary of accident black points.

Key words: traffic accident, accident black point, street-network spatial-temporal kernel density estimation

中图分类号: 

  • D917