计算机工程 ›› 2019, Vol. 45 ›› Issue (5): 291-297.doi: 10.19678/j.issn.1000-3428.0052124

• 开发研究与工程应用 • 上一篇    下一篇

基于稀疏采样GPS轨迹数据的路口识别方法

陈亚玲,范太华   

  1. 西南科技大学 计算机科学与技术学院,四川 绵阳 621010
  • 收稿日期:2018-07-16 出版日期:2019-05-15 发布日期:2019-05-15
  • 作者简介:陈亚玲(1993—),女,硕士研究生,主研方向为数据挖掘、智慧交通;范太华,副教授。
  • 基金项目:

    教育部人文社科基金“基于社交网络的涉核舆情挖掘关键技术研究”(17YJCZH260)。

Intersection identification method based on sparse sampling GPS trajectory data

CHEN Yaling,FAN Taihua   

  1. School of Computer Science and Technology,Southwest University of Science and Technology,Mianyang,Sichuan 621010,China
  • Received:2018-07-16 Online:2019-05-15 Published:2019-05-15

摘要:

目前的路口识别算法多数在密集采样的全球定位系统(GPS)轨迹数据基础上,采用车辆行驶方向转变作为依据进行路口识别。但稀疏采样的GPS轨迹数据由于采样时间间隔较长,无法准确模拟出车辆行驶方向。为此,针对交叉路口容易发生停车现象的特点,提出一种GPS轨迹数据的路口识别方法。从稀疏采样的GPS轨迹数据中提取出行驶轨迹的停车点及其后续点作为研究对象,依据改进GDBSCAN算法对提取出的停车点进行聚类,判断停车事件发生的热区。运用提取出的后续点对热区进行连通性计算,并根据连通性确定是否存在路口。实验结果表明,该路口识别方法具有较好的识别能力,且与DBSCAN算法相比,聚类速度明显提高。

关键词: 稀疏采样, 出租车, 全球定位系统, 轨迹数据, 停车点, 路口识别

Abstract:

Most of the current intersection identification algorithms are based on densely sampled Global Positioning System(GPS) trajectory data,using the direction of vehicle travel as the basis for intersection identification.However,the sparsely sampled GPS trajectory data cannot accurately simulate the direction of travel of the vehicle due to the increase of sampling time.Therefore,this paper proposes an intersection identification method for sparsely sampled GPS trajectory data.It extracts the parking point and its subsequent points of the driving track are extracted from the sparsely sampled GPS trajectory data as the research object,and clusters the extracted parking points are clustered according to the improved GDBSCAN algorithm to determine the hot zone where the parking event occurs.This paper uses the extracted subseguent points are to calculate the connectivity of the hot zone,and determined the connectivity according to the connectivity.Experimental results show that the intersection identification method has better recognition ability,and the clustering speed is significantly improved compared with the DBSCAN algorithm.

Key words: sparse sampling, taxi, Global Positioning System(GPS), trajectory data, parking spot, intersection identification

中图分类号: