基于轨迹聚类的路网构造

Road Network Construction Based on Trajectory Clustering

随着定位设备的普及以及大规模移动轨迹数据的快速涌现,使得基于移动轨迹的数据挖掘和知识发现研究变得越来越热,其中一个有价值的研究方向是从GPS轨迹中抽取路网结构.然而如何有效地从轨迹数据中探测交叉口并进而构造路网却是一个难以解决的问题,针对此问题提出了一个基于轨迹点聚类的交叉口探测算法,以及在此基础上的路网快速构造方法.实验结果表明,算法能够有效地确定交叉口的位置及数量,且利用探测到的交叉口能够实现路网结构的快速构造.

With the popularization of the tracking devices and the emergence of the vast moving trajectory data,the research of the trajectory data mining and knowledge discovery is becoming more prevalent in recent years;one of the valuable researches is the road network extraction from GPS trajectory. However,How to effectively detect road intersections and construct road network is a very complicated problem. For this issue,this paper proposed an intersection detection algorithm based on trajectory points clustering and a rapid road network construction method. Experimental results demonstrate that our algorithms can find most of the road intersections effectively,and can construct road network conveniently.