计算机应用 ›› 2013, Vol. 33 ›› Issue (06): 1604-1607.DOI: 10.3724/SP.J.1087.2013.01604

• 人工智能 • 上一篇    下一篇

局部多层网格划分下的轨迹数据概化方法

杨光,张磊,李帆   

  1. 中国矿业大学 计算机科学与技术学院,江苏 徐州 221116
  • 收稿日期:2012-12-25 修回日期:2013-02-18 出版日期:2013-06-01 发布日期:2013-06-05
  • 通讯作者: 杨光
  • 作者简介:杨光(1988-),女,山东济宁人,硕士研究生,主要研究方向:移动对象轨迹数据挖掘;张磊(1977-),男,江苏沛县人,副教授,博士,主要研究方向:移动对象轨迹数据挖掘;李帆(1986-),男,四川南充人,硕士,主要研究方向:移动对象轨迹数据挖掘。
  • 基金资助:

    江苏省博士后基金资助项目(0802023C);中国矿业大学青年科技基金资助项目(2008A040);中国矿业大学研究生实践与科研创新课题专项基金资助项目(GSF122111);教育部博士点基金资助项目(20110095110010)

Trajectory data generalization based on local multi-hierarchy grid

YANG Guang,ZHANG Lei,LI Fan   

  1. School of Computer Science and Technology, China University of Mining and Technology, Xuzhou Jiangsu 221116, China
  • Received:2012-12-25 Revised:2013-02-18 Online:2013-06-05 Published:2013-06-01
  • Contact: YANG Guang
  • Supported by:

    ;the youth foundation of China University of Mining and Technology;the China University of Mining and Technology Special Funds for the Graduate Practice and Research Innovation

摘要: 针对轨迹数据概化中空间划分的区域范围不能有效控制以及覆盖网格尺度难以合理选择的问题,提出局部多层网格划分方法,对样本密集的区域进行迭代划分。在此基础上提出一种轨迹数据概化方法,在局部多层网格划分的基础上,考虑时间约束合并轨迹连续往复通过的邻接区域,生成概化轨迹。真实数据的实验表明该算法得到的概化轨迹较同类算法保持了更多轨迹特性,更加适合后续数据挖掘,如聚类处理。

关键词: 轨迹概化, 局部多层网格, 时间约束, 轨迹数据

Abstract: For current trajectory data generalization methods, the scope of the generalized regions cannot be controlled effectively, and the parameters of the grids can hardly be selected logically. This paper proposed the method of Local Multi-hierarchy Grid (LMG), so that the region with dense trajectory points would be divided iteratively. And then a method for trajectory data generalization named TRAGenLMG was proposed, which was based on LMG, and time-constraint was used to merge some adjoining grids, and finally the generalized trajectory was got. The experiments with real open dataset show that the generalized trajectories generated by TRAGenLMG can well maintain the temporal feature of the trajectory data and can be efficiently applied into further data analysis.

Key words: trajectory generalization, Local Multi-hierarchy Grid (LMG), time-constraint, trajectory data

中图分类号: