作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (9): 270-272. doi: 10.3969/j.issn.1000-3428.2011.09.094

• 开发研究与设计技术 • 上一篇    下一篇

改进的基于密度的航迹聚类算法

赵恩来1,郝文宁1,赵 飞2,陈 刚1,邵校莎莎1   

  1. (1. 解放军理工大学工程兵工程学院,南京 210007;2. 中国海上卫星测控部,江苏 无锡 214400)
  • 出版日期:2011-05-05 发布日期:2011-05-12
  • 作者简介:赵恩来(1985-),男,硕士研究生,主研方向:数据工程,聚类分析,时间序列相似性分析;郝文宁,副教授、博士;赵 飞,助理工程师;陈 刚,副教授、硕士;邵校莎莎,硕士研究生

Improved Track Clustering Algorithm Based on Density

ZHAO En-lai 1, HAO Wen-ning  1, ZHAO Fei  2, CHEN Gang  1, SHAO Xiao-sha-sha  1   

  1. (1. Engineering Institute of Corps of Engineers, PLA University of Science & Technology, Nanjing 210007, China; 2. China Spacecraft Maritime Tracking and Control Department, Wuxi 214400, China)
  • Online:2011-05-05 Published:2011-05-12

摘要: 为解决雷达站观测数据的分类问题,提出一种改进的基于密度的航迹聚类算法。采用加权Manhattan距离与惩罚系数相结合的距离度量,根据目标运动的特征自定义点的邻域,利用时间裁剪提高算法运行效率。实验结果表明,该算法能高效准确地对数据进行聚类,形成运动目标的航迹。

关键词: 聚类, 航迹, 密度, 邻域, 时间序列

Abstract: In order to classify the data of radar, this paper proposes an improved track clustering algorithm based on density. Considering concrete application, the algorithm adopts the distance measure by Manhattan distance and penalty coefficient, newly defined the definition of neighborhood using the character of moving object, and improves the efficiency by time clipping. Experimental result shows that the improved algorithm can cluster the observation data accurately and form the tracks efficiently.

Key words: clustering, track, density, neighborhood, time series

中图分类号: