Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (11): 7-10.

Previous Articles     Next Articles

Research of spatial clustering of discrete points in direction

CHEN Yingxian   

  1. College of Mines, Liaoning Technical University, Fuxin, Liaoning 123000, China
  • Online:2012-04-11 Published:2012-04-16



  1. 辽宁工程技术大学 矿业学院,辽宁 阜新 123000

Abstract: The cluster similarity of classical clustering methods(such as the classification method, hierarchical method, etc.) is determined by the distance between points. However, according to the distribution of spatial data, the change of data direction tends to produce the different classes clusters. The direction of space discrete points is in any direction. With the direction of triangular face expressing the direction of discrete points, it designs and implements the spatial clustering of discrete points in the direction algorithm in a certain direction change threshold. It designs a pyramid-shaped experimental data and successfully achieves the experimental data clustering by this algorithm. By the clustering algorithm, the actual measurement points of an open pit coal mine in Inner Mongolia are successfully clustering.

Key words: spatial clustering, discrete points, direction

摘要: 经典的聚类方法(如划分方法、层次方法等)的聚类相似度由点与点之间的距离决定。但在空间数据的分布中,数据间的方向变化往往产生不同的类簇。空间离散点的方向可以是空间的任意方向,通过将空间离散点的方向转换为其所在三角面的方向,设计并实现了空间离散点在一定方向变化阈值内的方向聚类。设计各点间等间距的金字塔形实验数据,使用空间离散点方向聚类算法成功实现了对实验数据进行聚类。以内蒙古某露天煤矿的实际测量点数据为例,使用该聚类算法成功实现对露天矿采场的测量点进行聚类。

关键词: 空间聚类, 离散点, 方向