摘要: 针对现有空间对象多尺度索引结构聚簇性不高的问题,在R树索引的基础上提出一种基于聚类的空间数据多比例尺索引结构。利用树的层次结构反映空间数据的多比例尺特性,用k-means算法对相同等级的空间对象进行聚类分组,减少空间区域覆盖和重叠。实验结果表明,该方法与基于四叉树的多比例尺索引相比,能有效提高空间数据多比例尺显示的性能。
关键词:
多比例尺,
R树,
聚类算法,
空间索引,
地理信息系统
Abstract: Due to the problem about clustering of multi-scale index structure for the existing space object is not high, this paper represents a cluster-based multi-scale spatial data indexing structure based on R-tree index, which uses hierarchical tree to reflect the multi-scale features of spatial data, and uses k-means algorithm to cluster groups on the same level of spatial objects, reducing coverage and overlapping regions of space. Experiments result shows that compared with other ways, the algorithm has a distinct superiority in the speed of multi-scale display of spatial data.
Key words:
multi-scale,
R-tree,
clustering algorithm,
spatial index,
Graphic Information System(GIS)
中图分类号:
邹志文, 费洪哲, 李根. 基于聚类的空间数据多比例尺索引树[J]. 计算机工程, 2011, 37(14): 56-58.
JU Zhi-Wen, BI Hong-Zhe, LI Gen. Multi-scale Index Tree of Spatial Data Based on Clustering[J]. Computer Engineering, 2011, 37(14): 56-58.