摘要
确定重点监测区域是地灾应急响应中的一个重要环节。为了快速圈定监测范围,可以使用数据挖掘中的聚类方法。以粗略的监测区域作为工作区域,利用基于层次四叉树的面要素聚类方法缩小救灾范围。结合层次划分与四叉树剖分方法,将工作区域划分为包含面要素和相交面要素的单元格集合,在单元格集合中遍历搜索四方向邻近单元格,并将其聚合成多边形,从而实现面要素聚类。阐述了算法的可用性,通过实验分析了算法的复杂度,并对比分析了算法性能,进而利用面积差指标分析了层高对聚类多边形形状特征的影响。
A hierarchical quadtree-based clustering algorithm for polygons(HQBCP for short) was studied in this paper, to find and mark the important monitoring areas rapidly in the geologic hazard emergency response. Using the hierarchical idea and the quadtree partition principle, the HQBCP divided the work area into some grids, and formed a grids set at first. Each grid contained or intersected a polygon. And then, the HQBCP searched for each grid's adjacent neighbors according to the rook criterion, and merged them into a cluster polygon. Lastly, some experiments were carried out to show HQBCP's feasibility and analyze its complexity. Via the area difference coeffi cient, the effect of the hierarchical number on the clustering results was discussed, too.
出处
《地理空间信息》
2015年第5期1-3,7,共3页
Geospatial Information
基金
国家高技术研究发展计划资助项目(2012AA121301)
