中文摘要：

目前大多数空间聚类算法主要是针对关系数据，并且没有考虑空间拓扑关系的相似性，为此，对基于空间拓扑关系的空间聚类方法进行研究，提出两种基于点面包含关系的GML空间聚类算法SCGML_IR、SCGML—IR＊．两个算法将GML文档中点面空间对象的包含关系作为空间对象相似性度量准则，并用CLOPE算法对空间对象进行聚类．SCGML—IR＊算法在SCGML_IR的基础上，采用空间包含索引机制来提高空间包含关系的求解效率．实验结果表明，算法SCGML_IR和SCGML_IR＊能实现GML数据的空间聚类，并具有较高的效率．

英文摘要：

At present,most of the spatial clustering algorithms deal with the relational data,and do not consider the similarity of the spatial topological relations.Therefore,this paper researches on the method for spatial clustering based on spatial topological relations,and algorithms SCGML-IR and SCGML-IR for spatial clustering in GML data based on inclusion relations are proposed,which consider the inclusion relations between point and region spatial objects as the similarity measurement criteria,and CLOPE algorithm is used for clustering of spatial objects.Based on SCGML-IR,SCGML-IR is improved which uses the spatial inclusion index mechanism to enhance the efficiency of spatial inclusion relation computations.The experimental results show that SCGML-IR and SCGML-IR are effective and efficient.