通过扩展KD树索引结构,提出了一种新的多维空间数据索引结构——KDT树,给出了数据结构和算法描述,并通过与当前流行的空间数据索引结构——R树的对比,对其性能进行了测试与评估。实验表明,作为一种主存索引结构,KDT树在时间效率方面明显优于R树,并且此种优势随着索引记录数量的增多而越加明显。此外,KDT树亦能较好地解决常规KD树在索引占据一定空间范围的空间对象(如:线、面、体等)时存在的问题。
Authors firstly propose an innovative multi-dimensional index structure for spatial data,namely KD Ternary tree (KDT),with the expansion of traditional KD-tree.Then,the data structure and algorithms of KDT are stated.Finally,the performance of KDT is tested and evaluated by comparing with R-tree which is currently the most popular spatial data index structure.The results reveal that,as a main memory index strueture,KDT surpasses R tree over temporal efficiency,and this characteristic of KDT can be enhanced with the increasing data records indexed.In addition,the results also indicate that KDT has better capacity of indexing spatial data objects(i.e.lit uea,volume,etc.) coveting a certain space compared to KD tree.