Index technology of spatial database is the key technology to improve the performance of spatial database, which have effected directly on the storage efficiency of spatial datas and the performance of spatial retrieval. So it becomes key contents of spatial database to seek a good spatial index mechanism. Based on the demand of time and space validity of spatial index, how to improved the query efficiency when dealing with massive high-dimensional spatial datas is researched. emphatically. It is very important in theory and have reality significance on spatial database.Firstly, aiming at the low query efficiency caused by massive spatial datas, spatial index of QR-tree based on adaptive K-means is proposed. Structure of QR-tree is improved through the introduction of supernodes. And the clustering initial values are determined automatically to improve the quality and the speed. Also novel calculation formulas of clustering center are constructed to deal with kinds of space targets. Meanwhile, the validity of adaptive K-means was certified to enhance the performance of the index by experiments.Secondly, aiming at the low query efficiency caused by high-dimensional spatial datas, approximate compression of high-dimensional datas is adopted to reduce disk I/O cost, thus improve the query efficiency. Meanwhile, QAAR-tree index is proposed to deal with high-dimensional spatial datas, also insertion algorithm, deletion algorithm, and query algorithm of QAAR-tree are presented.Finally, experimental results shows that the feasibility, time and space validity of QAAR-tree was improved. The storage cost of QAAR-tree was reduced, while the performance of insertion, deletion and query algorithm were enhanced greatly in dealing with massive high-dimensional spatial datas. |