The Nearest Neighbor Query Based on the DSB-Tree

The nearest neighbor query, to And the spatial entity whose distance to the given object, is a kind of basic operation on spatial dataset and used frequently in GIS. The DSB-tree adopts minimum bounding circle and maximum enclosed circle together to express the spatial entity approximately as a variation of SS-tree. Massive computing and comparisons on distance are necessary to the nearest neighbor query. The characteristic of circle can reduce the complexity of these operations, and the additive maximum enclosed circle enhances the precision of the rough query result, so that the cost of the nearest neighbor query is saved. The nearest neighbor query based on DSB-tree is implemented with the branch and bound algorithm in our paper. The metrics of the branch and bound algorithm are improved to suit to the structure of the DSB-tree. Finally, the extensive experiments are given and discussed to illustrate the performance of the DSB-tree in the nearest neighbor query.