Research On The Object-Orientation-Based Direction Relation Query In Three-Dimensional Space

Direction relation, which is one of the important spatial relations in spatial database, is used widely in many domains, including GIS, multimedia system and image database. It is also frequently used as selection conditions in spatial queries. And in recent years, the efficiency of direction relation query has always been the hot topic of the academia. At present, the research on the processing of direction relation queries is mainly in two-dimensional space, while there is relatively little research in three-dimensional space. Based on these, the processing of object-orientation-based direction relations query in three-dimensional space is studied.First of all, the object-orientation-based direction relation model in two-dimensional space is extended, the corresponding direction relation model in three- dimensional space is proposed by using SDS. At the same time, single direction relation reasoning is explored and some laws of the direction relations reasoning are obtained. Furthermore, the theoretical proofs for the derived laws are presented.Secondly, in this paper, spatial objects are approximated by MBBs, and the base class of OpenShape in two-dimensional space is extended. An open shape model based on the 27 direction relations model in three-dimensional space is proposed. The basic idea is that the actual direction region is modeled as an open shape by constructing a new spatial ADT OpenShape, the computation related to the world boundary of spatial direction region is eliminated, and the processing of the direction predicates is converted into the processing of topological operations between open shapes and closed geometry objects. Meanwhile, the algorithms of topological operations between open shapes and closed geometry objects are presented and the theoretical proof for the correctness and completeness of the algorithms is performed. Finally, the processing of object-orientation-based direction relations is performed according to the proposed open shape model by using the R*-tree index structure, which is compared with the classical RQS. And the corresponding experiment is conducted. The experimental result shows that the new proposed open shape outperforms the classical RQS.