The Research On Qualitative Reasoning And Query Technology Of Directional Relationships In Three-Dimensional Space

With the development of spatial database technology, remote sensing technology and information technology, geographic information systems began to be used to solve the universal and globally problem. In order to meet the needs of a new generation of intelligent application system, the next generation of spatial information processing system must meet the qualitative description of the three-dimensional spatial information, commonsense reasoning processing, interaction with qualitative description language. Spatial relationships between objects include the topological relationships, the direction relationships and the distance relationships. In the three relationships, the direction of the relationship between qualitative reasoning and inquiry is the current hot topic issues of concern at home and abroad, multi-disciplinary field of spatial database. In two-dimensional space, it can’t express the accurate information when tailing after or querying flying objects and can’t give people the real sense of nature. In fact, the real objects in space include the three dimensional spatial information. This paper mainly focus on the relationship between the direction of the three-dimensional space modeling, qualitative reasoning, the direction of the relationship between queries and other related theories and key technologies, to carry out research related content.Firstly, with the reference object’s projections in xoy and xoz planes, each plane is divided into nine areas by the axes forming the minimum bounding rectangular of reference object. The directions are expressed by the intersecting areas of reference region and target region in two planes. The character “1” means intersectant while “0” means not. Therefore, directional relation matrix model based on double projections is given. The combinational reasoning method is researched using the calculation capability of matrix. After summing up the above combinational reasoning methods, the algorithm was given. And the algorithm was analized.Secondly, double projections and interior-division matrix model in 3D space is proposed by dividing the reference object within interior rectangle box. The expression of direction relations is combined with interior, boundary and exterior direction relations. 12 transforming rules are proposed by the constraint between directional and topological relationships. The model is evaluated by Roop K.Goyal’s five required properties. And the shortcomings of topological expression were sloved.Then, in view of bad performance of single index technology in 3D space, the advantages and disadvantages of octree and R-tree were analyzed. A hybrid index technology combining octree with R-tree was proposed which was named ORSI. The principle and structure description of ORSI were given. Atter that, R-tree segmentation clustering classification algorithm, ORSI insert algorithm, ORSI delete algorithm and ORSI retrieval algorithm were researched.Finally, the design process of direction relations reasoning prototype system in 3D space is given. The design principles of the system, the functions of the system and the algorithm flows of main module are introduced. At last, it was verified by test.