The Research On Visible Query On Three-dimensional Spatial Objects

A great number of spatial data is being generated with the wide application of Geographic Information System(GIS),traffic control probes and virtual reality(VR).Efficient query processing technology of spatial data has become the current research hotspot in spatial database domain.As a spatial query closing to practical application relatively,visible query(VQ)in obstacle space has received extensive attention.However,the existing research on VQ is mainly concentrated on 2D space.Therefore,the existing VQ algorithms have defects in the visible query processing of 3D spatial objects.This paper takes the problems of VQ of 3D spatial object in 3D space as its research object,implementing the research from visibility test of 3D spatial object to continuous visible query algorithm.The main research contents are as follows:Firstly,with the indication of 3D spatial object by adopting the oriented bounding box(OBB),the direction of data object is different from the obstacle object,which leads to the lack of a unified projection plane.Thus the traditional dual projection method cannot be used to handle the visibility testing of 3D spatial object.A test method based on horizon-projected angle curvilinear is put forward to solve the visibility testing problem of 3D spatial object,which can build a unified coordinate system to test the occlusion relationship between data objects and obstacle objects.The result shows that this method can make a comparison accurately and quickly about the occlusion relation of 3D data and obstacle objects,thus the visual query result of 3D data can be acquired.Secondly,the definition of visible field is provided and visible query algorithm based on visible field is raised on the basis of comprehensively considering spatial property of the concept of “visible”.For the data object,of which the exact location is not available,its relative spatial relations with query points can be acquired by using qualitative spatial reasoning,judging initially whether this data object belongs to potential visible object;while for the data object having accurate location,visible field can be used to pruning the potential visible object set.Experimental result indicates that VQ algorithm based on visible field can effectively improve the efficiency of VQ,reduce I/O resource consumption and quicken the query responding time.Thirdly,the complete processing mechanism and algorithm for continuous visible nearest neighbor query of 3D spatial object are put forward.This paper analyzes the changes of occlusion relation of 3D spatial object in continuous visible query,defines the control point and point change and provides the principle of dividing query intervals;and separately offers the series of algorithms such as data object pruning algorithm,obstacle object pruning algorithm,visible subinterval algorithm and result updating algorithm in continuous visible query processing.On the basis of these algorithms,we propose a Continuous Visible Nearest Neighbor query on 3D(3D-CVNN)algorithm.Experimental result shows that the 3D-CVNN algorithm effectively solves the problem of continuous visible nearest neighbor query of 3D spatial object.Fourthly,aiming at the the problem of dual attributes that 3D spatial object is both data object and obstacle object,continuous visible range query based on visible field algorithm of 3D spatial object is proposed.In this algorithm,the visible 3D spatial object can be used as an obstacle to participate in the subsequent visual computation.Thus the obstacle properties of the 3D object is reserved by dynamically updating the obstacle object set.The effective obstacle interval is calculated according to the visual field of query point.The efficiency of continuous visible range query efficiency is improved with the effective reduction of obstacle object scale in effective obstacle interval.Last but not least,the paper puts forward a dynamic loading frame based on visible query,and makes it suitable for the network application of the large-scale 3D virtual reality scene.This paper gains the Visible Relationship Table(VRT)by taking segment as query interval to do continuous visible query,thus building the corresponding visibility relationship between road net and objects.Users can acquire visible object by linear query in VRT according to their location.This frame avoids the long time online visible query and meets the dynamic loading requirement based on visibility.For the visible computing and query processing which are time consuming will be completed at the phase of off line.