Research On The Visual Group K-Nearest Neighbor And Group Inverse K-Nearest Neighbor Query Of Multi-Source Objects In Three-Dimensional Space

With the application of intelligent monitoring equipment technology,the 3D augmented reality virtual scene technology has generated a large amount of systematic spatial data.Efficient technology for processing data queries has become research hotspot.Among them,the research of visual group k-nearest neighbor query and visual group reverse k-nearest neighbor query mainly used in biometric identification and online games.However,most of the existing researches used in two-dimensional space.In the threedimensional space,it is only for a single query point without considering multiple query points.But there are many problems of visual k-nearest neighbor query for multiple locations in real life.Such as multiple traffic monitoring equipment in different locations have query requirements for road condition information.Therefore,this paper provides the visual group k-nearest neighbor and visual group reverse k-nearest neighbor query algorithm of multi-source object in three-dimensional space,the main research content is as follows.First,a visual group k-nearest neighbor query algorithm based on gradient descent and visual rate is proposed.The query algorithm combines the gradient descent algorithm to find the center of the query point group.Then,define the visual rate and propose a general visual group query algorithm.Then,the k-limit algorithm is proposed to reduce the amount of calculation by pruning the data set.Finally,a method of using spatial distance to refine the result set is proposed.Secondly,a visual group reverse k-nearest neighbor query algorithm of gradient descent sphere and double Voronoi diagram is proposed.The algorithm uses the gradient descent algorithm to find the centroid of the query point group,and then finds the k-th target object which is completely visible in the target object set.Use the distance to pruning to get the first-level candidate set.Then,using the double Voronoi diagram method to screen the first-level candidate set,take(x,y)and(x,z)in the three coordinates of the data object to construct.Finally,the spatial distance between the target object and the query point group is used to filter the secondary target object candidate set to obtain the anti-k result set of the visible group.Finally,experiments are carried out to compare the above algorithms respectively.In terms of object size,the k values,and the spend time,experiments analysis and comparative results show that the proposed algorithms have good performance when executing paper queries.