R-tree shape and location multi-target node splitting method for n-dimension mass point clouds

Abstract

The invention provides an R-tree shape and location multi-target node splitting method for n-dimension mass point clouds, belongs to the field of reverse product engineering, and aims at solving the problem for constructing an R-tree index for the n-dimension mass point clouds. The R-tree shape and location multi-target node splitting method for the n-dimension mass point clouds is characterized by comprising the steps that point data in a point cloud file is read into a linear table storage structure; the point data in a linear table is inserted into the R-tree index one by one; if node overflow occurs, an overflowed node is split; the axis of which a location distribution function and a shape distribution function are both minimized is selected as the splitting axis; child nodes of the overflow node are subjected to ascending sorting according to coordinate components of bounding box center points of the child nodes in the direction of the splitting axis, and a candidate splitting solution set Q is generated by taking the number of minimum nodes permitted in non-root nodes as the limiting condition; a Pareto optimal solution set P of the Q is obtained, and the maximum Silhouette value in the P serves as the node splitting result. The method can construct the R-tree index for mass sampling data of the surface of a real object in reverse engineering, and the constructed R-tree index has the advantages of being high in construction efficiency, high in data query efficiency and the like.

Description

technical field [0001] The invention provides an R tree-shaped multi-target node splitting method for an n-dimensional massive point cloud, which belongs to the field of product reverse engineering. Background technique [0002] The R-tree index structure can well meet the operation requirements of dynamic insertion, deletion, and query of data such as spatial points, triangular patches, and sliced ​​surfaces in the process of surface reconstruction. It is widely used in CAD / CAM, geographic information systems, and medicine. Image analysis, ancient building restoration and other fields. [0003] For the current literature search on the R-tree node splitting algorithm, Sellis et al. published in the academic journal " computer Science Department "The R+ tree proposed in the academic paper "The R+-tree: A dynamic index for multi-dimensional objects" published on the Internet stores a specific data object in multiple leaf index nodes, avoiding the conflict between sibling no...

AI Technical Summary

Problems solved by technology

Since the various optimization objectives involved in the node splitting process are not independent of each other, there is a certain correlation between them, the solution results of the cascaded single-objective optimization method may lead to the optimality only satisfying a certain main objective, while perform poorly on other targets

Images