A Study Of The Method Of Internal And External Organization Scheduling Of Massive Point Cloud Data

Three dimensional laser scanning(LIDAR)as a new large scale data acquisition measurement technology developing fastly,is playing an increasingly active role in daily application.While due to it's character,the data volume aquired by it is very large and the aquiring speed is rapid.The point cloud data acquired by three-dimensional laser scanned has a huge amount of data because the point cloud contains a wealth of information.If organizations,management and scheduling cannot be made valid for massive point cloud data,it will be restricted to the development of three-dimensional laser scanning technology.In order to solve the problem that the massive data are complicated and the usability is not high,the original point cloud data is managed according to certain rules and selected according to certain discriminant methods.As a result,this method is the most effective point and the key to the efficiency of cloud data utilization.Therefore,the organization and management of mass cloud has become an urgent problem in the field of laser scanning.Many experts and scholars have also studied this problem.In this paper,based on the research results of predecessors in the field of cloud management,organization,scheduling and other fields,this paper proposes a spatial organization index and LOD structure model based on database technology.Combined with the visibility judgment,this technology carries out internal and external orderly scheduling of the mass cloud management organization method,in order to achieve the effective scheduling of cloud and efficient visualization.The main work of this paper involves the following aspects:1.Massive point cloud data organization and storage: Through the summary of previous research results,in the way of data storage we decided to use an open source third-generation database Postgresql to store data management;In the construction of point cloud organization index,we first organize the spatial data of the space octree by comparing and analyzing the different spatial indexing methods,and then decided to construct a heap tree index structure by using the OBB of the real circumference of the data model to establish the tree root.Finally,according to the actual requirements of this paper,the appropriate table structure is set up in the database,and the effective storage and management of point cloud data and index structure is realized.2.Massive point cloud data of LOD model: This paper proposes a method to construct LOD hierarchical model structure based on octree spatial index structure hierarchical sampling.Among them,through the comparative analysis,the final decision to use random sampling method to different levels of nodes within the point cloud data precision sampling,and thus the entire tree structure filled with point cloud.At the same time according to the different node level,the LOD model is level divided.Finally,it implements the database storage management.After the hierarchical reorganization of the point cloud data,the point cloud is divided into different depth levels,combined with certain discriminant scheduling methods,to reduce the number of one-time point cloud drawing,which can achieve the purpose of improving the speed of the point cloud display.3.Inside and outside storage scheduling of the massive point cloud : A method of combining two kinds of scheduling strategies based on viewpoint distance selection LOD level and visibility judgment based on occlusion culling is used to display nodes.First of all,through the previous method to determine the selected LOD level at this time,and then based on the visibility of the judge,selectively read part of the data into the node into memory.This reduces the amount of data read from the database into memory,improves the efficiency of data entry and exit,reduces the occupancy rate of data memory,and finally achieves the goal of improving the efficiency of internal and external storage.4.Point cloud management system establishment: The secondary development of a database module is carried out on a three-dimensional laser point cloud modeling platform,the prototype of the system is designed and the experiment is carried out.Finally,the feasibility of this method is proved.