The Data Organization Method Of 3D City Model Based On Object Importance

The three-dimensional urban model can express the city's three-dimensional content in a very intuitive and detailed manner.As an important infrastructure for a smart city,it plays an important supporting role in building a smart city.In recent years,slanting photogrammetry,laser radar,and image-based large-scale three-dimensional reconstruction of scenes have developed rapidly,and they have strongly supported rapid 3D city modeling in terms of technology,and the amount of 3D city models produced has become increasingly large.How to efficiently and effectively organize large-scale 3D city model data with reasonable and compact 3D tile representation is an urgent problem to be solved.The current "Internet +" construction is in full swing.One of the inevitable trends in the development of three-dimensional cities is to face the Internet.And with the advent and maturity of WebGL,rendering 3D images in real time is no longer remote.However,the huge output of 3DCM data and limited network transmission speed and system memory have become a prominent contradiction in the development of the Internet-oriented 3D city.This paper analyzes the progress of data organization methods for three-dimensional urban models in recent years.It uses a three-dimensional building model data organization method based on the importance of objects to hierarchically organize buildings in the scene and has a gradual transmission of scenes taking into account the importance of objects and tiles.The advantages of relatively balanced data volume include the importance and metrics of building objects including the importance of semantic features and visual features,and their relationship with the construction of the Scene level of detail(SLOD).Based on the level of detail of the object's importance,it focuses on analyzing how to classify buildings based on the importance of buildings in a three-dimensional city.After that,it analyzes the phenomenon that the distribution of data is uneven and the object is fragmented due to the unbalanced spatial distribution of buildings when the rules are divided.Based on the improved R-tree,that is,DB-R tree,a spatial indexing method that takes into account the balance of data volume is proposed.Summarizes two kinds of progressive scene transmission modes:alternative progressive transmission and additive progressive transmission.Based on the two scenarios progressive transmission mode,a three-dimensional tile set suitable for the mobile terminal and the desktop browser is constructed.Using three-dimensional model data of more than 100,000 buildings in a certain city as experimental data,three-dimensional tile data organization experiments were conducted.The three-dimensional city model data organization experiment first constructs a three-dimensional city scene level of detail based on object importance metrics,and then uses a DB-R tree spatial index structure that takes into account an equal amount of data to organize the scene detail level,thereby generating a three-dimensional tile and a corresponding index structure..The experiment analyzes the advantages of constructing the level of detail of the scene based on the importance of the object,and then uses the improved R-tree based on the level of scene detail and the space and data volume equalization to perform the three-dimensional city model tiles using the alternative progressive transmission mode and the additive progressive transmission mode.Data organization.The research results show that the data organization method of 3D city model based on the importance of objects divides the scene detail level according to the importance of objects,and can effectively highlight the characteristics of 3D scenes during browsing.The level of scene details solves the 3D city model to a certain extent,especially buildings.Problems with transmission rendering.The R-tree construction algorithm is improved by increasing the data volume adjustment factor.Taking into account the DB-R tree with balanced data volume,the data amount of the tile nodes is relatively balanced,which provides more reasonable and uniform data transmission and visualization.The data tiles block.However,the methods in this article also have shortcomings.In this paper,when building 3D city model building object data according to the standard of object importance,the main considerations are the visual features of a large number of ordinary buildings,without taking into account particularly complex building objects such as building information models,and constructing scenarios.At the level of detail,the level of detail at the same time taking into account particularly complex buildings is a problem that needs to be further explored in the future.