| In 2019,the Ministry of Natural Resources proposed in the "General Plan for the Informatization Construction of the Ministry of Natural Resources" to establish a fullycovered three-dimensional natural resources database and promote the construction of three-dimensional reality databases.Three-dimensional reality is an important new infrastructure of the country,providing a unified spatial positioning framework and analysis basis for digital China.Three-dimensional reality models can be combined with basic geographic information data to serve urban planning,construction,management,and other aspects.As an important component of three-dimensional reality models,three-dimensional building models present fine and complex data features and larger data volumes with the rapid development of data collection technology,posing new challenges for the processing,storage,and data transmission of high-resolution three-dimensional building model data.Currently,the main solution in the academic community is to use the Level of Detail(LOD)model with multiple detail levels to express three-dimensional building models,and then use data organization and management methods that conform to the structure of threedimensional building models for data management to solve the problems of low query indexing and calculation efficiency of large-scale urban three-dimensional building model data.The key to generating LOD is to simplify the model to reduce model complexity and thus achieve the goal of reducing model data volume.However,existing three-dimensional building model simplification methods often ignore texture factors,leading to texture distortion and loss of detail features in the simplified model.How to efficiently store,manage,query,and calculate multi-resolution threedimensional building model data in large-scale urban scenes is also one of the challenges faced.To address the above issues,this paper conducted research on 3D building model simplification technology that takes texture features into account,with a focus on model simplification and model data organization and management.Specifically,this includes a 3D building model mesh simplification algorithm based on "local-vertex" texture features and a multi-resolution 3D building model data organization and management method based on a global position grid.In the process of simplifying 3D building models,texture elements were introduced,and a method that considers texture features was developed to retain important geometric and visual features of the model while minimizing its data volume and complexity.Based on this,focusing on the data organization and management needs of 3D building models,this paper further introduces the GeoSOT(Geographic Coordinate Subdividing Grid with OneDimension Integer Coding on 2n Tree)global position grid to organize and manage multi-resolution 3D building models,thereby achieving hierarchical management of multi-resolution 3D building model data and improving query retrieval and calculation efficiency of building models.The main research progress of this paper is as follows:(1)Propose a texture-aware simplification algorithm for 3D building models.To address the problem of existing 3D building model simplification algorithms not being able to preserve model details well and causing texture distortion,we propose a "localvertex" texture feature-based mesh simplification algorithm for 3D building models.This algorithm uses geometric and texture factors to optimize the quadratic error metric,which can maximize the preservation of the original appearance of the 3D building model and make the rich texture information of the 3D building model’s regions more complete,thus achieving better visualization effects.(2)Design a multi-resolution three-dimensional building model data organization and management method based on the global position grid.By utilizing the recursive and globally unique GeoSOT grid encoding,a three-layer structure of urban feature set,regional feature set,and independent feature set is established to achieve the organization and management of large-scale three-dimensional building models.Based on this method,a three-dimensional building model database is established using the relational database Postgre SQL for experimental validation.(3)Build a Web GIS visualization experimental system based on a global position grid.Using the method of organizing and managing multi-resolution 3D building model data based on the GeoSOT grid as the foundation,design a Web GIS visualization system for 3D building model query,retrieval,measurement,and analysis.This system provides support for the visualization of large-scale 3D urban scenes.The proposed algorithm for high-quality simplification of 3D architectural models and the multi-resolution organization and management method of 3D architectural models can meet the requirements of real-world 3D construction at various scales and complexities.It provides a solution for the collaborative updating of heterogeneous data sources and systematic management of urban 3D model data,supporting the management,storage,integration,and visualization of geospatial data. |
