| With the development of mobile communication technology,3D graphics rendering technology and geographic information system(GIS),mobile 3D technology has gradually become a hot topic in GIS research.Due to the limited situation of the mobile device,many mobile 3D applications have adopted the model of client/server.Most research of 3D mobile GIS server are based on the major platform in the market for secondary development,whose portability and flexibility is not good.Therefore,based on the B/S architecture,relying on the characteristics of Web server platform independence,we can make full use of rich two-dimensional GIS data resources,study 3D building model data modeling and multi-resolution generation technology,and establish a three-dimensional building model Server system for mobile client to achieve three-dimensional building model server-side data storage and scheduling,reduce the client’s storage and computing pressure.The main contents of this paper are as follows:The three-dimensional building data model and multi-resolution generation algorithm are studied.Studying the existing three-dimensional data model theory,the object-oriented 3D building data model is presented.Using the existing two-dimensional vector data,combined with SketcuUp and ArcGIS,the 3D building data model is constructed.Analyzing the advantages and disadvantages of various multi-resolution generation algorithms,Based on the obj file storage triangular mesh model data,the half-side folding algorithm is used to simplify the geometric data of the 3D building model,and the vertex index number is not updated so that the multi-resolution 3D building model share a set of vertex index data,which make the multi-resolution 3D Building models easier to store and schedule.The three-dimensional building model data database system are designed.Considering the’Reduce IOE’ to the demand of the times and wide range of city,the method of dividing the city areas by grid is put forward.According to the characteristics of the three-dimensional building model file of the city,the storage structure of the three-dimensional building model data database is designed by using the data of different range of data,and the grid storage of 3D building model data is realized.Data scheduling technology on the server-side is researched.The spatial indexing methods at home and abroad are studied.Considering the spatial stereoscopic characteristics of building model,the KD tree based on the center point of the model plane MBR is constructed.Through the experimental analysis,The indexing strategy that combines the grid index with the KD tree index is presented.Based on the three-dimensional building model database storage structure,a cache strategy combining server-side memory cache and disk cache is proposed.The data scheduling algorithm for database server system is designed,which realizes the reuse of retrieved data of 3D building model and the efficient scheduling of the target model data.The three-dimensional building model Web server architecture is designed.Combined with the software system architecture,the 3D architecture model Web server architecture is designed,which is based on the Java Web architecture and MVC model.Considering that the common Java EE framework maintains the memory overhead of database connections,the lightweight database access framework for this article is developed to simplify the database operation.Due to the B/S architecture is only for the browser client,the Web server communication module is designed on the research of http communication technology,which completes the response to the mobile client data request.By analyzing the time-consuming experimental data,it is verified that the scheduling strategy and caching technology under the Web server architecture can reduce the response time by half.The research on key technology of data storage and management of meshing 3D building model on the server is realized in the field of mobile 3D GIS research.Based on B/S architecture,the server architecture is designed and developed with efficient data storage and scheduling efficiency,and the function is working well. |
