Construction of realistic 3D model of Shandong based on multi-source data fusion
To address the challenges posed by inconsistent foundations,poor compatibility of multi-source heterogeneous data,inefficient processing of complex data during high-frequency updates,weak correlation and low accuracy of ultra-large-scale real-life 3D data,and limited storage,transmission,sharing,and application of real-life 3D data,this paper proposes a methodology for on-demand fusion of multi-source heterogeneous data.The approach encompasses several pivotal technical aspects:(1)Construction of a real-scene 3D cube model,which involves building a comprehensive 3D cube model across the dimensions of time,space,and attributes;(2)Consistency processing of the multi-source data base,aiming to unify the coordinate format of multi-source data;(3)On-demand scheduling for GPA-based geographic data modeling,where tasks are scheduled based on geospatial processing arrays(GPA);(4)Establishing a one-code interconnected data spectrum with geographic entity coding as a link,thereby fusing business data,indexes,and rules into an interconnected data spectrum map;(5)Lightweight processing of multi-source heterogeneous real 3D data to enhance layer loading efficiency and visual effects;(6)Integrated rendering of heterogeneous 3D information from multiple sources,achieved through centralized management at the server side and deployment of lightweight plug-ins at the client side.The research algorithm presented in this paper effectively solves the problem of efficient fusion and rendering of multi-type heterogeneous,large-scale real-time 3D data.It improves the efficiency of data fusion and establishes a"full spatial,full convergence,full fusion"framework for real-time 3D Shandong,achieving high-quality real-time 3D coverage of the province's land and sea integration scenarios.Specifically:(1)Based on the method of dynamic fusion of multidimensional information of geographic scenes and entities,the province has realized efficient fusion of real-life three-dimensional data and governmental thematic data.Over 560 TB of inclined photography data and 447 types of thematic data from 16 cities and municipalities within the province have been seamlessly integrated.(2)Employing lightweight and integrated rendering technology,we have enhanced the storage compression ratio and access loading performance for super-large-scale real-view 3D data.The province's 560 TB of inclined photography data is compressed to 160 TB(a compression ratio of 70.74％without compromising data accuracy),reducing memory occupation by 40％,doubling loading performance,and achieving a high-performance service response of 300 milliseconds under 100 concurrency.This significantly boosts the storage and access performance of Shandong's real-life three-dimensional big data.These outcomes provide a robust three-dimensional spatial foundation for decision-making across various domains,substantially enhancing the efficiency and service levels of planning and management activities.
realistic 3Dcube modeldata fusiongeographic data modelinglightweight processinggeographic entity codeGPA
NETL
NSTL
万方数据
