Research On Large-scale Geographic Vector Data Organization Technology For Visual Analysis

In the era of big data,with the development of spatial data collection means and source-producing technologies such as basic mapping means,satellite remote sensing observation and location service systems,the scale of spatial data has shown an explosive growth.As an important part of spatial data,geographic vector data represents the location and shape of geographic entities in the real world through spatial geometric entities such as points,lines and polygons and their collections,which can intuitively and accurately reflect and express geographic entities and their spatial distribution characteristics.Spatial analysis of large-scale geographic vector data is an important means of mining and capturing the value of vector data,considering that in most cases users are more interested in quickly visualizing and viewing the analysis results on screen than in obtaining substantive data analysis results.To meet the needs of this scenario,this paper focuses on efficient data organization techniques to support real-time visual analysis of large-scale geographic vector data,aims at the problem that the current geographic vector data organization and analysis methods can’t support real-time visual analysis of large-scale geographic vector data,explores the organization and analysis methods that directly uses screen display pixels as the organizing and computing unit,proposes a display-oriented model for organizing and analyzing large-scale geographic vector data,and designs LGTQ-tree,a lightweight Geohash-based tile quadtree that is specifically adapted to the computational characteristics of screen display.The pixel generation algorithm is designed to quickly determine the visualization analysis results,which is successfully applied to several scenarios such as visualization and spatial buffers analysis of large-scale geographic vector data,enabling significant performance improvements in data organization and analysis compared to existing methods,and realizing fast response in data organization and analysis processes for visualization analysis tasks.The research content and main contributions of this paper include the following:(1)Di OA: display-demand-oriented organization and analysis model for largescale geographic vector data.Aiming at the problem that the current geographic vector data organization and analysis methods are difficult to support and realize the efficient organization and real-time visual analysis of large-scale geographic vector data,this paper proposes Di OA for the organization and analysis of large-scale geographic vector data.In this model,pixels are taken as the computational objects for organization and analysis,and a specific LGTQ-tree index structure,pixel generation algorithms based on the LGTQ-tree,and a display-demand-oriented integrated-organization-analysis framework are designed respectively according to the computational characteristics of displaydemand-oriented.Di OA features fast index construction,small index generation size,small analysis computation and easy expansion parallelism,which significantly improves the response efficiency of large-scale geographic vector data in the organization and analysis process and has good application prospects in the field of spatial big data analysis.(2)The LGTQ-tree indexing technique for large-scale geographic vector data.Aiming at the problem that the computation amount and space occupation of existing geographic vector data organization methods increase rapidly with the increase of vector element scale,it is difficult to support real-time visual analysis of large-scale geographic vector data,this paper analyses the computational characteristics of display-oriented demands and designs the LGTQ-tree index structure by taking the spatial extent of pixels representing geographic vector elements as the organizing unit.Experiments show that the method is significantly better than existing spatial indexing techniques in terms of index construction and index generation size performance,and can be used as data organization support for real-time visual analysis of large-scale geographic vector data.(3)Real-time interactive visualization of large-scale geographic vector data.In view of the existing mainstream visualization methods are difficult to achieve real-time visualization of large-scale geographic vector data,the paper applies Di OA to the geographic vector elements visualization,converts the visualization problem into pixel calculation problem based on LGTQ-tree index under display requirements,and designs a visualization algorithm based on LGTQ-tree index.Experiments show that the index structure makes the algorithm is practical and performance were significantly better than the existing visualization methods,at the same time,the algorithm has good visual effect which can support billion-scale geographical vector elements of real-time interactive visualization and has a good application prospect on the exploratory analysis of largescale vector dataset.(4)Real-time interactive spatial buffer analysis of large-scale geographic vector data.In view of the existing mainstream spatial buffer analysis methods are difficult to achieve real-time spatial buffer analysis of large-scale geographic vector data,this paper applies Di OA to the geographic vector element spatial buffer analysis scenario,converts the buffer analysis problem into a display-demand oriented per-pixel calculation problem based on LGTQ-tree index,and designs a buffer analysis algorithm based on LGTQ-tree index.Experiments show that the index structure makes the algorithm is practical and performance were significantly better than the existing buffer analysis methods,and the algorithm has good visualization and real-time interactive effects,which can support realtime interactive buffer analysis of billion-scale geographic vector elements and has good application prospects in urban computing such as school siting and road construction.