A Two-level Tiled Vector Data Model With Considering Cartographic Representation And Feature Integrity

With the development of Internet and communication technology,electronic maps have become an indispensable part of people’s daily life.However,with the upgrading demands of users,the electronic maps represented by raster maps have emerged some limitations,including interactivity,style flexibility,and resolution of pictures and so on.Combining vector data and tile technology to construct vector tile map can make up the deficiencies of raster tile map.Nowadays,the research on vector tile map has become one of the most important directions and research hotspots in the development of network electronic maps.However,the existing vector tile maps also have some shortcomings.For example,tile segmentation is easy to destroy the original structure of geographical features,especially in the stage of cartographic representation,tile edges may appear the problems,including symbol graphics conflicts and losses,which would affect the visualization effects of vector tile maps.In addition,the interactive operations of the vector tile feature and the visualization efficiency on the client side directly influence the experience of users and the applications of the vector tile maps,while the related research has been little.Therefore,based on the basic theory and methods of cartography,this paper conducts in-depth research and analysis on the main problems existing in vector tile map.Specifically,this paper establish a two-level tiled vector feature model(TLTVFM)with considerations of cartographic representation and feature integrity.Based on TLTVFM,this paper proposes the feature reconstruction algorithms based on vector tile and the parallel visualization framework,which can improve the calculation efficiency of the vector tile map and promote the development of the vector tile map,including the following contents:(1)A two-level tiled vector feature model is proposed in the view of data model.Starting from the basic theory and methods of cartography,this paper abstracts the related process and object of geographical features mapping expression into algebraic structures.Then,according to the relationships among the geographical factors,cartographic representation and map features,this paper defines the “addition operators” of geographical features and map features and uses algebraic language to describe the slicing process of vector tiles,which can transform the problem of splicing map features into a problem of mathematical equation.Besides,examples are given to explain and illustrate the visualization problem of traditional tile feature model.Finally,according to the proposed the definition of “addition operation”,this paper puts forward a two-level structure model for three types of geographical features and proves that the model can give consideration to both cartographic representation integrity and feature integrity by means of model derivation.The tile feature model is constructed to ensure the correctness of visualization results of vector tile map.(2)The clipping algorithm and reconstruction algorithm of tile features based on TLTVFM are implemented.The structure of TLTVFM is more complex than the traditional vector tile feature model,which makes the clipping algorithm and reconstruction algorithm of TLTVFM face new challenges.This paper analyzes the structures of TLTVFM for different types of geographical features,and makes full use of the characteristic that the clipping frame of vector tile features is rectangle,and proposes an efficient vector tile clipping algorithm that fits TLTVFM.In addition,in order to ensure that users can better achieve interactive operation based on vector tile data,it is necessary to quickly reconstruct the complete geographical features.Therefore,by designing reasonable tile file structure and index file structure,this paper realizes rapid reconstruction for geographical features based on TLTVFM.(3)A method for estimating the visualization computation intensity of geographical features is designed.The efficiency of map visualization directly affects the users’ experience of vector tile map application.Therefore,improving the visualization efficiency of vector tile map is very important for its applications.The proposed TLTVFM can ensure the correctness of the cartographic representation after the visualization of tile subdivision,and is the basis and guarantee of parallel computing of tile visualization.Parallel visualization of vector tile map is an embarrassing parallelism operation.The key to improve the visualization efficiency is to accurately estimate the workload of each computing unit.In order to accurately estimate the visualization computing time of each tile,it is necessary to estimate the visualization computing time of geographical features in advance.In this paper,the visualization computing intensity is used to estimate the visualization computing time,and the visualization of geographical features is divided into three steps: feature retrieving,feature symbolization and feature rendering.This paper analyzes the influencing factors of the intensity of various types of geographical features in different visualization processes,constructs the computing intensity functions of different visualization processes of geographical features,and fitting the regression coefficients of the visualization computing intensity functions of each type of geographical features by linear regression method.(4)The parallel visualization load balancing strategy based on tile computing intensity is proposed.Since parallel visualization of vector tile map is an embarrassing parallelism,optimizing its load balancing strategy can effectively improve the visualization efficiency.Considering that the spatial distribution of geographical features is often uneven,this paper proposes a load balancing strategy based on the computing intensity of tiles.The load balancing strategy relies on the visualization computing intensity of tiles,while has no relationships with the spatial distribution of tiles.In this paper,the intensity of each tile is accurately estimated,and then the intensity is evenly distributed to the comuting units according to the load balancing algorithm.Therefore,the total intensity of each tile is roughly equal,which can reduce the waste of resources and improving the overall visualization efficiency.(5)The corresponding evaluation principles are put forward to evaluate the effectiveness and rationality of the proposed model and methods in this paper.The effectiveness of the proposed TLTVFM model in tile visualization,feature reconstruction,tile slice efficiency and data redundancy are evaluated by experimental method.Meanwhile,the applicability and feasibility of the proposed geographical visualizaiton intensity computing function and the load balancing strategy based on tile computing intensity are also verified by experiments.