Generalization Methods Of Buildings And Pois On Large-scale Urban Maps Under The Constraint Of Road Network

Automatic map generalization is always a hot and difficult research field of cartography.After decades of development,domestic and foreign scholars have designed a large number of algorithms and models for the generalization of point,line and surface features.Most of these algorithms deal with a single feature,and do not consider the automatic synthesis as a whole.A large number of generalization algorithms can only deal with the specific problems in a specific environment,and lack of data structures and models to support automatic generalization operations.With the development of cartographic generalization theory and cartography,automatic generalization considering geographical feature constraint has become the focus of cartographic generalization.In this paper,the generalization methods of large scale point of interest and building under the constraint of urban road network are studied.(l)Research on data structure to support automatic map generalization processIn the process of automatic map generalization,the rules and expert knowledge of cartography generalization must be transformed into the comPuter proeessing rules on geometric information,attribute information and topological information of geographical elements.In this paper,we propose a topological structure and a triangular network based structure with rich semantic information,whieh ean provide support for the automatic map generalization process.The application of data structure in automatic map generalization mainly includes the following aspects:To describe and express the geometric information and attribute information;To describe and express the spatial relationship between geographical elements;Support the basic geometric transformation,achieve map generalization operations.(2)Study on pretreatment method of urban road dataRoad network plays an important role in the formation of urban structure,as well as in the generalization process of other spatial data,such as imposing constraints on spatial distribution,dividing a greater range of data into smaller reigional data sets,working as space obstacle factors and so on.It is necessary to preprocess the road data before it is being used as constraint data base for the generalizaiton of the POI data and the building data.The characteristics of road data are analyzed from three aspects:geometric features,topological features and semantic features.The constraint effect of road network on the generalization process of other geographical elements in urban area is summarized.This paper studies the road data structure,and puts forward 3 structured approaches,which are the triangular network based structural approach,the topological structural approach and the stroke based structural approach.The related road data preprocessing algorithms are proposed:a road centerline extraction and optimization algorithm based on the triangular network based structure,to achieve the road dimension reduction;an intelligent road evaluation model using stroke based structure approach,to realize automatic road selection;a road simplification algorithm based on intelligent recognition and constraint of spatial features using node-arc topological structure.(3)Research on the generalization method of urban POI under road network constraintsAfter summarizing the similarities and differences between the urban POIs and the general point elements,it 1s pointed out that the generalization algorithms for general point elements are not suitable for the urban POIs;From the view of map cognition,this paper analyzes the characteristics of POI spatial information expression in three levels:the global clustering level,the local structure level and the single element level;Based on the cluster result,selection model of urban POIs under road network constraints is proposed based on the distribution patterns:spacing selection method for linear distributed POIs;weighted V diagram based selection method for cluster distributed POIs;and significance based selection method for discrete distributed POIs;Besides,the conflicts between the road and the POI are analyzed,and the possible solutions are offered.The experiment and analysis are carried out with real data.(4)Study on the generalization method of urban building under road network constraintsAnalyze the data characteristics of the city building and the needs of generalization.The data dividing and spatial obstacle influence of road network extered in the process of building generalization is researched,and the merge method of urban building under road network constraints is proposed.Firstly,the larger block of building is divided into smaller blocks using road network;and then,the the triangular data structure based on method of building merge in each block is presented:Taking the road as the spatial barrier,6 kinds of measurement parameters are used to eliminate and repair each triangle in the constrained Delaunay triangulation network,to detect the spatial neighborhood relationship;The outer profile of the reserved triangle is automatically identificated and it is used as a bridge area between the spatial adjacent buildings;In the end,the merge operation of the bridge areas and the building polygons is performed with the right angle characteristics remained.The effectiveness of the algorithm is tested and analyzed with the actual data as an example.Based on the spatial relationship and spatial conflicts of eartographic elements,the intelligent representation and recognition method of spatial relationship between point features,line features and polygon features is proposed.The extended topological structure and triangular network structure of spatial features is proposed to support POIs and building generalization under road network constraints in large scale urban area.The research content of this paper enriches the research results of the automatic generalization of large scale urban geographical elements,and to some extent,promotes the practical application of automatic map generalization.