Segmentation of Urban Areas Using Vector-Based Model

Abstract

Urban areas are often segmented into sub-regions for indepth analysis and complexity reduction. This paper tries to use vector based model to segment urban areas into regions by adopting a graph theory approach. Vector-based model uses geometric primitives such as points, lines and polygons to denote spatial objects on the Cartesian coordinate system. Generally, we mainly store and analyze the vector data with Post GIS, in which a myriad of powerful functions are available. Here, we first find all intersections of road segments and turn vector data into a graph. Secondly, we simplify the graph by merging the redundant lines and removing the needless points. Lastly, dijkstra algorithm is applied to partition the areas into regions. In addition, we will present a case study of the Open Street Map data of Beijing to demonstrate the usability of the segmentation method.