A progressive approach to the generalization of an urban street network

Abstract

This paper proposes a progressive approach to selecting streets from an urban street network. This approach relies on the constraints in generalization of the street network and the ameliorated dynamic decision tree structure which represents the map data by a tree. The paper begins with a review of theory and methodology of constraint-based map generalization, of approaches to generalization of the road network and street network, and of some auxiliary structures which are useful to represent contextual topological relation. And then a conceptual framework for map generalization is presented. In the following, the methodology is described under this framework and the ameliorated dynamic decision tree structure is introduced in detail. An algorithm is developed and the proposed approach is validated by a case study. The paper concludes with some discussion on the benefits and shortcomings of the approach.