Virtual City Map Generation using Area Subdivision

Abstract

In this paper we present a new approach to build typical street networks for urban and rural areas automatically. Most street networks have a characteristic appearance which is a mixture between a kind of a regular spreading of streets and various perturbations of the regular structure, e.g., caused by the terrain. Our City Map Generation (CMG) algorithm gets as input an height field including watercourses and restricted areas such as nature protection areas or mountain areas. Furthermore, the user can specify the urbanization by the number of city centers or by a density map. The CMG algorithm produces a street map for the given landscape based on rules, which the user can also modify to obtain an individual appearance. Additionally, the regular structure is varied by a set of rules incorporating randomness. The results can be used to generate models for city areas, e.g., in movies and commercials. Depending on the demands, our flexible approach can be adjusted to produce a variety of road maps automatically. We show some examples and demonstrate that our algorithm is easy to implement and to use.