Procedural City Modeling for AR Applications

In this paper we present a procedural city mod-eling approach which combines real-world street data and the multi-nuclei theory to generate believable cities. Our approach performs a partitioning of a city into urban zones based on Perlin noise and different parameters that are adjustable by the user. Our method is efficient enough to be used for augmented reality applications to be run on devices with limited processing capabilities such as smartphones. The approach can be used to build applications for professional urban planners and entertainment purposes. We illustrate the usefulness of our approach by applying it to data of three cities. Moreover, we provide implementation details and discuss challenges and limitations of the technique in the light of several application scenarios.