Image-based region modelling focusing on template generation of road networks

Abstract

Urban regions are complex multi-functional dynamic systems which play a significant role in the development of modern societies. They are mirrors of historical, cultural, economic and social aspects of human settlements. Urban synthesis, which is to collect, compile and analyze city data, construct the model and visualize the result, thus plays a very important role in the creation of the Digital Earth. Modeling and visualizing the dynamic growth of cities has become an important area of research in many fields such as urban planning, visibility analysis, environmental control, urban engineering, communication design, shopping and market facilities, cultural centers, museum display and entertainment industries. One significant purpose of urban synthesis is to design and develop the real city of tomorrow in the optimum way and manner by using virtual cities to plan real cities. Virtual cities may be regarded as working models for the city of the future, for further development of the city and building-up volume. It may also be a totally new model for some undeveloped areas. In this thesis, we develop an image-based 3D urban model with the main focus on transportation system generation which is a critical element in ensuring the growth and vitality within an urban community. A template-based road model is presented on analyzing several factors that will affect the transportation system. Given input maps such as land and water boundaries, population distribution and elevation, this system is able to generate road maps of various patterns. The model can also adjust itself intelligently to avoid illegal areas, follow along the direction of elevation and connect the dead-end roads to form a valid and efficient network. In our model, it is easy to incorporate future extensions of new road patterns by adding new road templates into a template library. The urban terrain is also constructed based on information extracted from maps. The roads and terrains are blended together to form an integrated surface model.