From urban terrain models to visible cities

We face the possibility and, in some cases, the results of acquiring accurate digital representations of our cities. The first challenge is to take data from multiple sources, which are often accurate but incomplete, and weave them together into comprehensive models. Because of the size and extent of the data that we can now obtain, this modeling task is daunting and must be accomplished in a semiautomated manner. Once we have comprehensive models, and especially if we can build them rapidly and extend them at will, the next question is what to do with them. Thus, the second challenge is making the models visible. In particular, they must be made interactively visible so users can explore, inspect, and analyze them. We discuss the nature of the acquired data and how we're beginning to meet these challenges and produce visually navigable models. We've developed 3D City, a semiautomated system that supports human intervention at key points to meet the challenge of constructing complete and extended urban models from several data sources. We've already built virtual environments (VEs) for urban planning and emergency response using 3D City.