Automating the construction of large-scale virtual worlds

Databases for large-scale virtual worlds have several critical applications. Automating their construction can improve fidelity and save considerable time. In this article, we focus on: (1) automated cartographic feature extraction (derivation of man-made objects from aerial imagery); (2) triangulated irregular network (TIN) generation (automated techniques for terrain skin generation); and (3) road correction (automated integration of complex transportation networks). The product of these components forms a major part of the input to the final database compilation via computer image generator-specific tools such as the S1000 or the Close Combat Tactical Trainer (CCTT) database formatter