Production of 3D terrain databases for visualization requires resolving tradeoffs between a complex set of constraints: time, resources, accuracy, resolution, hardware, limitations, etc. The relationship between these parameters and the quality of the final database for most applications is not well understood. There are few tools to measure quality and there is little understanding of the impacts of data uncertainty on end applications, so there is little incentive to expend additional resources to achieve an unappreciated and unmeasurable return. Our main objectives are to formulate a comprehensive approach for tracking fundamental quality attributes through the database production andscene generation processes, and to develop methods to evaluate database quality for various applications, especially for scene generation.The research will involve: defining fundamental quality attributes for terrain database generation and visualization; documenting the processes of database production and scene generation, with emphasis on establishing and tracking quality attributes and identifying processes that modify quality attributes; developing quality evaluation methods for scene generation and other simulation applications, establishing critical levels of quality for these applications; and developing a methodology to portray the resulting scene quality to an end-user. Our proposal is a collaboration of three companies, each with unique technical specialties to address key aspects of the problem.
Benefits: The potential benefits of this work are wide spread. For DOD applications,the main advantage will be for mission rehearsal, training and battlefield visualization, where understanding the differences between virtual and real worlds is of great importance. Commercial applications such as land-use and urban planning will also benefit from this approach. Coryphaeus Software will provide the vehicle for commercialization of the proposed research, with extensions to their existing COTS software for visual simulation.
