Developing a point selection strategy for elevation data modeling in synthetic environments

Very high resolution digital elevation models (DEM) are becoming more commonplace as source data to support the construction of terrain representations for use by simulations. Often however, the simulation systems are not capable of processing the large amounts of data found in the original elevation models. This is particularly true in the case of virtual simulations that, include a visualization component. Such systems must often dramatically reduce the data to meet processing limitations of graphics components responsible for rendering real-time, “out of the window” views of the environment for presentation to the simulation users who are to be immersed in a virtual world. Typically, the data reduction process relies on transformation of the dense, regular lattice of DEM elevation points into a network of irregularly sized, shaped, and oriented polygons. Point selection strategies are critical to this transformation. They define a small set of the original data points, according to a pre-specified polygon budget, that will represent the entire area described by the source DEM and thus determine the fidelity of the resulting polygonal representation. This paper describes a novel and developing point selection strategy and provides a brief comparison with a more traditional approach