Terrain modeling at variable resolution with fractal interpolation

In generating fine, realistic terrain surface from a coarse one, it is important to maintain the original natural features in the final results, while at the same time visualize the part of the scene that is close at a high level of detail and the part that is far at a low level of detail. In this paper, we present a new algorithm that can meet these two requirements. In order to meet the first requirement, we propose a fractal-based interpolation scheme for irregularly spaced data, where the fractal features are derived from the coarse sample data set. In this procedure, we add more detail to the area that is near the view point than to the area that is far away. Thus, the second demand is also satisfied. In addition, we use triangulated irregular network to visualize the generated terrain. A prime feature here is that it performs partial Delaunay triangulation at each interpolation step, so that the triangles with very small angles are avoided only at a tolerable computational cost.