Adaptive Terrain Triangulation Using The Representation Of Quad Trees By Vertex Textures And Wavelet Estimation of Vertex Significance

摘要

A method of adaptive terrain triangulation is proposed that can be implemented in hardware. The method is based on an estimate of the static error of a quad tree nodes using wavelet transforms and on the representation of the resulting quad tree by a vertex texture. The proposed method has the following characteristic features: the adjacent nodes of the generated adapted mesh can differ in any number of hierarchical levels; the triangulation process is not limited by the size of the decomposition segments, which solves the problem of joining segments without inserting additional nodes; the mulliscale terrain representation used in the method makes it possible to store the levels of detail in the graphics processor memory as a multilevel vertex texture; thus, the costliest part of the algorithm can be efficiently implemented using a vertex shader. When constructing the triangulation, the algorithm takes into account both local features of the terrain and the camera location; also, it has a natural support of geomorphing.