Study On Organization And Visualization Of The Massive Terrain Dataset

In this paper, study on organization and visualization of the massive terrain dataset mainly on Regular Square Grid data format are discussed. By building mass DEM terrain multi resolution model with quad-tree data structure, view port related visibility judgment and dynamic scheduling strategy mass data seamless walk-through is realized based on eliminating fraction the caused by different resolution terrain putting together. All the research work is list below:(1) Interpolation and simplifying methods of Digital Elevation Data is introducted here. This paper does a reach on the simplifying method of quad-tree data struct. Based on a given error, a simple way of judging whether the node is a leaf is proposed. By scaning the original data only once Using recursion, the nodes to be partitioned can be found quickly. The way of crack elimination is by setting the vertex of higher resolution node to the edge of conterminous lower resolution node.(2) The pyramid data models are achieved with the data process of data division and simplization on DEM and texture data. They can provid the data model for the rendering of massive mulit resolution terrain visualization.(3) The dynamic scheduling strategy of muliti resolution terrain model: the view-frustum culling and Level Of Detail Model are the key technology of massive terrain visualization. This paper discusses a clipping method of using linear quadtree data structure. In order to improve graphic display rate, view port related visibility judgment is introduced. By modifying the points of boundary of the higher detail block, the crack can be eliminated.(4) With Object Oriented method and Component Object Model technology, a massive terrain 3D visualization software for real time and seamless walk-through is built using Microsoft VC++6.0 and OpenGL. the function models and their relationships of the 3D visualization system built on the theory above are discussed.