Research Of Mesh Model Simplification For Large-Scale Terrain Dataset

As a hot topic in the field of computer graphics, the 3D visualization technology for large-scale terrain has become the focus of the study. It is applied to GIS, Virtual reality, aerospace and other areas. With the development of modern measuring and remote sensing technology, digital elevation model increases exponentially, the massive terrain data is far beyond the limited hardware conditions, which makes mesh simplification algorithm become one of the research directions in the visualization of 3D terrain. Mesh simplification algorithm refers to the use of as less as possible polygonal mesh to characterize the terrain model, and guarantees the facticity and reflect the local feature, enhancing the efficiency of terrain generation. Therefore, it is very important to study on the algorithm for mesh simplification technology.Massive terrain data is a major bottleneck restricting the development of 3D visualization for large-scale terrain. How to effectively implement the organization of data and the algorithm of mesh simplification, affects the performance of virtual walkthrough under the authenticity of terrain conditions. So, the paper studies the problems of terrain visualization from the data preprocessing operation and retaining the local detail characteristics. The organization of large-scale terrain data will affect the processing efficiency in the rendering stage. Therefore, the paper proposes an adaptive strategy of block division, which uses the standards of terrain block space range and topographic relief. The terrain entropy algorithm is improved to describe the roughness of the terrain block, which can reduce the different of terrain blocks in the recursive subdivision process time difference. Meanwhile, in order to avoid the loss of information feature, the calculation error of vertex degree is defined, which combines energy function and curvature. And the paper proposed a reserved terrain features block multi-resolution model adaptive parting method. Besides, in order to improve the speed of alleviate paging and reduce the recursive time consumption, the paper improves storage from implicit quad-tree structure which renders non-recursive. The improved divergence function representation method is used to measure terrain error calculation, which generates a local divergence criterion function. At last, the paper proposes a method of local divergence function based on view-dependent technology, combining with the viewpoint technology. The presented rule simultaneously keeps the local features of the terrain.