Accelerating Techniques in Rendering Terrain Surface Based on Image-Space Error
基于图象空间判据的地表模型加速绘制技术

LoD (level of detail) is one of the most important techniques in creating real-time rendering of virtual environment. As a special mesh model, terrain surface model has important and wide-ranged application background in virtual reality. In this paper, the authors present an algorithm for real-time LoD reduction and display of high-complexity terrain surface data. The algorithm allows the generation and computation of appropriate level of detail dynamically by using the compact and efficient regular grid representation. Through analysis of the special requirement for rendering virtual terrain surface, they combine size and focus criteria to estimate the importance of detail of terrain surface and provide a screen-space threshold to bound the maximum error of the projected image. Besides, the whole dataset is constructed into a hierarchy of block-based representations with their error bound precalculated. As a result, a coarse level of simplification could be firstly performed to choose discrete level of detail for the surface mesh, and further simplification is operated on individual mesh vertices within a block. The algorithm has been implemented for approximating and rendering digital terrain models and the experimental results show its high efficiency.