Research On 3D Visualization Technology Of Digital Elevation Model

With the advancement of computer technology and related fields,the production and analysis of Digital Elevation Model(DEM)have become more convenient and efficient.DEM find widespread applications in fields such as GIS,land-use planning,resource management,and natural disaster prediction.The 3D visualization of DEM is often realized by using level of detail(LOD)models based on quadtree,but there are some problems in these methods: data redundancy,feature missing and computational complexity.Therefore,in this study,a Lo D based on terrain features is proposed,employing a polygon-based accelerated occlusion culling technique to reduce rendering pressure.Firstly,to enhance the expressiveness of terrain factors in representing terrain features,an Intergated Terrain Feature Index(ITFI)is proposed.ITFI is obtained through several steps:(1)Qualitative analysis and selection of candidate terrain factors;(2)Calculation of the optimal analysis window for macro terrain factors to ensure data accuracy;(3)Determination of the best factor combination using Sheffield entropy;(4)Using extreme value entropy to obtain the ITFI.Secondly,in order to solve the problems of high tree height and repeated calculation in quadtree,a Lo D based on Terrain Features(Lo D-TF)is proposed.It utilizes a Progressive Node Evaluation System(PNES)to construct a quadtree of terrain blocks,limiting tree height while preserving basic terrain information,reducing redundant computations.Furthermore,feature points are selected based on the ITFI to improve model accuracy.Finally,to alleviate rendering pressure,an improved occlusion culling algorithm based on polygon acceleration is presented.By decomposing terrain blocks into different-sized sub-blocks and constructing occlusion polygons based on the screen space position of the terrain blocks,the algorithm determines if rays intersect with occlusion polygons after a certain number of frames.Non-occluded terrain blocks are rendered,and the occlusion polygons are dynamically updated for the next frame’s occlusion culling.