Study Of Rapid Geometry Rendering And Real-Time Texture Mapping For Large-Scaled Terrain

With the development of computer technology, people prefer three-dimensional topographical map to two-dimensional which lack of vivid performance. Along with the development of remote sensing technology, satellite technology, and measurement technology, people begin to no longer satisfy with the small-scaled three-dimensional terrain but to prefer large-scaled three-dimensional topography as the "Digital Earth". Today, large-scale three-dimensional terrain is playing key role in many applications such as geographic information systems, flight simulation and 3D games. So, how to organize large-scaled three-dimensional terrain data effectively, how to rendering terrain rapidly and vividly has become a very important research questions.The paper propose a framework for effective visualization of large-scaled three-dimensional terrain, our terrain data will be divided into geometric data and texture data, geometric data gets from digital elevation model (DEM), and texture data gets from remote sensing Image, both geometric data and texture data were multi-resolution organized, and based on the errors on the screen, we can select the appropriate resolution to rendering the terrain, the major contribution of this paper as follows:1. Building LOD models based on geometric terrain data. The terrain is divided into different blocks and further divided into many triangles to build binary tree data structure. The paper propose the triangle binary tree simplify algorithm based on our coding approach, the triangle simplify algorithm is the core of building LOD models. In our algorithm, each node of binary tree is encoded, and we calculate the coding of each node that we need in accordance with the formula of our paper. The algorithm eliminates the traditional structure of the binary tree that visits children or neighbors based on pointers, which saves lots of time and space.2. Data scheduling from out memory. the paper reorganize terrain data in external memory first, then open up a separate thread for scheduling data, which uses modern computer's two CPU feature, data scheduling thread and the main thread were assigned on their own CPU, and we sync the two threads based on Event to make a better use for out scheduling thread.3. Texture mapping for the large-scaled terrain. The paper build multi-resolution structure for texture data in external memory, then page the low resolution texture data into memory ahead of rendering, at last we map our texture data to the triangles. In this paper, our framework multi- resolution organized the geometric data and texture data, and the feature of the framework is that although the size of data far beyond the scale of memory capacity, the system can still rendering the terrain rapidly.