The Study And Realization About The 3D Terrain Reconstruction Based On Texture Mapping

Texture Mapping technology was proposed by Catmull in 1974. It is an efficient method to show the details of the surface. It constructs the one-to-one correspondence between each point on the embedded meshes and the one on the texture plane. The method can enhance the realistic feeling more greatly with less computing. For different cases, there are different algorithms for texture mapping. According to the types of texture used, there are Color Texture Mapping, Bump Mapping and Processing Texture Mapping. According to the correspondence, there are Forward Texture Mapping, Backward Texture Mapping and Two-step Texture Mapping. In the process of the mapping, because of the parameterization and the exchange of the perspective transformation from 3D to 2D, aliasing and distortion often happens. Hence, anti-aliasing technology must be used in the texture mapping. Because of texture mapping can well present the surface details of the objects, more and more researchers pay attention to this technology.By now the terrain models have been presented with single color or pseudo-color, so the visual scene is not realistic. In view that the texture mapping is an efficient algorithm to realize the objects, it is used into the 3D terrain reconstruction system. In the system, a kind of mapping relationship is set between the points of the 3D model and the pixel of the texture image, according to which the texture mapping is realized. Then a novel 3D terrain reconstruction system with good realistic visual scene is realized. Based on the performance of the original system, the novel system takes the terrain model as the processing object, and then maps the real texture image onto the terrain model according to the facts. The results perform good and real. This paper is a part of a rearch project. The experiment data is a set of discrete elevation data ditected from Protable High-resolution Multi-beam Echo Sounder at Songhua Lake of Jilin, so the image produced by the novel system is about the real sublacustrine landform.By the comparison of the performance between before and after improvement, we can see that the novel system remains good efficiency, accuracy and stability, and the user can do real-time human-computer operations. Above all, the novel system proposed has an ideal performance and wide application prospect.