A Synchronous Simplifed Method Of The River Network Line Vector And DEM Based On The Three Dimensional DP Algorithm

Level of detail(LOD) model is the foundation of multi-scale representation and operation in massive spatial data. At present, due to the approximation in format of image and DEM data, the synchronous LOD modeling of them are easier to achieve.but in the integrating approach of vector data and DEM, the method based on texture changes characteristic of vector data itself, and it is not easy to perform spatial analysis operations. while the method based on geometric superposition, LOD model of vector data and terrain grid data are established separately, so their operations are flexible and easy to exert the spatial analysis characteristic of vector data. But the LOD model of vector data and grid data are built separately, which exist a problem that simplifying factor(or standard) of them are inconsistencies, in general, the simplify of the vector data is based on the elements of geometric spatial characteristics, while the simplify of DEM according to the terrain roughness, etc. The geometric or topology contradiction will be happened in the scale operation and visual analysis, such as vector lines "through" and "hang", which would directly affect the accuracy of national large emergency decision and current trend of the public geographic information services.Therefore, through the application and modified the 3D_DP algorithm, this dissertation preliminarily explores the synchronous simplify and synchronous LOD modeling of river vector line data and DEM data. That is to say, in the synchronous simplify process of river network and DEM, different thresholds are set to control and select the discrete points set gradually, which can establish a dynamatic simplify data set. This data selection way is progressive data selection process and can provide different accuracy requirement for LOD model.The main works and the experiment conclusions include the following several aspects:(1) The select method of the first plane in 3D_DP algorithm is improved and the efficiency is improved. The generalization process of 3D_DP algorithm has the global features, it can simplify the tiny structure and remain the main landscape structure characteristics, while need not deliberately looking for the terrain structure lines. In addition, the way of progressive data selection can provide different accuracy requirement for LOD of the data set. In the 3D_DP algorithm, the determination of the first base surface plays a vital role in the whole algorithm, but the determination of the first base surface is very time-consuming in the original algorithm(we call the strict method). Considering the distribution characteristics of the earth's surface point set, the determination of the first base surface is improved and experiment is verified, it shows that: the result of generalization of 3D discrete point data using this improvement method is very similar to the generalized result by the strict method, but the time efficiency has improved dozens of times.(2) A bending adjustment index is designed to improved the 3D Douglas-Peucker algorithm, and this paper puts forward a synchronous simplified method of river network line vector dada and DEM data. Because the simplify of discrete point set by using 3D_DP algorithm mainly reflects on the height direction, if the river network line vector and DEM are simplified by 3D_DP algorithm, it is obviously that river's plane bending shape of itself will not be able to be reflected. In order to solve this problem, this paper introduces a "bending adjustment index" to improve 3D_DP algorithm, and puts forward a new simplified method of river network and DEM in three-dimensional space. In this method, river network line vector data are extracted into 3D discrete point data sets which are added elevation attributes, and then they are merged with the 3D discrete point data sets of DEM. The simplify operations are made by the improved 3D_DP algorithm after the hierarchical selection of river networks, in the end of this method, the river line segments which be seen as the implicit terrain topography lines were used as a mandatory constraint condition in the process of constructing the Delaunay Triangulation, this operation eliminates the "river climbing" phenomenon after synchronous simplify. the well experiment results have been achieved. Through the role of bending adjustment index, the main terrain features and the overall distribution form of the rivers can simultaneously be reserved reasonably on the synchronous simplify of the 3D river network and DEM. The river network and DEM data were generalized under the same simplified factor in this method. It improves the quality of the cartography generalization and this method expands the application scope of 3D_DP algorithm.(3)It puts forward a synchronous LOD modeling algorithm for the river network line vector dada and DEM data based on the improved 3D_DP algorithm. The key technology of LOD modeling is to simplify the model, this paper puts forward a simplified rule based on 3D Douglas-Peucker algorithm, and constructs the multi-resolution real-time synchronous LOD model of the river network line vector dada and DEM. In the data preprocessing stage, the main operations include: data block, boundary extraction, the merge data points set(include the three dimensional discrete point data set of river and the three dimensional discrete point data set of DEM) are sorted according to the importance degree by the improved 3D_DP algorithm, the model spatial error is calculated by the hausdorff distance, Delaunay triangulation are constructed by the river network line segment as the constraint conditions and the local triangulation process. In the real-time rendering stage, this paper takes a approach to simply the relationship between the screen projection error and model spatial error, the real-time synchronous LOD models are generated and invoked by the sight distance control. This algorithm ensures the renderer effect and efficiency.(4) The new algorithms are proved the rationality by the prototype system of the synchronous LOD modeling for the river network line vector data and DEM data based on the improved 3D_DP algorithm. The prototype system is designed and developed in the context of C# language and OpenGL 3D API, which achieves the functions such as the improved 3D_DP algorithm, the synchronous simplify of the river network line vector data and DEM data, the synchronous LOD modeling for the river network line vector data and DEM data, and the rendering, interoperation, drawing. The experiment datasets include ASTER GDEM elevation data in local area(30m×30m, Baoding region and the surrounding area of Hebei Province in China) and river network vector line data. Outcome and performance of experiments demonstrate the correctness, feasibility and high efficiency of the novel synchronous simplify and synchronous LOD modeling algorithm of the river network line vector data and DEM data. The results show that: the algorithm can enables to realize the real-time synchronous LOD modeling of river network elements and DEM to draw smoothly, at the same time it has a faster show speed.