Study On Canal System Model Automatic Generating Algorithm Based On Vector Data

The developing of three-dimensional (3D) virtual reality (VR) technology brings it to a brand new 3D age from 2D world in simulating the real world by computer. 3D virtual scene gives people immersion feeling which can not be given by the 2D image, and the immersion is based on 3D data, such as 3D model. But the producing of 3D data usually is a long-term and costly process, which is beyond the carrying capacity of most the company or unit.During the past few decades, people have already collected or surveyed a huge number of 2D data, such as paper map, contour lines, GIS vector and so on. But they can not be used by our 3D virtual reality system directly, mainly because lack of elevation information. Therefore, it becomes a critical problem that whether can we extend the traditional 2D data to a 3D data, and be used directly by our 3D system.The ad hoc algorithm of vector generation usually including the following two steps. Firstly, interpolate the elevation of each vector control point, according to the 3D terrain model. Secondly, generate vector 3D model, usually using a similar rectangle to represent the vector segment between two consecutive control points, and the linked rectangles representing the whole vector. And the resulting 3D vector model is not fit the 3D terrain well, there is significant elevation offset, higher or lower, between part of the vector and the terrain, of the same xy coordinate. Besides, the shape seems not nature.In this thesis, we use channel vector as example, discuss an improved algorithm for vector automatic generation. Besides the traditional steps, we provide a vector refine algorithm, based on TIN terrain model. This procedure improves the matching between 3D vector model and 3D terrain model greatly. Use the bisectrix buffer model concept from GIS to represent vector 3D model, makes appearance of the vector model more nature than traditional model. Besides, we discussed and applied vector compress algorithm to reduce the size of vector data, providing not cut the precision significantly.