A Study on dynamic resolution control for multi resolution geometric modeling

针对处理多种类型对象数据的几何造型问题，我们进行了一项研究，开发了一种动态控制数据分布的技术，并得到了以下结果.为了从大量的点数据中在线构造出与其分辨率分布相对应的网格数据，提出了一种在特定分辨率下发现点的方法。为此，我们开发了一种技术，生成一个多分辨率网格使用的曲面的逼近精度为标准，利用多分辨率分解方法，细分基本三角形。设计了一种基于三角形层次结构的金字塔数据树构造算法，并设计了一种适合于该数据结构的顶点和三角形搜索算法，并通过上机实验验证了该算法的有效性。我们还开发了一种控制空间网格分辨率的方法，假设对象作为简单物质存在。在这里，利用数据树中保存的空间局部性，可以保证不同分辨率网格存在时分辨率变化的连续性，提出了一种动态分辨率控制技术，以适应由事件运动引起的网格组成变化，这种变化决定了网格的分辨率分布。根据测距仪的构造和评价结果，提出了一种基于事件运动的动态分辨率控制技术，并给出了一种基于事件运动的动态分辨率控制方法为了获得本研究所用的真实的数据，我们使用独立购买的测距仪进行了实验。并对构建高效测距仪进行了初步研究。

We performed a research for developing a technique to control a data distribution dynamically for the geometric modeling problem which processes a lot of types of object formdata andobtained following results. The technique becomes important in CG (Computer Graphics) research.To construct a mesh data which corresponds to its resolution distribution from a huge number of point data on-line, we developed a technique for discovering points in a specific resolution level. For this aim, we developed a technique of generating a multi resolution mesh using an approximation accuracy of a curved surface as the standard utilizing the multi resolution decomposing method which subdivides basic triangles. We also designed an algorithm that constructs pyramid data tree based on the hierarchy of triangles and that searches vertexes and triangles which suit to the data structure.Moreover, we verified validity of the algorithm by mounting it on a computer and investigating its behavior. We also developed a method of controlling the spatial mesh resolution assuming that an object exists as a simple substance. Here, we could guarantee the continuity of the resolution change for the problem that meshes with different resolution have to exist using spatial locality saved in the data tree.We developed a dynamic resolution control technique for corresponding to changes of composition of a mesh which will be caused by movements of events that determines the resolution distribution of the mesh.From the result of the construction and evaluation of a range finder for acquiring the real data used by this research, we conducted experiments using a range finder purchased independently. We also performed a preliminary research for building a highly efficient range finder.