AN ADAPTIVE SUBDIVISION METHOD FOR RECONSTRUCTING FREE FORM SURFACES FROM MEASURED DATA POINTS

本研究提出了一种从一组实测数据点重构细分曲面的方法。提出了两种方法。首先，给出细分曲面的初始控制网格，并对控制点进行调整，使细分曲面与数据点拟合。根据曲面与数据点之间的间隙，自适应细分控制网格。然后在第二种方法中，通过简化由测量数据点生成的密集三角形网格来生成初始控制网格。在此简化过程中，评估最终细分曲面与密集网格之间的间隙，并用于确定网格的哪一部分必须首先进行简化。我们通过几个例子实现了原型软件来评估生成的细分表面的准确性。这些方法生成的细分曲面与数据点具有良好的拟合性。我们进一步开发了一种将这种细分曲面转换为具有G1连通性的b样条补丁集的方法。我们现在正在研究如何基于这些方法在两个细分曲面之间产生平滑过渡。

In this research a method is proposed for reconstructing subdivision surfaces from a set of measured data points. Two methods are developed. First, an initial control mesh for the subdivision surface is given and the control points are adapted so as to fit the surface to the data points. Based on the gap between the surface and the data points, the control mesh is adaptively subdivided. Then in the second method, such an initial control mesh is generated by simplifying the dense triangular mesh which is generated from measured data points. In this simplification process, the gap between the final subdivision surface and the dense mesh is evaluated and used to determine which portion of the mesh must be simplified first. We implemented prototype software to evaluate the accuracy of the generated subdivision surfaces using several examples. The subdivision surface generated by those methods showed a good fitness to the data points. We further developed a method to convert such subdivision surface to a set of B-spline patches which have G1 connectivity. We are now studying how to generate smooth transition between two subdivision surfaces base on those methods.