中文摘要：

介绍了一种快速的散乱数据点曲面重建算法（切面投影三角网法）,该算法不需要已知数据点的几何、拓扑信息以及是否存在边界等先验知识。算法利用邻近点集反映出的局部几何和拓扑信息,基于切平面投影方式计算每个数据点的邻域,从而完成每个数据点的局部拓扑重构。重构中物体表面数据点的降采样或不均匀采样可能会产生伪洞,因此,在重建后进行洞的检测,进而根据洞的大小来区分物体表面上实际存在的洞和重构过程中生成的伪洞,并对伪洞进行填充。利用多组散乱数据点进行重建的结果表明,切面投影三角网法高效、稳定,可以快速、自动地重构出复杂拓扑结构物体的三角网表面模型。

英文摘要：

A fast surface reconstruction algorithm for unorganized points,without any priori knowledge about geometry or topology information or the existence of boundary of the data,was proposed in this paper.The algorithm used the local geometry and topology information obtained from the neighbor points to calculate every sample point＇s neighborhood based on tangent plane projection,implementing the local topology reconstruction of every point.Sparse or non-uniform sampling from the surface of a object may generate pseudo hole during the surface reconstruction process.So detection of holes after surface reconstruction was required.The real holes and the pseudo holes can be distinguished according to the size of them,and then we filled the pseudo hole.The experiment result with multiple sets of unorganized points proves that the algorithm presented in this paper is efficient and robust,and can reconstruct the triangular mesh model of objects with complex topology structure quickly and automatically.