中文摘要：

针对零件数据采集过程中可能出现的测量数据边界不完整的情况,提出一种点云部分缺失条件下边界曲面的直接拟合算法。对于两种不同的边界曲面类型,选择相应的方法自动构造贴近点云的基面;对点云中的点进行基面投影参数化的同时,运用改进的病态指标准则确定基面的病态控制顶点;在曲面拟合过程中,通过对病态控制顶点的抑制和加入离散边界约束条件,使得拟合得到的边界曲面在逼近点云的同时保持了良好的曲面内部品质,且和周围曲面之间达到近似G1连续。工程应用实例验证了提出算法的有效性及实际应用价值。

英文摘要：

During the process of data acquisition in reverse engineering, the measuring data close to the boundary edges, under the influence of various factors, are sometimes missing. A practical algorithm for boundary surface straight fitting with incomplete point cloud is presented. The base surface is constructed automatically according to its boundary type. The points are projected to the base surface to get the parameters and meanwhile the weakness of each control point is calculated by using a new criterion. The weakly defined control points are constrained and discrete boundary constraints are included during the surface fitting process. The fitting surface is close to the point cloud while preserving good quality of its own. Furthermore, it satisfies approximate G1 continuity with adjacent surfaces. Practical application shows that the proposed algorithm is valid and valuable.