曲面的自组织拟合技术研究

【作者】 周会成； 邓春梅； 师汉民； 曾理湛；

【机构】 华中理工大学国家数控系统工程技术研究中心；

【摘要】 本文介绍了一种基于自组织拟合的曲面数据云建模方法。该方法以型值点集合而不是解析式表达曲面,不仅能满足曲面数字化模型的基本要求,完成插值、法矢、曲率的计算,而且避免了传统建模方法中的四边域、三角域的拓扑结构划分问题和子曲面片之间的几何连续拼接问题。本文对曲面自组织拟合建模方法中的局部规则、拟合窗口大小、迭代收敛特性以及局部可修改性等问题进行了研究,提出了自组织建模的计算方法和相应的数控编程算法。采用本文的方法,对激光测头测得的散乱数据云进行了建模实验,通过编制相应的数控加工代码,在数控机床上完成了蜡模切削实验,验证了该方法的可行性。更多还原

【Abstract】 A surface modeling method from datacloud based on self-organization fitting is introduced in this paper. Sampling points instead of analytical equations are used to express surfaces. In this way, those basic requirements of digitized surface model such as interpolation, normal vector and curvature calculation can be accomplished. Furthermore, the topology structure division problem of triangles and quadrangles and the problem of geometrical continuous blending sub-surfaces is avoided. Local function, fitting window width, iterative conjunction properly and local surface modifiability are addressed while the modeling method based on self-organization mechanism and corresponding NC tool path generation algorithm are presented. With the method presented in this paper, an experiment on reverse engineering is carried out. Surface is reconstructed from scattered datacloud measured by laser scanner, NC code is generated and a waxen mould is machined on a milling machine. The experiment shows that the surface modeling method based on self-organization is feasible and can be applied in industry use.更多还原