摘要
点云特征型面边界曲线的拟合质量直接决定了对应曲面裁剪的质量。以机械零件点云模型为研究对象,针对点云分割型面的边界点识别、排序、拟合和曲面裁剪问题展开研究。基于场力法提取的边界点集,提出了基于最短距离法的边界点集排序和方向夹角阈值限定的边界点集去噪等预处理算法;采用局部二次多项式函数拟合估算边界点集曲率,提出了基于曲率突变点归属性判断的边界点集分段和拟合算法;通过控制点投影变换求取二维裁剪边界,提出了基于特征点集多边形凹凸性的二维边界曲线方向判断法。使用机械零件模型典型型面进行实验,验证了各阶段算法。该算法具有很好的适用性,可进一步扩展应用至复杂曲面边界裁剪。
Quality of surface trimming is directly determined by the quality of curve fitting in point cloud(PC)feature surface.Employing point cloud model of mechanical part as research object,problems of boundary points recognition,sorting,fitting and surface trimming for segmented PC feature surface are studied. Upon boundary PC extracted by field force,preprocessing algorithms like boundary points sorting via shortest distance method and denoising under threshold limitation of direction angle are advanced. Curvature of boundary PC is estimated through local quadratic polynomial function fitting,and then boundary PC segmentation and fitting algorithms based on belongingness judgment of curvature mutation points are proposed. Two dimensional boundary is calculated by projection transform of control points,and direction of the correspondent curve is determined based on polygonal concavity of feature PC. Typical feature surface of mechanical part model is occupied for experiment,which validates the different stage of the algorithm. This algorithm has good applicability that can be further extended to boundary trimming of more complicated surfaces.
作者
陈华伟
袁小翠
伍权
徐卫平
CHEN Hua-wei;YUAN Xiao-cui;WU Quan;XU Wei-ping(School of Mechanical and Electrical Engineering,Guizhou Normal University,Guizhou Guiyang550025,China;Jiangxi Province Key Laboratory of Precision Drive&Control,Nanchang Institute of Technology,Jiangxi Nanchang330099,China)
出处
《机械设计与制造》
北大核心
2020年第1期5-8,13,共5页
Machinery Design & Manufacture
基金
国家自然科学基金资助项目（51365037）
关键词
点云特征
边界拟合
曲面拟合
曲面裁剪
Point Cloud Feature
Boundary Fitting
Surface Fitting
Surface Trimming
