摘要
为解决残缺点云模型数据精简时边界特征容易失真的问题,以汽车覆盖件中的薄壁类零件为研究对象。提出一种保留残缺点云边界特征的数据精简方法。借助KD-tree建立数据索引结构,获取数据点最近邻,并通过邻域点拟合出微切平面的方法,计算出点云数据的法向量。利用法向量夹角大小关系,选取边界以及孔洞特征点的初始种子点。再根据欧氏距离实现初始种子点的邻域搜索,从而完成边界以及孔洞邻域特征点的提取。根据曲率精简的方法,对非特征点进行数据精简,最后,合并特征点云与非特征点云,实现对残缺点云模型的数据精简。将随机精简法、曲率精简法分别用于点云模型精简处理,结果表明:相比于其他两种方法,所提方法更好地保留了模型边界以及孔洞邻域特征数据点,其标准偏差、曲面表面积变化率优于其他两种方法且变化相对稳定。
In order to solve the problem that the boundary features are easily distorted when the data of the residual defect cloud model is reduced,the thin-walled parts in the car panel were taken as the research object.A data reduction method that preserved the boundary features of residual faulty clouds was proposed.The data index structure was established with the help of KD-tree,the nearest neighbors of the data points were obtained,and the normal vector of the point cloud data was calculated by the method of fitting the micro-tangent plane through the neighborhood points.The relationship between the angle and the normal vector was used to select the initial seed points of the boundary and hole feature points.Then,the neighborhood search of the initial seed point was realized according to the Euclidean distance,so as to complete the extraction of boundary and hole neighborhood feature points.According to the method of curvature reduction,the data of the non-feature points was reduced.Finally,the characteristic point cloud and the non-feature point cloud were merged to realize the data reduction of the residual defect cloud model.The random reduction method and the curvature reduction method were respectively used for the reduction of the point cloud model.The results show that compared with the other two methods,the proposed method better preserves the model boundary and hole neighborhood feature data points,and its standard deviation and surface area change rate are better than the other two methods,and the changes are relatively stable.
作者
尹金林
王春香
刘流
王齐超
潘杙成
YIN Jin-lin;WANG Chun-xiang;LIU Liu;WANG Qi-chao;PAN Yi-cheng(School of Mechanical Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,China)
出处
《科学技术与工程》
北大核心
2023年第4期1607-1614,共8页
Science Technology and Engineering
基金
内蒙古科技大学科研项目(0341008001)。
关键词
残缺点云
邻域特征提取
法向量
平均曲率
数据精简
residual fault cloud
neighborhood feature extraction
normal vector
average curvature
data reduction
