基于无序点云的叶片截面特征参数提取

Cross-sectional feature parameters extraction of blade based on unorganized point cloud

针对提取航空发动机叶片截面特征参数的实用性要求,研究了基于无序点云数据的叶片截面特征参数提取方法.综合距离法和二分法的优点,采用基于矩形腐蚀法的距离-二分法对点云数据排序,基于最小包容区域直线和最小二乘圆拟合,提出了将整条叶片截面点云数据分割成前缘、后缘、叶盆和叶背4部分的自动分区方法,对特征参数提取方法做了研究并用VC++进行算法实现,使用UG/OpenGrip生成UG中叶片截面上的点云数据进行实验运算,计算精度达到10-4 mm,表明在实际测量和参数提取中算法误差可以忽略.

According to the practical requirements for extracting the cross-sectional feature parameters of aircraft engine blade, algorithms based on unorganized point cloud were studied. For the point cloud sorting, the distance-dichotomy based on rectangular corrosion was presented in consideration of the advantages of distance method and dichotomy. The approach based on minimum contain regional linear fitting and least square circle approximation was proposed for partitioning automatically the cross-sectional data of blade into four parts of leading edge, trailing edge, pressure face and suction face. The algorithms of feature parameters extraction were studied as well, and realized with VC＋＋ and simulated using the unorganized point data generated by UG/OpenGrip. The experiments show that the precision of the calculation can reach 10^-4 mm, indicating that algorithm calculation error can be ignored in the practical measurement and parameter extraction.