Algorithms On Preprocessing And Boundary Extraction From 3D Point Cloud Data Of Aircraft Engine Blade

Engine blade is one of the most important components of aircrafts,study on blade repairing technology can bring us huge economic benefits.In order to repair the damaged blade automatically,we need reconstruct the complete geometric model of the lossless blade from the point cloud data accessed by scanning the blade using reverse engineering technology,in which the preprocessing and the boundary extraction from point cloud data are the key steps.In this paper,we studied the problems about preprocessing and boundary extraction from point cloud for the purpose of recovery of damaged aircraft engine blades.The main contributions are as follows.1.Computing the K-Nearest neighbors of a point in the point cloud is a basic operation on point cloud.To address this issue,we proposed an improved algorithm for searching K-Nearest neighbor points based on space rasterization.Distinguished from the existed space rasterization method,the proposed method determines the succeeding searching lattices according to the location of the point in the lattice.Experimental results show that our method is much faster than the existed space rasterization method.2.For the denoising of the point cloud,we design an algorithm for removing the external flying spots based on rasterization of point cloud,and an algorithm for removing clustered outliers based on spread of neighborhood.The Experimental results show the effectiveness of the proposed algorithms.3.An improved algorithm for reducing the number of points in point cloud is proposed,which combined the plane fitting method and random sampling method.The experimental results show that this algorithm can reduce the number of points very fast meanwhile keeping the details of the point cloud.4.An algorithm for extraction of boundary points from the point cloud based on curvature is presented.For each point in the point cloud,we fit a ball to the given point and its K neighbor points and then use the curvature of the ball as the curvature at the given point.The experimental results show that the boundary extracted by this algorithm coincide the real boundary of the blade very well.