Research On Surface Feature Recognition And Domain Processing Technology In Reverse Engineering

With the development of scientific technology,the combination of reverse engineering and computer technology has been widely used in many fields such as automotive modeling,mold manufacturing,aerospace,architectural design,precision parts testing,artificial organs,archaeological restoration,etc.Nowadays,the process of reverse engineering research can be roughly divided into digitization of product models,pre-processing of point cloud data,reconstruction of complex surfaces,and optimal reproduction of models.Compared with traditional design methods,reverse engineering can not only invert more complex free-form surfaces,but also complete the re-repair of existing parts with higher R&D efficiency.So it is of great importance to increase,strengthen and deepen the research on reverse engineering,both theoretically and economically.At the same time,in order to address the problem of low overall processing efficiencies of free-form surfaces and hard to guarantee the processing quality,this paper not only simply reconstructs the original free-form surface,but also re-divides the reconstructed free-form model into areas in order to improve the processing efficiency and quality of free-form surfaces,and the main research contents are as follows:(1)Acquisition and pre-processing of surface model data.The first step is to introduce contact and non-contact measuring instruments systematically and to explain their advantages and disadvantages.The point cloud data of the free-form surface is acquired by using the self-constructed three-axis machine combined with the contact sensor.Secondly,we introduce the common point cloud data noise reduction filtering method,and optimize the statistical filtering method based on it,and introduce the voxel idea to achieve more efficient point cloud noise reduction filtering.Finally,the common point cloud refinement algorithms are introduced,and the k-means clustering algorithm combined with the refinement idea of octree is proposed to compare with the traditional method,which can more effectively refine the point cloud data while preserving the point cloud features.(2)Fitting of surface models and surface area delineation.The developments of each spline curve describing a free-form surface are introduced,and the use of non-uniform rational B spline curves(NURBS)to fit already processed point cloud data to reconstruct a NURBS surface.Then factors affecting surface characteristics are introduced,and the values of Gaussian curvature and mean curvature are used to classify surfaces into three types:convex,concave,and saddle.The classification results are optimized by combining the fuzzy c-means clustering algorithm,and finally the boundary given to surface pieces is determined using Voronoi diagrams combined with clustering centers to complete the surface classification.(3)Determination of CNC machining parameters and simulation verification.Firstly,the size and type of tool,tool path,tool path generation method and tool distance are introduced,and the machining parameters are determined accordingly.Then,the free-form surfaces are transferred from MATLAB to CATIA,and the simulation is verified by using the CNC machining module in CATIA,and the results show that the method in this paper can improve the machining efficiency and quality of free-form surfaces effectively.