Research And Implement Of Manufacturing Error Analysis Method Based On Point Cloud Segmentation

With the development of digital design and manufacture,the manufacturing precision and production efficiency of parts have been greatly improved.The traditional detection method measures the shape and dimension of the parts manually,obviously can not meet the needs of large-scale production.At present,the digital inspection method for manufacturing errors is to register part’s point cloud and standard model,and then evaluate the manufacturing errors of parts according to the accuracy of registration.However,in the design phase of a part,the various dimensions of the part usually have different precisions and references.Using registration quality to evaluate the manufacturing error of the part can neither analyze the error of the opposite piece,but also ignore the difference of dimensions of the part.So the pertinence and difference of error detection can not satisfy the demand very well,even can not reach the purpose of error detection.To address these issues,a kind of error detection method of mechanical parts based on point cloud segmentation is proposed,which research with the point cloud data of the parts and the Initial Graphics Exchange Specification(IGES)model and aim at the pertinence of the error analysis and improving the precision of the error analysis.The main research contents are as follows:(1)The IGES model was reconstructed to get information about the standard model of the part.Considering the interface of Non-Uniform Rational B-Splines(NURBS)in Open Graphics Library(Open GL),surface in IGES is transformed into NURBS surface.In order to analyze different kinds of surface more specifically,the NURBS surface and curve in IGES model are sampled and fitted,and the specific type of surface is determined.The sampling point set of the IGES model is further processed,which provides for the registration of the part point cloud and the IGES model.(2)In order to realize the segmentation of point cloud,the registration of point cloud and IGES model is carried out.After registration,the Euclidean distance between the points in the point cloud and the surface in the standard model of the part is calculated,the relationship between the point and the boundary of each surface is judged,and the corresponding relation between the points in the point cloud and the surface of the IGES model is determined,so that the point cloud is segmented for the first time.According to the result of the first segmentation,the registration result is adjusted by the reference,and the registration result and the point cloud segmentation effect are optimized.(3)According to the result of point cloud segmentation,the segmentation result is fitted.Different error analysis methods are designed.According to the specified parts for different error types,the errors of each dimension of each element of the parts are analyzed.The content and research results of this paper have been developed and implemented in the Manufacturing Error Analysis Aystem of Pechanical Parts.The validity of this method in the manufacturing error detection is verified,which provides a reference for the product designer to improve the products better.