Research And Implementation Of Software For Detecting Accuracy Of Parts Based On Point Cloud Registration

With the development of the laser ranging technology,3D laser measuring instrument has been widely used in the current application in the manufacture of products.We can directly get real space sampling points,that is,Data Points,on the surface of the object by using 3D laser scanning technology.The Data Points obtained by the Non-contact Measurement based on optics are charactered by large amount of data and random order.A processing system will be introduced in this paper,which can effectively analyze and calculate the sampling data of the parts especially which with continuous and complex curve shape.The function of this system is for fitting the Data Points,and reconstruct the manufacturing characteristic of the sample.Calculating and evaluating the manufacturing errors,such as Planeness,depth of Parallelism,and Profile Controls,according to the fitting results,which can promote the optimization of products designing and manufacturing,improve the integration in design and manufacture of products,and eventually improve the quality of products.The research content of the paper is as follows:The initial denoising after collecting point cloud data and the calibration and registration techniques for two point cloud data were studied.The calibration technique based on point cloud geometry is studied.The coarse registration based on the PCA algorithm was studied to transform the two point clouds to a roughly overlapping position.The registration process based on the ICP algorithm is then performed.Thereby reducing the number of iterations of the algorithm and improving the efficiency and accuracy of the registration.The fitting algorithm of point cloud data is researched.The fitting module in the system reconstructs the product shape by measuring the point cloud data on the product surface,which can visually reflect the shape characteristics of the product.The fitting result is used to calculate the error evaluation parameters of the product.The least square method is applied to fit the line,plane and quadric surface(including cylindrical surface and conical surface)on the surface of the product to obtain the parameters of each object to be processed.A software system with error assessment of the manufacturing error of the product shape after acquiring data and standard point cloud data by registering the point cloud is designed and implemented.This software uses the powerful data analysis and numerical calculation from MATLAB,improve the efficiency of data processing,builds the interface by Microsoft Visual Studio 2010.In the system,users can intuitively see the 3D model obtained by fitting Data Points.Besides,the system can directly reflect manufacturing errors size by color chart,which makes it easy for users to get a more intuitive understanding of the overall structure and error distribution of products.