Research On Engineering Model Reconstruction Based On Spline Surface Theory

The current mechanical product design is described by three-dimensional CAD model.Most new product designs are based on existing examples.Some existing examples are 2D engineering drawings or product objects.The point cloud can be obtained from the product through 3D scanning.The design efficiency of the product will be greatly improved if the product 3D CAD model can be quickly established based on the physical product or 2D engineering drawings.This paper intends to study a fast reconstruction technique of engineering surfaces for 3D point cloud data or 2D engineering drawings.The main research contents are as follows:According to the theoretical knowledge of B-spline curve and surface and the theory of spline energy,the function of B-spline is improved.The algorithm of B-spline back-calculation and fitting is compiled,which provides algorithm basis for 3D CAD model reconstruction.In this paper,different data processing methods are designed for the original data,ie the scanned point cloud data and the 2D CAD drawings in the engineering design.For the point cloud data,the density clustering method is used to eliminate the noise points.A point cloud data preprocessed algorithm based on the adaptive minimum distance method and the octree-based non-uniform grid method is proposed to exclude the data points with small curvature change.The quality and efficiency of the reconstruction are guaranteed by excluding the noise points that have a negative influence on the reconstruction and the bloated points that have less influence on the reconstruction effect.For 2D engineering drawings,the DWGDirectX plug-in is used to extract and analyze the dwg drawings,and divide the information into regions by means of element bounding boxes.According to the characteristics of the engineering drawings,the categories of information in the engineering drawings are corresponding to the divided areas,realizing the purpose of judging the information category according to the position coordinate of the bounding box of the information in the drawings and establishing the semantic recognition database.Different drawings can establish different semantic recognition databases due to their different categories and characteristics.The purpose of judging the type of engineering drawing and extracting information from the drawing by imported 2D engineering drawing in program is achieved by comparing the semantics to identify the content in the database.The optimized fitting algorithm of data point reduction is proposed.In the case of ensuring the fitting precision,the data points are reduced during the fitting process.The number of data points participating in the curve and surface fitted is reduced,the fitting efficiency is improved.Verify that the algorithm can effectively improve the efficiency of the fit.Based on the above research content,this paper uses C++ language to develop a program on VS2010 platform,which realizes the function of automatically and efficiently converting point cloud or 2D engineering drawing into 3D CAD model.