Research On Reconstruction Method Of Large Complex Surface Based On Point Cloud Data

With the development of measurement technology and equipment,reverse engineering technology of three-dimensional reconstruction has received more and more attention.However,it is inevitable that the acquired point cloud data will have problems such as disunity of position and pose of point cloud data,including noise points and outliers,large amount of point cloud data and low processing efficiency when measuring large complex surface.In view of the above problems,this paper studies the preprocessing and reconstruction methods of three-dimensional scattered point cloud of large wind power blades.The main research contents of this paper include:(1)The three-dimensional data acquisition method of large complex surface is studied,and its measurement principle,measurement efficiency and precision,measurement cost and applicable range are discussed.On this basis,the measurement scheme suitable for large scale wind power blade profile is determined.According to the characteristics of measured data,the pre-processing methods such as point cloud registration,denoising and simplification of point cloud data are studied respectively,and an improved ICP registration algorithm based on feature point selection is proposed,which can automatically realize the registration of two perspective point cloud data with large pose deviation.A statistical filtering method is used to remove outliers and noise points and the voxel grid is used to simplified the large-scale point cloud data.(2)The surface reconstruction methods of scattered point cloud are studied,including Lawson algorithm,Bowyer/Watson algorithm and Poisson algorithm.On this basis,a modified Delaunay triangulation algorithm based on normal vector is proposed to generate the reconstruction surface of large wind-power blade.The surface reconstruction technique based on Poisson equation is also studied and realized by programming.(3)A software prototype system for three-dimensional point cloud data reconstruction of large-scale wind power blades is designed and developed,which can realize point cloud data reading and visualization,point cloud registration,point cloud simplification and denoising,and point cloud surface reconstruction and other functions.The point cloud reconstruction software system is used to carry out point cloud registration,point cloud simplification and denoising,and point cloud surface reconstruction experiments to verify the effectiveness of the reconstruction system.Meanwhile,a reverse engineering software Imageware is used to conduct experimental comparison of point cloud model reconstruction of wind power blades.The experimental results show that the surface reconstructed by the development system can fit the original point cloud data well and generate a closed surface.