Research On The Key Technologies For Surface Reconstruction From Discrete Point Cloud

Reverse Engineering that is applied in kinds of field and provide technical support for the designand manufacture of a product in recent years, has become an important approach to digest and absorban advanced technique and to develop a new product quickly. Reverse Engineering rebuild the surfacemorphology based on the scanned point cloud data form a model, modify it and generate the finaldigital models. So how to rebuild the surface morphology form the scanned point cloud data form amodel is the key of Reverse Engineering, and have high research value. This paper study the relatedtechnologies of the process from point cloud to triangular mesh surface.1. This paper propose a point denoising algorithm based on Space Cell for solving the difficulty indistinguishing noise that raises in local area. The algorithm firstly use the Space Cell technology tosegment the spatial box bounded by the point cloud, and then denoises based on the topologyinformation of each point and each cell. The result of the algorithm is good.2. This paper improves point cloud simplify algorithm base on curvature sample using the SpaceCell technology to improve the efficiency of the algorithm. The algorithm can improve the efficiencyof five times while keeping the characteristic of point clouds, and is suitable for the simplifying oflarge number of point cloud data.3. The ICP algorithm is improved to adapt models that lack significant characteristic. We propose aself-adapting sampling along “slide” direction method and a new way to determine matching point setto promise a high quality registration.4. Triangular mesh reconstruction is also studied to solve the Delaunay triangulation’s complicationand low efficiency. A triangulation based on region growing method is realized using the Delaunaytriangulation optimization criterion and achieved good triangular mesh reconstruction effect. Thenumber of unstability triangular patch is chotrolled within one-in-one-thousand.The algorithm is achieved on the prototype system based on Microdoft Visual Studio.NET2008development platform and HOOPS/3dAF display platform. The result shows that the algorithmachieves satisfactory effect.