Research On Key Technologies Of Surface Reconstruction Based On Scattered Point Cloud Data

With the wide application of 3D reconstruction in the fields of reverse engineering,industrial design and virtual reality,3D reconstruction has become an important means to quickly acquire physical digital models.In the process of data acquisition of 3D reconstruction,point cloud is often used to express the model information of objects because of its simplicity,convenient collection and easy storage.The point cloud is often used to represent the model information of the object,so it is still a problem that how to reconstruct the 3D geometric model efficiently and accurately from the point cloud model expressed by the mass scattered point cloud data.This thesis is about the key technology of the surface reconstruction of scattered point cloud data.It is the geometric topology of the reduction point on the basis of the existing point cloud data,restoring the shape of the surface of the object model,making the surface of the reconstructed surface reflect the shape features of the surface as much as possible.The thesis studies the process from space scattered point cloud to 3D mesh model,and studies the key technologies such as point cloud spatial dissection,point cloud simplification,point cloud grid and so on.The following research results are obtained:1?The algorithm of point cloud spatial dissection is studied.In view of the inefficiency of the backtracking and the space enlargement of the octree subdivision of the k-d tree,a mixed subdivision strategy is adopted,which can not only reduce the dissection space,but also automatically adjust the depth of the dissection according to the density distribution of the point cloud.2?The algorithm of point cloud simplification is studied.In view of the problem of ignoring the point cloud feature and the low computation efficiency in the existing point cloud simplification algorithm,a new method of point cloud adaptive reduction based on mixed dissection is proposed.This method improves the estimation of point cloud density,and can effectively control the degree of simplification,and its simplification results can keep the point cloud evenly distributed while maintaining local characteristics.3?A point cloud meshing method is studied.A direct mesh generation method based on Delaunay criterion and growth method is proposed.In this method,various restrictive rules are formulated,vertex evaluation function is designed to select the best points,and the effect of grid can be controlled artificially.This method can directly mesh the point cloud in space,avoiding the complex computation in other methods,and greatly improving the efficiency of grid.4?The surface reconstruction platform of point cloud is realized.The platform is mainly used to display the experimental results,and the platform implements all kinds of algorithms proposed in this paper.Finally,according to the results of platform reconstruction,the results are compared with those of other algorithms,and the results are analyzed.