Font Size: a A A

Research On Three Dimensional Reconstruction Based On Laser Radar

Posted on:2019-02-28Degree:MasterType:Thesis
Country:ChinaCandidate:W N CaiFull Text:PDF
GTID:2348330566464275Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Due to the needs of reverse engineering,heritage preservation,virtual reality,medical assistance and other industries,as well as the development of three-dimensional laser scanning technology,3D reconstruction has become one of the hot topics in computer vision and reverse engineering.Three dimensional reconstruction is the process of point cloud preprocessing,point cloud registration,surface reconstruction and texture mapping for point cloud data acquired by laser scanner,so as to form the three-dimensional digital model of target object.Based on the research status and development trend of 3D reconstruction technology at home and abroad,this paper mainly focuses on point cloud registration and surface reconstruction algorithms.The specific work is as follows:In view of point cloud registration,a feature point extraction method based on normal vector angle and neighborhood is proposed in this paper.Points in a radius range are chosen as the neighborhood,and the local plane is fitted by the total least square method.The principal element analysis method is used to calculate the normal vector,and the direction of the normal vector is adjusted accordingly.Then the feature points are extracted by using the judgment of the angle of the normal vector and the number of the neighborhood points.Finally,the 4PCS algorithm is carried out on the set of feature points.Experimental results show that the proposed method can effectively improve the efficiency and accuracy of point cloud registration.For surface reconstruction,this paper mainly studies the Crust surface reconstruction algorithm based on Delaunay triangulation and the ratio of the radius of the circumscribed sphere of the tetrahedron and the shortest edge length is used for evaluating the obtained 3D mesh model quality.Unqualified Delaunay tetrahedron were screened out.Then the mesh model is optimized by inserting the center of circumscribed sphere and performing local Delaunay triangulation.The experimental results show that this method can reconstruct the mesh model of the target object completely.In order to meet the needs of practical application,the Loop algorithm is further used to subdivide the grid model.Eventually smooth 3D digital model is generated.
Keywords/Search Tags:3D reconstruction, point cloud registration, normal vector feature, surface reconstruction, Crust algorithm
PDF Full Text Request
Related items