Research On Delaunay-based Surface Reconstruction Algorithms From 3D Unorganized Point Clouds

With the development of Data Acquisition and Computer Science, the Reverse Engineering (RE) becomes a rising technology for product design. Compared with traditional CAD (Computer Aided Design) systems which built geometry models from points, curves, faces or features,the Reverse Engineering minimizes the highly dependency of designers during the design procedure, benefits a lot to rapid product developing and Rapid Prototype Manufacturing(RPM). It also can be used in various fields. Surface reconstruction which plays an important role in RE usually acquires coordinate data from discrete points (point clouds) which lie on the surfaces of the geometry entity and reconstructs surfaces by some algorithms to express the geometry entity. The reconstructed surface is an interpolation or approximation of the discrete point clouds. Furthermore, the geometrical and topological correctness must be satisfied.This thesis was studied with theories and methods of surface reconstruction. Firstly, we introduced Delaunay Triangulation and developed a two dimensional Delaunay triangulation program which can deal with large amount of data. We also achieved an efficient and robust three dimensional Delaunay Triangulation by using CGAL(Computational Geometry Algorithms Library) which is supported by the Europe Union. Moreover, we proposed a region-growing surface reconstruction algorithm. The filtering of the candidate triangles was improved a lot by constructing an influence region during the incrementally region-growing extraction of the surfaces. The construction of the Influence Region also contributes to make sure that the surfaces are geometrically and topologically correct. The proposed algorithm is capable of handling with kinds of point clouds data, such as three dimensional unorganized point clouds, point clouds acquired from organized CAD models or point clouds acquired from finite element analysis meshes. Finally, the reconstructed surfaces were used in computer aided visualization of electrical design of airborne radome. On contrary to the traditional way of computing the angles of incidence by solving the problem that where the ray of electromagnetic waves intersected with the surface of radome, our method can express the distribution of the value of the angle of incidence on airborne radome succinctly and efficiently.