Research On Surface Reconstructing Technology Of Point Cloud Data

Driven by the application of industrial、military、medical、virtual reality、reverse engineering etc, and with the development of3D scanning technology and the processing of computer graphics, research on surface reconstructing technology of point cloud data has become a hotspot in computer graphics. This paper studied the relevant content of this subject, researched and analyzed some key issues in this field.1. Analyze the type and visualization technology of point cloud data, and use Java3D to achieve the visualization of3D space, some interactive operations included; Study the methods of establishing topology relationships in3D space, describe the principle and evaluate the performance of related algorithms; Summarize the theory of triangulation, and analyze the advantages and disadvantages of some classical algorithms. Test the Crust algorithm using matlab, point out the advantages and the defects though the test data.2. Propose a fast mesh reconstruction algorithm of point cloud data based on incremental algorithm. There are two contributions:one is import the dynamic ball strategy into3D space segmentation, can locate the candidate points around the extended edge quickly. Another is design a function to evaluate vertices, based on the detailed analysis of Delaunay triangulation, this function can solve the not consistency between the criterion of minimum vector and the maximal-minimal angle. This algorithm derived from incremental computing, combined with the constraint criterion and the designed function, can reconstruct the surface in space from one base triangle.3. Construct simulation system. The result shows that it can generate triangular mesh gradually in3D space based on the morphological changes of surface, and can ensure the great smooth and better triangulation performance. Secondly, the algorithm uses an adaptive dynamic ball strategy, combined with the segmentation in space to partition point cloud data, can reduce the scale of the candidate points, and improve the efficiency of data processing. The algorithm is proved to be highly and efficient.Finally, summarize the related work in this thesis, and briefly introduced the main points in future work.