Technology Research And System Development Of Data Processing In Reverse Engineering Based On UGII

Since there are some shortcomings in reverse engineering systems and UGII, part of reverse engineeringsystem based on UGII is researched and developed with Visual Studio.NET 2003. The achievements of thispaper are as follows:(1) With the research of the interface technique about UG/Open API and the UDO method, the pointcloud class is designed, the point cloud object, the UDO class and object are created, and a large number ofpoints are displayed in the main window of UGII as pixel points with the pure virtual functions of the C++.(2)Through obtaining the main window pointer of UGII by functions of windows API and UG/OPEN CAPI, capturing the mouse coordinates in the screen by setting a mouse hook, and projecting the points of theworld coordinate system by calling OpenGL API functions, a point cloud is partitioned into two regions inUGII based on an algorithm of point in-out polygon test.(3) With the research of the spatial data structure called R-tree which supports dynamic data-managementand the selection of its variation — R*-tree as the spacial data structure of this project, a searching algorithmof K nearest neighbor points in R*-tree is proposed.(4) Through researching the algorithms of fitting fixed-sphere, transforming coordinates and dataintegration, the algorithm of registration and integration based on least-square sphere is proposed andintegrated with UGII, which can registrate and integrate multi-view point cloud data rapidly and accurately.(5) Through researching the algorithm of curvature estimate and the principle of data reduction, the datareduction algorithm based on curvature estimation is proposed and integrated with UGII, which can beapplied to any kind of measured data and reduces the mass and dense points directly and effectively.(6) Through analyzing several previous algorithms of triangulation for scattered data, and establishing theeffective criterion for candidate points, the search and optimizing criterions for match points and fourmethods of data updating, a universal and fast mesh generation algorithm named helix-wave expandingmethod is proposed, which can generate high-quality triangular mesh quickly.(7) Four algorithms of mesh postprocessing are realized, including inside and outside boundary extraction,data segment, adaptive mesh refinement and mesh model hollowing, which establish the groundwork forfeature line construction, finite element analysis and rapid prototyping.