Research On B-Spline Surface Segmentation Fitting Technique For Complicated Section Curve Data

Tomography scanning data is an important data source in the reverse engineering.The surfacereconstruction from section curve in the tomographic scanning images is an important way of3Dreconstruction.With the development of reverse engineering applied in biomedicine field,surfacereconstruction from section curve data has become one of the fundamental technology in reverseengineering.This thesis deeply studied into the surface reconstruction from complicated section curvedata from the angle of quadrangle.The main content and creative achievement of this thesis isdescribed as follows:Firstly,the complicated section curve data processing technique is studied,compared threemethods of resampling,method of accumulating chord length,the method of curve fitting and methodof searching by public slope.An algorithm based on angle segmentation for construct optimalmatching points is proposed.It can be done while re-sampling section curve data.Secondly,the topology reconstruction technique from section curve data is studied,and all thepossible results of the complicated section curve data is analyzed.An algorithm to construct groupedcriterion based on the complicated section curve data has been proposed by using the followingideas:from global to local and from preliminary classification to refinement classification.Finally,we have studied the surface reconstruction technique from section curve data,andreconstructed surface from segmental section curve data by using B-spline surface skinning.On thisbasis, this thesis further studied the bifurcation of port’s section curve data fitting techniques. analgorithm is presented to construct branch surfaces by trimming surrounding surface and filling thehole by a Coons patchwork,it can finish blending surface from B-spline surface skinning.These algorithms in this thesis have been programmed,and a platform for surface reconstructionof the complicated section curve data has been established.Many numerical tests show that algorithmsin this thesis are effective and stable.