Research On Constraint And Feature Based Structural Parts Solid Model Reconstruction

In order to reconstruct the solid model for structural parts and to improve the qualityof reconstruction, two different ways of constraint and feature based solid modelreconstruction are studied in this paper. One is based on constrained surface fitting, andthe other is feature based solid model reconstruction method. In order to implement theabove mentioned two ways,　 some critical technologies have been researchedsystematically and deeply, including data capture and preprocessing, planar point datasegmentation and constraint-based sketch profile generation, simultaneously surfacefitting of several surfaces to point data with constraints, boundary representation solidmodel generation and the methods of feature reconstruction. Some new algorithms andmethods have been introduced and the examples demonstrate the flexibility andefficiency of these methods. The main subjects and achievements of this paper aresummarized as following: Several kinds of data capture methods are introduced and some measurement rulesare given for the data acquisition of structural parts. Some experimental results are alsogiven to demonstrate the effect of such principles. Some data preprocessing methods for structural parts point data are discussed andimplemented in this paper, including point data space region partition, normal vector andcurvature estimation and data decimation. Afterwards, an improved ICP algorithm ispresented for multi-view registration and integration of different views. Also anintent-based segmentation method is given to obtain segmented point data which isconducive to modeling. Sketch profile generation method based on planar point data is studied and twodifferent methods of segmentation and type identification are given to find segmentedplanar point data, and a new method on how to reject redundant and contradictoryconstraints between curves has been researched and applied for simultaneous constrainedfitting in order to get a constraint-satisfied sketch profile. The method of simultaneously surface fitting of several surfaces to point data withconstraints is studied for B-rep solid model reconstruction. A surface fitting andautomatic type identification method is studied and a method is given to calculate quadricsurface parameters. Constrained fitting method is implemented to create an enclosedsurface model and then a B-rep solid model. Several different design-intent based feature recreation methods are presented, suchas extrude feature, revolve feature and constant radius blending feature and so on. Amongwhich the parameter extraction for extrude and revolve feature is achieved by the solvingof corresponding constraint optimization problem. Finally, the feature models of twostructural parts have been reconstructed to demonstrate the feasibility of such methods.