Research On Optimal Method For Complex Contour Fitting Based On Gaussian Imaging

As the need of the market grows up, the appearance of the taillight area becomes more and more three-dimensional and complex. The area of Auto-body taillight is a signature for a sedan. It is assembled at the end of assembly process, which concerns with assembly variation and final appearance of the sedan. This thesis combined the traditional fitting method and Gaussian Imaging Method, along with finite elements analysis of the taillight assembly structure, to analyze the fitting problems of complex contour. A fitting prototype system was constructed to optimize the taillight area of a certain type of sedan and confirmed the effectiveness and efficiency of the proposed method. The main contents are as follows:(1) The extraction of complex contour boundaryThe boundary of complex contour was defined to meet the need of the extraction from the point clouds. The extraction method was revised according to the taillight area’s fitting, while keeping the surface information for Gaussian Imaging Method. It matches the need of complex contour’s fitting and provides the basis for three-dimensional complex contour fitting optimization.(2) Gaussian Imaging Method based fitting optimizationOn one hand, this thesis introduced the Gaussian Imaging Method to the traditional fitting method, and separated the fitting process into two main procedures, attitude fitting and location fitting. In the first procedure, attitude fitting was based on Gaussian Imaging Method. And in the second procedure, location fitting was based on traditional Hausdorff fitting method. In 2-D and 3-D case study, Gaussian Imaging Method based mixture model was proved to be effective and efficient.On the other hand, weighted Gaussian Imaging Method was directly used to solve the 3-D real complex contour fitting problem. Surface information extracted from point clouds was used as a reference to evaluate the attitude fitting procedure, combined with contour boundary coordinates. The location fitting procedure was optimized by the definition of Gap and Flushness. The weighted Gaussian Imaging Method formed the basis of the evaluation of the complex contour fitting.(3) Finite element analysis of assembly structureConsidering of the influence of assembly structure deformation, a FEM(Finite element analysis) model was constructed to analyze the taillight area. It concerned with metal taillight bracket, rubber sealer and taillight body in the deformation process. Compared with in-situ experiments and assembly process, the model was proved to be effective. The results indicated that the assembly sequences of had slightly influence on assembly quality, but the deformation of the assembly structure could be complied to the improvement of the fitting quality.(4) Development of fitting prototype systemAccording to the characteristics of the taillight assembly and the proposed fitting method, a fitting prototype system was established for auto body taillight fitting optimization. The system integrated scanning process, boundary extraction and fitting optimization. It was a systematic solution to the improvement of taillight fitting. A case study of a real taillight was conducted within the prototype system to confirm the proposed method’s feasibility.