三角網格分割與曲面重建於汽車鈑金件之應用

Application of Triangular Mesh Segmentation and Surface Reconstruction for Automobile Sheet Metal

曾凱威(Kai-Wei Tzeng)

指導教授 ： 尤春風

國立臺灣大學/工學院/機械工程學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.02300

摘要

在汽車鈑金件的逆向開發流程中，點資料的曲面重建是最為耗時的程序之一。若欲縮短開發所需的時程，勢必得朝著曲面重建的自動化來著手。在曲面重建的過程中，倘若直接以整筆點資料進行曲面嵌合，曲面將無法滿足點資料的誤差要求。因此，在進行曲面嵌合以前，必須依循原始鈑金件的外形特徵，將點資料予以分割。

為了提昇汽車鈑金件逆向開發的效率，本研究提出一套適用於汽車鈑金件掃描資料的網格分割法則，作為點資料進行曲面嵌合的前處理。整體流程以STL網格檔案為主要輸入格式，首先利用半邊結構建立網格資料的拓樸關聯，並以離散曲面的觀點計算網格點的微分幾何性質，接著藉由網格點主曲率進行特徵網格面的判別，再透過網格面區域成長的擴展與分群，達到網格分割的效果。最後將分割所得的網格塊予以相關的後處理，使分割結果能更符合原始外形特徵的趨勢。

關鍵字

汽車鈑金件 ； 逆向工程 ； 三角網格 ； 區域成長 ； 網格分割

並列摘要

Surface Reconstruction is one of the most time-consuming process during the development process for automobile sheet metal in reverse engineering. In order to decrease the time cost of development process, surface reconstruction from point sets should be proceeded automatically. Taking the residual error of resulting surface into account, surface fitting from the whole point sets will be inadequate to satisfy the required accuracy; consequently, as a preprocessor, mesh segmentation should be proceeded before surface fitting to improve the residual error of resulting surface.

In this paper, a triangular mesh segmentation method adapted to automobile sheet metal is provided as the preprocessor of surface fitting. The proposed segmentation scheme takes STL mesh file as input data and is composed of 4 successive processes：creation of topological relations among mesh primitive elements based on half-edge data structure, approximation for differential properties of mesh vertex, feature property verification of mesh facets and clusters of mesh facets via region growing. By proceeding the processes above and some related post-processes, the original mesh data will be divided into several mesh patches according to the feature shapes and curvature variations.

並列關鍵字

automobile sheet metal ； reverse engineering ； triangular mesh ； region growing ； mesh segmentation

參考文獻

[1] Cohen-Steiner, D. and Morvan, J.-M., “Restricted delaunay triangulations and normal cycle”, 19th Annual ACM Symposium on Computational Geometry, pp.237-246, 2003.

[2] Guo, Y. W., Peng, Q. S., Hu, G. F. and Wang, J., “Smooth feature line detection for meshes”, The Journal of Zhejiang University SCIENCE, Vol.6A, No.5, pp.460-468, 2005.

[3] Hubeli, A. and Gross, M., “Multiresolution feature extraction for unstructured meshes”, Proceedings of the conference on Visualization 2001, pp.287-294, 2001.

[4] Huang, J. and Menq, C. H., “Automatic data segmentation for geometric feature extraction from unorganized 3D coordinate points”, IEEE Transactions on Robotics and Automation, Vol.17, No.3, 2001.

[5] Huang, J. and Menq, C. H., “Combinatorial manifold mesh reconstruction and optimization from unorganized points with arbitrary topology”, Computer-Aided Design, Vol.34, No.2, pp.149-165, 2002.