Direct segmentation of smooth, multiple point regions

The purpose of reverse engineering is to convert a large point cloud into an accurate, fair and consistent CAD model. For a class of conventional engineering objects we have the 'a priori' assumption that the object is bounded exclusively by simple, analytic surfaces. In this case it is possible to generate the model with a minimal amount of user interaction. The key issue is segmentation, i.e., to separate the point cloud into smaller regions, where each can be approximated by a single surface. While this is relatively simple, where the regions are bounded by sharp edges, problems arise when smoothly connected regions need to be separated. The direct segmentation method described in this paper is based on a special sequence of tests, by means of which a large point cloud can be robustly splitted into smaller subregions until no further subdivision is possible. Surfaces of linear extrusion and revolution are also detected. The structure of the smooth, multiple regions is the basis of constrained surface fitting in the final model building phase.

[1]  Paul J. Besl,et al.  Surfaces in Range Image Understanding , 1988, Springer Series in Perception Engineering.

[2]  Ramesh C. Jain,et al.  Segmentation through Variable-Order Surface Fitting , 1988, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  Paul J. Besl,et al.  Direct construction of polynomial surfaces from dense range images through region growing , 1995, TOGS.

[4]  Dmitry B. Goldgof,et al.  Extracting a Valid Boundary Representation from a Segmented Range Image , 1995, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  Andrew W. Fitzgibbon,et al.  High-level model acquisition from range images , 1997, Comput. Aided Des..

[6]  Ralph R. Martin,et al.  Reverse engineering of geometric models - an introduction , 1997, Comput. Aided Des..

[7]  Franc Solina,et al.  Superquadrics for Segmenting and Modeling Range Data , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Gábor Renner,et al.  Segmentation and Surface Fitting in Reverse Engineering , 1998, SSM.

[9]  Josef Hoschek,et al.  A geometric concept of reverse engineering of shape: approximation and feature lines , 1998 .

[10]  Robert B. Fisher,et al.  Object reconstruction by incorporating geometric constraints in reverse engineering , 1999, Comput. Aided Des..

[11]  Ralph R. Martin,et al.  Methods to recover constant radius rolling ball blends in reverse engineering , 1999, Comput. Aided Geom. Des..

[12]  Tamás Várady,et al.  Best Fit Translational and Rotational Surfaces for Reverse Engineering Shapes , 2000, IMA Conference on the Mathematics of Surfaces.

[13]  Ralph R. Martin,et al.  Algorithms for reverse engineering boundary representation models , 2001, Comput. Aided Des..

[14]  Ralph R. Martin,et al.  Constrained fitting in reverse engineering , 2002, Comput. Aided Geom. Des..