## Algorithms for reverse engineering boundary representation models (2001)

Venue: | Computer-Aided Design |

Citations: | 53 - 10 self |

### BibTeX

@ARTICLE{Benkő01algorithmsfor,

author = {Pál Benkő and Ralph R. Martin and Tamás Várady},

title = {Algorithms for reverse engineering boundary representation models},

journal = {Computer-Aided Design},

year = {2001},

volume = {33},

pages = {839--851}

}

### Years of Citing Articles

### OpenURL

### Abstract

Aprocedure for reconstructing solid models of conventional engineering objects from a multiple-view, 3D point cloud is described. (Conventional means bounded by simple analytical surfaces, swept surfaces and blends.) Emphasis is put on producing accurate and topologically consistent boundary representation models, ready to be used in computer aided design and manufacture. The basic phases of our approach to reverse engineering are summarised, and related computational difficulties are analysed. Four key algorithmic components are presented in more detail: efficiently segmenting point data into regions; creating translational and rotational surfaces with smooth, constrained profiles; creating the topology of B-rep models; and finally adding blends. The application of these algorithms in an integrated system is illustrated by means of various examples, including a well-known reverse engineering benchmark. 1.

### Citations

469 | Three-dimensional alpha shapes
- Edelsbrunner, Mücke
- 1994
(Show Context)
Citation Context ...es and disjoint parts. There are various techniques to build a consistent triangulation, which differ in their basic assumptions, their generality and their computational efficiency — see for example =-=[2, 14, 23]-=-. Methods based on variants of «-shapes are popular but require careful implementation and the proper selection of parameters to achieve successful results [18]. As an alternative, we use an efficient... |

243 |
The Representation, Recognition, and Locating of 3D Objects
- Faugeras, Hebert
- 1986
(Show Context)
Citation Context ...tiple faces) of the B-rep model. For each face a single surface is fitted to the related points within a given tolerance. 4sVarious methods for fitting simple analytic surfaces have been described in =-=[8, 15, 16, 19, 32, 40]-=-. We believe that the use of geometric, or geometrically faithful, distances is more satisfactory for surface reconstruction than algebraic distances. In general, segmentation is a complex process [38... |

208 | Survey of polygonal surface simplification algorithms
- Heckbert, Garland
- 1997
(Show Context)
Citation Context ...carrier surface. Decimation needs to be an adaptive process taking into account the local surface geometry; it is a widely studied problem to control the amount of data displayed in computer graphics =-=[20]-=-. The carrier surface reflects the original topology of the point cloud; each data point is either a vertex of the triangulation or associated with the closest triangle of the carrier. The carrier sur... |

207 |
Segmentation Through Variable-Order Surface Fitting
- Besl
(Show Context)
Citation Context ...is illustrated by means of various examples, including a well-known reverse engineering benchmark. 1. Introduction Reconstructing objects has been a central problem in computer vision for a long time =-=[7, 8, 22]-=-. Technological advances in laser scanning now make it possible to produce multiple 3D point clouds with high density and high accuracy, and in this way it has become a realistic expectation to genera... |

145 | Automatic reconstruction of b-spline surfaces of arbitrary topological type
- Eck, Hoppe
- 1996
(Show Context)
Citation Context ...AD/CAM systems must use standard representations for interoperability. Approaches exist which use standard free-form surfaces, such as B-splines, to reconstruct conventional engineering objects — see =-=[13, 17, 26, 37]-=-. Two difficulties arise. If multiple surfaces are used, it is hard to find good segmenting curves, and to assure � or higher continuity between the adjacent surfaces. On the other hand, fitting a sin... |

135 |
Towards a General Multi-view Registration Techniques
- Bergevin, Soucy, et al.
- 1996
(Show Context)
Citation Context ...ations need to be found by means of which the different data sets can be represented in a common coordinate system. This is often called registration, which is an important problem in computer vision =-=[5, 12]-=-. Iterative closest point methods match nearest points in several datasets. Such methods tend to be slow, and require relatively good initial estimates of the registration. Some progress has been made... |

111 |
On Three-Dimensional Surface Reconstruction Methods
- Bolle, Vemuri
- 1991
(Show Context)
Citation Context ...is illustrated by means of various examples, including a well-known reverse engineering benchmark. 1. Introduction Reconstructing objects has been a central problem in computer vision for a long time =-=[7, 8, 22]-=-. Technological advances in laser scanning now make it possible to produce multiple 3D point clouds with high density and high accuracy, and in this way it has become a realistic expectation to genera... |

54 | Registration and integration of multiple object views for 3d model construction
- Dorai, Wang, et al.
- 1998
(Show Context)
Citation Context ...ations need to be found by means of which the different data sets can be represented in a common coordinate system. This is often called registration, which is an important problem in computer vision =-=[5, 12]-=-. Iterative closest point methods match nearest points in several datasets. Such methods tend to be slow, and require relatively good initial estimates of the registration. Some progress has been made... |

33 | High-level CAD model acquisition from range images
- Fitzgibbon, Eggert, et al.
- 1997
(Show Context)
Citation Context ...tage or as a postprocess. In comparison, the majority of previous reverse engineering approaches only considered single view range data, and only concerned itself with simple surfaces—see for example =-=[10, 16, 21, 34]-=-. In the remainder of the paper, in Section 2, we outline our automatic reverse engineering approach for conventional objects, surveying relevant ideas and identifying the successive computational ste... |

26 | Interpolating and approximating scattered 3d data with hierarchical tensor product splines. Surface Fitting and Multiresolution Methods
- Greiner, Hormann
- 1996
(Show Context)
Citation Context ...AD/CAM systems must use standard representations for interoperability. Approaches exist which use standard free-form surfaces, such as B-splines, to reconstruct conventional engineering objects — see =-=[13, 17, 26, 37]-=-. Two difficulties arise. If multiple surfaces are used, it is hard to find good segmenting curves, and to assure � or higher continuity between the adjacent surfaces. On the other hand, fitting a sin... |

24 |
Polyhedral metrics in surface reconstruction: tight triangulation[R
- Alboul, Damme
- 1995
(Show Context)
Citation Context ...ater. There are other alternatives to identify highly curved parts using the triangulation; for example, the angular variation of normals was suggested in [26], or curvature estimates can be computed =-=[1]-=-. After choosing an appropriate threshold value, planarity filtering will remove those triangles which are classified as belonging to highly curved regions of the point cloud. For example, in Figure 2... |

23 | On Editability of Feature-based Design
- Chen, Hoffmann
- 1996
(Show Context)
Citation Context ... perpendicular, the points of the two related surfaces can be refitted under a perpendicularity constraint, which will lead to a better model. There are many papers on constrained geometric modelling =-=[9]-=-, but its relevance to reverse engineering has been recognised only 5srecently [4, 41]. Having concluded the overview of the reconstruction process, we continue by providing more details of particular... |

13 |
Schikore D. Automatic Reconstruction of 3D Cad Models
- Bajaj, Bernardini, et al.
- 1996
(Show Context)
Citation Context ...oximations suitable for graphics (see e.g. Hoppe et al. [23]), while others represent the boundary of objects using special surface representations such as A-patches or superquadrics (see for example =-=[2, 31]-=-). Such representations are of limited use as commercial CAD/CAM systems must use standard representations for interoperability. Approaches exist which use standard free-form surfaces, such as B-splin... |

12 |
A New Method For Segmenting 3-D Scenes Into Primitives
- Hebert, Ponce
(Show Context)
Citation Context ...tiple faces) of the B-rep model. For each face a single surface is fitted to the related points within a given tolerance. 4sVarious methods for fitting simple analytic surfaces have been described in =-=[8, 15, 16, 19, 32, 40]-=-. We believe that the use of geometric, or geometrically faithful, distances is more satisfactory for surface reconstruction than algebraic distances. In general, segmentation is a complex process [38... |

9 |
Solid-model generation from measured point data,” Computer-Aided Design
- Chivate, Jablokow
- 1993
(Show Context)
Citation Context ...tage or as a postprocess. In comparison, the majority of previous reverse engineering approaches only considered single view range data, and only concerned itself with simple surfaces—see for example =-=[10, 16, 21, 34]-=-. In the remainder of the paper, in Section 2, we outline our automatic reverse engineering approach for conventional objects, surveying relevant ideas and identifying the successive computational ste... |

6 |
Surface reconstruction using alpha shapes
- Guo, Menoo, et al.
- 1997
(Show Context)
Citation Context ...nal efficiency — see for example [2, 14, 23]. Methods based on variants of «-shapes are popular but require careful implementation and the proper selection of parameters to achieve successful results =-=[18]-=-. As an alternative, we use an efficient method suggested by Kós [28], in which locally created Delaunay triangulations are merged in a consistent manner. For efficiency reasons, the original triangul... |

2 |
Algorithms to create B-rep models from measured point clouds
- Benko, Lukacs, et al.
- 1998
(Show Context)
Citation Context ...sistent and accurate to be able to combine them using Boolean operations to produce a final model. The pros and cons of point cloud merging and combining single-view models are discussed in detail in =-=[3]-=-. Even a single point cloud may contain at least � and perhaps as many as � data points. In order to determine the surface of an object, first the neighbourhoods and connectivity of the data points mu... |

1 |
Constrained fitting in reverse engineering”, submitted to Computer Aided Geometric Design
- Benkő, Andor, et al.
- 2000
(Show Context)
Citation Context ...pendicularity constraint, which will lead to a better model. There are many papers on constrained geometric modelling [9], but its relevance to reverse engineering has been recognised only 5srecently =-=[4, 41]-=-. Having concluded the overview of the reconstruction process, we continue by providing more details of particular algorithmic phases within the above general framework. 3. Direct segmentation In this... |

1 |
3D Modeling using the ACIS Kernel and Toolkit
- Corney
(Show Context)
Citation Context ...ner system (3D Scanners, London, UK) is used for data acquisition. The program is written in C++ and uses the VTK graphics library; it runs under LINUX and Windows NT. The ACIS solid modelling kernel =-=[11]-=- is used for solid model creation and blending. The original simulated data set for our benchmark test object shown in Figure 1, contained 18970 points. The size of the bounding box was ¢ � ¢ � . The ... |