Image analysis – Image transformation or preprocessing – Changing the image coordinates
Reexamination Certificate
2001-03-16
2004-10-12
Patel, Kanji (Department: 2625)
Image analysis
Image transformation or preprocessing
Changing the image coordinates
C382S159000, C345S651000
Reexamination Certificate
active
06804416
ABSTRACT:
RESERVATION OF COPYRIGHT
This patent document contains information subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent, as it appears in the U.S. Patent and Trademark Office files or records, but otherwise reserves all copyright rights whatsoever.
BACKGROUND
1. Field of the Invention
The present invention, in certain respects, relates to the field of machine vision. In other respects, the present invention relates to a method and system that aligns a geometric object model with an image.
2. Description of Background Information
Various machine vision systems perform alignment as a first step for various tasks. For example, a machine vision inspection system will identify the correct alignment of an object before the object appearing in an image is inspected. Another example is that a robot aligns its field of view before it decides the direction in which to proceed and identifies those what obstacles to avoid. Such alignment often requires training based on an image of the object to be recognized, together with, for instance, a specification of the origin of the object as well as the dimension information of the object. While such image based training for alignment can be effective, there are certain situations where it is impractical.
Training an alignment tool based on object images is tedious and time consuming. This becomes especially a problem for manufacturing processes, where there may be a wide variety of products or objects that need to be inspected using machine vision inspection. Furthermore, product designs may frequently change. Even a minor revision to an object, for example, its shape, may require retraining.
Parameterized geometric object models may be used, instead of using object images, to train machine vision inspection systems. For purposes of the disclosure herein, a geometric model of an object comprises parameter representations of the geometry of the object. For example, parameterized geometric models are created and employed in machine vision inspection systems when aligning fiducial marks on electronic components. For a simple object, its geometric model may be created by manually entering the shape and dimension information about the object. For more complicated objects, however, creating a geometric object model may be even more difficult than training using images. That is, the complexity associated with creating a geometric model for a complicated object may outweigh the advantage of using a geometric object model for training an alignment tool.
SUMMARY OF THE INVENTION
There is a need to make direct use of available geometric object models for alignment purposes in machine vision systems so that the training process can be more efficient and such trained alignment tools can effectively align a geometric object model with an image.
The present invention is provided to improve upon techniques for obtaining a geometric object model used in a machine vision application and for performing alignment with such models. Improved methods and features are presented that address the need identified above and provide methods and apparatuses for aligning an object model with an image. Aspects of the present invention can improve both the efficiency of a training phase as well as the effectiveness of an on-line alignment phase. Existing geometric object models, such as CAD models, are utilized directly but revised with respect to the needs for alignment. A revised geometric object model is constructed especially for alignment purpose. It retains only the geometric features, selected from the original geometric object model, that are considered salient and useful with respect to the underlying task of aligning a geometric object model with an image. An alignment tool can be trained based on a revised geometric object model. Both a revised geometric object model and its correspondingly trained alignment tool may be further refined using the feedback information from an on-line alignment operation so that the revised model and the corresponding alignment tool can be adjusted to better fit the real scenarios of different machine vision applications.
REFERENCES:
patent: 4980971 (1991-01-01), Bartschat et al.
patent: 5060276 (1991-10-01), Morris et al.
patent: 5111516 (1992-05-01), Nakano et al.
patent: 5113565 (1992-05-01), Cipolla et al.
patent: 5226095 (1993-07-01), Okumura et al.
patent: 5268999 (1993-12-01), Yokoyama
patent: 5343028 (1994-08-01), Figarella et al.
patent: 5371690 (1994-12-01), Engel et al.
patent: 5471541 (1995-11-01), Burtnyk et al.
patent: 5495537 (1996-02-01), Bedrosian et al.
patent: 5497451 (1996-03-01), Holmes
patent: 5500906 (1996-03-01), Picard et al.
patent: 5545887 (1996-08-01), Smith et al.
patent: 5602937 (1997-02-01), Bedrosian et al.
patent: 5621807 (1997-04-01), Eibert et al.
patent: 5625715 (1997-04-01), Trew et al.
patent: 5627912 (1997-05-01), Matsumoto
patent: 5627915 (1997-05-01), Rosser et al.
patent: 5663809 (1997-09-01), Miyaza et al.
patent: 5828769 (1998-10-01), Burns
patent: 5850469 (1998-12-01), Martin et al.
patent: 5974169 (1999-10-01), Bachelder
patent: 6078700 (2000-06-01), Sarachik
patent: 6151406 (2000-11-01), Chang et al.
patent: 6324298 (2001-11-01), O'Dell et al.
patent: 6421458 (2002-07-01), Michael et al.
patent: 6462751 (2002-10-01), Felser et al.
Cognex Corporation, Cognex 3000/4000/5000 Vision Tool, Revision 7.6, Chapter 4, Caliper Tool, 1996.
Cognex Corporation, Cognex 3000/4000/5000 Vision Tool, Revision 7.6, Chapter 5, Inspection, 1996.
Cognex Corporation, Cognex 4000/5000 SMD Placement Guidance Package, User's Manual Release 3.8.00, Chapter 7, Rectilinear Device Inspection, 1998.
Cognex Corporation, Cognex 4000/5000 SMD Placement Guidance Package, User's Manual Release 3.8.00, Chapter 8, Large-Leaded Device Inspection, 1998.
Cognex Corporation, Cognex 4000/5000 SMD Placement Guidance Package, User's Manual Release 3.8.00, Chapter 15, MFOV-LL Device Inspection, 1998.
Cognex Corporation, Cognex 3000/4400 SMD Tools Release 5.2, SMD 2, 1994.
Timothy S. Newman, Anil K. Jain and H.R. Keshavan, 3D CAD-Based Inspection I: Coarse Verification, pp. 49-52, IEEE, 1992.
Anthony Hoogs and Ruzena Bajcsy, Model-based Learning of Segmentations, pp. 494-499, IEEE, 1996.
Shimon Ullman, Aligning pictorial descriptions: An approach to object recognition, Cognition, vol. 32, No. 3, pp. 193-254, Aug. 1989.
Cognex Corporation, Cognex MVS-8000 Series, GDE User's Guide, Revision 1.1, Apr. 7, 2000.
Andre Fischer, Thomas Kolbe and Felicitas Lang,On the Use of Geometric and Semantic Models for Component-Based Building Reconstruction: Institue for Photogrammetry, University of Bonn, pp. 101-119, 1999.
George Vosselman,Interactive Aligment of Parameterised Object Models to Images: Delft University of Technology, Faculty of Civil Engineering and Geosciences, 1998.
D.W. Jacobs,The Use of Grouping in Visual Object Recognition, MIT Artifical Intelligence Laboratory, Office of Naval Research, pp. 1-162, Oct. 1988.
Bachelder Ivan
Jacobson Lowell
Negro Jay
Segal Leonid
Cognex Corporation
Patel Kanji
Weinzimmer Russ
LandOfFree
Method and system for aligning geometric object models with... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for aligning geometric object models with..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for aligning geometric object models with... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3329706