Method for refining geometric description models using images

• US 6,963,338 B1
• Filed: 07/31/1998
• Issued: 11/08/2005
• Est. Priority Date: 07/31/1998
• Status: Expired due to Term

• Current Assignee: Cognex Corporation
• Original Assignee: Cognex Corporation
• Inventors: Bachelder, Ivan A., Marrion, Cyril C. Jr.
• Primary Examiner(s): Nguyen, Phu K.
• Application Number:
  US09/127,676
• Time in Patent Office: 2,657 Days
• Field of Search: 345/433, 345/418, 345/419, 345/420, 382/151, 382/154, 382/219, 382/236
• US Class Current: 345/420
• CPC Class Codes: G06F 30/00 Computer-aided design [CAD]
• CPC Class Codes: G06T 2207/30141 Printed circuit board [PCB]
• CPC Class Codes: G06T 7/001 using an image reference approach
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06V 20/64 Three-dimensional objects

Abstract

The creation of accurate geometric description-based models of objects in a machine vision system can be performed in a computer-assisted process. A rough model of an image of an object forms the basis of the model. The model is refined through the application of machine vision tools and techniques so as to provide a model including a geometric description. The model can be used in machine vision applications that may be used, for example, to compare images of objects under inspection to the model.

Claims

1. A method for refining a geometric description model of an object using an image of that object, the method comprising:

providing an image of the object;

providing a rough geometric description model of the object;

extracting sub-models from the rough geometric description model;

finding poses of the sub-models in the image;

using the found sub-models to construct a refined geometric description model;

using constraints to further refine the refined geometric description model; and

refitting the further refined geometric description model to the found sub-model poses to provide a new pose of the further refined geometric description model.

2. The method of claim 1, wherein one of the constraints is an angular constraint.

3. The method of claim 1, wherein the model includes a wire frame structure.

4. The method of claim 3, wherein the wire frame structure includes vertices and segments connected thereto.

5. The method of claim 1, wherein the refined model for the object is required to satisfy certain user-defined geometric constraints.

6. The method of claim 5, wherein the geometric constraints specify that segments must form 90 degree angles at vertices.

7. The method of claim 1, wherein finding poses of the sub-models in the image includes:

validating the poses of the sub-models.

8. The method of claim 1, wherein providing a rough model of the object includes also providing a rough model pose.

9. A method for refining a geometric description model of an object using an image of that object, the method comprising:

providing an image of the object;

providing a rough geometric description model of the object;

constructing a refined geometric description model using the image;

using constraints to further refine the refined geometric description model; and

refitting the further refined geometric description model to the found sub-model poses to provide a new pose of the further refined model.