Method for fast, robust, multi-dimensional pattern recognition

• US 8,363,956 B1
• Filed: 12/30/2004
• Issued: 01/29/2013
• Est. Priority Date: 07/13/1998
• Status: Expired due to Fees

• Current Assignee: Cognex Technology & Investment Corporation (Cognex Corporation)
• Original Assignee: Cognex Corporation
• Inventors: Silver, William M., McGarry, E. John, Hill, Matthew L., Foster, Nigel J., Nichani, Sanjay, Foster, Willard P., Wagman, Adam
• Primary Examiner(s): Koziol, Stephen R
• Application Number:
  US11/028,008
• Time in Patent Office: 2,952 Days
• Field of Search: 382/118, 382/141, 382/149, 382/181, 382/190, 382/197, 382/209, 382/276, 348/92, 348/125
• US Class Current: 382/209
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06V 10/469 Contour-based spatial representations, e.g. vector-coding
• CPC Class Codes: G06V 10/752 Contour matching

Abstract

Disclosed is a method for determining the absence or presence of one or more instances of a predetermined pattern in an image, and for determining the location of each found instance within a multidimensional space. A model represents the pattern to be found, the model including a plurality of probes. Each probe represents a relative position at which a test is performed in an image at a given pose, each such test contributing evidence that the pattern exists at the pose. The method further includes a comparison of the model with a run-time image at each of a plurality of poses. A match score is computed at each pose to provide a match score surface. Then, the match score is compared with an accept threshold, and used to provide the location any instances of the pattern in the image.

Claims

1. An apparatus for inspecting a found instance of a pattern in an image, the apparatus comprising:

a training module computer program, on a data processing device, configured to receive a training image, and provide a model pattern, the model pattern including a geometric description of the expected shape of the object, the geometric description including a plurality of probes, each probe having a position with respect to the model pattern at which at least one test is to be performed upon an image at a found pose of the model pattern, each such test contributing evidence for inspection; and

a run-time module computer program, on a data processing device, configured to receive;

the model pattern,

a list of generalized degrees-of-freedom (DOFs), and

a run-time image,

the run-time module computer program is further configured to provide a list of results, each result of the list of results containing a match score and individual probe rating information for inspection purposes.

2. The apparatus of claim 1, wherein some of the probes of the plurality of probes have a position with respect to the model that is along a boundary of the model.

3. The apparatus of claim 1, wherein some of the probes of the plurality of probes have a position with respect to the model that is along an imaginary line extending from a boundary of the model.

4. The apparatus of claim 1, wherein each result of the list of results contains a match score at the found pose of the model, the match score providing an overall measure of quality of the candidate instance of the pattern found in the run-time image.

5. The apparatus of claim 4, wherein the match score is obtained by computing a direction rating factor for each probe.

6. The apparatus of claim 4, wherein the match score is obtained by computing a magnitude rating factor for each probe.

7. The apparatus of claim 4, wherein the match score is obtained by computing a product of a magnitude rating factor and a direction rating factor for each probe.

8. The apparatus of claim 7, wherein the magnitude rating factor is computed such that it can assume a value of 1.0 for gradient magnitudes above a first value, a value of 0 for magnitudes below a second value, and proportionally between 0 and 1.0 for values between the first and second values.

9. The apparatus of claim 4, wherein each probe has a probe weight and a probe rating, and the match score is a weighted sum of all the probe ratings of the model divided by a total of all the weights of all the respective probes of the model.

10. The apparatus of claim 1, wherein the individual probe rating information provides detailed information about local differences between the model and the candidate instance of the pattern in the run-time image.

11. The apparatus of claim 1, wherein the individual probe rating information is computed by:

computing a direction rating factor for each probe; and

assigning a probe rating to each probe that is the product of the direction rating factor and the gradient magnitude under the probe.

12. The apparatus of claim 1, wherein the individual probe rating information is computed by:

computing a direction rating factor for each probe;

computing a magnitude rating factor that is 1.0 for gradient magnitudes above a first value, 0 for magnitudes below a second value, and proportionally between 0 and 1.0 for values between the first and second values; and

assigning a probe rating to each probe that is the product of the direction rating factor and the magnitude rating factor under the probe.

13. The apparatus of claim 12, wherein computing a direction rating factor includes:

computing gradient direction error for each probe;

assigning a direction rating factor of 1.0 for a probe that has a gradient direction error below a first value, assigning a direction rating factor of 0 for a probe that has a gradient direction error above a second value, and assigning a direction rating factor that is proportionally between 0 and 1.0 for a probe that has a gradient direction error that falls between the first and second values.

14. The apparatus of claim 13, wherein computing gradient direction error for each probe includes:

computing a first gradient direction of the model at the probe position, and computing a second gradient direction at a corresponding position in the run-time image, and subtracting the gradient directions so as to provide gradient direction error.

15. An apparatus for inspecting a found instance of a pattern in an image, the apparatus comprising:

a training module including a data processing device configured to receive a training image, and provide a model pattern, the model pattern including a geometric description of the expected shape of the object, the geometric description including a plurality of probes, each probe having a position with respect to the model pattern at which at least one test is to be performed upon an image at a found pose of the model pattern, each such test contributing evidence for inspection; and

a run-time module including a data processing device configured to receive;

the model pattern,

a list of generalized degrees-of-freedom (DOFs), and

a run-time image,

the data processing device included in the run-time module is further configured to provide a list of results, each result of the list of results containing a match score and individual probe rating information for inspection purposes.

16. The apparatus of claim 15 wherein the data processing device included in the training module and the data processing device in the run-time module are the same data processing device.