Method and apparatus for training a probe model based machine vision system

• US 8,705,851 B2
• Filed: 01/03/2013
• Issued: 04/22/2014
• Est. Priority Date: 10/10/2008
• Status: Active Grant

• Current Assignee: Cognex Corporation
• Original Assignee: Cognex Corporation
• Inventors: Barker, Simon, Wagman, Adam, Wallack, Aaron, Michael, David J
• Primary Examiner(s): Park, Edward
• Application Number:
  US13/733,685
• Publication Number:
  US 20130182948A1
• Time in Patent Office: 474 Days
• Field of Search: 382/155, 382/159, 382/161
• US Class Current: 382/159
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30208 Marker matrix
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06V 10/243 by compensating for image skew or non-uniform image deformations
• CPC Class Codes: G06V 30/1914 Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches

Abstract

A method for training a pattern recognition algorithm including the steps of identifying the known location of the pattern that includes repeating elements within a fine resolution image, using the fine resolution image to train a model associated with the fine image, using the model to examine the fine image resolution image to generate a score space, examining the score space to identify a repeating pattern frequency, using a coarse image that is coarser than the finest image resolution image to train a model associated with the coarse image, using the model associated with the coarse image to examine the coarse image thereby generating a location error, comparing the location error to the repeating pattern frequency and determining if the coarse image resolution is suitable for locating the pattern within a fraction of one pitch of the repeating elements.

Claims

1. A method for training a pattern recognition algorithm for a machine vision system that uses models of a pattern to be located, the method comprising the steps of:

(a) identifying the known location of the pattern that includes repeating elements within a fine image resolution image;

(b) using the fine image resolution image to train a model associated with the fine image;

(c) using the model associated with the fine image to examine the fine image resolution image and to generate a score space;

(d) examining the score space to identify a repeating pattern frequency;

(e) using a coarse image at a coarse image resolution that is coarser than the finest image resolution image to train a model associated with the coarse image;

(f) using the model associated with the coarse image to examine the coarse image thereby generating a location error;

(g) comparing the location error to the repeating pattern frequency; and

(h) based on the comparison, determining if the coarse image resolution is suitable for locating the pattern within a fraction of one pitch of the repeating elements.

2. The method of claim 1 further including the step of using the coarse image to generate a plurality of lower resolution images at different coarse subpixel phases and wherein steps (f) through (h) are performed for each of the lower resolution images.

3. The method of claim 1 further including the steps of (i) using the model associated with the coarse image to examine the coarse image thereby generating a score space associated with the coarse image and (j) examining the score space associated with the coarse image to identify a sparsest scan code that is robust at the coarse image resolution.

4. The method of claim 3 further including the step of using the coarse image to generate a plurality of lower resolution images at different coarse subpixel phases and wherein steps (i) and (j) are performed for each of the lower resolution images.

5. The method of claim 4 wherein the fine image resolution is a finest image resolution.

6. The method of claim 3 wherein each model includes a plurality of probes that are each a measure of similarity of at least one of a run-time image feature and a run time image region to at least one of a pattern feature and a pattern region, the step of using the model associated with the fine image to examine the fine image resolution image and to generate a score space including the steps of applying the plurality of probes in the model associated with the fine image at a plurality of poses to the fine image resulting in a plurality of score spaces at each pose where the score spaces include peaks, the peaks indicating likely locations of occurrences of the pattern in the fine image at each pose.

7. The method of claim 6 wherein the step of using the model associated with the coarse image to examine the coarse image resolution image and to generate a score space includes the steps of applying the plurality of probes in the model associated with the coarse image at a plurality of poses to the coarse image resulting in a plurality of score spaces at each pose where the score spaces include peaks, the peaks indicating likely locations of occurrences of the pattern in the coarse image at each pose.

8. The apparatus of claim 3 wherein the fine image resolution is a finest image resolution.

9. The method of claim 1 wherein the fine image resolution is a finest image resolution.

10. The method of claim 1 wherein each model includes a plurality of probes that are each a measure of similarity of at least one of a run-time image feature and a run time image region to at least one of a pattern feature and a pattern region, respectively, at a specific location.

11. An apparatus for training a pattern recognition algorithm for a machine vision system that uses models of a pattern to be located, the apparatus comprising:

a processor programmed to perform the steps of;

(a) identifying the known location of the pattern that includes repeating elements within a finest image resolution image;

(b) using the finest image resolution image to train a model;

(c) using the model to examine the finest image resolution image and to generate a score space;

(d) examining the score space to identify a repeating pattern frequency;

(e) using a coarse image at a coarse image resolution that is coarser than the finest image resolution image to train a model;

(f) using the model associated with the coarse image to examine the coarse image thereby generating a location error;

(g) comparing the location error to the repeating pattern frequency; and

(h) based on the comparison, determining if the coarse image resolution is suitable for locating the pattern within a fraction of one pitch of the repeating elements.

12. The apparatus of claim 11 wherein the processor is programmed to further perform the steps of using the coarse image to generate a plurality of lower resolution images at different coarse subpixel phases and wherein steps (f) through (h) are performed for each of the lower resolution images.

13. The apparatus of claim 11 wherein the processor is further programmed to perform the steps of (i) using the model associated with the coarse image to examine the coarse image thereby generating a score space associated with the coarse image and (j) examining the score space associated with the coarse image to identify a sparsest scan code that is robust at the coarse image resolution.

14. The apparatus of claim 13 wherein the processor is further programmed to perform the step of using the coarse image to generate a plurality of lower resolution images at different coarse subpixel phases and wherein steps (i) and (j) are performed for each of the lower resolution images.

15. The apparatus of claim 13 wherein each model includes a plurality of probes that are each a measure of similarity of at least one of a run-time image feature and a run time image region to at least one of a pattern feature and a pattern region, the step of using the model associated with the fine image to examine the fine image resolution image and to generate a score space including the steps of applying the plurality of probes in the model associated with the fine image at a plurality of poses to the fine image resulting in a plurality of score spaces at each pose where the score spaces include peaks, the peaks indicating likely locations of occurrences of the pattern in the fine image at each pose.

16. The apparatus of claim 15 wherein the step of using the model associated with the coarse image to examine the coarse image resolution image and to generate a score space includes the steps of applying the plurality of probes in the model associated with the coarse image at a plurality of poses to the coarse image resulting in a plurality of score spaces at each pose where the score spaces include peaks, the peaks indicating likely locations of occurrences of the pattern in the coarse image at each pose.

17. The apparatus of claim 11 wherein the fine image resolution is a finest image resolution.

18. The apparatus of claim 11 wherein each model includes a plurality of probes that are each a measure of similarity of at least one of a run-time image feature and a run time image region to at least one of a pattern feature and a pattern region, respectively, at a specific location.