System and method for using normalized gray scale pattern find

• US 20040136611A1
• Filed: 01/09/2003
• Published: 07/15/2004
• Est. Priority Date: 01/09/2003
• Status: Active Grant

• Current Assignee: Banner Engineering Corporation
• Original Assignee: Banner Engineering Corporation
• Inventors: Reiners, Lawrence Lee
• Granted Patent: US 7,113,652 B2
• Time in Patent Office: Days
• US Class Current: 382/291
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors

Abstract

A computer controlled system, method, and corresponding computer data product, for utilizing a normalized gray scale pattern find within a machine vision system. The system and method (a) correlates a reference image having a first spatial resolution having no particular degree of rotation within a first search region of interest of an acquired image to find a potential match location, (b) correlates a rotated reference image having the first spatial resolution at a particular degree of rotation within the first search region of interest to find a potential match location, (c) selects the potential match location and the particular degree of rotation corresponding to the highest correlation value, (d) increases the spatial resolution of the reference image at the particular degree of rotation to generate a new reference image and correlating the new reference image within a second search region of interest centered at the selected potential match location, (e) selects an updated potential match location corresponding to a location having the highest correlation value that is completely within the search region of interest, (f) repeats steps (d) and (e) for all available resolutions for the reference image, and (g) selects a match location and orientation of a reference image within the acquired image corresponding to the location having the highest correlation value.

Claims

1. A system for utilizing a normalized gray scale pattern find within a machine vision system, the system comprising:

(a) means for correlating a reference image having a first spatial resolution having no particular degree of rotation within a first search region of interest of an acquired image to find a potential match location;

(b) means for correlating a rotated reference image having the first spatial resolution at a particular degree of rotation within the first search region of interest to find a potential match location;

(c) means for selecting the potential match location and the particular degree of rotation corresponding to the highest correlation value;

(d) means for increasing the spatial resolution of the reference image at the particular degree of rotation to generate a new reference image and correlating the new reference image within a second search region of interest centered at the selected potential match location;

(e) means for selecting an updated potential match location corresponding to a location having the highest correlation value that is completely within the search region of interest;

(f) means for repeating steps (d) and (e) for all available resolutions for the reference image.

(g) means for selecting a match location and orientation of a reference image within the acquired image corresponding to the location having the highest correlation value.

2. The system according to claim 1, wherein the first spatial resolution corresponds to a minimum decimated resolution for a reference image generated from a high resolution reference image.

3. The system according to claim 2, wherein the increased spatial resolution of a reference image corresponds to the next highest decimated image resolution for a reference image generated from a high resolution reference image.

4. The system according to claim 1, wherein the first search region of interest corresponds to the entire acquired image.

5. The system according to claim 1, wherein the second search region of interest corresponds to a 6×

5 pixel region of interest center around the potential match location with the search region increasing as necessary to completely enclose the peak correlation value.

6. The system according to claim 2, wherein processing within the correlating means performs the processing using a checkerboard search when the processing the first spatial resolution.

7. The system according to claim 6, wherein the processing within the correlating means performs the correlation operation to fill in missing correlation values around the potential match location.

8. A system for utilizing a normalized gray scale pattern find within a machine vision system, the system comprising:

an image acquisition interface module for obtaining an acquired image from a image capture device;

a image reference data module for providing a set of reference modules for use in locating a reference mark within the acquired image;

a image processing module for correlating the set of reference images within a set of search-regions of interest to find the best match location;

wherein the image processing module uses a normalized gray scale pattern find process.

9. The system according to claim 8, wherein the image processing module comprises:

(a) means for correlating a reference image having a first spatial resolution having no particular degree of rotation within a first search region of interest of an acquired image to find a potential match location;

(b) means for correlating a rotated reference image having the first spatial resolution at a particular degree of rotation within the first search region of interest to find a potential match location;

(c) means for selecting the potential match location and the particular degree of rotation corresponding to the highest correlation value;

(d) means for increasing the spatial resolution of the reference image at the particular degree of rotation to generate a new reference image and correlating the new reference image within a second search region of interest centered at the selected potential match location;

(e) means for selecting an updated potential match location corresponding to a location having the highest correlation value that is completely within the search region of interest;

(f) means for repeating steps (d) and (e) for all available resolutions for the reference image.

(g) means for selecting a match location and orientation of a reference image within the acquired image corresponding to the location having the highest correlation value.

10. The system according to claim 8, wherein the first spatial resolution corresponds to a minimum decimated resolution for a reference image generated from a high resolution reference image and the increased spatial resolution of a reference image corresponds to the next highest decimated image resolution for a reference image generated from a high resolution reference image.

11. The system according to claim 9, wherein the first search region of interest corresponds to the entire acquired image.

12. The system according to claim 9, wherein the second search region of interest corresponds to a 6×

5 pixel region of interest center around the potential match location.

13. A method for utilizing a normalized gray scale pattern find within a machine vision system, the method comprising:

(a) correlating a reference image having a first spatial resolution having no particular degree of rotation within a first search region of interest of an acquired image to find a potential match location;

(b) correlating a rotated reference image having the first spatial resolution at a particular degree of rotation within the first search region of interest to find a potential match location;

(c) selecting the potential match location and the particular degree of rotation corresponding to the highest correlation value;

(d) increasing the spatial resolution of the reference image at the particular degree of rotation to generate a new reference image and correlating the new reference image within a second search region of interest centered at the selected potential match location;

(e) selecting an updated potential match location corresponding to a location having the highest correlation value that is completely within the search region of interest;

(f) repeating steps (d) and (e) for all available resolutions for the reference image.

(g) select a match location and orientation of a reference image within the acquired image corresponding to the location having the highest correlation value.

14. The method according to claim 13, wherein the first spatial resolution corresponds to a minimum decimated resolution for a reference image generated from a high resolution reference image.

15. The method according to claim 14, wherein the increased spatial resolution of a reference image corresponds to the next highest decimated image resolution for a reference image generated from a high resolution reference image.

16. The method according to claim 13, wherein the first search region of interest corresponds to the entire acquired image.

17. The method according to claim 13, wherein the second search region of interest corresponds to a 6×

5 pixel region of interest center around the potential match location.

18. A computer data product readable by a computing system and encoding instructions for implementing a computer method for utilizing a normalized gray scale pattern find within a machine vision system, the method comprising:

(a) correlating a reference image having a first spatial resolution having no particular degree of rotation within a first search region of interest of an acquired image to find a potential match location;

(b) correlating a rotated reference image having the first spatial resolution at a particular degree of rotation within the first search region of interest to find a potential match location;

(c) selecting the potential match location and the particular degree of rotation corresponding to the highest correlation value;

(d) increasing the spatial resolution of the reference image at the particular degree of rotation to generate a new reference image and correlating the new reference image within a second search region of interest centered at the selected potential match location;

(e) selecting an updated potential match location corresponding to a location having the highest correlation value that is completely within the search region of interest;

(f) repeating steps (d) and (e) for all available resolutions for the reference image.

(g) select a match location and orientation of a reference image within the acquired image corresponding to the location having the highest correlation value.

19. The computer data product according to claim 18, wherein the first spatial resolution corresponds to a minimal decimated resolution for a reference image generated from a high resolution reference image.

20. The computer data product according to claim 19, wherein the increased spatial resolution or a reference image corresponds to the next highest decimated image resolution for a reference image generated from a high resolution reference image.

21. The computer data product according to claim 18, wherein the first search region of interest corresponds to the entire acquired image.

22. The computer data product according to claim 18, wherein the second search region of interest corresponds to a 6×

5 pixel region of interest center around the potential match location.