System and method for recognizing markers on printed circuit boards

• US 7,110,591 B2
• Filed: 03/28/2001
• Issued: 09/19/2006
• Est. Priority Date: 03/28/2001
• Status: Active Grant

• Current Assignee: Siemens AG
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Fang, Ming, Neubauer, Claus
• Primary Examiner(s): Ahmed, Samir
• Assistant Examiner(s): KIM, CHONG R
• Application Number:
  US09/819,211
• Publication Number:
  US 20020168097A1
• Time in Patent Office: 2,001 Days
• Field of Search: 382/141, 382/144, 382/145, 382/147, 382148-151, 382/159, 382/160, 382/168, 382/170, 382/209, 382/278
• US Class Current: 382/147
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/507 Summing image-intensity values; Histogram projection analysis
• CPC Class Codes: G06V 10/758 Involving statistics of pixels or of feature values, e.g. histogram matching
• CPC Class Codes: H05K 1/0269 for visual or optical inspection
• CPC Class Codes: H05K 13/0812 the monitoring devices being integrated in the mounting machine, e.g. for monitoring components, leads, component placement
• CPC Class Codes: H05K 2201/09936 Marks, inscriptions, etc. for information

Abstract

A system and method for recognizing markers on, e.g., a PCB (printed circuit board). In one aspect, a system for recognizing a marker in an image comprises an image capture module (14) for extracting image features associated with an input image of a ROI (region of interest) captured through a lens 15, an image processor (16) comprising a first marker recognition processor (17) for recognizing a marker in the input image based on a normalized correlation and a second marker recognition processor (18) for recognizing a marker in the input image based on gray value histograms; a training module 19 for building template images and histograms that are used by the image processor (16) to detect a marker in the input image and a database (20) for indexing and storing trained template images and trained histograms.

Claims

1. A method for detecting a marker in an image, comprising the steps of:

selecting one of a first marker recognition process based on a normalized correlation and a second marker recognition process based on gray value histograms;

extracting image features associated with an input image of a ROI (region of interest); and

comparing the extracted image features with a trained model associated with the selected recognition process to determine if a marker is present in the input image,

wherein if the first marker recognition process is selected, the trained model comprises at least one template image for a first template class and at least one template image for a second template class, wherein the at least one template image for the first template class is trained from one or more images of the ROI having a maker, and wherein the at least one template image for the second template class is trained from one or more images of the ROI not having the marker, and wherein the step of comparing comprises the steps of;

normalizing the template images and input image with respect to brightness;

computing a correlation between the input image and each normalized template image; and

determining if a marker is present in the input image based on the computed correlations.

2. The method of claim 1, wherein the step of normalizing comprises computing

$I$

( i ) = ( I ⁡

( i ) - μ

) σ

, ⁢

where I(i) is the gray value of pixel I and where μ and

σ

denote the average brightness and contrast, respectively.

3. The method of claim 1, wherein the step of computing a correlation comprises computing

$\rho$

= ∑

allpixels ⁢

I ⁡

( i ) * T ⁡

( i ) ) ⁢

where ρ

comprises the correlation coefficient, I comprises the input image, and T comprises the template image.

4. The method of claim 1, further comprising the step of computing the at least one template image for the first or second template class from an average of a plurality of template images of the first or second template class.

5. The method of claim 1, wherein the step of determining if a marker is present in the input image is based on a maximum computed correlation.

6. The method of claim 1, further comprising the step of reducing the resolution of the input image and the template image by a predetermined factor prior to the comparing step.

7. A method for detecting a marker in an image, comprising the steps of:

selecting one of a first marker recognition process based on a normalized correlation and a second marker recognition process based on gray value histograms;

extracting image features associated with an input image of a ROI (region of interest); and

comparing the extracted image features with a trained model associated with the selected recognition process to determine if a marker is present in the input image,

wherein if the second marker recognition process is selected, the trained model comprises a sample image histogram comprising a gray value distribution of a sample image and the step of comparing comprises the steps of;

generating a local image histogram comprising a gray value distribution for each of a plurality of equal-sized sub-regions of the input image;

combining the local image histograms to form a single histogram representation;

computing a distance measure between the single histogram representation and the sample image histogram; and

determining if a marker is present in the input image based on the computed distance measure.

8. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for detecting a marker in an image, the method comprising the steps of:

selecting one of a first marker recognition process based on a normalized correlation and a second marker recognition process based on gray value histograms;

extracting image features associated with an input image of a ROI (region of interest); and

comparing the extracted image features with a trained model associated with the selected recognition process to determine if a marker is present in the input image,

wherein if the first marker recognition process is selected, the trained model comprises at least one template image for a first template class and at least one template image for a second template class, wherein the at least one template image for the first template class is trained from one or more images of the ROI having a maker, and wherein the at least one template image for the second template class is trained from one or more images of the ROI not having the marker, and wherein the step of comparing comprises the steps of;

normalizing the template images and input image with respect to brightness;

computing a correlation between the input image and each normalized template image; and

determining if a marker is present in the input image based on the computed correlations.

9. The program storage device of claim 8, wherein the instructions for normalizing comprise instructions for computing

$I$

( i ) = ( I ⁡

( i ) - μ

) σ

,

where I(i) is the gray value of pixel I and where μ and

σ

denote the average brightness and contrast, respectively.

10. The program storage device of claim 8, wherein the instructions for computing a correlation comprise instructions for computing

$\rho$

= ∑

allpixels ⁢

I ⁡

( i ) * T ⁡

( i ) )

where ρ

comprises the correlation coefficient, I comprises the input image, and T comprises the template image.

11. The program storage device of claim 8, further comprising instructions for computing the at least one template image for the first or second template class from an average of a plurality of template images of the first or second template class.

12. The program storage device of claim 8, wherein the step of determining if a marker is present in the input image is based on a maximum computed correlation.

13. The program storage device of claim 8, further comprising instructions for the step of reducing the resolution of the input image and the template image by a predetermined factor prior to the comparing step.

14. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for detecting a marker in an image, the method comprising the steps of:

selecting one of a first marker recognition process based on a normalized correlation and a second marker recognition process based on gray value histograms;

extracting image features associated with an input image of a ROI (region of interest); and

comparing the extracted image features with a trained model associated with the selected recognition process to determine if a marker is present in the input image,

wherein if the second marker recognition process is selected, the trained model comprises a sample image histogram comprising a gray value distribution of a sample image and the instructions for the step of comparing comprise instructions for;

generating a local image histogram comprising a gray value distribution for each of a plurality of equal-sized sub-regions of the input image;

combining the local image histograms to form a single histogram representation;

computing a distance measure between the single histogram representation and the sample image histogram; and

determining if a marker is present in the input image based on the computed distance measure.