System and method for color normalization of board images

• US 6,570,612 B1
• Filed: 09/21/1998
• Issued: 05/27/2003
• Est. Priority Date: 09/21/1998
• Status: Expired due to Term

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Bank One Insurance Holdings Incorporated (JP Morgan Chase & Co)
• Inventors: Marimont, David H., Saund, Eric
• Primary Examiner(s): HO, TUAN V
• Application Number:
  US09/158,443
• Time in Patent Office: 1,709 Days
• Field of Search: 348/207.99, 348/207.1, 348/207.11, 348/218.1, 348/222.1, 348/239, 348/335, 382/282, 382/283, 382/284
• US Class Current: 348/218.1
• CPC Class Codes: H04N 1/3876 Recombination of partial images to recreate the original image
• CPC Class Codes: H04N 1/56 Processing of colour picture signals H04N1/52 takes precedence
• CPC Class Codes: H04N 9/12 Picture reproducers H04N9/11 takes precedence
• CPC Class Codes: H04N 9/3147 Multi-projection systems displays in general H04N9/12; video walls G06F3/1446, G09G2300/026

Abstract

A system and method are provided to normalize colors in an image of a board surface so as to eliminate effects of variable illumination, and in the case of mosaic images, to eliminate visible seams between tiles. Each pixel is classified as foreground or background, where foreground pixels contain ink marks. An efficient smoothing process estimates the ideal background color of all pixels, including the foreground pixels. The estimated background color is used to normalize each pixel'"'"'s color such that in the final image, artifacts arising from camera gain and uneven lighting of the board surface are eliminated.

Claims

1. A method for color normalization of images of a surface, the surface having a background with foreground markings placed thereon, the method comprising steps of:

inputting a surface image comprised of pixels;

generating image arrays for each RGB color channel representing the image;

determining for each image array which of the pixels of the image are background pixels representing the background of the surface and which of the pixels of the image are foreground pixels representing the foreground markings based on pixel intensity and proximity to the white axis in RGB color space;

generating a background image array representing the background pixels for each RGB color channel based on the determining;

modifying each background image array by replacing values corresponding to each of the background pixels by a weighted average of its own value and a value of its predecessor and replacing values corresponding to each pixel located in a position formerly occupied by a foreground pixel are replaced by a value of a predecessor pixel; and

,

dividing the image arrays by a corresponding modified background image array.

2. The method of claim 1 wherein the determining step includes applying a center-surround filter to each pixel.

3. The method of claim 2 wherein the determining step includes applying a thresholding routine to each pixel.

4. The method of claim 3 wherein the determining step includes applying a morphological dilation.

5. The method of claim 1 wherein generating the background array includes removing the foreground pixels from the image array.

6. The method of claim 1 wherein the filtering is performed using a filter of the following form:

$V_x^{\mathrm{\prime <br><br>}}\leftarrow <br><br>\{\text{\hspace{1em}<br><br>}<br><br>\begin{array}{cc}\mathrm{\alpha <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>V_x+\left(1-\mathrm{\alpha <br><br>}\right)<br><br>V_{x-1}^{\mathrm{\prime <br><br>}}& \mathrm{if}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{background}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{pixel}\\ V_{x-1}^{\mathrm{\prime <br><br>}}& \mathrm{if}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{fore}<br><br>\mathrm{ground}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{pixel}\end{array}<br><br>\text{\hspace{1em}<br><br>}$

where V represents pixel values, the index x is a pixel location, α

is a constant, the index x−

1 indicates the previous pixel position along the direction of the filter'"'"'s pass and the prime notation indicates a pixel modified in the pass.

7. A system adapted to normalize color images of a surface, the surface having a background with foreground markings placed thereon, the system comprising:

means for inputting a surface image comprised of pixels;

means for generating image arrays for each RGB color channel representing the image;

means determining for each image array which of the pixels of the image are background pixels representing the background of the surface and which of the pixels of the image are foreground pixels representing the foreground markings;

means for generating a background array representing the background pixels for each RGB color channel based on the determining;

means for filtering each background array in a plurality of directions wherein values corresponding to each of the background pixels are replaced by a weighted average of its own value and a value of its predecessor and values corresponding to each pixel located in a position formerly occupied by a foreground pixel are replaced by a value of a predecessor pixel; and

,

means for dividing the image arrays by a corresponding filtered background array.

8. The system of claim 7 wherein the inputting means comprises a video camera.

9. A method for color normalization of images of a surface, the surface having a background with foreground markings placed thereon, the method comprising steps of:

inputting a surface image comprised of image tiles formed of pixels;

generating image arrays for each RGB color channel of each image tile representing the image;

determining for each image array which of the pixels of the image are background pixels representing the background of the surface and which of the pixels of the image are foreground pixels representing the foreground markings based on pixel intensity and proximity to the white axis in RGB color space;

generating a background array representing the background pixels for each RGB color channel of each image tile based on the determining;

filtering each background array in a plurality of directions wherein values corresponding to each of the background pixels are replaced by a weighted average of its own value and a value of its predecessor and values corresponding to each pixel located in a position formerly occupied by a foreground pixel are replaced by a value of a predecessor pixel;

dividing the image arrays by a corresponding filtered background array to obtain output image tiles; and

,

generating an output image based on the output image tiles.

10. The method of claim 9 wherein the determining step includes applying a center-surround filter to each pixel.

11. The method of claim 10 wherein the determining step includes applying a thresholding routine to each pixel.

12. The method of claim 11 wherein the determining step includes applying a morphological dilation.

13. The method of claim 9 wherein generating the background array includes removing the foreground pixels from the image array.

14. The method of claim 9 wherein the filtering is performed using a filter of the following form:

$V_x^{\mathrm{\prime <br><br>}}\leftarrow <br><br>\{\text{\hspace{1em}<br><br>}<br><br>\begin{array}{cc}\mathrm{\alpha <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>V_x+\left(1-\mathrm{\alpha <br><br>}\right)<br><br>V_{x-1}^{\mathrm{\prime <br><br>}}& \mathrm{if}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{background}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{pixel}\\ V_{x-1}^{\mathrm{\prime <br><br>}}& \mathrm{if}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{fore}<br><br>\mathrm{ground}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{pixel}\end{array}<br><br>\text{\hspace{1em}<br><br>}$

where V represents pixel values, the index x represents a pixel location, α

is a constant, the index x−

1 indicates the previous pixel position along the direction of the filter'"'"'s pass and the prime notation indicates a pixel modification in the pass.