Auto white balance apparatus and white balance adjusting method

• US 20060159336A1
• Filed: 10/21/2005
• Published: 07/20/2006
• Est. Priority Date: 01/19/2005
• Status: Active Grant

• Current Assignee: Monument Peak Ventures, LLC (Dominion Harbor Enterprises, LLC)
• Original Assignee: Eastman Kodak Company
• Inventors: Uezono, Tetsuji
• Granted Patent: US 7,705,892 B2
• Time in Patent Office: Days
• US Class Current: 382/167
• CPC Class Codes: H04N 23/88 for colour balance, e.g. white-balance circuits or colour temperature control

Abstract

To provide an auto white balance apparatus and a white balance adjusting method capable of performing more stable white balance adjustment. For every block, a characteristic pixel determination circuit 106 determines a characteristic pixel from among a pixel group in a block and a light-source reflection determination circuit 108 determines whether the typical value of a block reflects the light source of an input image, by comparing the color difference of a characteristic pixel of the block with the typical color difference of the block. Then, as a result of the determination, a white balance evaluating circuit 110 determines the light source of an image by using blocks that reflect the light source.

Claims

1. An auto white balance apparatus comprising:

block-dividing means for dividing an input image into a plurality of blocks;

typical-value calculating means for calculating, for every block, a typical value including luminance and color difference representing a block in accordance with each color value in the block;

characteristic pixel determination means for determining, for every block, a characteristic pixel from a pixel group constituting the block;

light-source reflection determination means for determining, for every block, whether the typical value of the block reflects a light source of an input image, by comparing a typical color difference of the block with a color difference of a characteristic pixel of the block;

white balance evaluating means for estimating the color of the light source of the input image in accordance with the typical luminance and the typical color difference of a block for which the light-source reflection determination means determines that the block reflects a light source and the object luminance of the input image;

white balance gain calculating means for calculating a white balance gain to be applied to the image in accordance with the estimated light-source color; and

white balance adjusting means for adjusting white balance to the image by use of the calculated white balance gain.

2. The auto white balance apparatus according to claim 1, wherein:

the characteristic pixel determination means determines as a characteristic pixel of a block a pixel having the highest luminance from among a pixel group constituting the block.

3. The auto white balance apparatus according to claim 1, wherein:

the characteristic pixel determination means determines as a characteristic pixel of a block a pixel having the largest color difference from among a pixel group constituting the block.

4. The auto white balance apparatus according to claim 1, wherein:

the light-source reflection determination means determines whether the typical value of a block reflects the color of a light source, in accordance with an angle θ

formed between a vector T0 connecting the typical color difference of the block with the origin of a color difference space and a vector T1 connecting the color difference of a characteristic pixel of the block with the origin of the color difference space, and a scalar ratio R between the vector T0 and the vector T1.

5. The auto white balance apparatus according to claim 4, wherein:

the light-source reflection determination means determines that the typical value of a block reflects the color of a light source when the pixel having the highest luminance is a characteristic pixel and conditions of cos θ

>

α

(α

;

constant) and β

₁<

R<

β

₂ (β

₁ and β

₂ are constants) are satisfied.

6. The auto white balance apparatus according to claim 4, wherein:

the light-source reflection determination means determines that the typical value of a block does not reflect the color of a light source when a pixel having the largest color difference is a characteristic pixel and conditions of cos θ

>

α

(α

;

constant) and β

₃<

R<

β

₄ (β

₃ and β

₄ are constants) are satisfied.

7. A white balance adjusting method for performing white balance adjustment for a picked-up image, comprising:

a block-dividing step of dividing an input image into a plurality of blocks;

a typical value calculating step of calculating, for every block, a typical value including luminance and color difference representing a block in accordance with each color value in the block;

a characteristic pixel determination step of determining, for every block, a characteristic pixel from a pixel group constituting the block;

a light-source reflection determination step of determining, for every block, whether the typical value of the block indicates reflection of a light source of the input image, by comparing a typical color difference of the block with a color difference of a characteristic pixel of the block;

a white balance evaluating step of estimating the color of the light source of the image in accordance with a typical luminance of the block for which, in the light-source reflection determination step, a determination is made that the block reflects a light source, and the object luminance of the input image;

a white balance gain calculating step of calculating a white balance gain to be applied to the image in accordance with the estimated color of the light source; and

a white balance adjusting step of adjusting white balance for the image by using the calculated white balance gain.

8. The white balance adjusting method according to claim 7, wherein:

in the characteristic pixel determination step, a pixel having the highest luminance is determined from among a pixel group constituting the block as a characteristic pixel in the block.

9. The white balance adjusting method according to claim 7, wherein:

in the characteristic pixel determination step, a pixel having the largest color difference is detected from a pixel group constituting a block as a characteristic pixel of the block.

10. The white balance adjusting method according to claim 7, wherein:

in the light-source reflection determination step, for every block a determination is made as to whether the typical value of a block indicates reflection of the color of a light source, in accordance with an angle θ

formed between a vector T0 connecting the typical color difference of the block with the origin of a color difference space and a vector T1 connecting the color difference of a characteristic pixel of the block with the origin of the color difference space, and a scalar ratio R between the vector T0 and the vector T1.

11. The white balance adjusting method according to claim 10, wherein:

in the light-source reflection determination step, a determination is made that the typical value of a block reflects the color of a light source when a pixel having the largest color difference is a characteristic pixel and conditions of cos θ

>

α

(α

;

constant) and β

₁<

R<

β

₂ (β

₁ and β

₂ are constants) are satisfied.

12. The white balance adjusting method according to claim 10, wherein:

in the light-source reflection determination step, a determination is made that the typical value of a block does not reflect the color of a light source when a pixel having the largest color difference is a characteristic pixel and conditions of cos θ

>

α

(α

;

constant) and β

₃<

R<

β

₄ (β

₃ and β

₄ are constants) are satisfied.