Fixed pattern noise removal in CMOS imagers across various operational conditions

• US 7,443,431 B2
• Filed: 04/25/2003
• Issued: 10/28/2008
• Est. Priority Date: 09/13/2002
• Status: Active Grant

• Current Assignee: Omnivision Technologies Incorporated (Will Semiconductor Co., Ltd. Shanghai)
• Original Assignee: Eastman Kodak Company
• Inventors: Pillman, Bruce H., Guidash, Robert M., Kelly, Sean C.
• Primary Examiner(s): Ho; Tuan V
• Application Number:
  US10/423,279
• Publication Number:
  US 20040051796A1
• Time in Patent Office: 2,013 Days
• Field of Search: 348/187, 348/188, 348241-251, 348/262
• US Class Current: 348/244
• CPC Class Codes: H04N 25/63 applied to dark current
• CPC Class Codes: H04N 25/671 for non-uniformity detection or correction

Abstract

A method of minimizing noise in an image produced by an electronic imager comprising: determining a correction system for a range of imager integration times and a range of imager temperatures for an electronic imager which has taken a series of dark capture images and a series of flat field capture images in a calibration mode; and applying the correction system to an image produced by the electronic imager in an image capture mode.

Claims

1. A method of minimizing noise in an image produced by an electronic imager comprising:

generating a correction system having a range of imager integration times and a range of imager temperatures for an electronic imager which has taken a series of dark capture images and a series of flat field capture images in a calibration mode; and

applying said correction system to a plurality of images having different integration times or imager temperatures produced by said electronic imager in an image capture mode without reentering the calibration mode by using the correction system to perform regression analysis for the different integration times or imager temperatures.

2. The method of claim 1 wherein said generating a correction system includes generating a single fixed pattern noise (FPN) dark map and a singe FPN pixel response non-unifonnity (PRNU) map based on said series of dark capture images and said series of flat field capture images.

3. The method of claim 1 wherein said generating a correction system includes taking a series of dark capture images and a series of flat field images for each temperature and imager integration time of a matrix of a plurality of temperatures and imager integration times, averaging the dark capture images for each said temperature and imager integration time to remove temporal noise, connecting each flat field image by subtracting said averaged dark capture image, and averaging said flat field images to produce a pixel response non-uniformity correction.

4. The method of claim 1 wherein said electronic imager has three different color channels and wherein said correction system effects an aim response in each said color channel to which all pixels will be gained and per pixel gain is then computed.

5. The method of claim 4 wherein said three color channels are R(Red), (G)Green, (B)Blue color channels.

6. A method of minimizing noise in an image produced by an electronic imager comprising:

generating a correction system having a range of imager integration times and a range of imager temperatures for an electronic imager which has taken a series of dark capture images and a series of flat field capture images in a calibration mode;

applying said correction system to a plurality of images having different integration times or imager temperatures produced by said electronic imager in an image capture mode without reentering the calibration mode;

wherein said electronic imager has three different color channels and wherein said correction system effects an aim response in each said color channel to which all pixels will be gained and per pixel gain is then computed; and

said aim response is defined as the average of each color channel in the center region of the imager'"'"'s format.

7. The method of claim 6 wherein said center region is ¼

of the center of the imager'"'"'s format.

8. A method of minimizing noise in an image produced by an electronic imager having image capture pixels and dark pixels comprising:

generating correction maps for a range of integration times, each map associated with different imager dark pixel values during a calibration mode;

capturing a plurality of images and for each image adjusting the correction maps as a function of the captured image dark pixel values using regression at the time of image capture without reentering the calibration mode; and

correcting each captured image with the adjusted correction maps.

9. The method of claim 8, wherein adjusting the correction maps further comprises adjusting a dark fixed pattern noise map and pixel response nonuniformity map.

10. The method of claim 8, wherein adjusting the correction maps further comprises adjusting a dark fixed pattern noise map.