Method and system of noise removal for a sparsely sampled extended dynamic range image sensing device

• US 6,646,246 B1
• Filed: 11/21/2000
• Issued: 11/11/2003
• Est. Priority Date: 11/21/2000
• Status: Expired due to Term

• Current Assignee: Apple Inc.
• Original Assignee: Eastman Kodak Company
• Inventors: Gallagher, Andrew C., Gindele, Edward B.
• Primary Examiner(s): Kim, Robert H.
• Assistant Examiner(s): SONG, HOON K
• Application Number:
  US09/718,882
• Time in Patent Office: 1,085 Days
• Field of Search: 348/241, 348/607, 348/242, 348/256, 348/279, 250/208.1
• US Class Current: 250/208.1
• CPC Class Codes: H04N 23/70 Circuitry for compensating brightness variation in the scene
• CPC Class Codes: H04N 23/843 Demosaicing, e.g. interpolating colour pixel values
• CPC Class Codes: H04N 25/134 based on three different wavelength filter elements
• CPC Class Codes: H04N 25/60 Noise processing, e.g. detecting, correcting, reducing or removing noise

Abstract

A method of removing noise from a sparsely sampled extended dynamic range image produced by a sparsely sampled extended dynamic range image sensing device having fast photosites with a predetermined response to a light exposure for producing fast pixel values and slow photosites with a slower response to the same light exposure for producing slow pixel values, includes the steps of: providing a sparsely sampled extended dynamic range digital image having fast and slow pixel values; using a noise filter method to remove noise from the sparsely sampled extended dynamic range digital image by using only slow pixel values to generate slow noise reduced pixel values and using only fast pixel values to generate fast noise reduced pixel values; and combining the slow noise reduced pixel values and the fast noise reduced pixel values to generate a noise reduced sparsely sampled digital image.

Claims

1. A method of removing noise from a sparsely sampled extended dynamic range image produced by a sparsely sampled extended dynamic range image sensing device having fast photosites with a predetermined response to a light exposure for producing fast pixel values and slow photosites with a slower response to the same light exposure for producing slow pixel values, comprising the steps of:

a) providing a sparsely sampled extended dynamic range digital image having fast and slow pixel values;

b) using a noise filter method to remove noise from the sparsely sampled extended dynamic range digital image by using only slow pixel values to generate slow noise reduced pixel values and using only fast pixel values to generate fast noise reduced pixel values; and

c) combining the slow noise reduced pixel values and the fast noise reduced pixel values to generate a noise reduced sparsely sampled digital image.

2. The method of claim 1 wherein the noise filter method includes the steps of:

identifying a pixel of interest and a local neighborhood of pixels located about the pixel of interest;

calculating a difference pixel value for pixels in the local neighborhood of pixels based on the absolute difference between the value of the pixel of interest and the individual values of pixels included in the local neighborhood of pixels; and

using the absolute difference pixel values to calculate a noise reduced pixel value.

3. The method of claim 2, wherein the step of using the absolute difference pixel values, further includes the steps of:

comparing the absolute difference pixel values to a threshold value; and

using only the values of pixels included in the local neighborhood of pixels for which the corresponding absolute difference pixel values are less than the threshold value to calculate the noise reduced pixel value.

4. The method of claim 3, wherein the threshold value is determined by:

providing a table of two or more noise characteristic values;

using the values of pixels in the local neighborhood of pixels to select one of the noise characteristic values; and

using the selected noise characteristic value as the threshold value.

5. The method of claim 4, wherein different tables of noise characteristic values are provided for the slow pixel values and the fast pixel values.

6. The method of claim 5, wherein the sparsely sampled extended dynamic range digital image includes red, green, and blue slow pixel values and red, green, and blue fast pixel values and wherein different tables of noise characteristic values are provided respectively for the red slow, green slow and blue slow pixel values and the red fast, green fast and blue fast pixel values.

7. The method of claim 1 wherein the noise filter method employs a median filter.

8. The method of claim 1, further comprising the step of:

d) expanding the dynamic range of the noise reduced fast and slow pixel values in the noise reduced sparsely sampled digital image.

9. The method of claim 8, wherein the sparsely sampled extended dynamic range digital image includes red, green and blue slow pixel values and red, green and blue fast pixel values, and further including the step of:

using an interpolation spatial filter to generate missing pixel values to produce a noise reduced digital image.

19. A computer program product for performing the method of claim 1.

20. A digital image produced by the method of claim 1.

21. The method of claim 1, wherein the method is performed in a digital camera.

10. A system for removing noise from a sparsely sampled extended dynamic range image produced by a sparsely sampled extended dynamic range image sensing device having fast photosites with a predetermined response to a light exposure for producing fast pixel values and slow photosites with a slower response to the same light exposure for producing slow pixel values, comprising:

a) means for providing a sparsely sampled extended dynamic range digital image having fast and slow pixel values;

b) a noise filter for removing noise from the sparsely sampled extended dynamic range digital image by using only slow pixel values to generate slow noise reduced pixel values and using only fast pixel values to generate fast noise reduced pixel values; and

c) means for combining the slow noise reduced pixel values and the fast noise reduced pixel values to generate a noise reduced sparsely sampled digital image.

11. The system of claim 10 wherein the noise filter includes:

means for identifying a pixel of interest and a local neighborhood of pixels located about the pixel of interest;

means for calculating a difference pixel value for pixels in the local neighborhood of pixels based on the absolute difference between the value of the pixel of interest and the individual values of pixels included in the local neighborhood of pixels; and

means for using the absolute difference pixel values to calculate a noise reduced pixel value.

12. The system of claim 11, wherein the means for using the absolute difference pixel values, further includes:

means for comparing the absolute difference pixel values to a threshold value; and

means for using only the values of pixels included in the local neighborhood of pixels for which the corresponding absolute difference pixel values are less than the threshold value to calculate the noise reduced pixel value.

13. The system of claim 12, further comprising means for determining the threshold value comprising:

a table of two or more noise characteristic values;

means for using the values of pixels in the local neighborhood of pixels to select one of the noise characteristic values; and

means for using the selected noise characteristic value as the threshold value.

14. The system of claim 13, including different tables of noise characteristic values for the slow pixel values and the fast pixel values.

15. The system of claim 14, wherein the sparsely sampled extended dynamic range digital image includes red, green and blue slow pixel values and red, green and blue fast pixel values and including different tables of noise characteristic values respectively for the red slow, green slow and blue slow pixel values and the red fast, green fast and blue fast pixel values.

16. The system of claim 10 wherein the noise filter includes a median filter.

17. The system of claim 10, further comprising:

d) means for expanding the dynamic range of the noise reduced fast and slow pixel values in the noise reduced sparsely sampled digital image.

18. The system of claim 17, wherein the sparsely sampled extended dynamic range digital image includes red, green and blue slow pixel values and red, green and blue fast pixel values, and further including:

an interpolation spatial filter for generating missing pixel values to produce a noise reduced digital image.