Image noise reduction using spectral transforms
Various techniques are provided for reducing noise in captured image frames. In one example, a method includes determining row values for image frames comprising scene information and noise information. The method also includes performing first spectral transforms in a first domain on corresponding subsets of the row values to determine first spectral coefficients. The method also includes performing second spectral transforms in a second domain on corresponding subsets of the first spectral coefficients to determine second spectral coefficients. The method also includes selectively adjusting the second spectral coefficients. The method also includes determining row correction terms based on the adjusted second spectral coefficients to reduce the noise information of the image frames. Additional methods and systems are also provided.
1. A method comprising:
determining row values for image frames comprising scene information and noise information;
performing first spectral transforms in a first domain on corresponding subsets of the row values to determine first spectral coefficients;
performing second spectral transforms in a second domain on corresponding subsets of the first spectral coefficients to determine second spectral coefficients;
selectively adjusting the second spectral coefficients;
performing reverse second spectral transforms in the second domain on corresponding subsets of the adjusted second spectral coefficients to determine adjusted first spectral coefficients;
performing reverse first spectral transforms in the first domain on corresponding subsets of the adjusted first spectral coefficients to determine adjusted row values; and
determining the row correction terms based on the adjusted row values.
2. The method of claim 1 , wherein the adjusting comprises comparing the second spectral coefficients to a set of noise coefficients calculated from additional image frames comprising only the noise information.
3. The method of claim 1 , further comprising:
determining a plurality of adjusted row values for each row; and
wherein the row correction terms are further based on a weighted average of the adjusted row values.
4. The method of claim 1 , further comprising:
identifying matching subsets of the row values in the image frames; and
wherein the first spectral transforms are applied to the matching subsets to compensate for motion.
5. The method of claim 1 , wherein:
the first domain is a spatial domain taken along the row values associated with individual ones of the image frames; and
the second domain is a temporal domain taken along corresponding ones of the first spectral coefficients associated with different ones of the image frames.
6. The method of claim 1 , wherein the row values are row mean values.
7. The method of claim 1 , further comprising repeating the method for column values of the image frames to determine column correction terms.
8. The method of claim 1 , wherein the image frames are thermal image frames.
9. The method of claim 1 , further comprising:
capturing the image frames from a real world scene; and
applying the row correction terms to the image frames to reduce the noise of the image frames.
10. A system comprising:
a memory component storing machine-executable instructions; and
a processor configured to execute the instructions to cause the system to:
determine row values for image frames comprising scene information and noise information,
perform first spectral transforms in a first domain on corresponding subsets of the row values to determine first spectral coefficients,
perform second spectral transforms in a second domain on corresponding subsets of the first spectral coefficients to determine second spectral coefficients,
selectively adjust the second spectral coefficients,
perform reverse second spectral transforms in the second domain on corresponding subsets of the adjusted second spectral coefficients to determine adjusted first spectral coefficients,
perform reverse first spectral transforms in the first domain on corresponding subsets of the adjusted first spectral coefficients to determine adjusted row values, and
determine the row correction terms based on the adjusted row values.
11. The system of claim 10 , wherein the processor is configured to execute the instructions to cause the system to compare the second spectral coefficients to a set of noise coefficients calculated from additional image frames comprising only the noise information to adjust the second spectral coefficients.
12. The system of claim 10 , wherein:
the processor is configured to execute the instructions to cause the system to determine a plurality of adjusted row values for each row; and
the row correction terms are further based on a weighted average of the adjusted row values.
13. The system of claim 10 , wherein:
the processor is configured to execute the instructions to cause the system to identify matching subsets of the row values in the image frames; and
the first spectral transforms are applied to the matching subsets to compensate for motion.
14. The system of claim 10 , wherein:
the first domain is a spatial domain taken along the row values associated with individual ones of the image frames; and
the second domain is a temporal domain taken along corresponding ones of the first spectral coefficients associated with different ones of the image frames.
15. The system of claim 10 , wherein the row values are row mean values.
16. The system of claim 10 , wherein the processor is configured to execute the instructions to cause the system to determine column correction terms.
17. The system of claim 10 , wherein the image frames are thermal image frames.
18. The system of claim 10 , further comprising:
an image capture component configured to capture the image frames from a real world scene; and
wherein the processor is configured to execute the instructions to cause the system to apply the row correction terms to the image frames to reduce the noise of the image frames.