Techniques for scene-based nonuniformity correction in shutterless FIR cameras

1. A method for correcting nonuniformity in far-infrared (FIR) images captured by a shutterless FIR camera, comprising:

determining a noise of a current image based on updating a noise estimate of a previous image with a noise estimate of a current image;

determining a weight mask matrix of the current image, where the weight mask matrix includes high values corresponding to pixels of the current image in which noise estimation is facilitated, and low values corresponding to pixels of the current image in which noise estimation is inhibited, wherein the low values inhibit pixels representing regions with limited changes between the current image and the previous image;

applying the weight mask matrix to the current image; and

correcting the nonuniformity of the current image incrementally based on the determined noise of the current image and the applied weight mask matrix.

2. The method of claim 1 , wherein the noise estimate of the current image is calculated by applying a nonlinear high-pass filter to the current image.

3. The method of claim 2 , wherein the nonlinear high-pass filter is configured to expose, within the current image, at least one of: a smooth area, a homogenous area and a dynamic area.

4. The method of claim 1 , further comprising:

determining if the shutterless FIR camera is in motion; and

correcting the nonuniformity of the current image if it is determined that the shutterless FIR camera is in motion.

5. The method of claim 4 , wherein the shutterless FIR camera is determined to be in motion if a percentage of pixels relative to a previous image have changed above a predetermined threshold.

6. The method of claim 1 , wherein the weight mask matrix contains low values that inhibit oversaturated pixels.

7. The method of claim 1 , wherein the weight mask matrix contains low values that inhibit pixels that are determined to come from regions within the current image that have strong edges.

8. The method of claim 7 , wherein pixels with strong edges are determined by comparing pixels to neighboring pixels and determining if there is a significant gap in pixel value.

9. The method of claim 1 , wherein the weight mask matrix contains low values that inhibit pixels within the current image that are known a-priori to be defective.

10. The method of claim 1 , wherein the shutterless FIR camera is mounted on a moving vehicle.

11. A non-transitory computer readable medium having stored thereon instructions for causing a processing circuitry to perform a process, the process comprising:

determining a noise of a current image based on updating a noise estimate of a previous image with a noise estimate of a current image;

determining a weight mask matrix of the current image, where the weight mask matrix includes high values corresponding to pixels of the current image in which noise estimation is facilitated, and low values corresponding to pixels of the current image in which noise estimation is inhibited, wherein the low values inhibit pixels representing regions with limited changes between the current image and the previous image;

applying the weight mask matrix to the current image; and

correcting a nonuniformity of the current image incrementally based on the determined noise of the current image and the applied weight mask matrix.

12. An electronic circuit integrated in a shutterless far-infrared (FIR) camera and configured to process FIR images, comprising:

a processing circuitry; and

a memory, the memory containing instructions that, when executed by the processing circuitry, configure the electronic circuit to:

determine a noise of a current image based on updating a noise estimate of a previous image with a noise estimate of a current image;

determine a weight mask matrix of the current image, where the weight mask matrix includes high values corresponding to pixels of the current image in which noise estimation is facilitated, and low values corresponding to pixels of the current image in which noise estimation is inhibited, wherein the low values inhibit pixels representing regions with limited changes between the current image and the previous image;

apply the weight mask matrix to the current image; and

correct a nonuniformity of the current image incrementally based on the determined noise of the current image and the applied weight mask matrix.

13. The electronic circuit of claim 12 , wherein the noise estimate of the current image is calculated by applying a nonlinear high-pass filter to the current image.

14. The electronic circuit of claim 13 , wherein the nonlinear high-pass filter is configured to expose, within the current image, at least one of: a smooth area, a homogenous area and a dynamic area.

15. The electronic circuit of claim 12 , wherein the processing circuitry is further configured to:

determine if the shutterless FIR camera is in motion; and

correct the nonuniformity of the current image if it is determined that the shutterless FIR camera is in motion.

16. The electronic circuit of claim 15 , wherein the shutterless FIR camera is determined to be in motion if a percentage of pixels relative to a previous image have changed above a predetermined threshold.

17. The electronic circuit of claim 12 , wherein the weight mask matrix contains low values that inhibit oversaturated pixels.

18. The electronic circuit of claim 12 , wherein the weight mask matrix contains low values that inhibit pixels that are determined to come from regions within the current image that have strong edges.

19. The electronic circuit of claim 18 , wherein pixels with strong edges are determined by comparing pixels to neighboring pixels and determining if there is a significant gap in pixel value.

20. The electronic circuit of claim 12 , wherein the weight mask matrix contains low values that inhibit pixels within the current image that are known a-priori to be defective.

21. The electronic circuit of claim 12 , wherein the shutterless FIR camera is mounted on a moving vehicle.