Method and system for high accuracy and reliability registration of multi modal imagery

1. A non-transitory computer-readable storage medium having instructions stored therein, the instructions executable by a processor to cause the processor to perform operations comprising:

receiving a plurality of target images;

receiving a reference image that overlaps the plurality of target images;

preprocessing the plurality of target images, wherein the preprocessing includes:

rejecting one or more first target images of the plurality of target images with a shadow region above a shadow threshold while keeping one or more second target images of the plurality of target images with a shadow region below the shadow threshold;

for each second target image:

transforming the second target image into an atlas projection to match the reference image; and

partitioning the transformed target image into a plurality of sub-regions;

determining a matching statistic for each sub-region to determine a location for each sub-region relative to a corresponding portion of the reference image; and

determining an alignment of each sub-region with the corresponding portion of the reference image based on the determined matching statistic for each sub-region.

2. The non-transitory computer readable storage medium of claim 1 , further comprising determining an offset of the second target image relative to the reference image and an associated uncertainty for the determined offset.

3. The non-transitory computer readable storage medium of claim 1 , further comprising:

determining an uncertainty for the reference image; and

interpolating the reference image to match the second target image orientation and pixel spacing based on the determined uncertainty for the reference image.

4. The non-transitory computer readable storage medium of claim 1 , wherein determining the matching statistic for each sub-region includes:

selecting one or more basis functions to map a target image amplitude for the sub-region to an amplitude of the reference image.

5. The non-transitory computer readable storage medium of claim 1 , further comprising:

providing a model for mapping between the second target image and the reference image; and

determining an uncertainty in a parameter of the model used for mapping.

6. The non-transitory computer readable storage medium of claim 1 , wherein the target image includes one of a synthetic aperture radar (SAR) image, an optical image, an infrared image, a millimeter image, an ultrasonic image, a magnetic resonance imaging (MRI) image, or an x-ray image.

7. The non-transitory computer readable storage medium of claim 1 , wherein the reference image includes one of a SAR image, an optical image, an infrared image, a millimeter image, an ultrasonic image, a MRI image, or an X-Ray image.

8. An image processing system, comprising:

an image acquisition device configured to acquire a plurality of target images and a reference image; and

a processor structured to pre-process the plurality of target images, wherein the pre-processing of the target image includes:

receiving the plurality of target images;

rejecting one or more first target images with a shadow region above a shadow threshold while keeping one or more second target images with a shadow region below the shadow threshold to create a kept target image;

providing a model for mapping between the kept target image and the reference image;

transforming the kept target image into an atlas projection to match the reference image;

partitioning the transformed target image into sub-regions; and

determining a fractional area of each sub-region that is shadowed;

wherein the processor is structured to pre-process the reference image, wherein the pre-processing of the reference image includes:

receiving the reference image;

providing a digital elevation map (DEM) of the reference image;

projecting the reference image onto a target slant plane, wherein the projection includes a shadow region; and

interpolating the projected reference image to match orientation and pixel spacing of the kept target image;

wherein the processor is structured to determine a matching statistic for each target image sub-region to determine a location and an alignment for each sub-region relative to a corresponding portion of the received reference image.

9. The image processing system of claim 8 , wherein the processor is further structured to determine a shadow area for the received reference image.

10. The image processing system of claim 9 , wherein the processor is structured to:

select one or more basis functions to map a target image amplitude for each sub-region to an amplitude of the reference image.

11. The image processing system of claim 8 , wherein the matching statistic is based on a shadow content of the target image sub-region matching a shadow content of the corresponding portion of the received reference image within a specified tolerance.

12. The image processing system of claim 8 , wherein the processor is further structured to determine a model parameter for mapping between the kept target image and the received reference image and to determine an uncertainty with the model parameter.

13. The image processing system of claim 8 , wherein the processor is structured to determine a down sampling ratio for the kept target image.

14. The image processing system of claim 8 , wherein the received target image includes one of a synthetic aperture radar (SAR) image, an optical image, an infrared image, a millimeter image, an ultrasonic image, a magnetic resonance imaging (MRI) image, or an x-ray image.

15. A method of mapping a synthetic aperture radar (SAR) image to a reference image, the method comprising:

pre-processing, by a processor, a plurality of target images including:

rejecting one or more first target images with a shadow region above a shadow threshold while keeping one or more second target images with a shadow region below the shadow threshold to create a kept target image;

partitioning the kept target image into a plurality of sub-regions; and

determining a fractional area of each of the plurality of kept target sub-regions that is shadowed;

pre-processing, by the processor, the reference image including:

projecting the reference image onto a target slant plane, wherein the projection includes a shadow region; and

determining a fraction area of the projected reference image that is shadowed;

determining, by the processor, a matching statistic for each sub-region that includes:

selecting one or more basis functions to map a target image amplitude for each sub-region to a reference image amplitude;

providing an upper limit and tolerance for fractional coverage, and providing an upper limit beyond which the sub-region is not considered due to having an excessive shadow region; and

determining a location and a normalized mean square error (NMSE) for each considered sub-region based on the matching statistic for each considered sub-region;

wherein the determined location of the sub-region corresponds with a shadow content of the sub-region matching a shadow content of the projected reference image within a specified tolerance.

16. The method of claim 15 , wherein pre-processing of the kept target image further includes excluding a pixel from the target image based on the pixel representing a bright spot in the target image.

17. The method of claim 15 , wherein a target image with an excessive shadow region is a target image that has a greater than or equal to fifty-percent contiguous SAR shadow.

18. The method of claim 15 , further comprising determining a model parameter for an offset of the kept target image relative to the projected reference image and an uncertainty in the model parameter.

19. The method of claim 15 , wherein the target image includes one of a synthetic aperture radar (SAR) image, an optical image, an infrared image, a millimeter image, an ultrasonic image, a magnetic resonance imaging (MRI) image, or an x-ray image.