Change subtraction of synthetic aperture radar data

• US 6,650,273 B1
• Filed: 05/06/2002
• Issued: 11/18/2003
• Est. Priority Date: 05/06/2002
• Status: Active Grant

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Obenshain, Kenneth Fell
• Primary Examiner(s): Tarcza, Thomas H.
• Assistant Examiner(s): ALSOMIRI, ISAM A
• Application Number:
  US10/139,953
• Publication Number:
  US 20030206129A1
• Time in Patent Office: 561 Days
• Field of Search: 342/25, 342/160-162, 382/103, 382/106, 382/107, 382/294
• US Class Current: 342/25.00R
• CPC Class Codes: G01S 13/9023 combined with interferometric techniques
• CPC Class Codes: G01S 13/9029 specially adapted for moving target detection within a single SAR image or within multiple SAR images taken at the same time

Abstract

A method (100) for coherent change subtraction of mission and reference synthetic aperture radar (SAR) data (20′,20″) is provided. The method (100) forms (102) mission and reference images (22′,22″) from the mission and reference SAR data (20′,20″), registers (122) the mission and reference images (22′,22″) on a common plane to form registered mission and reference images (24′,24″), and forms (124) the registered mission and reference images (24′,24″) into at least one patch (26) containing mission and reference data (28′,28″). The method (100) then processes (126) each patch (26) by removing (130) linear phase terms (34) from the mission data (28′), trimming (142) non-overlapping spectra of the mission and reference data (28′,28″), and balancing (144) phases and amplitudes of the mission data (28′). The method (100) then concatenates (160) the patches (26) to produce processed mission and reference images (52′,52″), and subtracts (162) the processed mission and reference images (52′,52″) to form a delta image (54). The method (100) then postprocesses (164,170) the delta image (54) for specific applications.

Claims

1. A method effects coherent change subtraction of mission synthetic aperture radar (SAR) data and reference SAR data for suppression of combined noise and backscatter energy, said method comprising:

a) forming a mission image from said mission SAR data;

b) forming a reference image from said reference SAR data;

c) registering said mission and reference images to form registered mission and reference images;

d) forming said registered mission and reference images into a plurality of patches of said registered mission and reference images, wherein each of said patches contains mission data, being data from said registered mission image, and reference data, being data from said registered reference image, and wherein each of said patches is substantially equal to a separate portion of said registered mission and reference images;

e) processing said mission and reference data in each patch of said plurality of patches to produce processed mission and reference data;

f) producing a processed mission image and a processed reference image by concatenating said processed mission and reference data from each of said plurality of patches;

g) subtracting a first one of said processed mission and reference images from a second one of said processed mission and reference images to form a delta image, wherein said mission image, said reference image, and said delta image each have substantially a like number of pixels; and

h) postprocessing said delta image, said postprocessing activity h) comprising;

determining, for each pixel of said delta image, if said each pixel has a clutter-to-noise ratio of less than one; and

substituting, when said determining activity determines said each pixel has a clutter-to-noise ratio of less than one, a replacement pixel for said each pixel, wherein said replacement pixel is a coincident pixel from one of said processed mission image, said processed reference image, and said delta image having a lowest amplitude ratio.

2. A method as claimed in claim 1 wherein:

said forming activity a) forms said mission image from said mission SAR data so that said mission image bears a predetermined relationship to said mission SAR data; and

said forming activity b) forms said reference image from said reference SAR data so that said reference image bears substantially said predetermined relationship to said reference SAR data.

3. A method as claimed in claim 1 wherein said forming activity a) comprises:

flattening a range spectrum of said mission SAR data;

producing said mission image in response to said mission SAR data;

flattening an azimuth spectrum of said mission image.

4. A method as claimed in claim 3 wherein:

said range spectrum contains holes; and

said forming activity a) additionally comprises interpolating over said holes prior to said producing activity.

5. A method as claimed in claim 1 wherein said registering activity c) registers said mission and reference images to better than 0.5 pixel.

6. A method for coherent change subtraction of mission synthetic aperture radar (SAR) data and reference SAR data, said method comprising:

forming a mission image from said mission SAR data;

forming a reference image from said reference SAR data;

registering said mission and reference images to form registered mission and reference images;

forming at least one patch of said registered mission and reference images, wherein said at least one patch contains mission data, being data from said registered mission image, and reference data, being data from said registered reference image;

processing said at least one patch, said processing activity e) comprising;

a) removing linear phase terms from at least said mission data in said at least one patch in a frequency domain;

b) removing linear phase terms from at least said mission data in said at least one patch in an image domain;

c) trimming non-overlapping spectrum from said mission data and reference data in said at least one patch in said frequency domain;

d) balancing phases and amplitudes of said mission data in said at least one patch in said frequency domain; and

e) balancing phases and amplitudes of said mission data in said at least one patch in said image domain producing a processed mission image and a processed reference image from said mission data and said reference data, respectively;

subtracting a first one of said processed mission and reference images from a second one of said processed mission and reference images to form a delta image; and

postprocessing said delta image.

7. A method as claimed in claim 6 wherein said method effects said coherent change subtraction for detection of a change in one of an amplitude and a phase of backscatter, and wherein:

said forming activity d) forms said registered mission and reference images into one patch substantially equal to the whole of said registered mission and reference image;

said processing activity e) process said mission and reference data in said one patch to produce processed mission and reference data;

said producing activity f) produces said processed mission and reference images from said processed mission and reference data;

said postprocessing activity h) comprises;

computing one of an amplitude and a phase of said delta image; and

detecting a target in said delta image in response to said one of said amplitude and said phase.

8. A method as claimed in claim 7 wherein said detecting activity detects said target using constant false alarm rate detection.

9. A method as claimed in claim 6 wherein each of said removing activities a) and b) comprises:

determining phase differences between phases of said mission data and phases of said reference data;

computing said linear phase terms of said phase differences; and

extracting said linear phase terms from at least said mission data.

10. A method as claimed in claim 9 wherein said extracting activity extracts substantially all of said linear phase terms from said mission data.

11. A method as claimed in claim 9 wherein said extracting activity extracts substantially a first half of said linear phase terms from said mission data and a second half of said linear phase terms from said reference data.

12. A method as claimed in claim 6 wherein each of said balancing activities d and e comprises:

determining phase differences between phases of said mission data and phases of said reference data;

determining amplitude ratios between amplitudes of said mission data and amplitudes of said reference data;

subtracting said phase differences from phases of said mission data; and

multiplying amplitudes of said mission data by said amplitude ratios.

13. A method for coherent change subtraction of mission synthetic aperture radar (SAR) data and reference SAR data for detection of a change in one of an amplitude and a phase of backscatter, said method comprising:

a) forming a mission image from said mission SAR data;

b) forming a reference image from said reference SAR data;

c) registering said mission and reference images on a common plane to produce registered mission and reference images;

d) removing linear phase terms from at least said registered mission image in a frequency domain;

e) removing linear phase terms from at least said registered mission image in an image domain;

f) trimming non-overlapping spectra of said registered mission and reference images;

g) balancing phases and amplitudes of said mission and reference images in said frequency domain;

h) balancing phases and amplitudes of said mission and reference images in said image domain;

i) subtracting a first one of said processed mission and reference images from a second one of said processed mission and reference images to form a delta image; and

j) computing one of an amplitude and a phase of said delta image; and

k) detecting, in response to said one of said amplitude and said phase of said delta image, a target in said delta image using constant false alarm rate detection.

14. A method as claimed in claim 13 wherein said forming activity a) comprises:

reading said mission SAR data;

flattening a range spectrum of said mission SAR data;

producing said mission image in response to said mission SAR data; and

flattening an azimuth spectrum of said mission image.

15. A method as claimed in claim 13 wherein said forming activity b) comprises:

reading said reference SAR data;

flattening a range spectrum of said reference SAR data;

producing said reference image in response to said reference. SAR data; and

flattening an azimuth spectrum of said reference image.

16. A method as claimed in claim 13 wherein each of said removing activities d) and e) comprises:

determining phase differences between phases of said mission image and phases of said reference image;

computing linear phase terms of said phase differences; and

extracting said linear phase terms from at least said mission image.

17. A method as claimed in claim 13 wherein each of said balancing activities g) and h) comprises:

determining phase difference between phases of said mission data and phases of said reference images;

determining amplitude ratios between amplitudes of said mission image and amplitudes of said reference image;

subtracting said phase differences from phases of said mission image; and

multiplying amplitudes of said mission image by said amplitude ratios.

18. A method for coherent change subtraction of mission synthetic aperture radar (SAR) data and reference SAR data for suppression of combined noise and backscatter energy, said method comprising:

a) forming a mission image from said mission SAR data;

b) forming a reference image from said reference SAR data;

c) registering said mission and reference images on a common plane to form registered mission and reference images;

d) forming a plurality of patches containing mission data, being data from said registered mission image, and reference data, being data from said registered reference image, wherein each of said patches in said plurality of patches is substantially equal to separate substantially coincident portions of said registered mission and reference images;

e) removing linear phase terms from at least said mission data in each of said patches in each of a frequency domain and an image domain;

f) trimming non-overlapping spectra of said mission data and said reference data in each of said patches;

g) balancing phases and amplitudes of said mission and reference data in each of said patches in each of said frequency and image domains;

h) concatenating each of said patches to produce a processed mission image and a processed reference image from said mission data and said reference data, respectively;

i) subtracting a first one of said processed mission and reference images from a second one of said processed mission and reference images to form a delta image, wherein said processed mission image, said processed reference image, and said delta image each have substantially a like number of pixels;

j) determining, for each pixel of said processed reference image, if said each pixel has a clutter-to-noise ratio of less than one; and

k) substituting, when said determining activity j) determines said each pixel has a clutter-to-noise ratio of less than one, a replacement pixel for said each pixel, wherein said replacement pixel is a coincident pixel from one of said processed mission image, said processed reference image, and said delta image having a lowest amplitude ratio.

19. A method as claimed in claim 18 wherein, in each of said frequency and image domains, said removing activity e) comprises:

determining phase differences between a phase of said mission data and a phase of said reference data;

computing said linear phase terms of said phase difference; and

extracting said linear phase terms from at least said mission data.

20. A method as claimed in claim 19 wherein said extracting activity extracts a first portion of said linear phase terms from said mission data and a second portion of said linear phase terms from said reference data.

21. A method as claimed in claim 18 wherein, in each of said frequency and image domains, said balancing activity g) comprises:

determining phase differences between phases of said mission data and phases of said reference data;

filtering said phase differences to produce filtered phase differences;

determining amplitude ratios between amplitudes of said mission data and amplitudes of said reference data;

filtering said amplitude ratios to produce filtered amplitude ratios;

subtracting said filtered phase differences from phases of said mission data; and

multiplying amplitudes of said mission data by said filtered amplitude ratios.

22. A computer program for coherent change subtraction of mission synthetic aperture radar (SAR) data and reference SAR data for one of detection of a change in one of an amplitude and a phase of backscatter and suppression of combined noise and backscatter energy, said program comprising:

a first program segment configured to produce a mission image from said mission SAR data;

a second program segment configured to produce a reference image from said reference SAR data;

a third program segment configured to register each of said mission and reference images on a common plane to produce registered mission and reference images;

a fourth program segment configured to form at least one patch containing mission data, being data from said mission image, and reference data, being data from said reference image, wherein;

when said program performs said coherent change subtraction for said detection of a change in one of an amplitude and a phase of backscatter, said fourth program segment forms one patch substantially equal to the whole of said mission and reference complex images; and

when said program performs said coherent change subtraction for said suppression of combined noise and backscatter energy, said fourth program segment forms a plurality of patches, wherein each of said patches in said plurality of patches is substantially equal to separate substantially coincident portions of said registered mission and reference images;

a fifth program segment configured to process said at least one patch to produce at least one processed patch;

a sixth program segment configured to concatenate said at least one processed patch to produce processed mission and reference images;

a seventh program segment configured to subtract a first one of said processed mission and reference images from a second one of said processed mission and reference images to form a delta image, wherein said processed mission image, said processed complex image, and said delta image each have substantially a like number of pixels; and

an eighth program segment wherein;

when said program performs said coherent change subtraction detection of a change in one of an amplitude and a phase of backscatter, said eighth program segment is configured to compute an amplitude of said delta complex image, and detect said moving target in said delta complex image in response to said amplitude; and

when said program performs said coherent change subtraction for said suppression of combined noise and backscatter energy, said eighth program segment is configured to substitute, for each pixel of said delta complex image, when a coincident pixel of said processed reference image has a clutter-to-noise ratio of less than one, a replacement pixel, wherein said replacement pixel is a coincident pixel from one of said mission, reference, and delta complex images having a lowest amplitude ratio.

23. A computer program as claimed in claim 22 wherein said first program segment comprises:

a first routine configured to read said mission SAR data;

a second, routine configured to flatten a range spectrum of said mission SAR data;

a third routine configured to interpolate over holes in said range spectrum;

a fourth routine configured to produce said mission image in response to said mission SAR data; and

a fifth routine configured to flatten an azimuth spectrum of said mission image.

24. A computer program as claimed in claim 22 wherein said second program segment comprises:

a first routine configured to read said reference SAR data;

a second routine configured to flatten a range spectrum of said reference SAR data;

a third routine configured to interpolate over holes in said range spectrum;

a fourth routine configured to produce said reference image in response to said reference SAR data; and

a fifth routine configured to flatten an azimuth spectrum of said reference image.

25. A computer program as claimed in claim 22 wherein said fifth program segment comprises:

a first routine configured to remove linear phase terms from said mission data in said at least one patch in each of a frequency domain and an image domain;

a second routine configured to trim non-overlapping spectra from said mission data and said reference data in said at least one patch; and

a third routine configured to balance phases and amplitudes of said mission data in said at least one patch in each of said frequency and image domains.