Method for phase unwrapping in imaging systems

• US 6,011,625 A
• Filed: 07/08/1998
• Issued: 01/04/2000
• Est. Priority Date: 07/08/1998
• Status: Expired due to Fees

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Glass, Carter M.
• Primary Examiner(s): Kim, Robert H.
• Assistant Examiner(s): Lee, Andrew H.
• Application Number:
  US09/112,049
• Time in Patent Office: 545 Days
• Field of Search: 356/357, 356/359, 356/360, 342/25
• US Class Current: 356/496
• CPC Class Codes: G01S 13/9023 combined with interferometric techniques

Abstract

An improved method is provided for phase unwrapping in coherent imaging systems which use two complex images to form an interferogram. In the improved method, the interferogram of phase differences is divided into a relatively large plurality of transform blocks, wherein each transform block comprises a matrix of o×p data samples, with o and p=2ⁿ +1, and n being an integer greater than 2. Further, adjacent transform blocks are preferably defined to partially overlap. The wrapped phase values of adjacent data samples within each transform block are then compared and if the difference therebetween exceeds a predetermined value, the entire block of data samples may be discarded for further use. Data samples of the retained transform blocks are then unwrapped to obtain phase values via an unweighted, least-squares technique implemented by Fast Fourier Transform. Using a known height value corresponding with a single data sample, an integration constant can be determined for a corresponding first transform block. Path-following may then be used to determine an integration constant for the other retained transform blocks. The phase values and integration constants may then be employed to determine height values for the interferogram. The disclosed method yields enhanced accuracy with reduced computational burden.

Claims

1. A method for determining height information from an interferogram of complex data samples generated from two complex images of an imaged terrain region acquired by an interferometric synthetic aperture radar system, wherein each of said complex data samples comprises a wrapped phase value, comprising:

dividing said interferogram into a plurality of transform blocks, wherein each transform block comprises a plurality of complex data samples, and wherein each of said transform blocks partially overlaps at least one adjacent, block in a corresponding overlapping region;

determining a difference value between the wrapped phase values of each set of adjacent complex data samples comprising a given transform block;

comparing each said difference value to a predetermined value, wherein all complex data samples comprising a given transform block are discarded when any single difference value corresponding with the transform block exceeds said predetermined value, and wherein all complex data samples comprising a transform block are retained when all difference values corresponding with the transform block are less than the predetermined value;

unwrapping a phase difference value for each of a plurality of retained data samples comprising an overlapping region corresponding with first and second transform blocks; and

using a known height value corresponding with one of said plurality of complex data samples comprising said first transform block to establish an integration constant for the first transform block; and

employing said integration constant for said first transform block and said unwrapped phase difference values to determine an integration constant for at least said second transform block.

2. A method as claimed in claim 1, said employing step comprising:

averaging the unwrapped phase difference values for the overlapping region of the first transform block to obtain a first average;

averaging the unwrapped phase difference values for the overlapping region of the second transform block to obtain a second average; and

using the first average and the second average to determine the integration constant for the second transform block.

3. A method as claimed in claim 1, further comprising:

generating an integration constant for each of a first plurality of said transform blocks using said integration constraints for said first and second transform blocks and path-following.

4. A method as claimed in claim 3, further comprising:

smoothing the complex data samples comprising the interferogram using a spanning filter of a predetermined size, wherein said predetermined size is less than about one-half of said plurality of complex data samples comprising each said transform block.

5. A method as claimed in claim 1, wherein the total number of data samples comprising any one of said plurality of transform blocks is at least about 100.

6. A method as claimed in claim 5, wherein each transform block comprises a matrix of o×

p data samples, where o and p=2ⁿ 1, and wherein n is an integer greater than 2.

7. A method as claimed in claim 5, wherein said overlapping region of comprises a total number of complex data samples that is between about 5% and 50% of the total number of data samples in the first transform block.

8. A method as claimed in claim 7, wherein said overlapping region is at least four complex data samples wide.

9. A method as claimed in claim 1, said unwrapping step comprising:

using a least-squares algorithm.

10. A method as claimed in claim 9, wherein said least squares algorithm is implemented by Fast Fourier Transform.

11. A method as claimed in claim 1, wherein said predetermined value is less than about 135°

.

12. A method as claimed in claim 1, wherein said predetermined value is less than about 90°

.

13. A method as claimed in claim 1, further comprising:

generating an irtegration constant for each of a first plurality of said retained transform blocks using said integration constants for said first and second transform blocks and path-following;

using a known height value corresponding with one of said plurality of complex data samples comprising a third transform block to establish an integration constant for the third transform block, wherein said third transform block is separated from said first plurality of said retained transform blocks;

generating an integration constant for each of a second plurality of said retained transform blocks using said integration constant for said third transform block and path-following.

14. A method as claimed in claim 1, further comprising:

using said integration constant for said second transform block to determine height values for the data samples comprising the second transform block; and

employing a known height value corresponding with one of said plurality of complex data samples comprising one of said transform blocks, other than said first transform block, to determine a height adjustment value for application to said heights determined for said complex data samples comprising said second transform block.

15. A method for determining height information from an interferogram of complex data samples generated from two complex images of an imaged terrain region acquired by an interferometric synthetic aperture radar system, comprising:

dividing said irterferogram into a plurality of transform blocks, wherein each transform block comprises a plurality of complex data samples that each include a wrapped phase value;

determining a difference value between the wrapped phase values of each set of adjacent complex data samples comprising a transform block;

comparing each said difference value to a predetermined value, wherein all complex data samples comprising a transform block are discarded from further use in the method when any single difference value corresponding with the transform block exceeds said predetermined value, and wherein all complex data samples comprising a transform block are retained for further use in the method when all difference values corresponding with the transform block are less, than the predetermined value;

unwrapping a phase difference value for each of the retained complex data samples; and

determining height values in corresponding relation to at least a portion of said plurality of retained complex data samples utilizing corresponding unwrapped phase difference values.

16. A method as claimed in claim 15, said unwrapping step comprising:

using a least-squares algorithm.

17. A method as claimed in claim 16, wherein said least-squares algorithm is implemented by Fast-Fourier Transform.

18. A method as claimed in claim 15, wherein each transform block comprises a matrix of o×

p data samples, where o and p=2_n+ 1, and wherein n is an integer greater than 2.

19. A method as claimed in claim 15, said determining step comprising:

using a known height value corresponding with one of said plurality of complex data samples comprising a first retained transform block to establish an integration constant for the first retained transform block; and

employing said integration constant for said first retained transform block and said unwrapped phase difference values to determine an integration constant for at least a second retained transform block.

20. A method as claimed in claim 19, said determining step further comprising:

using a predetermined phase-to-height conversion factor.

21. A method for determining height information from an interferogram of complex data samples generated from two complex images of an imaged terrain region acquired by an interferometric synthetic aperture radar system, comprising:

dividing said interferogram into a plurality of transform blocks, wherein each transform block comprises a plurality of complex data samples, and wherein at least a first of said transform blocks partially overlaps an adjacent, second transform block in an overlapping region, wherein said overlapping region includes complex data samples that partially comprise each of said first and second transform blocks;

unwrapping a phase difference value for each of the complex data samples comprising said overlapping region of said first and second transform blocks;

using a known height value corresponding with one of said plurality of complex data samples comprising said first transform block to establish an integration constant for the first transform block; and

employing said integration constant for the first transform block and said unwrapped phase difference values to determine an integration constant for at least said second transform block.

22. A method as claimed in claim 21, said employing step comprising:

averaging the unwrapped phase difference values for the overlapping region of the first transform block to obtain a first average;

averaging the unwrapped phase difference values for the overlapping region of the second transform block to obtain a second average; and

using the first average and the second average to determine the integration constant for the second transform block.

23. A method as claimed in claim 21, said unwrapping step comprising:

use of a least-squares algorithm implemented by Fast Fourier Transform.

24. A method as claimed in claim 21, further comprising:

analyzing of each of said plurality of transform blocks on a transform block-by-transform block basis to identify any inconsistencies in the complex data samples comprising each given transform block, wherein the only complex data samples utilized in the analysis of a given transform block are those corresponding with the given transform block being analyzed, and wherein all complex data samples comprising a given transform block are discarded when an inconsistency is identified with respect to the transform block and wherein all complex data samples comprising a given transform block are retained in the absence of the identification of an inconsistency corresponding with the transform block.

25. A method as claimed in claim 24, wherein each of said complex data samples comprises a wrapped phase value, and wherein for each of said plurality of transform blocks said analyzing step comprises:

determining a difference value between the wrapped phase values of each set of adjacent complex data samples comprising a transform block;

comparing each said difference value to a predetermine value, wherein all complex data samples comprising a transform block are discarded from further use in the method when any single difference value corresponding with the transform block exceeds said predetermined value, and wherein all complex data samples comprising a transform block are retained for further use in the method when all difference values corresponding with the transform block are less than the predetermined value.

26. A method as claimed in claim 24, wherein said analyzing step is completed after said unwrapping step and comprises:

rewrapping the unwrapped, complex data samples comprising each of said transform blocks to obtain rewrapped pixel data;

using said rewrapped pixel data for a given transform block and pixel data corresponding with said complex data samples for said given transform block to obtain a difference function;

wrapping difference data corresponding with said difference function to identify any said inconsistency.

27. A method as claimed in claim 24, further comprising:

generating an integration constant for each of a first plurality of said retained transform blocks using said integration constants for said first and second transform blocks and path-following;

using a known height value corresponding with one of said plurality of complex data samples comprising a third transform block to establish an integrating constant for the third transform block, wherein said third transform block is separated from said first plurality of said retained transform blocks;

generating an integration constant for each of a second plurality of said retained transform blocks using said integration constant for said third transform block and path-following.