System and method for comparing signals

• US 6,622,118 B1
• Filed: 06/12/2001
• Issued: 09/16/2003
• Est. Priority Date: 03/13/2001
• Status: Expired due to Term

• Current Assignee: Aido LLC (Endpoint IP LLC)
• Original Assignee: Alphatech, Inc. (BAE Systems Plc)
• Inventors: Verbout, Shawn M., Crooks, Steven M.
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): Desta, Elias
• Application Number:
  US09/879,736
• Time in Patent Office: 826 Days
• Field of Search: 702/71-77, 702/106, 702/107, 702/124, 702/126, 702/183, 702/189, 702/193, 702/194, 702/197, 702/FOR.103, 702/104, 702/108, 702/110, 702/134, 702/135, 702/164, 702/168, 702/170, 702/171, 342/25, 342/91, 342/173, 342/174, 708/3-5, 708/8-9, 708/200, 708/201, 708/205, 708/277, 708/300, 708/314, 708/315, 708/403, 708/420, 708/514, 708/520, 382/190, 382/191, 382/260, 382/263, 382/264, 382/270, 382/278, 382/280, 382/295, 382/296
• US Class Current: 702/190
• CPC Class Codes: G01S 13/9027 Pattern recognition for feature extraction
• CPC Class Codes: G01S 13/904 SAR modes
• CPC Class Codes: G01S 7/292 Extracting wanted echo-signals

Abstract

A method and system that include a first measurement signal and a second measurement signal that can be input to first and second filters. The filters can be subject to a first constraint to minimize the energy difference between the first and second measurement signals on a per frequency basis, and subject to a second constraint that includes a model frequency and phase response. By adapting the filters subject to the two constraints, coherent differences between the two measurement signals can be identified. In one embodiment, the system can be applied to Synthetic Aperture Radar (SAR) data.

Claims

1. A method for comparing a first signal and a second signal, the method comprising,

applying a first filter to the first signal,

applying a second filter to the second signal,

constraining the first filter and the second filter to minimize the energy difference between the filtered first signal and the filtered second signal and based on a model response having a model magnitude and a model phase, and,

determining signal components for which the energy difference exceeds a threshold.

2. A method according to claim 1, wherein constraining the first filter and the second filter further includes constraining the first filter and the second filter to provide unity magnitude.

3. A method according to claim 1, wherein applying the first filter to the first signal and the second filter to the second signal further includes,

constraining the first filter to provide unity magnitude and zero phase, and,

constraining the second filter to provide unity magnitude and variable phase.

4. A method according to claim 1, wherein constraining the first filter and the second filter further includes minimizing the energy based on phase.

5. A method according to claim 1, wherein constraining the first filter and the second filter further includes constraining based on the phase of the model response.

6. A method according to claim 1, wherein providing a first signal includes providing a first vector.

7. A method according to claim 1, wherein providing a second signal includes providing a second vector.

8. A method according to claim 1, wherein providing a first signal includes,

providing image data,

computing a two-dimensional, inverse discrete Fourier transform using the image data,

converting the inverse Fourier transform data to a one-dimensional vector.

9. A method according to claim 8, further including removing an extant aperture weighting from the inverse Fourier transform data.

10. A method according to claim 1, wherein providing a second signal includes,

providing image data,

computing a two-dimensional, inverse discrete Fourier transform using the image data,

converting the inverse Fourier transform data to a one-dimensional vector.

11. A method according to claim 1, further including computing an adaptive threshold, and wherein determining signal components includes determining signal components for which the mean energy difference exceeds the adaptive threshold.

12. A method according to claim 1, wherein determining the signal components for which the energy difference exceeds a threshold further includes determining at least one frequency for which the mean energy difference exceeds a threshold.

13. A method according to claim 1, wherein minimizing the energy difference includes minimizing the mean energy difference.

14. A method according to claim 1, wherein applying a first filter to the first signal and applying a second filter to the second signal includes applying at least one finite impulse response filter.

15. A method according to claim 1, wherein applying a first filter to the first signal and applying a second filter to the second signal includes,

computing a first inner product to generate a first scalar, and,

computing a second inner product to generate a second scalar.

16. A method according to claim 1, wherein constraining the first filter and the second filter to minimize the energy difference includes minimizing at least one compensation term based on the first filter and the second filter.

17. A method according to claim 1, wherein constraining the first filter and the second filter to minimize the energy difference includes minimizing a summation that includes a difference between the first filtered signal and the second filtered signal, and at least one compensation term.

18. A method according to claim 1, wherein constraining the first filter and the second filter to minimize the energy difference includes,

selecting a compensation term based on the first filter,

selecting a compensation term based on the second filter, and,

minimizing a sum including a product of the first compensation term and the first filter'"'"'s squared Euclidean norm, and a product of the second compensation term and the second filter'"'"'s squared Euclidean norm.

19. A method according to claim 1, wherein constraining the first filter and the second filter to minimize the energy difference includes minimizing the energy on a per frequency basis.

20. A method of providing coherent change detection, the method comprising,

applying a first filter to the first signal,

applying a second filter to the second signal,

constraining the first filter and the second filter to minimize the energy difference between the filtered first signal and the filtered second signal, based on at least one model response, and,

determining signal components for which the mean energy difference exceeds a threshold.

21. A method according to claim 20, wherein constraining based on at least one model response further includes constraining at least one of the first filter and the second filter to provide unity gain.

22. A method according to claim 20, wherein constraining based on at least one model response further includes constraining at least one of the first filter and the second filter to provide variable phase.

23. A method according to claim 20, wherein constraining further includes minimizing the energy difference based on phase.

24. A method according to claim 20, wherein constraining further includes minimizing the mean energy difference.

25. A method according to claim 20, further including discretizing the first signal and the second signal.

26. A method according to claim 20, further including computing an adaptive threshold, and wherein determining signal components includes determining signal components for which the energy difference exceeds the adaptive threshold.

27. A method according to claim 20, wherein constraining the first filter and the second filter to minimize the energy difference includes minimizing at least one compensation term based on the first filter and the second filter.

28. A method according to claim 20, wherein constraining the first filter and the second filter to minimize the energy difference includes minimizing the energy difference on a per frequency basis.

29. A system for comparing a first signal and a second signal, comprising,

a first filter to filter the first signal,

a second filter to filter the second signal, and,

a constraint module to adapt at least one of the first filter and the second filter to minimize an energy difference between the first-filtered signal and the second filtered signal, based on at least one model response having at least one model phase and at least one model magnitude.

30. A system according to claim 29, further including,

a first model filter response to which the first filter output is constrained, and,

a second model filter response to which the second filter output is constrained.

31. A system according to claim 29, wherein

the first model filter response includes unity magnitude and zero phase, and,

the second model filter response includes unity magnitude and variable phase.

32. A system according to claim 29, further including a threshold to which the energy signal can be compared.

33. A method according to claim 29, wherein at least one of the first signal and the second signal are derived from at least one of a first image and a second image.

34. A method of comparing a first data signal and a second data signal, the first and second data signals from a coherent sensor, the method comprising,

providing a first one-dimensional vector of length M based on the first data signal,

providing a second one-dimensional vector-of length M based on the second data signal,

providing a first filter vector of length M,

providing a second filter vector of length M,

computing at least one first inner product between the first vector and the first filter vector,

computing at least one second inner product between the second vector and the second filter vector,

squaring the differences of the at least one first scalar products and the at least one second inner products,

determining coefficients for the first filter vector and the second filter vector based on minimizing the squared differences, and,

comparing the squared differences to a threshold.

35. A method according to claim 34, further including,

computing a first squared Euclidean norm based on the first filter vector,

computing a second squared Euclidean norm based on the second filter vector,

determining a first multiplier for the first squared Euclidean norm,

determining a second multiplier for the second squared Euclidean norm,

and wherein minimizing the squared differences includes minimizing the additive quantity that includes the first multiplier multiplied by the first squared Euclidean norm, added to the second multiplier multiplied by the second Euclidean norm.

36. A method according to claim 34, wherein providing a first one-dimensional data vector includes providing the first data signal based on a first image.

37. A method according to claim 34, wherein providing a second one-dimensional data vector includes providing the second data signal based on a second image.

38. A method according to claim 34, wherein at least one of the first data signal and the second data signal are of length N, wherein N is greater than M.

39. A system for comparing a first signal and a second signal, comprising,

first filter means to filter the first signal,

second filter means to filter the second signal, and,

means to constrain at least one of the first filter and the second filter to minimize an energy difference between the first filtered signal and the second filtered signal, based on at least one model response having at least one model phase and at least one model magnitude.

40. A system according to claim 39, further including,

means to provide a first model filter response to which the first filter output is constrained, and,

means to provide a second model filter response to which the second filter output is constrained.

41. A system according to claim 39, wherein

the first model filter response includes unity magnitude and zero phase, and,

the second model filter response includes unity magnitude and variable phase.

42. A system according to claim 39, further including a threshold means for providing a threshold to which the energy signal can be compared.

43. A computer program product disposed on a computer readable medium, for comparing a first signal and a second signal, the computer program product having a processor with instructions for causing the processor to,

apply a first filter to the first signal,

apply a second filter to the second signal,

constrain the first filter and the second filter to minimize the energy difference between the filtered first signal and the filtered second signal and based on a model response having a model magnitude and a model phase, and,

determine signal components for which the energy difference exceeds a threshold.

44. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter further include instructions to constrain the first filter and the second filter to provide unity magnitude.

45. A computer program product according to claim 43, wherein instructions to apply the first filter to the first signal and the second filter to the second signal further include instructions to,

constrain the first filter to provide unity magnitude and zero phase, and,

constrain the second filter to provide unity magnitude and variable phase.

46. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter further include instructions to minimize the energy based on phase.

47. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter further include instructions to constrain based on the phase of the model response.

48. A computer program product according to claim 43, wherein instructions to provide a first signal include instructions to provide a first vector.

49. A computer program product according to claim 43, wherein instructions to provide a second signal include instructions to provide a second vector.

50. A computer program product according to claim 43, wherein instructions to provide a first signal include instructions to,

provide image data,

compute a two-dimensional, inverse discrete Fourier transform using the image data,

convert the inverse Fourier transform data to a one-dimensional vector.

51. A computer program product according to claim 43, wherein instructions to provide a second signal include instructions to,

provide image data,

compute a two-dimensional, inverse discrete Fourier transform using the image data,

convert the inverse Fourier transform data to a one-dimensional vector.

52. A computer program product according to claim 43, further including instructions to compute an adaptive threshold, and wherein instructions to determine signal components include instructions to determine signal components for which the mean energy difference exceeds the adaptive threshold.

53. A computer program product according to claim 43, wherein instructions to determine the signal components for which the energy difference exceeds a threshold further include instructions to determine at least one frequency for which the mean energy difference exceeds a threshold.

54. A computer program product according to claim 43, wherein instructions to minimize the energy difference include instructions to minimize the mean energy difference.

55. A computer program product according to claim 43, wherein instructions to apply a first filter to the first signal and applying a second filter to the second signal include instructions to apply at least one finite impulse response filter.

56. A computer program product according to claim 43, wherein instructions to apply a first filter to the first signal and applying a second filter to the second signal include instructions to,

compute a first inner product to generate a first scalar, and,

compute a second inner product to generate a second scalar.

57. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter to minimize the energy difference include instructions to minimize at least one compensation term based on the first filter and the second filter.

58. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter to minimize the energy difference include instructions to minimize a summation that includes a difference between the first filtered signal and the second filtered signal, and at least one compensation term.

59. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter to minimize the energy difference include instructions to,

select a compensation term based on the first filter,

select a compensation term based on the second filter, and,

minimize a sum including a product of the first compensation term and the first filter'"'"'s squared Euclidean norm, and

a product of the second compensation term and the second filter'"'"'s squared Euclidean norm.

60. A computer program product according to claim 43, wherein instructions to constrain the first filter and the second filter to minimize the energy difference include instructions to minimize the energy on a per frequency basis.