Communications: directive radio wave systems and devices (e.g. – Synthetic aperture radar
Reexamination Certificate
2000-02-22
2001-01-30
Sotomayor, John B. (Department: 3662)
Communications: directive radio wave systems and devices (e.g.,
Synthetic aperture radar
C342S174000, C342S189000, C342S196000
Reexamination Certificate
active
06181270
ABSTRACT:
FIELD OF THE INVENTION
This invention related generally to signal processing and, in particular, to the removal of phase error in an uncompensated dataset, including datasets representative of synthetic aperture radar (SAR) images.
BACKGROUND OF THE INVENTION
In signal-processing systems having two partially correlated data sets with relative spectral errors, phase-related problems may arise. Such is the case with interferometric synthetic aperture radar (IFSAR) systems operating in a bistatic mode. In a bistatic configuration, only one antenna of a pair transmits, but both antennas receive simultaneously. This is in contrast to monostatic IFSAR operation, wherein both antennas transmit and receive essentially independently on alternate pulses. In a monostatic mode, the same local oscillator (LO) is used for both the transmit and receive functions. As such, the LO phase noise which degrades the SAR output is the result of a first difference operation on the phase noise with a time offset equal to the roundtrip range delay. This provides significant cancellation for the low-frequency components of the LO phase noise (common-mode cancellation).
In a bistatic mode of operation, however, the transmit and receive functions may use independent local oscillators. Accordingly, there is no phase noise cancellation in the bistatic channel of such an interferometric system. In fact, the two noise sources add, and the phase noise requirements on both LOs are much more severe. (see, for example, J. Autennan, “Phase Stability Requirements for a Bistatic SAR”, Proceedings of the IEEE National Radar Conference, March 1994).
These LO errors become phase errors in the demodulated SAR data. These, in turn, contribute to a loss of resolution, high sidelobes and smearing of the image data. In an IFSAR system, LO errors also cause loss of correlation in the interferometer comparison. The phase information contained in the IFSAR correlation is normally converted to elevation map data via several additional processing steps. Due to the loss of correlation, however, the phase will be noisy, and the elevation data accuracy will also be degraded.
In a bistatic mode of operation, extreme stability or some other compensation technique is required to mitigate IFSAR bistatic phase errors. In terms of stability, one approach is to use high-quality atomic clocks; another approach is to actively lock the two oscillators together. These solutions could have potentially significant cost impact, particularly in a spaceborne system. With high-quality oscillators, the problems with bistatic operation typically causes low-frequency phase errors and high-azimuth sidelobes in the bistatic image channel for modest aperture times (i.e., several seconds). Autofocus approaches, both quadratic and higher-order, can help to alleviate this problem, but the degree of compensation is only partial, and still requires high-quality clock stability. In addition, this technique may not be reliable in all terrain clutter environments.
SUMMARY OF THE INVENTION
The present invention improves upon the prior art by providing phase-correction procedures that may be implemented without the need for specialized spaceborne hardware. The technique, generally referred to as reference-based autofocusing, or RBA, is particularly suited to existing and future interferometric SAR systems, though the principles are generally applicable to signal processing systems for which there are two partially correlated data sets having relative spectral errors. In an airbourne SAR system, the technique only requires some additional steps in the ground processor functions.
Broadly, the invention takes advantage of the fact that in a bistatic system, one antenna phase center both transmits and receives with the usual common-mode cancellation, and can thus be expected to form a reasonably well-focused image. In the second system, with the degraded phase error response, the image provided by the first system is used as a coherent reference to aid the estimation and removal of the relative phase errors between the two. Thus, the methodology uses the initially degraded image pair correlation data to help estimate and remove errors in a way which naturally maximizes the final correlation level obtained.
In an interferometric synthetic aperture radar (IFSAR) system producing complex images, A(x,y) and B(x,y), where A represents a monostatic image and B is a degraded bistatic image having an initial phase error, and where (x,y) are azimuth and range sample coordinates, respectively, a preferred method of reducing phase error would include the following important steps:
performing a normalized, complex, cross-correlation operation on A and B to obtain a result, &mgr;(x,y);
multiplying &mgr; and B on a pixel-by-pixel basis to obtain a result, B′(x,y);
performing one-dimensional FFT on A and B′ in the azimuth dimension to obtain a
y
(t) and b
y
′(t);
performing a complex, cross-correlation operation on a
y
(t) and b
y
′(t), on a sample-by-sample basis, and normalizing the result to obtain a phase-error correction signal, exp(j&phgr;(t));
performing a one-dimensional FFT on B to obtain b
y
(t);
complex multiplying b
y
(t) with exp(j&phgr;(t)) on a line-by-line basis to obtain a complex compensation signal, bc
y
(t);
performing a one-dimensional FFT on bc
y
(t) to obtain a compensated image, Bc(x,y); and
cross-correllating A and Bc to obtain a compensated correlation image, &mgr;c(x,y).
A and B may represent an entire image, or a selected subregion or subregions of the whole image, in which case the regions are preferably significantly larger in the azimuth dimension than the amount of smearing expected from the phase errors. The subregions are also preferably selected to contain statistically significant amounts of the best correlated data. The correlation magnitude or various possible linear or non-linear thresholded values of this correlation magnitude may also be used to automatically select the subregions of the whole image to be used. The correlation levels observed are typically a positively biased statistical distribution, in that the actual correlation of the underlying complex data is less than the average correlation level measured. This bias may be estimated by various methods and removed. The corrected correlation values may then be thresholded at some minimum value, and in minimum size blocks of pixels, to determine adjusted weights, w(x,y) and m(x,y), for the subsequent compensation steps according to an alternative embodiment.
REFERENCES:
patent: 4786906 (1988-11-01), Krogager
patent: 4924229 (1990-05-01), Eichel et al.
patent: 5248976 (1993-09-01), Hiho et al.
patent: 5327140 (1994-07-01), Buckreuss
patent: 5343204 (1994-08-01), Farmer et al.
patent: 5608404 (1997-03-01), Burns
patent: 5952955 (1999-07-01), Kennedy et al.
J. Auterman, “Phase Stability Requirements for a Bistatic SAR,” Proceedings of the IEEE National Radar Conference, Mar. 13-14, 1984.
Gifford Krass Groh Sprinkle Anderson & Citkowski PC
Sotomayor John B.
Veridian ERIM International, Inc.
LandOfFree
Reference-based autofocusing method for IFSAR and other... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Reference-based autofocusing method for IFSAR and other..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Reference-based autofocusing method for IFSAR and other... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2518143