Methods for two-dimensional autofocus in high resolution radar systems
First Claim
1. A synthetic aperture radar (SAR) system for automatically compensating on-axis and off-axis distortions, the system comprising:
- an image formation processor (IFP) configured to generate a complex image;
a processor configured to remove the on-axis and off-axis distortions in the complex image based on a two-dimensional autofocus algorithm to generate a corrected image;
a memory configured to store the algorithm and the corrected image; and
an output interface configured to output the corrected image to an external device,wherein the complex image is defined in two-dimensions with range information in a vertical direction and azimuth information in a horizontal direction.
12 Assignments
0 Petitions
Accused Products
Abstract
Provided are two-dimensional autofocus methods in a synthetic aperture radar (SAR) system which include: (1) two-dimensional pulse pair product algorithm including shear PGA, eigenvector phase history (“EPH”), shear PGA/EPH); (2) two-dimensional optimization algorithms including parametric one-dimensional estimate/two-dimensional correction, parametric two dimensional estimate/two-dimensional correction, unconstrained two-dimensional nonparametric and constrained two-dimensional nonparametric methods; (3) a two-dimensional geometry filter algorithm; (4) a two-dimensional prominent point processing algorithm; (5) a one-dimensional phase estimate of higher order two dimensional phase errors; and, (6) a fast SHARP parametric autofocus algorithm.
38 Citations
149 Claims
-
1. A synthetic aperture radar (SAR) system for automatically compensating on-axis and off-axis distortions, the system comprising:
-
an image formation processor (IFP) configured to generate a complex image; a processor configured to remove the on-axis and off-axis distortions in the complex image based on a two-dimensional autofocus algorithm to generate a corrected image; a memory configured to store the algorithm and the corrected image; and an output interface configured to output the corrected image to an external device, wherein the complex image is defined in two-dimensions with range information in a vertical direction and azimuth information in a horizontal direction. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A method for automatically compensating on-axis and off-axis distortions in a synthetic aperture radar (SAR) image, the method comprising:
-
creating a complex image based on information collected by a sensor; removing the on-axis and off-axis distortions in the complex image based on a two-dimensional autofocus algorithm; generating a corrected image; and outputting the corrected image to an external device, wherein the complex image is defined in two-dimensions with range information in a vertical direction and azimuth dispersed phase history information in a horizontal direction. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32)
-
-
33. A computer-readable medium having computer-executable instructions for automatically compensating on-axis and off-axis distortions in a synthetic aperture radar (SAR) image, the computer-executable instructions configured to perform a method comprising:
-
creating a complex image based on information collected by a sensor; removing the on-axis and off-axis distortions in the complex image based on a two-dimensional autofocus algorithm; generating a corrected image; and outputting the corrected image to an external device, wherein the complex image is defined in two-dimensions with range information in a vertical direction and azimuth dispersed phase history information in a horizontal direction. - View Dependent Claims (34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48)
-
-
49. A two-dimensional polar geometry model method used in a synthetic aperture radar (SAR) system for removing a two-dimensional phase error in a complex image, the method comprising:
-
creating the complex image with a sensor; modeling the two-dimensional phase error as a combination of a range error which is substantially identical from pulse to pulse and a slow-time error; fitting an unconstrained two-dimensional estimate to the modeled two-dimensional phase error; and correcting the two-dimensional phase error to generate a corrected image. - View Dependent Claims (50, 51, 52, 53, 55)
-
-
54. A computer-readable medium having computer-executable instructions for a two-dimensional polar geometry filter method used in a synthetic aperture radar (SAR) system for removing a two-dimensional phase error in a complex image, the computer-executable instructions configured to perform a method comprising:
-
creating the complex image with a sensor; modeling the two-dimensional phase error as a combination of a range error which is substantially identical from pulse to pulse and a slow-time error; fitting an unconstrained two-dimensional estimate to the modeled two-dimensional phase error; and correcting the two-dimensional phase error to generate a corrected image. - View Dependent Claims (56, 57, 58)
-
-
59. A two-dimensional prominent point processing method used in a synthetic aperture radar (SAR) system for removing a two-dimensional phase error in a complex image, the method comprising:
-
generating the complex image with a sensor; identifying a good quality unfocused isolated point target within the complex image; cropping a target and an immediate area surrounding the target; taking a two-dimensional inverse Fast Fourier Transform (FFT) to obtain a phase history; and taking a two-dimensional phase from the phase history of the cropped target as a measurement of the two-dimensional phase error on the image. - View Dependent Claims (60, 61, 62, 63)
-
-
64. A computer-readable medium having computer-executable instructions for a two-dimensional prominent point processing method used in a synthetic aperture radar (SAR) system for removing a two-dimensional phase error in a complex image, the computer-executable instructions configured to perform a method comprising:
-
generating the complex image with a sensor; identifying a good quality unfocused isolated point target within the complex image; cropping a target and an immediate area surrounding the target; taking a two-dimensional inverse Fast Fourier Transform (FFT) to obtain a phase history; and taking a two-dimensional phase from the phase history of the cropped target as a measurement of the two-dimensional phase error on the image. - View Dependent Claims (65, 66, 67, 68)
-
-
69. A method for slow-time error correction with one-dimensional phase estimate of higher order two-dimensional phase errors in an image generated by a synthetic aperture radar, the method comprising:
-
taking a portion in range of phase history which extends a full azimuth extent of the phase history to generate a coarse resolution range and fine resolution azimuth sub-band image; focusing the coarse resolution range and fine resolution azimuth sub-band image with a one-dimensional autofocus algorithm; mapping a resulting one-dimensional phase error estimate sample spacings as a function of angle, and scaling each of the phase estimate sample spacings to a center frequency using the ratio of a center frequency to an original phase history sample frequency; and applying to the phase history a two-dimensional correction mapping the one-dimensional phase equation φ
(θ
/θ
max) to a two-dimensional phase equation aswherein φ
denotes the phase, θ
max denotes a maximum dwell angle of a processed aperture with θ
=0°
at center aperture, ƒ
c denotes a pulse center frequency, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted frequency for the given phase history sample.- View Dependent Claims (70, 71)
-
-
72. A method for fast-time error correction with one-dimensional phase estimate of higher order two-dimensional phase errors in an image generated by a synthetic aperture radar, the method comprising:
-
taking a portion in azimuth of phase history data centered on a set theoretical pulse to generate a coarse resolution azimuth and fine resolution range sub-band image; focusing the coarse resolution azimuth and fine resolution range image with a one-dimensional autofocus algorithm in a range direction; mapping a phase estimate sample spacings to a function of frequency; applying to the phase history a two-dimensional correction mapping a one-dimensional phase equation φ
(f) to a two-dimensional phase equation aswherein φ
denotes the phase, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted range frequency for the given phase history sample.- View Dependent Claims (73, 74)
-
-
75. A method for slow-time error correction with one-dimensional amplitude estimate of higher order two-dimensional amplitude errors in an image generated by a synthetic aperture radar, the method comprising:
-
taking a portion in range of phase history which extends a full azimuth extent of the phase history to generate a coarse resolution range and fine resolution azimuth sub-band image; focusing the coarse resolution range and fine resolution azimuth sub-band image with a one-dimensional amplitude autofocus algorithm; mapping a resulting one-dimensional amplitude error estimate sample spacings as a function of angle; and applying to the phase history a two-dimensional correction mapping the one-dimensional amplitude equation α
(θ
/θ
max) to a two-dimensional amplitude equation as
α
2D(u,v)=α
(tan−
1(u/v)/θ
max),wherein α
denotes the amplitude, θ
max denotes a maximum dwell angle of a processed aperture with θ
=0°
at center aperture, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted frequency for the given phase history sample. - View Dependent Claims (76, 77)
-
-
78. A method for fast-time error correction with one-dimensional amplitude estimate of higher order two-dimensional amplitude errors in an image generated by a synthetic aperture radar, the method comprising:
-
taking a portion in azimuth of phase history data centered on a set theoretical pulse to generate a coarse resolution azimuth and fine resolution range sub-band image; focusing the coarse resolution azimuth and fine resolution range image with a one-dimensional amplitude autofocus algorithm in a range direction; mapping a amplitude estimate sample spacings to a function of frequency; and applying to the phase history a two-dimensional correction mapping a one-dimensional amplitude equation α
(f) to a two-dimensional amplitude equation aswherein α
denotes the amplitude, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted range frequency for the given phase history sample.- View Dependent Claims (79, 80)
-
-
81. A computer-readable medium having computer-executable instructions for slow-time error correction with one-dimensional phase estimate of higher order two-dimensional phase errors in an image generated by a synthetic aperture radar, the computer-executable instructions configured to perform a method comprising:
-
taking a portion in range of phase history which extends a full azimuth extent of the phase history to generate a coarse resolution range and fine resolution azimuth sub-band image; focusing the coarse resolution range and fine resolution azimuth sub-band image with a one-dimensional autofocus algorithm; mapping a resulting one-dimensional phase error estimate sample spacings as a function of angle, and scaling each of the phase estimate sample spacings to a center frequency using the ratio of a center frequency to an original phase history sample frequency; and applying to the phase history a two-dimensional correction mapping the one-dimensional phase equation φ
(θ
/θ
max) to a two-dimensional phase equation aswherein φ
denotes the phase, θ
max denotes a maximum dwell angle of a processed aperture with θ
=0°
at center aperture, ƒ
c denotes a pulse center frequency, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted frequency for the given phase history sample.- View Dependent Claims (82, 83)
-
-
84. A computer-readable medium having computer-executable instructions for fast-time error correction with one-dimensional phase estimate of higher order two-dimensional phase errors in an image generated by a synthetic aperture radar, the computer-executable instructions configured to perform a method comprising:
-
taking a portion in azimuth of phase history data centered on a set theoretical pulse to generate a coarse resolution azimuth and fine resolution range sub-band image; focusing coarse resolution azimuth and fine resolution range image with a one-dimensional autofocus algorithm in a range direction; mapping a phase estimate sample spacings to a function of frequency; and applying to the phase history a two-dimensional correction mapping a one-dimensional phase equation φ
(f) to a two-dimensional phase equation aswherein φ
denotes the phase, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted range frequency for the given phase history sample.- View Dependent Claims (85, 86)
-
-
87. A computer-readable medium having computer-executable instructions for slow-time error correction with one-dimensional amplitude estimate of higher order two-dimensional amplitude errors in an image generated by a synthetic aperture radar, the computer-executable instructions configured to perform a method comprising:
-
taking a portion in range of phase history which extends a full azimuth extent of the phase history to generate a coarse resolution range and fine resolution azimuth sub-band image; focusing the coarse resolution range and fine resolution azimuth sub-band image with a one-dimensional amplitude autofocus algorithm; mapping a resulting one-dimensional amplitude error estimate sample spacings as a function of angle; and applying to the phase history a two-dimensional correction mapping the one-dimensional amplitude equation α
(θ
/θ
max) to a two-dimensional amplitude equation as
α
2D(u,v)=α
(tan−
1(u/v)/θ
max),wherein α
denotes the amplitude, θ
max denotes a maximum dwell angle of a processed aperture with θ
=0°
at center aperture, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted frequency for the given phase history sample. - View Dependent Claims (88, 89)
-
-
90. A computer-readable medium having computer-executable instructions for fast-time error correction with one-dimensional amplitude estimate of higher order two-dimensional amplitude errors in an image generated by a synthetic aperture radar, the computer-executable instructions configured to perform a method comprising:
-
taking a portion in azimuth of phase history data centered on a set theoretical pulse to generate a coarse resolution azimuth and fine resolution range sub-band image; focusing the coarse resolution azimuth and fine resolution range image with a one-dimensional amplitude autofocus algorithm in a range direction; mapping a amplitude estimate sample spacings to a function of frequency; and applying to the phase history a two-dimensional correction mapping a one-dimensional amplitude equation α
(f) to a two-dimensional amplitude equation aswherein α
denotes the amplitude, u denotes a polar formatted azimuth frequency for a given phase history sample, and v denotes a polar formatted range frequency for the given phase history sample.- View Dependent Claims (91, 92)
-
-
93. A fast SHARP parametric autofocus method used in a synthetic aperture radar (SAR) system for removing a one-dimensional azimuth phase error in an image generated by the SAR system, the method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on a sign of the basis function derivative defined as a function of azimuth sample scaled to the interval [−
1 1];splitting an azimuth dispersed, range compressed phase history aperture into two sub-apertures using the positive mask and the negative mask; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional azimuth fast fourier transform (FFT) to form a complex cross correlate; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; and defining an estimated coefficient of the basis function based on a peak location offset from a DC sample, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (94, 95, 96)
-
-
97. A fast SHARP parametric autofocus method used in a synthetic aperture radar (SAR) system for removing a two-dimensional slow-time phase error in an image generated by the SAR system, the method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on the sign of the basis function derivative defined as a function of slow-time angle scaled to a range interval from negative one to positive one; splitting a fully dispersed phase history aperture into two sub-apertures using the positive mask and the negative mask; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional azimuth FFT to form a complex cross correlate; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; defining an estimated coefficient of the basis function based on a peak location offset from a DC sample; defining a final two-dimensional phase error using the basis function defined as a function of slow-time angle scaled to the range interval from negative one to positive one; scaling the final two-dimensional phase error by a ratio of fast-time frequency to center frequency; and multiplying the scaled final two-dimensional phase error by the estimated coefficient to obtain a final error, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (98, 99)
-
-
100. A fast SHARP parametric autofocus method used in a synthetic aperture radar (SAR) system for removing a two-dimensional fast-time phase error in an image generated by the SAR system, the method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on the sign of the basis function derivative defined as a function of fast-time frequency scaled to a range interval from negative one to positive one. splitting a fully dispersed phase history aperture into two sub-apertures using the positive mask and the negative mask; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; defining an estimated coefficient of the basis function based on a peak location offset from a DC sample; defining a final two-dimensional phase error using the basis function defined as a function of fast-time frequency scaled to the range interval from negative one to positive one; and multiplying the scaled final two-dimensional phase error by the estimated coefficient to obtain a final error, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (101, 102)
-
-
103. A computer-readable medium having computer-executable instructions for a fast SHARP parametric autofocus method used in a synthetic aperture radar (SAR) system for removing a one-dimensional azimuth phase error in an image generated by the SAR system, the computer-executable instructions configured to perform a method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on a sign of the basis function derivative defined as a function of azimuth sample scaled to a range interval from negative one to positive one; splitting an azimuth dispersed, range compressed phase history aperture into two sub-apertures using the positive mask and the negative mask; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional azimuth fast fourier transform (FFT) to form a complex cross correlate; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; and defining an estimated coefficient of the basis function based on a peak location offset from a DC sample, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (104, 105, 106)
-
-
107. A computer-readable medium having computer-executable instructions for a fast SHARP parametric autofocus algorithm used in a synthetic aperture radar (SAR) system for removing a two-dimensional slow-time phase error in an image generated by the SAR system, the computer-executable instructions configured to perform a method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on the sign of the basis function derivative defined as a function of slow-time angle scaled to a range interval from negative one to positive one; splitting a fully dispersed phase history aperture into two sub-apertures using the positive mask and the negative mask; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional azimuth FFT to form a complex cross correlate; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; defining an estimated coefficient of the basis function based on a peak location offset from a DC sample; defining a final two-dimensional phase error using the basis function defined as a function of slow-time angle scaled to the range interval from negative one to positive one; and scaling the final two-dimensional phase error by a ratio of fast-time frequency to center frequency; and multiplying the scaled final two-dimensional phase error by the estimated coefficient to obtain a final error, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (108, 109)
-
-
110. A computer-readable medium having computer-executable instructions for a fast SHARP parametric autofocus algorithm used in a synthetic aperture radar (SAR) system for removing a two-dimensional fast-time phase error in an image generated by the SAR system, the computer-executable instructions configured to perform a method comprising:
-
choosing a basis function from an orthonormal set; defining a positive mask and a negative mask based on the sign of the basis function derivative defined as a function of fast-time frequency scaled to a range interval from negative one to positive one; splitting a fully dispersed phase history aperture into two sub-apertures using the positive mask and the negative mask; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; collapsing the two sub-apertures and removing zeros; multiplying element-by-element each of the collapsed sub-apertures by a complex conjugate of the other sub-aperture; performing a one-dimensional range fast fourier transform (FFT) for each sub-aperture; detecting the complex cross correlate; summing cross correlate samples in a range direction and forming a single vector; defining an estimated coefficient of the basis function based on a peak location offset from a DC sample; defining a final two-dimensional phase error using the basis function defined as a function of fast-time frequency scaled to the range interval from negative one to positive one; multiplying the scaled final two-dimensional phase error by the estimated coefficient to obtain a final error, wherein the DC sample is a 0th order Fourier Series component. - View Dependent Claims (111, 112)
-
-
113. A two-dimensional optimization method used in a synthetic aperture radar, the optimization method comprising:
-
generating a complex image with a sensor; defining a relative quality of a focus of the complex image with an optimization metric; searching for a phase error correction best suited to correct a phase error in the complex image with an algorithm; and correcting the phase error based on the phase error correction, wherein the phase error correction optimizes the optimization metric. - View Dependent Claims (114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131)
-
-
132. A computer-readable medium having computer-executable instructions for a two-dimensional optimization method used in a synthetic aperture radar, the computer-executable instructions configured to perform a method comprising:
-
generating a complex image with a sensor; defining a relative quality of a focus of the complex image with an optimization metric; searching for a phase error correction best suited to correct a phase error in the complex image with an algorithm; and correcting the phase error based on the phase error correction, wherein the phase error correction optimizes the optimization metric. - View Dependent Claims (133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149)
-
Specification