CN115792906A

CN115792906A - Satellite-borne large squint sliding bunching SAR imaging processing method

Info

Publication number: CN115792906A
Application number: CN202310051747.9A
Authority: CN
Inventors: 张衡; 任明山; 邓云凯; 禹卫东; 贾小雪
Original assignee: Aerospace Information Research Institute of CAS
Current assignee: Aerospace Information Research Institute of CAS
Priority date: 2023-02-02
Filing date: 2023-02-02
Publication date: 2023-03-14
Anticipated expiration: 2043-02-02
Also published as: CN115792906B

Abstract

The invention discloses a satellite-borne large squint sliding bunching SAR imaging processing method, which comprises the steps of arranging points in a scene, and calculating the slope distance process of each point; generating an azimuth deskew function, deskewing SAR data at the same time, and calculating the slope course and the slope course coefficient of each point after deskew; generating an azimuth reacquisition function, executing azimuth reacquisition on the SAR data at the same time, and fitting a slope course coefficient after the reacquisition; generating a two-dimensional matched filter, and performing matched filtering on the SAR data; calculating a distance frequency mapping function, and executing distance frequency mapping on the SAR data; calculating an azimuth frequency mapping function, and executing azimuth frequency mapping on the SAR data; calculating the distance migration of SAR data residue to correct in an RD domain; performing azimuth matched filtering on the SAR data; calculating an RD domain azimuth frequency mapping function according to the residual azimuth modulation; and executing orientation IFT on the SAR data to obtain the SAR image. The invention accurately and efficiently realizes the imaging processing of the high-resolution large squint spaceborne sliding gather SAR data.

Description

Satellite-borne large squint sliding bunching SAR imaging processing method

Technical Field

The invention belongs to the technical field of satellite-borne Synthetic Aperture Radars (SAR), and particularly relates to a satellite-borne large squint sliding bunching SAR imaging processing method.

Background

Three main problems are faced in the current high-resolution large squint spaceborne SAR imaging processing: under a large squint beam sliding mode, aliasing of an SAR data azimuth frequency domain cannot be effectively removed; the skew distance history and the corresponding frequency spectrum are difficult to accurately model; the pitch model parameters in large scenes become spatially severe along the distance and azimuth dimensions.

Ultra-high resolution satellite-borne SAR systems typically operate in either a beamforming or sliding beamforming mode. In this mode, the overall azimuth bandwidth of the satellite-borne SAR data is much higher than the Pulse Repetition Frequency (PRF), and if an efficient Frequency domain imaging algorithm is used to focus the satellite-borne SAR data, aliasing of the SAR data in the azimuth Frequency domain due to insufficient PRF must be solved through some processing. The current mainstream method for eliminating the aliasing of the SAR data in the azimuth frequency domain mainly comprises two methods, namely a sub-aperture method and a full-aperture method. The sub-aperture mode introduces phase discontinuity for the SAR data azimuth direction during aperture splicing, and causes a virtual image on a final imaging result. The full-aperture two-step method can improve the data volume of SAR data to an unacceptable ground step when squint bunching or sliding bunching SAR data is processed, and reduces the processing efficiency, which is particularly serious under the condition that the bandwidth of a transmitted signal is wide.

The hyperbolic slope distance model is a slope distance model commonly used in the satellite-borne SAR imaging processing, and is used for describing a slope distance process between an antenna phase center and a target. However, when the azimuth resolution of the space-borne SAR system is improved and the synthetic aperture time of the target is further increased, the bending orbit effect cannot be ignored, and the hyperbolic slope distance model is not accurate any more. In order to adapt to the slope distance history of the satellite-borne SAR signals under high resolution, a fourth-order slope distance model, an improved hyperbolic slope distance history, an equivalent acceleration model and other slope distance models are sequentially proposed, but the slope distance models are difficult to accurately describe the slope distance history of the radar and the targets in the scene under the conditions of ultrahigh resolution and large squint. And when the slant range model is determined, how to develop an accurate satellite-borne SAR signal spectrum is also a difficult problem.

The synthetic aperture time of the high-resolution large squint spaceborne SAR signal is long, so that the track bending effect is obvious, the signal characteristic of the SAR signal under a large scene is seriously subjected to two-dimensional space-variant, and the traditional direction translation invariance assumption for SAR signal processing is not satisfied any more. If the traditional frequency domain SAR imaging algorithm is used for processing SAR echo data of a large scene under high resolution and large squint, the direction space-variant problem of the SAR signal cannot be ignored. Ideas such as azimuth nonlinear scaling, azimuth time-frequency resampling, blocking and the like are provided for correcting the azimuth space-variant of the SAR signal. However, in the method, a new space variant term is introduced while the space variant is corrected by the azimuth nonlinear scaling, and two-dimensional distortion exists in an imaging result; azimuth time-frequency resampling can only correct the low-order space-variant of low-order Doppler parameters (Doppler centroid, azimuth modulation frequency and third-order Doppler parameters) in SAR signals (when the Doppler parameter azimuth space-variant is corrected, the parameter is generally at most assumed to be a second-order function of a target azimuth position), and the space-variant of higher-order parameters and the severe space-variant of low-order parameters do not have accurate correction capability; the block correction space-variant increases the time complexity of algorithm processing, and the blocks are sometimes difficult to splice after the block processing. In addition, coupling exists in two-dimensional space-variant of Doppler parameters of the satellite-borne high-resolution large squint SAR signals, and the characteristic causes that unified azimuth space-variant correction cannot solve azimuth space-variant in the whole scene, so that the difficulty of full-aperture processing of the high-resolution satellite-borne SAR signals is further aggravated. Therefore, a more accurate spaceborne SAR signal orientation space-variant correction method and a two-dimensional space-variant coupling correction method are developed, and the method has important significance for high-resolution large squint spaceborne SAR data imaging processing of a large scene.

In conclusion, aliasing of a two-dimensional frequency domain of the spaceborne sliding gather SAR signal is effectively removed, a slope model of the SAR signal is accurately designed, a corresponding frequency spectrum is developed simultaneously, and a correction method of the two-dimensional space-variant of the SAR signal characteristic is reasonably designed, so that the method has important significance for development of future high-resolution large squint spaceborne SAR imaging processing.

Disclosure of Invention

In order to solve the technical problems, the invention provides a satellite-borne large squint sliding bunching SAR imaging processing method, which can accurately and efficiently realize the imaging processing of high-resolution large squint satellite-borne sliding bunching SAR data.

In order to achieve the purpose, the invention adopts the technical scheme that:

a satellite-borne large squint sliding bunching SAR imaging processing method comprises the following steps:

step 101, arranging points in a scene, and simultaneously calculating the slope distance course of each point;

102, performing azimuth deskew processing on the SAR data in a distance frequency domain;

103, performing azimuth time domain resampling processing on the SAR data in a two-dimensional time domain;

104, performing reference point matching filtering on the SAR data in a two-dimensional frequency domain;

105, mapping range frequency on the SAR data in a two-dimensional frequency domain;

106, mapping azimuth frequency on the SAR data in a two-dimensional frequency domain;

step 107, performing residual range migration correction on the SAR data in a range-Doppler domain;

108, performing azimuth matching filtering on the SAR data in the RD domain;

step 109, performing azimuth frequency mapping on the SAR data in the RD domain;

and step 110, performing azimuth inverse Fourier transform on the SAR data to obtain a focused SAR image.

Further, the step 101 includes:

and points are distributed in the interested scene area according to the position, the attitude information and the scene information of the satellite, the distributed points meet the condition that the focus positions of the target after imaging processing are distributed in a uniform grid shape along the distance direction and the azimuth direction, and then the beam irradiation time of each point in the scene and the slant range course in the beam irradiation time are calculated.

Further, the step 102 includes:

calculating a virtual rotation point coordinate of the satellite-borne sliding spotlight SAR system according to the satellite position, the attitude information and the interested scene information, generating an azimuth deskew function by using the virtual rotation point, and multiplying the deskew function by the SAR data phase in a distance frequency domain to complete deskew; and calculating a new slope distance process of each point after the deskew is finished, setting orders, and calculating each slope distance process coefficient of each order of each point by utilizing polynomial fitting.

Further, the step 103 includes:

fitting an azimuth space-variant characteristic coefficient according to the distance, azimuth position and each-order slope distance history coefficient of each point in the scene, generating an azimuth resampling function according to the azimuth space-variant characteristic coefficient, and then performing azimuth time domain resampling processing on SAR data in a two-dimensional time domain according to the azimuth resampling function; and finally, fitting each step of slope distance process coefficient of each point after azimuth time domain resampling and calculating a two-dimensional space-variant characteristic function of each step of slope distance process coefficient.

Further, the step 104 includes:

and generating a matched filter function of a two-dimensional frequency domain by utilizing the beam irradiation time of the scene central point and the corresponding slope distance process, and multiplying the SAR data and the phase of the matched filter function in the two-dimensional frequency domain to complete reference point matched filtering.

Further, the step 105 includes:

using the two-dimensional space-variant characteristic function of each step of the slope distance process coefficient generated in the step 103 to lay points at the center of the azimuth of the scene along the distance and calculating each step of the slope distance process coefficient of each point, wherein the central point is the reference point in the step 104 when the points are laid; calculating distance frequency and azimuth frequency coupling terms of each point, calculating a corresponding mapping function in distance frequency mapping by using least squares, and performing distance frequency mapping on SAR data in a two-dimensional frequency domain according to the mapping function.

Further, the step 106 includes:

using the two-dimensional space-variant characteristic function of each step of the slope distance process coefficient generated in the step 103 to arrange points along the azimuth direction at the scene distance center and calculating each step of the slope distance process coefficient of each point, wherein the central point is the reference point in the step 104 when the points are arranged; calculating distance frequency and azimuth frequency coupling terms of each point, calculating a mapping function corresponding to azimuth frequency mapping in a two-dimensional frequency domain by using least squares, and performing azimuth frequency mapping on SAR data in the two-dimensional frequency domain according to the mapping function.

Further, the step 107 comprises:

bringing the inverse function of the corresponding mapping function in the distance frequency mapping into the coupling terms of the distance frequency and the azimuth frequency of each point distributed along the distance direction in the step 105, and deriving the coupling terms with respect to the distance frequency to obtain the residual distance migration of each point at different distance units; and performing two-dimensional interpolation on the residual range migration to be consistent with the data dimension of the SAR data, and then performing residual range migration correction on the SAR data in a range-Doppler domain.

Further, the step 108 includes:

the azimuth matched filter at each point along the range profile in the calculation step 105 is interpolated by two-dimensional interpolation to a dimension consistent with the data dimension of the SAR data, and then the azimuth matched filter is performed on the SAR data in the range-doppler domain.

Further, the step 109 includes:

and (3) performing two-dimensional interpolation on the range-doppler domain azimuth frequency resampling function to be consistent with the data dimension of the SAR data, and performing azimuth frequency mapping on the SAR data in the range-doppler domain.

Further, the step 110 includes: and performing azimuth inverse Fourier transform on the SAR data to obtain a focused SAR image.

Has the advantages that:

the imaging processing scheme can effectively remove the aliasing of the azimuth frequency domain of the SLR data, accurately model the SAR signal and simultaneously correct the two-dimensional space-variant characteristic of the SAR signal, and efficiently realize the accurate focusing processing of the satellite-borne large-squint SLR data.

Drawings

FIG. 1 is a schematic flow chart of a spaceborne large squint sliding bunching SAR imaging processing method of the invention;

FIG. 2 is points distributed in a scene for calculating the range-to-range coefficients and the space-variant characteristic function of the on-board SAR signal; wherein (a) is the distribution of the distributed points on the earth surface along the longitude and latitude, and (b) is the distribution of the distributed points on the slant range plane;

FIG. 3 is a graph of points distributed in a skew plane for computing a mapping function using a space-variant feature function, where (a) for computing points distributed along the distance center at the distance center for a distance-frequency mapping function in the two-dimensional frequency domain, a distance space-variant RCM in the RD domain, and an orientation-matched filter, (b) for computing points distributed along the orientation center at the distance center for an orientation-frequency mapping function in the two-dimensional frequency domain, and (c) for computing points distributed along both dimensions in the skew plane for an orientation-frequency mapping function in the RD domain;

FIG. 4 is a plot of points in a scene for generating echoes; wherein (a) is the distribution of the distributed points on the surface of the earth, and (b) is the distribution of the distributed points on the slant plane;

FIG. 5 is a processing result of simulated high-resolution large squint spaceborne SAR data; wherein, (a) is an Impulse Response Function (IRF) of an imaging result of a lower left corner point in a scene; (b) IRF which is the scene center point imaging result; and (c) IRF of the imaging result of the upper right corner in the scene.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

As shown in fig. 1, the satellite-borne large squint sliding bunching SAR imaging processing method of the present invention specifically includes the following steps:

step 101: and (3) arranging points in the scene, and simultaneously calculating the slope course of each point.

Points are distributed in the interested scene area according to the position, the attitude information, the scene information and the like of the satellite, and the distributed points should meet the condition that the focusing positions of the point targets after imaging processing are uniformly distributed in a grid shape along the distance direction and the azimuth direction as far as possible. Fig. 2 shows an example of distribution of points, which are distributed in a trapezoid shape on the earth's surface as shown in fig. 2 (a); as shown in fig. 2 (b), are distributed in a uniform grid along the distance and orientation on the slant plane.

And then calculating the beam irradiation time of each point in the scene and the slant distance from the point to the satellite in the time according to the attitude, the position and the like of the satellite.

Step 102: azimuth deskew processing is performed on the SAR data in the range frequency domain.

Sampling point number according to azimuth

And pulse repetition frequency PRF setting azimuth time axis

Comprises the following steps:

(1)

calculating the virtual rotation point coordinate of the spaceborne sliding gather SAR system according to the satellite position, the attitude information, the interesting scene information and the like, and recording the point coordinate as

The position coordinates of the satellite in the whole azimuth time are recorded as

Then the deskew function can be expressed as:

(2)

wherein the content of the first and second substances,

is a unit of an imaginary number, and is,

in order to be a function of the distance frequency,

in order to obtain the carrier frequency of the SAR system,

in order to be the speed of light,

in order to tune the frequency of the SAR signal,

in the form of a euclidean norm operator,

to be used for the course of the ramp for deskewing,

is composed of

At azimuth time

Take the value at 0. The deskew processing shown in equation (2) completes both azimuth deskew and range-matched filtering.

Will be provided with

And the SAR data is multiplied by the phase of the SAR data in the distance frequency domain to complete the deskew processing. The deskew backward sliding SAR data is not aliased in an azimuth frequency domain, and the SAR data can be directly transformed into the azimuth frequency domain through azimuth Fourier Transform (FT). After the phase multiplication is completed, the SAR data must be transformed back to the range time domain using the range-wise IFT.

Step 103: and performing azimuth time domain resampling processing on the SAR data in a two-dimensional time domain.

Assuming that there are L points in the scene, the ramp distance history after each point in the scene is deskewed as shown in FIG. 2

Comprises the following steps:

(3)

wherein, the first and the second end of the pipe are connected with each other,

is as follows

The slope history before the point declivity,

the sequence number of points in the scene ranges from 1 to L.

Setting a certain fitting order N, and fitting the slope distance course of each point in the scene by utilizing a polynomial

And (3) fitting:

(4)

wherein the content of the first and second substances,

is as follows

The point is the zero doppler time of the point,

the slope distance corresponding to zero doppler time, which to some extent characterize the position at which the target is focused,

and

are respectively the first

The start time and the end time of the spot being illuminated by the beam,

to fit the coefficients of each order of the slope history.

Coefficient of slope course

Modeling as target focus position

The binary function of (c):

(5)

wherein, in the formula (5)

Is coefficient of slope course

Is not a space-variant term of (a),

in order to obtain the space-variant coefficient of the distance,

in order to obtain the orientation space-variant coefficient,

is a coupling space-variant coefficient; n is a fitting order of the slope distance process, N is a serial number of the traversal fitting order, and the value is from 1 to N; u and V are respectively fitting orders set by fitting slope distance process coefficient azimuth space-variant and distance space-variant characteristics, and U and V are serial numbers traversing the fitting orders; and X and Y are fitting orders set by fitting the slope distance process coefficient coupling space-variant characteristics, and X and Y are corresponding serial numbers. r and

respectively representing the distances of point objectsAn off-directional position and an on-directional position,

represents the power of the r to the u power,

represents

To the power of v of (a),

x-power sum of r

The product of the power y of (c). Will be given in formula (4)

、

And

and (5) substituting the equation into the equation (5), and fitting a space-variant coefficient of the slope distance course coefficient.

Only the spatial variation of the slope history coefficient with the azimuth focal position is considered in the correction process of the azimuth spatial variation, so the equation in the formula (5) is updated as follows:

(6)

correction of

When the space-variant characteristic is adopted, firstly, the correction is carried out

The expression of the resampling function is shown in formula (7), and more specifically, when the fitting order V in formula (6) takes 4, taylor expansion may be further performed on formula (7):

(7)

wherein t is the new azimuth time after resampling,

is the original azimuth time;

is the coefficient of the slope fit in equation 6

The corresponding expression when n is taken to be 2,

to fit to

The 0-order coefficient of the orientation space-variant characteristic function,

to fit to

The n-order coefficient of the orientation space-variant characteristic function;

intermediate variables set for convenient representation;

is a function name used for convenience of representation.

Substituting t in equation (7) into equation (4) for t in equation

While recalculating

. Re-executing the fitting in the formulas (4) to (6), calculating new space-variant characteristic coefficients after re-sampling shown in the formula (7), and still recording the new space-variant coefficients as the space-variant characteristic coefficients for convenient representation

、

、

And

structural correction

The resampling function of the orientation space-variant characteristic is shown in formula (8):

(8)

where t is the original azimuth time

The new azimuth time obtained after the mapping shown in equation (7),

as the original azimuth time

Obtaining new azimuth time after mapping shown in formulas (7) and (8); in formula (8)

、

、

And

all are orientation space-variant characteristic coefficients obtained by recalculation after mapping shown in formula (7);

is for convenience of representationBut the function name used.

Repeating the above process, and replacing the original azimuth time in the formula (4) with the new azimuth time obtained by mapping

While recalculating

The fitting in equations (4) to (6) is re-performed. Using recalculation

、

、

And

structural correction

(n =4,5,6) function of the azimuthal space-variant characteristic

As shown in equations (9-11):

structural correction

The resampling function of the orientation space-variant characteristic of (3) is shown in equation (9):

(9)

wherein the content of the first and second substances,

as the original azimuth time

Obtaining new azimuth time after mapping shown in formulas (7), (8) and (9);

is the new azimuth time in equation (8); in the formula (9)

、

、

And

all are orientation space-variant characteristic coefficients obtained by recalculation after mapping shown in formula (8);

is the function name used for ease of representation.

Structural correction

The resampling function of the orientation space-variant characteristic of (2) is shown in equation (10):

(10)

wherein the content of the first and second substances,

as the original azimuth time

Obtaining new azimuth time after mapping shown in formulas (7), (8), (9) and (10);

is the new azimuth time in equation (9); in the formula (10)

、

、

And

all are orientation space-variant characteristic coefficients obtained by recalculation after mapping shown in formula (9);

is a function name used for convenience of representation.

Structural correction

The resampling function of the orientation space-variant characteristic of (1) is shown in equation (11):

(11)

wherein the content of the first and second substances,

as the original azimuth time

New azimuth time obtained after mapping shown in formulas (7), (8), (9), (10) and (11);

is the new azimuth time in equation (10); in the formula (11)

、

、

And

the azimuth space-variant characteristic coefficients are obtained by recalculation after mapping shown in formula (11);

is the function name used for ease of representation.

It should be noted that each time the resampling function is calculated, a new space-variant characteristic is calculated according to the last resampling function

And

specifically, the resampled azimuth time shown in the formulas (7) to (10) is substituted into the formulas (4-6) to recalculate

And

。

the final resampling function is expressed as:

(12)

wherein the content of the first and second substances,

the final azimuth time axis obtained for resampling.

And (4) completing the resampling of the SAR data azimuth time domain in the two-dimensional time domain according to the mapping relation shown in the formula (12).

After resampling, the method uses the formula (12)

Substituting the formula (4) to (5) to obtain a new space-variant characteristic function of the slope distance course coefficient:

(13)

for new target orientation after resamplingTowards the position of focus,

、

、

and

the new slope course space-variant characteristic coefficient is obtained by recalculation after the mapping shown in the formula (12). r and

respectively representing the distance position and the azimuth position of the point target after resampling,

represents the power of the r to the u power,

represents

To the power of v;

x-power sum of r

The product of the power y of (c).

Step 104: reference point matched filtering is performed on the SAR data in the two-dimensional frequency domain.

Selecting a scene central point as a reference point to generate a matched filtering function, wherein the expression of the matched filter is as follows:

(14)

wherein j is an imaginary unit,

is heavyThe new time of orientation after sampling,

for the corresponding new azimuth frequency,

is the phase of the frequency spectrum of the filter,

is the magnitude of the filter spectrum.

Is represented by formula (15):

(15)

wherein the content of the first and second substances,

for a reference point of zero doppler time,

the slope distance corresponding to zero Doppler time of the reference point, which to some extent characterize the position of the target focus,

to fit the coefficients of each order of the reference point ramp history. In the formula (15)

Is the distance frequency

And azimuth frequency

The azimuth time can be obtained by using the principle of stationary phase and the law of series inversion

Sum distance frequency

Azimuth frequency

The functional relationship of (1) is:

(16)

wherein the content of the first and second substances,

zero Doppler time for target point;

to fit coefficients of each order of the slope history, the quantity being a function of the focus position of the target, different point targets in the scene having different slope history coefficients

。

The expression of (a) is:

(17)

wherein the content of the first and second substances,

representing taking the absolute value.

Combining SAR data and reference signals in a two-dimensional frequency domain

And completing reference point matched filtering by matrix multiplication.

Step 105: range frequency mapping is performed on the SAR data in the two-dimensional frequency domain.

As shown in fig. 3 (a), along the distance cloth at the azimuth center

DotThe distance position of each point is

The azimuth position is

And ensuring that the distance direction range of the stationing is larger than that of the SAR data. The individual step slope history coefficients for each point are calculated using equation (18):

(18)

coupling terms of range frequency and azimuth frequency along range to each point

The expression of (a) is:

(19)

in the formula

Is represented by the formulas (15) and (16), wherein in the formula (15)

Calculated for substitution by equation (18)

The inversion of the series in equation (16) also needs to be performed again;

and matching the filtered residual SAR signal spectrum phase for the reference point.

To pair

Derivation of the distance to the focus position r, corresponding derivative function

I.e. the distance frequencyThe mapping function in the mapping process is,

is the mapped range frequency. The corresponding mapping relation is as follows:

(20)

further elaboration of equation (20) can yield:

(21)

calculating the distance frequency mapping function of formula (21) by numerical method

About the mapped range frequency

The inverse function:

(22)

formula (II)

The derivation shown can be achieved by least squares, in particular by constructing an observation matrix

And coefficient matrix A is shown as equation (23):

(23)

wherein the content of the first and second substances,

for the range-wise dimension of the SAR data,

for the azimuthal dimension of the SAR data,

is of dimension of

A is a dimension of

Of the matrix of (a). Mapping matrix

The expression of (a) is:

(24)

wherein, in the formula

Representing the ith row of the fetch matrix,

representing the transpose of the fetch matrix.

Obtaining a mapping matrix

And then, completing the mapping of the SAR data to the frequency axis by resampling according to the mapping relation shown in the formula (20), wherein the resampling can be realized by sinc interpolation.

Step 106: and performing azimuth frequency mapping on the SAR data in a two-dimensional frequency domain.

As shown in (b) of FIG. 3, along the azimuth line from the center

Points, each point being located at a distance from the position of the other point

The azimuth position is

And ensuring that the azimuth range of the stationing is larger than that of the SAR data. Calculated using equation 13Each step of slope distance course coefficient of each point:

(25)

after the range-frequency mapping shown in step 105, the range-frequency and azimuth-frequency coupling terms for each point along the azimuth direction

The expression of (a) is:

(26)

for is to

Focal position with respect to azimuth

Derivation, corresponding derivative functions

I.e. a mapping function in the process of azimuth frequency mapping. The corresponding mapping relation is as follows:

(27)

the derivation in equation 27 is achieved using least squares, specifically constructing an observation matrix

Sum coefficient matrix

As shown in equation (28):

(28)

wherein the content of the first and second substances,

is of dimension of

The matrix of (a) is a matrix of (b),

is of dimension of

Of the matrix of (a). Mapping matrix

The expression of (a) is:

(29)

obtaining a mapping matrix

And then, completing the mapping of the SAR data azimuth frequency by resampling according to the mapping relation shown in the formula (27), wherein the resampling can be realized by sinc interpolation.

Step 107: residual RCMC is performed on the SAR data in the RD (range-doppler) domain.

Substituting equation (22) into the coupling terms of the range frequency and the azimuth frequency in the spectral phase of each point along the distance direction in (a) in FIG. 3

The method comprises the following steps:

(30)

wherein the content of the first and second substances,

for the range frequency after the mapping, the distance frequency is,

the distance position of each point laid,

points arranged in a distance directionThe number i is the index number and has a value ranging from 1 to

。

For the function of equation (30)

With respect to the mapped range frequency

Derivation is carried out to obtain the range migration trajectory of each point

Comprises the following steps:

(31)

is the mapped coupling phase in equation 30,

the distance position of each point laid,

i is the index number of the point distributed in the distance direction, and the value range is 1 to

。

Function for correcting residual range migration

Comprises the following steps:

(32)

wherein the content of the first and second substances,

the residual RCMC is correctedThe function to be used is,

the distance position of each point laid,

i is the number of points distributed along the distance direction, i is the index number thereof, and the value range is 1 to

。

The data dimension of the term is far smaller than that of the SAR data, and the dimension of the term needs to be interpolated to be the same as that of the SAR data through two-dimensional interpolation. After the two-dimensional interpolation is completed, the residual RCM of the SAR data can be directly corrected through resampling in an RD domain. Step 108: azimuth matched filtering is performed on the SAR data in the RD domain.

And (4) after RCMC, azimuth modulation along distance space variation needs to be compensated, and azimuth matched filtering is completed. Specifically, the residual azimuth modulation after each point in (a) in fig. 3 is subjected to matching filtering shown in step 104 is calculated, and the expression of the azimuth modulation is:

(33）

of azimuthal matched filters

The expression of (a) is:

(34)

wherein the content of the first and second substances,

is the residual azimuth modulation in equation 34.

Regarding the azimuth modulation item of the SAR data residual, the data dimension of the item is far smaller than that of the SAR data, and the item needs to be interpolated to be the same as the data dimension of the SAR data through two-dimensional interpolation. After the two-dimensional interpolation is completed, SAR data can be combined with the data in the RD domain

And performing azimuth matched filtering by direct phase multiplication.

Step 109: and mapping the azimuth frequency of the SAR data in the RD domain.

As shown in (c) of fig. 3, arranged in the entire scene

Points, each point having a range position and an azimuth position

. The distributed points are ensured to cover a scene larger than the corresponding scene range of the SAR data, and the scene center point is the reference point in step 104. And (4) calculating the slope course coefficient of each point in the scene by using a formula (13) according to the position of the target, wherein the slope course coefficient is specifically shown in formulas (18) and (25).

At a certain fixed distance unit

After the azimuth matching filter shown in step 108, the residual azimuth modulation at each point along the azimuth is performed

Comprises the following steps:

(35)

to pair

Focal position with respect to azimuth

Derivation, corresponding derivation knotThe result is a mapping function of the RD domain orientation to resampling. The corresponding mapping relation is as follows:

(36)

the derivation in equation (36) is implemented using least squares, specifically to construct an observation matrix

Sum coefficient matrix

As shown in equation (37):

(37)

wherein the content of the first and second substances,

is of dimension of

The matrix of (a) is,

is of dimension of

Of the matrix of (a). Mapping matrix

The expression of (a) is:

(38)

for the residual azimuth space-variant azimuth modulation item of the SAR data, the data dimension of the item is far smaller than that of the SAR data, and the item needs to be interpolated to be the same as the data dimension of the SAR data through two-dimensional interpolation. After the two-dimensional interpolation is completed, the two-dimensional interpolation can be passed in the RD domainThe mapping relation shown in equation (37) directly corrects the azimuth space-variant azimuth modulation term of the residual SAR data by resampling.

Step 110: performing azimuth IFT on the SAR data results in a focused SAR image.

And directly executing azimuth IFT on the processed SAR data to obtain a focused SAR image. The technical solution of the present invention will be further described in detail with reference to the following specific examples.

Example 1

The effectiveness of the technical scheme of the invention is verified by adopting the imaging processing of satellite-borne large squint sliding bunching mode SAR data with system and scene parameters shown in the table 1.

TABLE 1 spaceborne sliding spotlight SAR System parameters and Scenario parameters

FIG. 4 is a diagram of the arrangement of point objects in a scene during generation of echoes, where (a) is the distribution of points on the surface of the earth and (b) is the distribution of points on a slope plane; wherein the field-to-field width of the scene in Table 1 is 50km

The 50km means that the length of the solid line and the broken line in fig. 4 is 50km on the earth's surface.

Fig. 5 shows the processing result of the echo data of the satellite-borne SAR shown in table 1 by using the method of the present invention. Wherein, (a) is an Impulse Response Function (IRF) of the focusing result of the near-edge point, which corresponds to the point target surrounded by the circle at the lower left corner in fig. 4; (b) The IRF of the focusing result of the scene center point corresponds to a point target surrounded by a circle in the center of the graph 4; (c) The IRF, which is the result of far-away edge point focusing, corresponds to the point object encircled by a circle in the upper right corner of fig. 4.

It can be seen from the figure that all point targets are well-focused, without significant defocus, and the IRF of all points is close to the ideal two-dimensional sinc function.

It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A satellite-borne large squint sliding spotlight SAR imaging processing method is characterized by comprising the following steps:

104, performing reference point matched filtering on the SAR data in a two-dimensional frequency domain;

108, performing azimuth matching filtering on the SAR data in the RD domain;

and 110, performing azimuth inverse Fourier transform on the SAR data to obtain a focused SAR image.

2. The method for processing spaceborne large squint sliding spotlight SAR imaging according to claim 1, wherein the step 101 comprises:

3. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 2, characterized in that the step 102 comprises:

calculating the virtual rotation point coordinates of the satellite-borne sliding spotlight SAR system according to the satellite position, the attitude information and the interesting scene information, generating an azimuth deskew function by using the virtual rotation point, and multiplying the deskew function and the SAR data phase in a distance frequency domain to complete deskew; and calculating a new slope distance process of each point after the deskew is finished, setting orders, and calculating each slope distance process coefficient of each order of each point by utilizing polynomial fitting.

4. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 3, characterized in that the step 103 comprises:

5. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 4, characterized in that the step 104 comprises:

and generating a matched filter function of a two-dimensional frequency domain by utilizing the beam irradiation time of the scene central point and the corresponding slant distance process, and multiplying the SAR data and the phase of the matched filter function in the two-dimensional frequency domain to complete reference point matched filtering.

6. The method for processing spaceborne large squint sliding spotlight SAR imaging according to claim 5, wherein the step 105 comprises:

7. The processing method of on-board large strabismus sliding spotlight SAR imaging according to claim 6, wherein the step 106 comprises:

utilizing the two-dimensional space-variant characteristic function of each-order slope distance process coefficient generated in the step 103 to point at the scene distance center along the azimuth direction and calculating each-order slope distance process coefficient of each point, wherein the central point is ensured to be the reference point in the step 104 when the point is placed; calculating distance frequency and azimuth frequency coupling terms of each point, calculating a mapping function corresponding to azimuth frequency mapping in a two-dimensional frequency domain by using least squares, and performing azimuth frequency mapping on SAR data in the two-dimensional frequency domain according to the mapping function.

8. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 7, characterized in that the step 107 comprises:

bringing the inverse function of the corresponding mapping function in the distance frequency mapping into the coupling terms of the distance frequency and the azimuth frequency of each point along the distance distribution in the step 105, and deriving the coupling terms with respect to the distance frequency to obtain the residual distance migration of each point at different distance units; the residual range migration is two-dimensionally interpolated to be consistent with the data dimensions of the SAR data, and then residual range migration correction is performed on the SAR data in the range-Doppler domain.

9. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 8, characterized in that the step 108 comprises:

10. The processing method of spaceborne large squint sliding spotlight SAR imaging according to claim 9, wherein the step 109 comprises:

using the two-dimensional space-variant characteristic function of each step of the slant range history coefficients generated in the step 103 to perform two-dimensional stationing on the scene and calculate each step of the slant range history coefficients of each point, calculating the residual azimuth modulation of each point after azimuth matching filtering, using least squares to calculate the azimuth space-variant characteristic function of the residual azimuth modulation at different distance units and generating the range-doppler domain azimuth frequency resampling function according to the azimuth space-variant characteristic function, interpolating the range-doppler domain azimuth frequency resampling function two-dimensionally to be consistent with the data dimension of the SAR data, and performing azimuth frequency mapping on the SAR data in the range-doppler domain.

11. The processing method of on-board large strabismus sliding spotlight SAR imaging according to claim 10, wherein the step 110 comprises:

and performing azimuth inverse Fourier transform on the SAR data to obtain a focused SAR image.