CN103278819B - Onboard high-resolution strabismus bunching synthetic aperture radar (SAR) imaging method based on sliding receiving window - Google Patents

Abstract

The invention discloses an onboard high-resolution strabismus bunching synthetic aperture radar (SAR) imaging method based on a sliding receiving window. The onboard high-resolution strabismus bunching SAR imaging method comprises the following steps of 1, reading original echo data and relevant imaging parameters; 2, performing azimuth linear frequency demodulation processing; 3, performing azimuth Fourier transformation processing; 4, performing azimuth linear frequency demodulation residual phase error compensation processing; 5, performing range Fourier transformation processing; 6, performing uniform compression processing; 7, performing stolt interpolation processing; 8, performing azimuth inverse Fourier transformation processing; 9, performing geometric correction processing; and 10, performing range Fourier transformation processing. The invention provides the onboard high-resolution strabismus bunching SAR imaging method based on the sliding receiving window in order to solve the problem that a method for imaging the onboard high-resolution strabismus bunching SAR original echo data based on a sliding receiving window technology does not exist at present.

Description

Based on the Airborne High-resolution stravismus Spotlight SAR Imaging formation method of slip receiver window

Technical field

The invention belongs to signal transacting field, particularly a kind of Airborne High-resolution squint bunching synthetic aperture radar synthetic-aperture radar based on slip receiver window (Synthetic Aperture Radar, SAR) formation method.

Background technology

SAR is a kind of active remote sensing device being operated in microwave frequency band, overcome the defect that optical imagery limits by weather and illumination condition, can round-the-clock, round-the-clock, carry out remote sensing of the earth observation at a distance, and natural vegetation, artificial camouflage etc. can be penetrated, substantially increase the information capture ability of radar.Therefore, SAR has become the popular research field of Radar Technology, by increasing country is paid attention to.Compared to traditional side-looking SAR system, Airborne Squint SAR imaging has very high dirigibility and maneuverability in actual applications, and by adjustment controlling antenna wave beam to point, SAR system can select observation area in freedom and flexibility ground, and heavily can visit sensitizing range fast, substantially increase the observing capacity of SAR.In addition, along with the raising of SAR resolution, SAR target reconnaissance and recognition capability are significantly improved.Therefore, in the last few years, the imaging of high resolving power stravismus had become an important developing direction.

But high resolving power high squint SAR imaging also brings new technological difficulties.On the one hand, along with the increase of stravismus angle, SAR original echo orientation, to more serious to coupling phenomenon with distance, causes high precision imaging more difficult; On the other hand, range migration amount forms geometric growth with the increase of angle of squint and the increase of resolution, causes the remarkable increase of SAR raw radar data amount, increases data and stores the difficulty with real time imagery.Slip receiver window technology refers to by changing echo window start-up time, eliminates the once item in range migration, i.e. range walk.Slip receiver window technology can reduce SAR original echo range migration amount, and then reduces SAR raw radar data amount.But, the slip receiver window technological adjustment admission initial time of echo data, in observation area the Doppler history of each target also with orientation to there occurs change, make conventional imaging method no longer applicable.

Summary of the invention

The object of the invention is to solve the problem, based on slip receiver window technical characterstic, in conjunction with conventional wave number field formation method, proposing a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method.

Based on an Airborne High-resolution stravismus Spotlight SAR Imaging formation method for slip receiver window, comprise following step:

Step one: read in raw radar data and dependent imaging parameter;

Read in the Airborne High-resolution stravismus Spotlight SAR Imaging two dimension original echo emulation complex data S based on slip receiver window _startand corresponding imaging parameters, specifically comprise: orientation is to sampling number N _a, distance is to sampling number N _r, signal sampling rate f _s, signal bandwidth Bw, pulse width τ, chirp rate b, pulse repetition rate PRF, with reference to oblique distance R _ref, doppler centroid fd ₀, doppler frequency rate f _r0, satellite velocities P _v, equivalent squint angle signal wavelength lambda, signal carrier frequency f ₀, signal velocity c;

Step 2: orientation is to the process of solution linear frequency modulation;

By two-dimentional echo simulation complex data S _startcarry out orientation to the process of solution linear frequency modulation: first complex data S _startwith factor H ₁be multiplied, obtain complex data S _{1_1}; Secondly to complex data S _{1_1}do orientation to Fourier transform, namely carry out fast Fourier change (FFT) along each distance to (by row), obtain complex data S _{1_2}; Finally by complex data S _{1_2}with factor H ₂be multiplied, obtain final orientation to the complex data S separated after linear frequency modulation ₁;

Step 3: orientation is to Fourier transform process;

The complex data S that step 2 is obtained ₁carry out Fast Fourier Transform (FFT) (FFT) along each distance to (by row), obtain azimuth spectrum complex data S ₂;

Step 4: orientation is to the compensation deals of solution linear frequency modulation residual phase offset;

The complex data S that step 3 is obtained ₂with the orientation in corresponding orientation moment to solution linear frequency modulation residual phase offset compensating factor Ω ₁be multiplied, the complex data S after being compensated ₃;

Step 5: distance is to Fourier transform process;

The complex data S that step 4 is obtained ₃carry out Fast Fourier Transform (FFT) (FFT) along each orientation moment (by row), obtain 2-d spectrum complex data S ₄;

Step 6: unanimously compress process;

The complex data S that step 5 is obtained ₄with corresponding consistent compressibility factor Ω ₂be multiplied, slightly focused on complex data S ₅;

Step 7: Stolt (stolt) interpolation processing;

For the complex data S that step 6 obtains ₅, utilize Singh (sinc) method of interpolation to carry out stolt interpolation processing, obtain the complex data S being mapped to two-dimentional wavenumber domain by two-dimensional frequency ₆;

Step 8: orientation is to inverse Fourier transform process;

The complex data S that step 7 is obtained ₆carry out inverse fast Fourier transform (IFFT) along each distance to (by row), obtain orientation time domain distance wavenumber domain complex data S ₇;

Step 9: geometry correction process;

The complex data S that step 8 is obtained ₇with geometry correction factor Ω ₃be multiplied, obtain the complex data S after geometry correction ₈;

Step 10: distance is to Fourier transform process;

The complex data S that step 9 is obtained ₉carry out inverse fast Fourier transform (IFFT) along each orientation moment (by row), obtain final imaging results S _end;

The invention has the advantages that:

(1) the present invention proposes a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, solve and look side ways based on the Airborne High-resolution of slip receiver window technology the present situation that Spotlight SAR Imaging raw radar data does not have formation method at present.

(2) the present invention proposes a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, there is the feature of high precision focal imaging.Formation method due to the present invention's proposition is a kind of wavenumber domain formation method of improvement, and the advantage of wavenumber domain formation method is, as long as meet this condition of constant airspeed (carried SAR meets this condition just), just can realize high precision and focus on.Therefore, the present invention is utilized can to realize scene high precision focal imaging.

(3) the present invention proposes a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, have the advantages that applicability is strong.On the one hand, the formation method due to the present invention's proposition is a kind of wavenumber domain formation method of improvement, and wavenumber domain formation method is by the restriction of stravismus angle, therefore, under the condition that stravismus angle is very large, the present invention can realize the vernier focusing of scene equally, obtains high-quality SAR image.On the other hand, the present invention can realize the imaging of 0.1m ultrahigh resolution, and obtains vernier focusing, and therefore, the present invention is applicable to the imaging requirements of current various resolution.

Accompanying drawing explanation

Fig. 1 is a kind of stravismus of the Airborne High-resolution based on slip receiver window Spotlight SAR Imaging formation method process flow diagram that the present invention proposes;

Fig. 2 is embodiment simulating scenes schematic diagram;

Fig. 3 is embodiment imaging results;

Fig. 4 is embodiment upper left point target sectional view;

Fig. 5 is embodiment intermediate point object profile figure;

Fig. 6 is embodiment lower-right most point object profile figure.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.

The present invention proposes a kind of airborne resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, process to as if based on the Airborne High-resolution stravismus Spotlight SAR Imaging raw radar data of slip receiver window, the result obtained is a panel height resolution oblique-view image.

Slip receiver window refers to that SAR is when receiving echo, and echo reception window start-up time changed with the orientation moment, thus reduced the range migration of target, eliminated the range walk (linear segment in range migration) of target in other words.Adopt SAR system each orientation moment echo window start-up time T'(i at work of slip receiver window) be:

$T^{\′}\left(i\right)=T_0-\frac{\mathrm{\λ}\·{\mathrm{fd}}_0\·t\left(i\right)}{c}---\left(1\right)$

Wherein, T ₀traditional SAR system fixed echo window start-up time under the same terms, fd ₀refer to doppler centroid, λ refers to signal wavelength, and c refers to signal velocity, and t (i) refers to the orientation moment, and , i=0,1,2 ..., N _a-1.Slip receiver window technology reduces the range migration amount of echo, and then reduces echo data amount, alleviates the pressure of data storage and real time imagery.

The present invention is a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, and idiographic flow as shown in Figure 1, comprises the following steps:

Step one: read in the Airborne High-resolution stravismus Spotlight SAR Imaging two dimension original echo complex data S based on slip receiver window _startand corresponding imaging parameters.Wherein, S _startbe a two-dimensional complex number group, size is N _a× N _r, and imaging parameters specifically comprises: orientation is to sampling number N _a, distance is to sampling number N _r, signal sampling rate f _s, signal bandwidth Bw, chirp rate b, pulse repetition rate PRF, with reference to oblique distance R _ref, doppler centroid fd ₀, doppler frequency rate f _r0, satellite velocities P _v, equivalent squint angle , signal wavelength lambda, signal carrier frequency f ₀, signal velocity c;

Step 2: by two-dimentional original echo complex data S _startcarry out orientation to the process of solution linear frequency modulation, specifically can be divided into following step:

(a) structure two one-dimensional sequence i, j, wherein i represents orientation to sequence (OK), and j represents distance to sequence (row);

i＝[1,2,…,N _a] （2）

j＝[1,2,…,N _r]

B () obtains two-dimentional original echo complex data S _startorientation moment t (i) that each row is corresponding;

$t\left(i\right)=\frac{i-N_a/2}{\mathrm{PRF}}---\left(3\right)$

C () is by two-dimentional original echo complex data S _startwith removing twiddle factor H ₁be multiplied, obtain data S _{1_1}(i, j), wherein factor H ₁i () is size is N _athe one dimension plural groups of × 1, formula is:

H ₁(i)＝exp{jπ(fr ₀·t ²(i)+2·fd ₀·t(i))} （4）

Then two-dimensional complex number group S _{1_1}can be drawn by following formula:

S _{1_1}(i,j)＝S _start(i,j)·H ₁(i) （5）

D () is to complex data S _{1_1}carry out fast Fourier change (FFT) along each distance to (by row), obtain complex data S _{1_2};

S _{1_2}(:,j)＝FFT(S _{1_1}(:,j)) （6）

Wherein, S _{1_2}(:, j) represent S _{1_2}n-th row, S _{1_1}(:, j) represent S _{1_1}n-th row, FFT () represent Fast Fourier Transform (FFT) is carried out to one-dimension array.

E () obtains and removes postrotational equivalent pulse repetition frequency PRF';

${\mathrm{PRF}}^{\′}=\frac{N_r\·{\mathrm{fr}}_0}{\mathrm{PRF}}---\left(7\right)$

G () convolution (7) obtains two-dimensional complex number according to often going corresponding orientation moment t ₁(i);

$t_1\left(i\right)=\frac{i-N_a/2}{{\mathrm{PRF}}^{\′}}---\left(8\right)$

H () is by complex data S _{1_2}with factor H ₂be multiplied, obtain final orientation to the data S separated after linear frequency modulation ₁.Wherein factor H ₂i () is size is N _athe one dimension plural groups of × 1, its formula is:

$H_2\left(i\right)=\exp \{\mathrm{j\π}\·{\mathrm{fr}}_0\·t_1^2\left(i\right)\}---\left(9\right)$

Then two-dimensional complex number group S ₁can be drawn by following formula:

S ₁(i,j)＝S _{1_2}(i,j)·H ₂(i) （10）

Step 3: the complex data S that step 2 is obtained ₁(i, j) carries out Fast Fourier Transform (FFT) (FFT) along each distance to (by row), obtains azimuth spectrum complex data S ₂(i, j);

S ₂(:,j)＝FFT(S ₁(:,j)) （11）

Wherein, S ₂(:, j) represent S ₂jth row, S ₁(:, j) represent S ₁jth row, FFT () represent Fast Fourier Transform (FFT) is carried out to one-dimension array.

Step 4: the complex data S that step 3 is obtained ₂(i, j) with the orientation in corresponding orientation moment to solution linear frequency modulation residual phase offset compensating factor Ω ₁i () is multiplied, the complex data S after being compensated ₃(i, j);

A () convolution (7) obtains two-dimentional orientation to frequency domain distance to time domain complex data S ₂(i, j) orientation frequency f that often row is corresponding _a(i);

$f_a\left(i\right)=\frac{i-N_a/2}{N_a}\·{\mathrm{PRF}}^{\′}---\left(12\right)$

B () convolution (12) obtains size is N _athe one dimension compensating factor Ω of × 1 ₁(i);

${\mathrm{\Ω}}_1\left(i\right)=\exp \{\mathrm{j\π}\·\frac{f_a^2\left(i\right)}{{\mathrm{fr}}_0}\}---\left(13\right)$

C () obtains the two-dimensional complex number after compensating according to S ₃(i, j);

S ₃(i,j)＝S ₂(i,j)·Ω ₁(i) （14）

Step 5: the complex data S that step 4 is obtained ₃(i, j) carries out Fast Fourier Transform (FFT) (FFT) along each orientation moment (by row), obtains 2-d spectrum complex data S ₄(i, j);

S ₄(i,:)＝FFT(S ₃(i,:)) （15）

Wherein, S ₃(i :) represent S ₃the i-th row, S ₄(i :) represent S ₄the i-th row.

Step 6: the complex data S that step 5 is obtained ₄(i, j) is with corresponding consistent compressibility factor Ω ₂(i, j) is multiplied, and is slightly focused on complex data S ₅(i, j).

A () is according to reference oblique distance R _refobtain the shortest oblique distance R _min;

$R_{\min }=R_{\mathrm{ref}}-\frac{c}{{2f}_s}\·\frac{N_r}{2}---\left(16\right)$

B () obtains two-dimensional frequency complex data S ₄(i, j) often arranges corresponding distance frequency domain f _τ(j);

$f_{\mathrm{\τ}}\left(j\right)=\frac{j-N_r/2}{\mathrm{Nr}}\·f_s---\left(17\right)$

C () convolution (12) and formula (17) obtain two-dimensional frequency complex data S ₄(i, j) often arranges corresponding orientation to wave number k _xi () and often capable corresponding distance are to wave number k _rc(j);

$k_x\left(i\right)=\frac{2\mathrm{\π}{f}_a\left(i\right)}{P_v}$ $\left(18\right)$

$k_{\mathrm{rc}}\left(j\right)=\frac{4\mathrm{\π}(f_0+f_{\mathrm{\τ}}\left(j\right))}{c}$

D () convolution (16) ~ formula (18) obtains size is N _a× N _rthe consistent compressibility factor Ω of two dimension ₂(i, j);

E () convolution (19) obtains the two-dimensional complex number after consistent compression according to S ₅(i, j);

S ₅(i,j)＝S ₄(i,j)·Ω ₂(i,j) （20）

Step 7: the complex data S that step 6 is obtained ₅(i, j), utilizes sinc method of interpolation to carry out stolt interpolation processing, obtains the complex data S being mapped to two-dimentional wavenumber domain by two-dimensional frequency ₆(i, j);

A (), according to the mapping relations of distance frequency domain to distance wavenumber domain, convolution (18) obtains distance wavenumber domain wave number k' _rc(i, j);

B () traversal obtains distance wavenumber domain wave number k' _rcthe maximal value k' of (i, j) _{rc, max}with minimum value k' _{rc, max}, and obtain point interval delta k' such as distance wavenumber domain wave number _rc;

$\mathrm{\Δ}{k}_{\mathrm{rc}}^{\′}=\frac{k_{\mathrm{rc},\max }^{\′}-k_{\mathrm{rc},\min }^{\′}}{N_r}---\left(22\right)$

C () obtains two-dimentional wavenumber domain data uniform distance wavenumber domain wave number

$k_{\mathrm{rc}}^e\left(j\right)=k_{\mathrm{rc}}^{\′}\min +j\·{\mathrm{\Δk}}_{\mathrm{rc}}^{\′}---\left(23\right)$

D () obtains each uniform distance wavenumber domain wave number of two-dimentional wavenumber domain complex data at the uneven k' that often row is corresponding _rcposition p (i, j) in (i :).

Method is for carry out following operation by row: to obtain the position p (1,1) of the first row first row for example, first obtain absolute difference n=[1,2 ..., N _r], obtain minimum absolute difference Δ k _minwith the n of correspondence position, if $k_{\mathrm{rc}}^e\left(1\right)<k_{\mathrm{rc}}^{\&prime;}(1,n),$ P (1,1)=n-1, if $k_{\mathrm{rc}}^e\left(1\right)\&GreaterEqual;k_{\mathrm{rc}}^{\&prime;}(1,n),$ p(1,1)＝n，

By that analogy, each position p (i, j) is obtained.

E () combines and above walks position p (i, j) obtained, obtain sampled point interval q (i, j, n) needed for sinc interpolation;

$q(i,j,n)=\frac{k_{\mathrm{rc}}^e\left(j\right)-k_{\mathrm{rc}}^{\&prime;}(i,(p(i,j)+n))}{\frac{4\mathrm{\&pi;}{f}_s}{c}\&CenterDot;\frac{1}{N_r}},$ n＝[-N/2,-N/2+1,…,N/2-1] （24）

F () convolution (24) utilizes sinc method of interpolation, obtain out the 2-D data S after stolt interpolation ₆, because 2-D data is complex data, need respectively to S ₆the real part S of (i, j) _{6_re}(i, j) and imaginary part S _{6_im}(i, j) carries out the acquisition of sinc method of interpolation respectively and draws.

Wherein, N is interpolation kernel length, and sinc () refers to interpolating function s _{5_re}(i, j) refers to 2-D data S _{5_re}the real part of the i-th row jth row, S _{5_im}(i, j) refers to 2-D data S _{5_im}the imaginary part of the i-th row jth row.

Step 8: the complex data S that step 7 is obtained ₆(i, j) carries out inverse fast Fourier transform (IFFT) along each distance to (by row), obtains orientation time domain distance wavenumber domain complex data S ₇(i, j);

S ₇(:,j)＝IFFT(S ₆(:,j)) （27）

Wherein, S ₆(:, j) represent S ₆jth row, S ₇(:, j) represent S ₇jth row, IFFT () represent inverse fast Fourier transform is carried out to one-dimension array.

Step 9: the complex data S that step 8 is obtained ₇(i, j) is with geometry correction factor Ω ₃(i, j) is multiplied, and obtains the complex data S after geometry correction ₈(i, j);

A () convolution (8) and formula (12), obtain geometry correction factor Ω ₄(i, j);

${\mathrm{\&Omega;}}_4(i,j)=\exp \{-j2\mathrm{\&pi;}\&CenterDot;f_a\left(i\right)\&CenterDot;\frac{\mathrm{\&lambda;}\&CenterDot;{\mathrm{fd}}_0\&CenterDot;t_1\left(i\right)}{c}\}---\left(28\right)$

B () utilizes formula (30) to obtain the complex data S after geometry correction ₈(i, j);

S ₈(i,j)＝S ₇(i,j)·Ω ₃(i,j) （29）

Step 10: the complex data S that step 9 is obtained ₈(i, j) carries out inverse fast Fourier transform (IFFT) along each orientation moment (by row), obtains final imaging results S _end(i, j);

S _end(i,:)＝IFFT(S ₈(i,:)) （30）

Wherein, S ₈(i :) represent S ₈the i-th row, S _end(i :) represent S _endthe i-th row.

Embodiment:

The present embodiment proposes a kind of Airborne High-resolution based on slip receiver window stravismus Spotlight SAR Imaging formation method, simulating scenes is 3 × 3 dot matrix, its point is spaced apart 100m with point, finally respectively upper left, centre, three, bottom right point in simulating scenes are assessed, as shown in Figure 2, the imaging parameters related in its imaging process is as shown in table 1 for concrete simulating scenes.

Table 1 embodiment parameter

The present embodiment specifically comprises the following steps:

Step one: read in the Airborne High-resolution stravismus Spotlight SAR Imaging two dimension original echo complex data S based on slip receiver window _startand corresponding imaging parameters.Wherein, S _startbe two-dimensional complex number group, size is 65536 × 16384, and concrete imaging parameters is as shown in table 1;

Step 2: by two-dimentional original echo complex data S _startcarry out orientation to the process of solution linear frequency modulation, concrete operation step is:

(a) structuring one-dimensional sequence, as the formula (2), i=[1,2 ..., 65536], j=[1,2 ..., 16384];

B () obtains two-dimentional original echo complex data S _startorientation moment t (i) that often row is corresponding, detailed process is undertaken by formula (3);

C () is by data S _startwith removing twiddle factor H ₁i () is multiplied, obtain data S _{1_1}(i, j).Wherein factor H ₁to be size be 65536 × 1 one dimension plural groups, concrete acquisition process is undertaken by formula (4), and two-dimensional complex number group S _{1_1}(i, j) acquisition process is undertaken by formula (5);

D () is to complex data S _{1_1}(i, j) along each distance to carrying out fast Fourier change (FFT), obtains complex data S by formula (6) _{1_2}(i, j);

E () obtains and removes postrotational equivalent pulse repetition frequency PRF', acquisition process is undertaken by formula (7);

G () convolution (7) obtains two-dimensional complex number according to often going corresponding orientation moment t ₁(i), acquisition process is undertaken by formula (8);

H () is by data S _{1_2}(i, j) is with factor H ₂i () is multiplied, obtain final orientation to the data S separated after linear frequency modulation ₁(i, j).Wherein factor H ₂(i) to be size be 65536 × 1 one dimension plural groups, its acquisition process is undertaken by formula (9), and two-dimensional complex number group S ₁acquisition process is undertaken by formula (10);

Step 3: the complex data S that step 2 is obtained ₁(i, j) carries out Fast Fourier Transform (FFT) (FFT) along each distance to (by row), obtains azimuth spectrum complex data S ₂(i, j), specific operation process is undertaken by formula (11);

Step 4: the complex data S that step 3 is obtained ₂(i, j) with the orientation in corresponding orientation moment to solution linear frequency modulation residual phase offset compensating factor Ω ₁i () is multiplied, the complex data S after being compensated ₃(i, j);

A () obtains two-dimentional orientation to frequency domain distance to time domain complex data S ₂(i, j) orientation frequency f that often row is corresponding _a, concrete acquisition process is undertaken by formula (12);

B () convolution (7) obtains the one dimension compensating factor Ω that size is 65536 × 1 ₁(i), concrete acquisition process is undertaken by formula (13);

C () obtains the two-dimensional complex number after compensating according to S ₃(i, j), concrete acquisition process is undertaken by formula (14);

Step 5: the complex data S that step 4 is obtained ₃(i, j) carries out Fast Fourier Transform (FFT) (FFT) along each orientation moment (by row), obtains 2-d spectrum complex data S ₄(i, j), specific operation process is undertaken by formula (15);

Step 6: the complex data S that step 5 is obtained ₄(i, j) is with corresponding consistent compressibility factor Ω ₂(i, j) is multiplied, and is slightly focused on complex data S ₅(i, j).

A () is according to reference oblique distance R _ref=19.31km obtains the shortest oblique distance R _min, concrete acquisition process is undertaken by formula (16);

B () obtains two-dimensional frequency complex data S ₄(i, j) often arranges corresponding distance frequency domain f _τ(j), concrete acquisition process is undertaken by formula (17);

C () convolution (7) and formula (11) obtain two-dimensional frequency complex data S ₄(i, j) often arranges corresponding orientation to wave number k _xi () and often capable corresponding distance are to wave number k _rc(j), concrete acquisition process is undertaken by formula (18);

D () convolution (10) ~ formula (12) obtains the consistent compressibility factor Ω of two dimension that size is 65536 × 16384 ₂(i, j), concrete acquisition process is undertaken by formula (19);

E () convolution (13) obtains the two-dimensional complex number after consistent compression according to S ₅(i, j), concrete acquisition process is undertaken by formula (20);

Step 7: the complex data S that step 6 is obtained ₅(i, j), utilizes sinc method of interpolation to carry out stolt interpolation processing, obtains the complex data S being mapped to two-dimentional wavenumber domain by two-dimensional frequency ₆(i, j);

A (), according to the mapping relations of distance frequency domain to distance wavenumber domain, convolution (18) obtains distance wavenumber domain wave number k' _rc(i, j), concrete acquisition process is undertaken by formula (21);

B () traversal obtains distance wavenumber domain wave number k' _rcthe maximal value k' of (i, j) _{rc, max}=69.12rad/s and minimum value k' _{rc, max}=-105.06rad/s, and obtain point interval delta k' such as distance wavenumber domain wave number _rc, concrete operations are undertaken by formula (22);

C () obtains two-dimentional wavenumber domain complex data uniform distance wavenumber domain wave number concrete operations are undertaken by formula (23);

D () obtains each uniform distance wavenumber domain wave number of two-dimentional wavenumber domain complex data at the uneven k' that often row is corresponding _rcposition p (i, j) in (i :).Method is for carry out following operation by row: the position p (1,1) first obtaining the first row first row, obtains absolute difference n=[1,2 ..., 16384], obtain minimum absolute difference Δ k _minwith the n of correspondence position, if $k_{\mathrm{rc}}^e\left(1\right)<k_{\mathrm{rc}}^{\&prime;}(1,n),$ P (1,1)=n-1, if $k_{\mathrm{rc}}^e\left(1\right)\&GreaterEqual;k_{\mathrm{rc}}^{\&prime;}(1,n),$ p(1,1)＝n，

By that analogy, each position p (i, j) is obtained.

E () combines and above walks position p (i, j) obtained, obtain sampled point interval q (i, j, n) needed for sinc interpolation, concrete operations are undertaken by formula (24);

F () convolution (18) utilizes sinc method of interpolation, select sinc interpolation kernel length to be N=8, obtain out the two-dimensional complex number after stolt interpolation according to S ₆(i, j), because 2-D data is complex data, needs respectively to S ₆the real part S of (i, j) _{6_re}(i, j) and imaginary part S _{6_im}(i, j) carries out the acquisition of sinc method of interpolation respectively and draws, concrete operations are undertaken by formula (25) (26).

Step 8: the complex data S that step 7 is obtained ₆(i, j) carries out inverse fast Fourier transform (IFFT) along each distance to (by row), obtains orientation time domain distance wavenumber domain complex data S ₇(i, j), concrete operations are undertaken by formula (27);

Step 9: the complex data S that step 8 is obtained ₇(i, j) is with geometry correction factor Ω ₃(i, j) is multiplied, and obtains the complex data S after geometry correction ₈(i, j);

A () convolution (8) and formula (12), obtain geometry correction factor Ω ₃(i, j), concrete operations are undertaken by formula (28);

B () utilizes formula (28) to obtain the complex data S after geometry correction ₈(i, j), concrete operations are undertaken by formula (29);

Step 10: the complex data S that step 9 is obtained ₈(i, j) carries out inverse fast Fourier transform (IFFT) along each orientation moment (by row), obtains final imaging results S _end(i, j), concrete operations are undertaken by formula (30);

Through the imaging processing of above-mentioned steps, obtain final scene imaging result as shown in Figure 3.Table 2 gives the Imaging Evaluation result of scene upper left, centre, three point targets in bottom right, and Fig. 4, Fig. 5, Fig. 6 sets forth the sectional view of scene upper left, centre, three point targets in bottom right.

Table two Imaging Evaluation result

Sectional view according to table 2 assessment result and Fig. 4 ~ Fig. 6, can draw: on the one hand, and this formation method still can vernier focusing when to look side ways angle be 70 degree, illustrates that method that the present invention proposes is not subject to look side ways the restriction of angle; On the other hand, this formation method still can vernier focusing for 0.1m super-resolution, illustrates that the method that the present invention proposes can realize vernier focusing to current various resolution.Therefore, method proposed by the invention can realize the Airborne High-resolution angle of squint Spotlight SAR Imaging accurately image based on slip receiver window, obtains high-precision imaging results.

Claims (1)

1., based on an Airborne High-resolution stravismus Spotlight SAR Imaging formation method for slip receiver window, comprise the following steps:

Step one: read in the Airborne High-resolution stravismus Spotlight SAR Imaging two dimension original echo complex data S based on slip receiver window _startand corresponding imaging parameters;

S _startbe a two-dimensional complex number group, size is N _a× N _r, imaging parameters comprises: orientation is to sampling number N _a, distance is to sampling number N _r, signal sampling rate f _s, signal bandwidth Bw, chirp rate b, pulse repetition rate PRF, with reference to oblique distance R _ref, doppler centroid fd ₀, doppler frequency rate f _r0, satellite velocities P _v, equivalent squint angle , signal wavelength lambda, signal carrier frequency f ₀, signal velocity c;

Step 2: by two-dimentional original echo complex data S _startcarry out orientation to the process of solution linear frequency modulation, specifically comprise following step:

(a) structure two one-dimensional sequence i, j, wherein i represents orientation to sequence, and j represents distance to sequence;

i＝[1,2,…,N _a]

(2)

j＝[1,2,…,N _r]

B () obtains two-dimentional original echo complex data S _startorientation moment t (i) that each row is corresponding;

$t\left(i\right)=\frac{i-N_a/2}{\mathrm{PRF}}---\left(3\right)$

C () is by two-dimentional original echo complex data S _startwith remove twiddle factor H ₁be multiplied, obtain data S _{1_1}(i, j), wherein factor H ₁i () is size is N _athe one dimension plural groups of × 1, formula is:

H ₁(i)＝exp{jπ(fr ₀·t ²(i)+2·fd ₀·t(i))} (4)

Then two-dimensional complex number group S _{1_1}drawn by following formula:

S _{1_1}(i,j)＝S _start(i,j)·H ₁(i) (5)

D () is to complex data S _{1_1}along each distance to carrying out fast Fourier change, obtain complex data S _{1_2};

S _{1_2}(:,j)＝FFT(S _{1_1}(:,j)) (6)

Wherein, S _{1_2}(:, j) represent S _{1_2}jth row, S _{1_1}(:, j) represent S _{1_1}jth row, FFT () represent Fast Fourier Transform (FFT) is carried out to one-dimension array;

E () obtains and removes postrotational equivalent pulse repetition frequency PRF';

$\mathrm{PR}{F}^{\&prime;}=\frac{N_r\&CenterDot;{\mathrm{fr}}_0}{\mathrm{PRF}}---\left(7\right)$

G () convolution (7) obtains two-dimensional complex number according to often going corresponding orientation moment t ₁(i);

$t_1\left(i\right)=\frac{i-N_a/2}{{\mathrm{PRF}}^{\&prime;}}---\left(8\right)$

H () is by complex data S _{1_2}with factor H ₂be multiplied, obtain final orientation to the data S separated after linear frequency modulation ₁; Wherein factor H ₂i () is size is N _athe one dimension plural groups of × 1, its formula is:

$H_2\left(i\right)=\exp \{\mathrm{j\&pi;}\&CenterDot;{\mathrm{fr}}_0\&CenterDot;t_1^2\left(i\right)\}---\left(9\right)$

Then two-dimensional complex number group S ₁drawn by following formula:

S ₁(i,j)＝S _{1_2}(i,j)·H ₂(i) (10)

Step 3: the complex data S that step 2 is obtained ₁(i, j), along each distance to carrying out Fast Fourier Transform (FFT), obtains azimuth spectrum complex data S ₂(i, j);

S ₂(:,j)＝FFT(S ₁(:,j)) (11)

Wherein, S ₂(:, j) represent S ₂jth row, S ₁(:, j) represent S ₁jth row, FFT () represent Fast Fourier Transform (FFT) is carried out to one-dimension array;

Step 4: the complex data S that step 3 is obtained ₂(i, j) with the orientation in corresponding orientation moment to solution linear frequency modulation residual phase offset compensating factor Ω ₁i () is multiplied, the complex data S after being compensated ₃(i, j);

A () convolution (7) obtains two-dimentional orientation to frequency domain distance to time domain complex data S ₂(i, j) orientation frequency f that often row is corresponding _a(i);

$f_a\left(i\right)=\frac{i-N_a/2}{N_a}\&CenterDot;{\mathrm{PRF}}^{\&prime;}---\left(12\right)$

B () convolution (12) obtains size is N _athe one dimension compensating factor Ω of × 1 ₁(i);

${\mathrm{\&Omega;}}_1\left(i\right)=\exp \{\mathrm{j\&pi;}\&CenterDot;\frac{f_a^2\left(i\right)}{{\mathrm{fr}}_0}\}---\left(13\right)$

C () obtains the two-dimensional complex number after compensating according to S ₃(i, j);

S ₃(i,j)＝S ₂(i,j)·Ω ₁(i) (14)

Step 5: the complex data S that step 4 is obtained ₃(i, j) carries out Fast Fourier Transform (FFT) along each orientation moment, obtains 2-d spectrum complex data S ₄(i, j);

S ₄(i,:)＝FFT(S ₃(i,:)) (15)

Wherein, S ₃(i :) represent S ₃the i-th row, S ₄(i :) represent S ₄the i-th row;

Step 6: the complex data S that step 5 is obtained ₄(i, j) is with corresponding consistent compressibility factor Ω ₂(i, j) is multiplied, and is slightly focused on complex data S ₅(i, j);

A () is according to reference oblique distance R _refobtain the shortest oblique distance R _min;

$R_{\min }=R_{\mathrm{ref}}-\frac{c}{2f_s}\&CenterDot;\frac{N_r}{2}---\left(16\right)$

B () obtains two-dimensional frequency complex data S ₄(i, j) often arranges corresponding distance frequency domain f _τ(j);

$f_{\mathrm{\&tau;}}\left(j\right)=\frac{j-N_r/2}{\mathrm{Nr}}\&CenterDot;f_s---\left(17\right)$

C () convolution (12) and formula (17) obtain two-dimensional frequency complex data S ₄(i, j) often arranges corresponding orientation to wave number k _xi () and often capable corresponding distance are to wave number k _rc(j);

$\begin{array}{c}{k}_x\left(i\right)=\frac{2\mathrm{\&pi;}{f}_a\left(i\right)}{P_v}\\ k_{\mathrm{rc}}\left(j\right)=\frac{4\mathrm{\&pi;}(f_0+f_{\mathrm{\&tau;}}\left(j\right))}{c}\end{array}---\left(18\right)$

D () convolution (16) ~ formula (18) obtains size is N _a× N _rthe consistent compressibility factor Ω of two dimension ₂(i, j);

E () convolution (19) obtains the two-dimensional complex number after consistent compression according to S ₅(i, j);

S ₅(i,j)＝S ₄(i,j)·Ω ₂(i,j) (20)

Step 7: the complex data S that step 6 is obtained ₅(i, j), utilizes sinc method of interpolation to carry out stolt interpolation processing, obtains the complex data S being mapped to two-dimentional wavenumber domain by two-dimensional frequency ₆(i, j);

A (), according to the mapping relations of distance frequency domain to distance wavenumber domain, convolution (18) obtains distance wavenumber domain wave number k' _rc(i, j);

B () traversal obtains distance wavenumber domain wave number k' _rcthe maximal value k' of (i, j) _{rc, max}with minimum value k' _{rc, max}, and obtain point interval delta k' such as distance wavenumber domain wave number _rc;

${\mathrm{\&Delta;k}}_{\mathrm{rc}}^{\&prime;}=\frac{k_{\mathrm{rc},\max }^{\&prime;}-k_{\mathrm{rc},\min }^{\&prime;}}{N_r}---\left(22\right)$

C () obtains two-dimentional wavenumber domain data uniform distance wavenumber domain wave number

$k_{\mathrm{rc}}^e\left(j\right)=k_{\mathrm{rc}}^{\&prime;}\min +j\&CenterDot;{\mathrm{\&Delta;k}}_{\mathrm{rc}}^{\&prime;}---\left(23\right)$

D () obtains each uniform distance wavenumber domain wave number of two-dimentional wavenumber domain complex data at the uneven k' that often row is corresponding _rcposition p (i, j) in (i :);

Be specially: for example, first obtain absolute difference with the position p (1,1) obtaining the first row first row n=[1,2 ..., N _r], obtain minimum absolute difference Δ k _minwith the n of correspondence position, if p (1,1)=n-1, if p (1,1)=n, in like manner, by that analogy, obtains each position p (i, j);

E () combines and above walks position p (i, j) obtained, obtain sampled point interval q (i, j, n) needed for sinc interpolation;

$q(i,j,n)=\frac{k_{\mathrm{rc}}^e\left(j\right)-k_{\mathrm{rc}}^{\&prime;}(i,(p(i,j)+n))}{\frac{4\mathrm{\&pi;}{f}_s}{c}\&CenterDot;\frac{1}{N_r}},n=[-N/2,-N/2+1,...,N/2-1]---\left(24\right)$

F () convolution (24) utilizes sinc method of interpolation, obtain out the 2-D data S after stolt interpolation ₆, because 2-D data is complex data, need respectively to S ₆the real part S of (i, j) _{6_re}(i, j) and imaginary part S _{6_im}(i, j) carries out the acquisition of sinc method of interpolation respectively and draws;

Wherein, N is interpolation kernel length, and sinc () refers to interpolating function s _{5_re}(i, j) refers to 2-D data S _{5_re}the real part of the i-th row jth row, S _{5_im}(i, j) refers to 2-D data S _{5_im}the imaginary part of the i-th row jth row;

Step 8: the complex data S that step 7 is obtained ₆(i, j), along each distance to carrying out inverse fast Fourier transform, obtains orientation time domain distance wavenumber domain complex data S ₇(i, j);

S ₇(:,j)＝IFFT(S ₆(:,j)) (27)

Wherein, S ₆(:, j) represent S ₆jth row, S ₇(:, j) represent S ₇jth row, IFFT () represent inverse fast Fourier transform is carried out to one-dimension array;

Step 9: the complex data S that step 8 is obtained ₇(i, j) is with geometry correction factor Ω ₃(i, j) is multiplied, and obtains the complex data S after geometry correction ₈(i, j);

A () convolution (8) and formula (12), obtain geometry correction factor Ω ₄(i, j);

${\mathrm{\&Omega;}}_4(i,j)=\exp \{-j2\mathrm{\&pi;}\&CenterDot;f_a\left(i\right)\&CenterDot;\frac{\mathrm{\&lambda;}\&CenterDot;{\mathrm{fd}}_0\&CenterDot;t_1\left(i\right)}{c}\}---\left(28\right)$

B () utilizes formula (30) to obtain the complex data S after geometry correction ₈(i, j);

S ₈(i,j)＝S ₇(i,j)·Ω ₃(i,j) (29)

Step 10: the complex data S that step 9 is obtained ₈(i, j) carries out inverse fast Fourier transform along each orientation moment, obtains final imaging results S _end(i, j);

S _end(i,:)＝IFFT(S ₈(i,:)) (30)

Wherein, S ₈(i :) represent S ₈the i-th row, S _end(i :) represent S _endthe i-th row.