CN103163520B - Frequency modulation continuous wave synthetic aperture radar (SAR) sweep frequency non-linear correction method - Google Patents

Abstract

The invention provides a frequency modulation continuous wave synthetic aperture radar (SAR) sweep frequency non-linear correction method. The frequency modulation continuous wave SAR sweep frequency non-linear correction method comprises the following steps: a full-aperture echo signal is divided into two sub-aperture signals with the same time length; sweep frequency non-linear rough evaluation is carried out by utilizing a phase gradient auto-focus algorithm, and rough correction is carried out on the full-aperture echo signal according to the result of rough evaluation; a residual video phase filter is designed according to work parameters and a model of a system; the full-aperture echo signal is enabled to pass through the residual video phase filter; and sweep frequency non-linear fine evaluation is carried out according to the signal after passing the residual video phase filter by means of the phase gradient auto-focus algorithm, and frequency modulation continuous wave SAR sweep frequency non-linear fine correction is achieved. The frequency modulation continuous wave SAR sweep frequency non-linear correction method carries out rough evaluation on the sweep frequency non-linear phase error by utilizing the relevant characteristics between the sub-aperture signals, and eliminates the limitation conditions to an action distance in a traditional method, and the evaluation result has the characteristic of stability.

Description

A kind of Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method

Technical field

The invention belongs to Radar Technology field, specifically, the present invention relates to a kind of Continuous Wave with frequency modulation SAR (Synthetic-aperture radar, synthetic-aperture radar) frequency sweep non-linear correction method.

Background technology

Continuous Wave with frequency modulation SAR is the high-resolution radar imaging device that a kind of volume is little, lightweight, cost is low and antijamming capability is strong, is a current topmost class miniaturization SAR, is widely used in unmanned plane etc. load is required on very harsh flying platform.

Continuous Wave with frequency modulation SAR obtains high range resolution by the linear FM signal in continuous transmitting broadband.In actual wideband radar system, due to the impact of the various factorss such as device property, the frequency transmitting always inevitably can depart from desirable frequency.That is to say, inevitably there is nonlinear problem in actual transmit signal.Paired pulses radar, because the duration of pulse is very short, the nonlinear impact of frequency sweep is not outstanding.And for Continuous Wave with frequency modulation SAR, the nonlinear impact of frequency sweep is very serious, and along with the raising of the indexs such as resolution and operating distance, the nonlinear impact of frequency sweep is also more and more outstanding.Therefore, frequency sweep nonlinear problem becomes and in engineering reality, restricts one of of paramount importance factor of frequency modulated continuous wave radar system performance.

For the frequency sweep nonlinear problem of Continuous Wave with frequency modulation SAR, normally utilize special aobvious point methods to proofread and correct.The maximum operating range that this method requires SAR is the scope of tens of meters.But in SAR application background, mapping swath width arrives thousands of meters hundreds of conventionally, and the frequency sweep that now cannot utilize special aobvious point methods to proofread and correct radar is non-linear.For the frequency sweep gamma correction in Continuous Wave with frequency modulation SAR, the most frequently used method is a kind of bearing calibration that is called three-step approach.This method, to the not restriction of the operating distance of SAR, is also convenient to be integrated in imaging algorithm very much.But it is non-linear that this method need to be measured frequency sweep in advance by hardware device, thereby increased system overhead, limited the dirigibility of its application.

Summary of the invention

The object of the invention is to, short and need the restrictive conditions such as additional hardware equipment for existing Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method operating distance, a kind of frequency sweep nonlinear estimation and bearing calibration based on echoed signal proposed, this method makes full use of the correlation properties of sub-aperture signal, eliminate the dependence that frequency sweep nonlinear phase error is adjusted the distance, realize the nonlinear accurate correction of frequency sweep.

For realizing described object, technical solution of the present invention is: a kind of Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method, the method comprises the following steps: 1. the full aperture echoed signal of reception is carried out to the division of sub-aperture, by orientation, be divided into two sub-aperture signals that time span equates to the time; 2. the correlation properties in distance according to two sub-aperture signals, utilize PGA(Phase Gradient Autofocus, phase gradient self-focusing) algorithm carries out the nonlinear rough estimate of frequency sweep, and according to rough estimate result, full aperture signal slightly proofreaied and correct; 3. according to the running parameter of system and modelling RVP(Residual Video Phase, residual video phase) wave filter; 4. will complete the thick full aperture signal afterwards of proofreading and correct of frequency sweep nonlinear phase error by above-mentioned RVP wave filter; 5. utilize PGA algorithm to carry out the nonlinear essence of frequency sweep according to the signal after RVP filtering and estimate, realize the nonlinear fine correction of Continuous Wave with frequency modulation SAR frequency sweep.

Compared with prior art, beneficial effect of the present invention is:

(1) the present invention only relies on echo data itself and estimates the phase error causing because frequency sweep is non-linear, and it is compensated, and without extra hardware device, is a kind of bearing calibration of pure software, thereby can saves system hardware expense; Be convenient to be integrated in imaging algorithm, there is very strong applicability.

(2) correlation properties that the present invention makes full use of between sub-aperture signal are carried out the rough estimate of frequency sweep nonlinear phase error, have eliminated the restrictive condition of classic method to operating distance;

(3) the present invention comprises to the estimation of frequency sweep nonlinear phase error the step that rough estimate and essence are estimated, therefore has very sane feature.

Accompanying drawing explanation

Fig. 1 is Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method process flow diagram of the present invention;

Fig. 2 is that schematic diagram is divided in the sub-aperture of the present invention;

Fig. 3 is the simulation parameter figure of one embodiment of the invention;

Fig. 4 be one embodiment of the invention distance to compression comparison diagram.

Embodiment

Below in conjunction with the drawings and specific embodiments, a kind of Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method provided by the invention is done to further detailed explanation.

1. the full aperture echoed signal of reception is carried out to the division of sub-aperture, by orientation, be divided into two sub-aperture signals that time span equates to the time.

Considering that in the nonlinear situation of frequency sweep, Continuous Wave with frequency modulation SAR point echoed signal can be written as:

$\mathrm{ss}(t_a,t_r;r_0)=A_0\exp (-j\frac{{4\mathrm{\πf}}_c}{c_0}r)\exp \left({j2\mathrm{\πf}}_d{t}_r\right)\·\exp [-j\frac{{4\mathrm{\πk}}_r}{c_0}(r-r_c)(t_r-\frac{{2r}_c}{c_0})]$

$\exp \left[j\frac{{4\mathrm{\πk}}_r}{c_0^2}{(r-r_c)}^2\right]\·\exp \{-j2\mathrm{\π}[\mathrm{\ϵ}(t_r-\frac{{2r}_c}{c_0})-\mathrm{\ϵ}(t_r-\frac{2r}{c_0})\left]\right\}$

T wherein _afor orientation is to the time, and-T _a/ 2≤t _a<T _a/ 2, T _afor synthetic aperture time, t _rfor distance is to the time, r ₀for the minimum oblique distance of point target, A ₀for echo signal amplitude, f _cfor the centre frequency that transmits, c ₀for the light velocity, f _dfor Doppler frequency, k _rfor chirp rate, r _cfor reference distance, r is instantaneous oblique distance, can be expressed as v is platform movement velocity, for because frequency sweep is non-linear the phase error causing, and corresponding function ε (t) is the integration of e (t).

The non-linear e of frequency sweep (t) can be modeled as:

$e\left(t\right)=\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}[E_{\mathrm{mc}}\cos \left(\frac{2\mathrm{\&pi;m}}{T_r}t\right)+E_{\mathrm{ms}}\sin \left(\frac{2\mathrm{\&pi;m}}{T_r}t\right)]$

T wherein _rfor distance is to data acquisition time, m is the nonlinear harmonic component of frequency sweep, and M is maximum harmonic component, E _mcand E _msthe amplitude that represents respectively cosine component and sinusoidal component, order can obtain for simplicity, conventionally use f _m=m/T _rrepresent the nonlinear harmonic frequency of frequency sweep.The nonlinear order of severity of frequency sweep is led to common linearity L=|e (t _r) | _max/ B _rcharacterize, wherein B _rfor transmitted signal bandwidth, subscript m ax represents to get maximal value, e (t)=LB _rsin (2 π f _mt).

For full aperture signal ss (t _a, t _r; r ₀) process of carrying out the division of sub-aperture is exactly by ss (t _a, t _r; r ₀) by orientation, be divided into surrounding time two parts equal in length to the time:

First sub-aperture signal ss _sub1(t _a, t _r; r ₀) be expressed as: ss _sub1(t _a, t _r; r ₀)=ss (t _a, t _r; r ₀) ,-T _a/ 2≤t _a< 0

Second sub-aperture signal ss _sub2(t _a, t _r; r ₀) be expressed as:

ss _sub2(t _a,t _r;r ₀)=ss((t _a+T _a/2),t _r;r ₀),-T _a/2≤t _a<0

Schematic diagram is divided as shown in Figure 2 in sub-aperture.The echoed signal N of full aperture _r* N _athis two-dimensional matrix represents, wherein N _rfor distance is to sampling number, N _afor orientation is to sampling number.First first half that sub-aperture is matrix, second latter half that sub-aperture is matrix, the size of two sub-aperture signals is N _r* N _a/ 2.

2. the correlation properties in distance according to two sub-aperture signals, utilize PGA(Phase Gradient Autofocus, phase gradient self-focusing) algorithm carries out the nonlinear rough estimate of frequency sweep, and according to rough estimate result, full aperture signal slightly proofreaied and correct.Specifically comprise following sub-step:

(2a) by first sub-aperture signal and distance corresponding to second sub-aperture signal to signal conjugate multiplication:

${\mathrm{ss}}_{\mathrm{diff}}(t_a,t_r;r_0)={\mathrm{ss}}_{\mathrm{sub}1}(t_a,t_r;r_0)\&times;{\mathrm{ss}}_{\mathrm{sub}2}^{*}(t_a,t_r;r_0),-T_a/2\&le;t_a<0$

(2b) utilize PGA algorithm to estimate ss _diff(t _a, t _r; r ₀) phase error, obtain the non-linear e (t of frequency sweep _r) rough estimate

(2c) right carry out integration, and utilize fitting of a polynomial to eliminate constant term error, obtain frequency sweep nonlinear phase error ε (t _r) rough estimate

(2d) to full aperture echoed signal ss (t _a, t _r; r ₀) carry out the thick correction of frequency sweep nonlinear phase error, obtain:

ss _NLC(t _a,t _r;r ₀)=ss(t _a,t _r;r ₀)·H _NLC(t _r)

Coarse phase error correction factor H wherein _nLC(t _r) be:

$H_{\mathrm{NLC}}\left(t_r\right)=\exp \left[j2\mathrm{\&pi;}{\widehat{\mathrm{\&epsiv;}}}_1(t_r-\frac{{2r}_c}{c_0})\right]$

So far complete the nonlinear rough estimate of frequency sweep and take into account thick correction, the realization flow of above-mentioned processing as shown in Figure 3.

3. according to the running parameter of system and modelling RVP wave filter.

Design following RVP wave filter:

$H_{\mathrm{RVPC}}\left(f_r\right)=\exp (-j\frac{{\mathrm{\&pi;f}}_r^2}{k_r})$

Wherein, f _rfor distance is to frequency.

4. by completing the thick full aperture signal afterwards of proofreading and correct of frequency sweep nonlinear phase error by above-mentioned RVP wave filter, make frequency sweep nonlinear phase error and instantaneous oblique distance irrelevant.

(4a) the thick full aperture signal ss afterwards that proofreaies and correct of frequency sweep nonlinear phase error will be completed _nLC(t _a, t _r; r ₀) carry out distance to FFT, obtain sSN _lC(t _a, f _r; r ₀);

(4b) by signal sS _nLC(t _a, f _r; r ₀) by RVP wave filter, obtain:

sS _RVPC(t _a,f _r;r ₀)=sSN _LC(t _a,f _r;r ₀)·H _R"C(f _r)；

(4c) by sS _rVPC(t _a, f _r; r ₀) carry out distance to contrary FFT, obtain two-dimensional time-domain signal ss _rVPC(t _a, t _r; r ₀), the frequency sweep nonlinear phase error of this signal is no longer relevant with instantaneous oblique distance, and only relevant with reference distance.

5. utilize PGA algorithm to carry out the nonlinear essence of frequency sweep according to the signal after RVP filtering and estimate, realize the nonlinear fine correction of Continuous Wave with frequency modulation SAR frequency sweep.

(5a) utilize PGA algorithm to estimate ss _rVPC(t _a, t _r; r ₀) phase error, the essence that obtains frequency sweep nonlinear phase error is estimated

(5b) to ss _rVPC(t _a, t _r; r ₀) carry out the nonlinear fine correction of frequency sweep, obtain:

ss _NLR(t _a,t _r;r ₀)=ss _RVPC(t _a,t _r;r ₀)·H _NLR(t _r)

Smart phase error correction factor H wherein _nLR(t _r) be:

$H_{\mathrm{NLR}}\left(t_r\right)=\exp [-j2\mathrm{\&pi;}{\widehat{\mathrm{\&epsiv;}}}_2(t_r-\frac{{2r}_c}{c_0})]$

So far completing the nonlinear essence of frequency sweep estimates and fine correction.

Method below by emulation is done further checking to the present invention.Simulation parameter as shown in Figure 3, centre frequency f _ccorresponding wavelength X is 0.0086m, chirp rate k _rbe 5 * 10 ¹⁰hz/s.Distance is to sampling number N _rbe 1024 points, Data in Azimuth Direction acquisition time T _abe 2 seconds, orientation is to points N _abe 2048 points.In emulation, point target used is placed on r ₀=4.6km place, and suppose echo amplitude A ₀be 1.Emulation middle distance is to time variable t _rwith orientation to time variable t _avalue be respectively:

$t_r=\frac{n_r}{f_s}=1\&times;{10}^{-6}\&times;n_r,n_r=\mathrm{0,1},\&CenterDot;\&CenterDot;\&CenterDot;1023$

$t_a=\frac{n_a-N_a/2}{f_p}={1\&times;10}^{-3}\&times;(n_a-1024),n_a=\mathrm{0,1},\&CenterDot;\&CenterDot;\&CenterDot;2047$

Non-linear e (the t of frequency sweep used in emulation _r) be:

e(t _r)=L·B _r·sin(2πf _mt _r)=2.5×10 ⁶×sin(2000π×t _r)

Do not proofreading and correct in the nonlinear situation of frequency sweep, distance to compression result as shown in phantom in Figure 4.In figure, result can obviously be found out the video stretching causing because frequency sweep is non-linear, and in the present embodiment, the size of broadening is 20kHz, is about 20 range units.During sub-aperture is divided, first sub-aperture signal is 1024 row above of full aperture echo matrix, and second sub-aperture signal is for 1024 being listed as below.Utilize the present invention carry out distance after frequency sweep gamma correction to compression result as shown in solid line in Fig. 4.By figure, can find out that the Range compress effect after proofreading and correct approaches ideal situation.

Claims (1)

1. a Continuous Wave with frequency modulation SAR frequency sweep non-linear correction method, is characterized in that, comprises the following steps: by the full aperture echoed signal ss (t receiving _a, t _r; r ₀) carry out the division of sub-aperture, by orientation, be divided into two sub-aperture signals that time span equates to the time, obtain first sub-aperture signal ss _sub1(t _a, t _r; r ₀) and second sub-aperture signal ss _sub2(t _a, t _r; r ₀); Correlation properties according to two sub-aperture signals in distance, utilize Phase-gradient autofocus algorithem to carry out the nonlinear rough estimate of frequency sweep, and according to rough estimate result, full aperture signal are slightly proofreaied and correct, and concrete steps are as follows:

By first sub-aperture signal and distance corresponding to second sub-aperture signal to signal conjugate multiplication:

${\mathrm{ss}}_{\mathrm{diff}}(t_a,t_r;r_0)={\mathrm{ss}}_{\mathrm{sub}1}(t_a,t_r;r_0)\&times;{\mathrm{ss}}_{\mathrm{sub}2}^{*}(t_a,t_r;r_0),-T_a/2\&le;t_a<0$

Wherein, t _afor orientation is to the time, and-T _a/ 2≤t _a< T _a2, T _afor synthetic aperture time, t _rfor distance is to the time, r ₀for the minimum oblique distance of point target;

Utilize PGA algorithm to estimate ss _diff(t _a, t _r; r ₀) phase error, obtain the non-linear e (t of frequency sweep _r) rough estimate

Right carry out integration, and utilize fitting of a polynomial to eliminate constant term error, obtain frequency sweep nonlinear phase error ε (t _r) rough estimate

To full aperture echoed signal ss (t _a, t _r; r ₀) carry out the thick correction of frequency sweep nonlinear phase error, obtain:

ss _NLC(t _a,t _r；r ₀)＝ss(t _a,t _r；r ₀)·H _NLC(t _r)，

Wherein, coarse phase error correction factor H _nLC(t _r) be:

$H_{\mathrm{NLC}}\left(t_r\right)=\exp \left[j2\mathrm{\&pi;}{\widehat{\mathrm{\&epsiv;}}}_1(t_r-\frac{{2r}_c}{c_0})\right];$

In above formula, r _cfor reference distance, c ₀for the light velocity;

According to the running parameter of system and modelling residual video phase wave filter, the residual video phase wave filter of design is as follows:

$H_{\mathrm{RVPC}}\left(f_r\right)=\exp (-j\frac{\mathrm{\&pi;}{f}_r^2}{k_r}),$

Wherein, f _rfor distance is to frequency, k _rfor chirp rate;

To complete the thick full aperture signal afterwards of proofreading and correct of frequency sweep nonlinear phase error by above-mentioned residual video phase wave filter, concrete steps are as follows:

The thick full aperture signal ss afterwards that proofreaies and correct of frequency sweep nonlinear phase error will be completed _nLC(t _a, t _r; r ₀) carry out distance to FFT, obtain sS _nLC(t _a, f _r; r ₀);

By signal sS _nLC(t _a, f _r; r ₀) by residual video phase wave filter, obtain:

sS _RVPC(t _a,f _r；r ₀)＝sS _NLC(t _a,f _r；r ₀)·H _RVPC(f _r)；

By sS _rVPC(t _a, f _r; r ₀) carry out distance to contrary FFT, obtain two-dimensional time-domain signal ss _rVPC(t _a, t _r; r ₀); Utilize Phase-gradient autofocus algorithem to carry out the nonlinear essence of frequency sweep according to the signal after residual video phase filtering and estimate, realize the nonlinear fine correction of Continuous Wave with frequency modulation synthetic-aperture radar frequency sweep.