CN100357759C - A self-focusing processing method for synthetic aperture sonar imaging - Google Patents

Abstract

The present invention relates to an auto-focus processing method for the sonar imaging of synthetic apertures, which mainly comprises the following procedures that a broadband analytic signal rp (t) is obtained when an echo wave signal is demodulated by a complex radical band and is processed by matched filtering each time; adjacent echo wave analytic signals are processed relatively; phase error estimation between every two adjacent echo wave analytic signals is obtained by solving a complex-number phase angle; a cumulative phase error caused by a kinematic error is obtained by the cumulative processing of space positions; compensation is carried out by utilizing an obtained phase error sequence { epsilon <p>} so as to realize the correction of echo wave data; a corresponding image reconstruction algorithm is applied by the data after compensation so that a focused sonar image of a synthetic aperture can be obtained. The auto-focus processing method for the sonar imaging of synthetic apertures has the advantages that iterative operation is not needed and the auto-focus processing method for the sonar imaging of synthetic apertures has better real-time operation performance and better capability of macroscopic image correction. In addition, the influence of strong target scattering points and noise on kinematic error estimation is further weakened by alpha weighting. Thus, application conditions are widened.

Description

A kind of self-focusing processing method of synthetic aperture sonar imaging

Technical field

The present invention relates to the signal Processing of synthetic aperture sonar imaging, mainly is a kind of self-focusing processing method of synthetic aperture sonar imaging, is used to obtain the acoustic imaging of meticulousr sea-bed topography mapping and hypostasis detection.

Background technology

Synthetic aperture sonar (Synthetic Aperture Sonar is called for short SAS) is a kind of high resolution acoustics imaging equipment.It is to utilize small-bore basic matrix linear uniform motion, sequential transmissions and received signal in motion process, and with the superposition that is concerned with of received signal in one section motion course, thereby obtain being equivalent to the synthetic virtual aperture battle array of several times to tens times of actual physics acoustic matrixs, obtain very high resolving power and spatial gain.Yet this condition does not satisfy in the real navigation process, makes synthetic aperture image produce distortion in intensity and geometric distortion.The self-focusing processing method of realizing the kinematic error compensation is a gordian technique of synthetic aperture sonar.

The estimation of synthetic aperture sonar and bearing calibration are divided into two classes usually: a class is for adopting the hardware compensating of sensor, and another kind of is that autofocus algorithm compensates.Because sensor accuracy is difficult to satisfy the requirement of synthetic aperture sonar, synthetic aperture sonar generally all needs to adopt autofocus algorithm to realize the kinematic error compensation.Autofocus algorithm can be divided into two kinds of data field autofocus algorithm (as redundant phase center autofocus algorithm) and image area autofocus algorithm (as the phase gradient self-focusing) algorithms usually again.

The autofocus algorithm that external synthetic aperture sonar is adopted mainly is redundant phase center (Redundance Phase Center, be called for short RPC) autofocus algorithm, it is that the redundancy that utilize to receive echoed signal is handled, and needs to adopt redundant submatrix design or sacrifices spatial sampling and realize.Classical phase gradient self-focusing (Phase Gradient Autofocus is called for short PGA) algorithm need carry out iterative processing and just have good estimated performance, and classical self-focusing processing method need be selected strong target scattering point simultaneously.

Summary of the invention

The objective of the invention is in order to overcome the deficiency of said method, and a kind of self-focusing processing method of synthetic aperture sonar imaging is provided.

The technical solution adopted for the present invention to solve the technical problems.The self-focusing processing method of this synthetic aperture sonar imaging mainly may further comprise the steps:

(1), each echoed signal obtains broadband analytic signal r through complex radical band demodulation and matched filter processing _p(t);

(2), relevant treatment between the adjacent echo analytic signal; The broadband analytic signal r of adjacent two transponder pulse echoes _p(t), r _P-1(t) computing cross-correlation is:

${\mathrm{rr}}_p={\&Integral;}_0^T{r}_p^{*}\left(t\right)r_{p-1}\left(t\right)\mathrm{dt}---\left(5\right)$

Represent complex conjugate in the formula;

(3), by asking plural phase angle, obtain the phase error estimation and phase error between the adjacent echo analytic signal:

$\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p=\mathrm{Arg}\left\{{\mathrm{rr}}_p\right\}---\left(6\right)$

(4), obtain the accumulated phase error that kinematic error causes by the locus accumulated process:

${\widehat{\mathrm{\&epsiv;}}}_p={\widehat{\mathrm{\&epsiv;}}}_{p-1}+\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p;$ ${\widehat{\mathrm{\&epsiv;}}}_0\&equiv;0---\left(7\right)$

(5), utilize the sequence of phase errors that obtains

Compensate and realize the echo data correction:

$R_p^{\mathrm{ldeal}}\left(f\right)=R_p\left(f\right)\exp \left(j{\widehat{\mathrm{\&epsiv;}}}_p\right)---\left(8\right)$

In the formula, R _p(f) be r _p(t) Fourier transform;

(6), compensation back data are used corresponding image reconstruction algorithm, the synthetic aperture sonar picture that then can obtain to focus on.

Useful effect of the present invention is: it is a kind of kinematic error method of estimation of utilizing the data field relevant treatment of redundant information in the synthetic aperture sonar reverberation.Employing with equal autofocus algorithm be core, undertaken by the phasing degree of measuring broadband analytic signal multiple correlation that the synthetic aperture sonar kinematic error is estimated and compensation, thereby the self-focusing that obtains synthetic aperture sonar imaging is handled.It does not need interative computation, and has real-time operation performance preferably and better image macroscopic view calibration capability.And further weaken strong target scattering point and noise to the kinematic error estimation effect by the α weighting, the broad beam SAS that makes the average autofocus algorithm of shear both be adapted to the gradual scene in space handles; Also be applicable to the scene that has strong target scattering point, thereby widened application conditions.

Description of drawings

The process flow diagram of the average self-focusing processing method of α weighting shear of Fig. 1 embodiment of the invention 2;

Embodiment

The invention will be described further below in conjunction with drawings and Examples:

Embodiment 1:

One, the kinematic error model of estimating based on reverberation

(reverberation is occupied an leading position for a reverberation restriction, and noise is relatively weak) slowly changing field scape seabed model, the size of each scatterer of supposing the seabed is much smaller than wavelength, they can be regarded as " point source " target, in this approximate frequency and the beam direction influence to scattering of can not considering down, reverberation model can be seen as the linear sum of a lot of Point Target echoes like this.If total Q scatterer in the mapping zone, then p exomonental echo can be described as:

$e_p\left(t\right)=\underset{q=1}{\overset{Q}{\mathrm{\&Sigma;}}}{s}_p\left(t\right)\&CircleTimes;h(t,{\mathrm{\&tau;}}_{\mathrm{pq}})+n_p\left(t\right)---\left(1\right)$

 represents convolution operation in the formula, s _p(t) be the broadband emission signal, h (t, τ _Pq) be point spread function (comprising propagator in the medium, scatterer scattering response and basic matrix response function), n _p(t) be noise item.When not having kinematic error, echoed signal obtains analytic signal r through complex radical band demodulation and matched filter processing _p(t), be expressed as:

$r_p^{\mathrm{ldeal}}\left(t\right)=\left|r_p^{\mathrm{ldeal}}\left(t\right)\right|\exp \left(j{\mathrm{\&psi;}}_p^{\mathrm{ldeal}}\left(t\right)\right)---\left(2\right)$

Because scatterer is obeyed evenly and distributed in the reverberation field, receiving echo is each scatterer scattered signal sum, so ψ _p ^Ldeal(t) be equally distributed stochastic variable in [π, π].

Suppose at acoustic transducer basic matrix generation stagger x _p, and the main body of scatterer is approximate constant, then the echo delay time error for each scatterer in the scene all equals ε _p≈ 2x _p/ c, p echo can be expressed as

$r_p\left(t\right)=r_p^{\mathrm{ldeal}}(t-{\mathrm{\&epsiv;}}_p)\exp (-j2\mathrm{\&pi;}{f}_0{\mathrm{\&epsiv;}}_p)$

$=\left|r_p^{\mathrm{ldeal}}(t-{\mathrm{\&epsiv;}}_p)\right|\exp \left(j{\mathrm{\&psi;}}_p^{\mathrm{ldeal}}(t-{\mathrm{\&epsiv;}}_p)\right)\exp (-j2\mathrm{\&pi;}{f}_0{\mathrm{\&epsiv;}}_p)---\left(3\right)$

So obtain Measurement Phase be

${\mathrm{\&psi;}}_p\left(t\right)=\mathrm{Arg}\left\{r_p\left(t\right)\right\}=[{\mathrm{\&psi;}}_p^{\mathrm{ldeal}}(t-{\mathrm{\&epsiv;}}_p)-\mathrm{2\&pi;}{f}_0{\mathrm{\&epsiv;}}_p]\mathrm{mod}2\mathrm{\&pi;}---\left(4\right)$

Arg{} is for getting plural phase operation in the formula.

If we can be to each ψ _p(t) go the processing of reeling, then obtain a constant phase shift 2 π f ₀ε _pWith a stochastic variable ψ _p ^Ldeal(t-ε _p), further just can obtain the constant phase shift by on time t, doing average treatment, estimate thereby obtain the error time delay

Two, the average self-focusing processing method of the shear of synthetic aperture sonar imaging

(1) each echoed signal obtains broadband analytic signal r through complex radical band demodulation and matched filter processing _p(t).

(2) relevant treatment between the adjacent echo analytic signal.

The individual pulse echo is the constant error phase 2 π f that average treatment obtains in (4) formula of utilization on time t ₀ε _pThereby generally can cause phase fuzzy problem greater than 2 π.The method of a this phase place coiling of feasible solution is exactly the phase differential of estimating between adjacent two emission echoed signals, and its core is a kind of relevant treatment.The broadband analytic signal r of adjacent two transponder pulse echoes _p(t), r _P-1(t) computing cross-correlation is:

${\mathrm{rr}}_p={\&Integral;}_0^T{r}_p^{*}\left(t\right)r_{p-1}\left(t\right)\mathrm{dt}---\left(5\right)$

In the formula ^*The expression complex conjugate.

(3), obtain the phase error estimation and phase error between the adjacent echo analytic signal by asking plural phase angle:

$\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p=\mathrm{Arg}\left\{{\mathrm{rr}}_p\right\}---\left(6\right)$

(4) obtain the accumulated phase error that kinematic error causes by the locus accumulated process

${\widehat{\mathrm{\&epsiv;}}}_p={\widehat{\mathrm{\&epsiv;}}}_{p-1}+\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p;$ ${\widehat{\mathrm{\&epsiv;}}}_0\&equiv;0---\left(7\right)$

(5) utilize the sequence of phase errors that obtains

Compensate and just can realize the echo data correction:

$R_p^{\mathrm{ldeal}}\left(f\right)=R_p\left(f\right)\exp \left(j{\widehat{\mathrm{\&epsiv;}}}_p\right)---\left(8\right)$

In the formula, R _p(f) be r _p(t) Fourier transform.

(6) compensation back data are used corresponding image reconstruction algorithm, the synthetic aperture sonar picture that then can obtain to focus on.

Embodiment 2: the average self-focusing processing method of α weighting shear

In order to improve the precision of error phase variance estimation, people such as K.A.Johnson propose lowest mean square meaning weighted mean core, select a kind of weighting function ω (t)=r _p(t) | make it lay particular emphasis on the bigger phase differential of signal to noise ratio (S/N ratio), thereby effectively suppress the influence of noise phase.Use method of weighting to reappraise as follows to (5) formula related function:

$\mathrm{rr}=\left\{\frac{{\&Integral;}_0^T{\mathrm{\&omega;}}^2\left(t\right)r_p^{*}\left(t\right)r_{p-1}\left(t\right)\mathrm{dt}}{{\&Integral;}_0^T{\mathrm{\&omega;}}^2\left(t\right)\mathrm{dt}}\right\}---\left(9\right)$

But the K.A.Johnson method of weighting has also increased the weight of the contribution of strong target scattering body to phase function when suppressing noise effect.When strong target exists, the ψ in (4) formula _p ^Ldeal(t) no longer can be considered as [π, a π] interior uniformly distributed random variable, this moment, it mainly showed as the range migration phase place of strong target, so be subjected to the influence of range migration variable quantity by the time delay variable quantity of (6) formula estimation.Simulation result and real data processing also proof make kinematic error estimate at partially when the strong scattering body exists, and must adopt rational method to suppress strong target scattering body influences phase estimation.

We propose a kind of new α weighting weighting function, in (9) formula,

${\mathrm{\&omega;}}^2\left(t\right)=\frac{\left|r_p^{*}\left(t\right)r_{p-1}\left(t\right)\right|}{{\mathrm{\&alpha;}}^4+{\left|r_p^{*}\left(t\right)r_{p-1}\left(t\right)\right|}^2}---\left(10\right)$

But alpha parameter is an adjustment amount in the formula, can trade off control strong target scattering body and noise to the error phase estimation effect by choose reasonable α value.Consider that from suppressing the strong scattering target α value is the smaller the better, and consider that from suppressing noise the α value is the bigger the better; Can optimize the α value in conjunction with the priori of reality.

The average self-focusing processing method of α weighting shear of this synthetic aperture sonar imaging as shown in Figure 1.At first echo data is carried out complex demodulation and matched filtering, carry out α weighting shear average treatment then and obtain the accumulation of kinematic error time delay variable quantity locus, and compensate echo data with kinematic error, last application image restructing algorithm carries out synthetic aperture imaging.

The test findings of this self-focusing processing method: use the average self-focusing processing method of α weighting shear that actual synthetic aperture sonar lake examination data processed result is represented, self-focusing is handled back T font gas-jar target and is focused to three bright spots, and the letter of self-focusing processing back image mixes than obviously enhancing.Test findings shows that the average autofocus algorithm of improved α weighting shear can better estimate kinematic error, thereby suppresses image blurring effectively.

Claims (3)

1, a kind of self-focusing processing method of synthetic aperture sonar imaging is characterized in that: mainly may further comprise the steps:

(1), each echoed signal obtains broadband analytic signal r through complex radical band demodulation and matched filter processing _p(t);

(2), relevant treatment between the adjacent echo analytic signal; The broadband analytic signal r of adjacent two transponder pulse echoes _p(t), r _P-1(t) computing cross-correlation is:

${\mathrm{rr}}_p={\&Integral;}_0^T{r}_p^{*}\left(t\right)r_{p-1}\left(t\right)\mathrm{dt}---\left(5\right)$

* represents complex conjugate in the formula;

(3), by asking plural phase angle, obtain the phase error estimation and phase error between the adjacent echo analytic signal:

$\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p=\mathrm{Arg}\left\{{\mathrm{rr}}_p\right\}---\left(6\right)$

(4), obtain the accumulated phase error that kinematic error causes by the locus accumulated process:

${\widehat{\mathrm{\&epsiv;}}}_p={\widehat{\mathrm{\&epsiv;}}}_{p-1}+\mathrm{\&Delta;}{\widehat{\mathrm{\&phi;}}}_p;$ ${\widehat{\mathrm{\&epsiv;}}}_0\&equiv;0---\left(7\right)$

(5), utilize the sequence of phase errors that obtains

Compensate and realize the echo data correction:

$R_p^{\mathrm{Ideal}}\left(f\right)=R_p\left(f\right)\exp \left(j{\widehat{\mathrm{\&epsiv;}}}_p\right)---\left(8\right)$

In the formula, R _p(f) be r _p(t) Fourier transform;

(6), compensation back data are used corresponding image reconstruction algorithm, the synthetic aperture sonar picture that then can obtain to focus on.

2, the self-focusing processing method of synthetic aperture sonar imaging according to claim 1 is characterized in that:

Adopt α weighting weighting function, in (9) formula,

$\mathrm{rr}=\left\{\frac{{\&Integral;}_o^T{\mathrm{\&omega;}}^2\left(t\right)r_p^{*}\left(t\right)r_{p-1}\left(t\right)\mathrm{dt}}{{\&Integral;}_o^T{\mathrm{\&omega;}}^2\left(t\right)\mathrm{dt}}\right\}---\left(9\right)$

${\mathrm{\&omega;}}^2\left(t\right)=\frac{\left|r_p^{*}\left(t\right)r_{p-1}\left(t\right)\right|}{{\mathrm{\&alpha;}}^4+{\left|r_p^{*}\left(t\right)\right|r_{p-1}\left(t\right)}^2}---\left(10\right)$

But alpha parameter is an adjustment amount in the formula, controls strong target scattering body and noise to the error phase estimation effect by selecting the α value.

3, the self-focusing processing method of synthetic aperture sonar imaging according to claim 1 and 2 is characterized in that: suppose at acoustic transducer basic matrix generation stagger x _p, and the main body of scatterer is approximate constant, then the echo delay time error for each scatterer in the scene all equals ε _p≈ 2x _p/ c, p echo can be expressed as

$r_p\left(t\right)=r_p^{\mathrm{Ideal}}(t-{\mathrm{\&epsiv;}}_p)\exp (-j2\mathrm{\&pi;}{f}_0{\mathrm{\&epsiv;}}_p)$

$=\left|r_p^{\mathrm{Ideal}}(t-{\mathrm{\&epsiv;}}_p)\right|\exp \left(j{\mathrm{\&psi;}}_p^{\mathrm{Ideal}}(t-{\mathrm{\&epsiv;}}_p)\right)\exp (-j2\mathrm{\&pi;}{f}_0{\mathrm{\&epsiv;}}_p)---\left(3\right)$

Ψ _p ^IdealBe equally distributed stochastic variable in [π, π]; f ₀Be signal center frequency; When not having kinematic error, echoed signal obtains analytic signal r through complex radical band demodulation and matched filter processing _p(t), be expressed as r _p ^Ideal(t).