CN105068072A - Speed compensation method of one-dimensional distance images of moving object - Google Patents

Abstract

The invention discloses a speed compensation method of one-dimensional distance images of a moving object. The method comprises the steps of: caching one-dimensional distance images continuously obtained by a radar in preset time intervals to a buffer; carrying out pre-processing on the one-dimensional distance images cached in the buffer, obtaining an analysis window of the one-dimensional distance images, and removing noises; then carrying out cross-correlation processing on all the one-dimensional distance images, arranged in the analysis window, of the buffer; estimating an object envelopment moving amount, adding a reference signal time-delay corresponding position, and obtaining the practical position of the object; obtaining a speed estimation value of the one-dimensional distance images by a primary curve fitting method; according to the speed estimation value of the one-dimensional distance images, carrying out speed compensation on the one-dimensional distance images; and carrying out variance estimation on each one-dimensional distance image after the speed compensation, and determining whether a pulse number used for the speed compensation meets error requirements. By adopting the method provided by the invention, unfavorable influences caused by high-speed movement can be effectively eliminated.

Description

A kind of speed compensation method of one-dimensional range profile of moving target

Technical field

The present invention relates to radar target feature identification technique, particularly a kind of speed compensation method of one-dimensional range profile of moving target.

Background technology

Inverse synthetic aperture radar (ISAR) (ISAR) imaging technique grows up in Synthetic Aperture Radar Technique.ISAR imaging system is that radar is static, the imaging system of target travel, is different from conventional radar, and ISAR is a kind of high-resolution imaging radar, can obtain the exact image of non-cooperative moving targets at a distance.

ISAR imaging technique as the valuable detection means of one in the past few decades in obtain and pay attention to widely.Be dispersed with spacecraft in-orbit in running order in a large number in space, also having the space junk of enormous amount, aerolite and inefficacy spacecraft etc. in addition, is important contents of space exploration to the imaging of these targets.But the research of ISAR technology up to now mainly concentrates on the lower aerial target of speed (as aircraft etc.), also less to the research of the ISAR imaging of the higher Space Object in Orbit of speed.

The remarkable difference of extraterrestrial target and conventional I SAR imaging object is that target speed is very high, an order of magnitude is improve than the movement velocity of conventional target, target distance image is made to produce fuzzy and distortion, and then cause ISAR image radial direction to defocus, therefore must consider on extraterrestrial target imaging the impact that high-speed motion brings.

When measured target radial motion speed is lower, the echo after De-chirp method is approximately sinusoidal signal, just can be obtained a Range Profile of target by pulse compression; But when object of observation is the high-speed moving objects such as space orbit spacecraft, because target is very high relative to the radial velocity of radar, making to accept scattering point echo is linear frequency modulation (LFM) signal (being no longer the simple signal under low-speed motion), the Range Profile Severe distortion of the high-speed moving object obtained after process of pulse-compression.

Summary of the invention

In view of this, the invention provides a kind of speed compensation method of one-dimensional range profile of moving target, thus effectively can eliminate the impact that high-speed motion brings.

Technical scheme of the present invention is specifically achieved in that

A speed compensation method for the one-dimensional range profile of moving target, the method comprises:

Steps A, the one-dimensional range profile sequence obtained continuously within a preset time interval by radar are cached in buffer, and carry out pre-service to the one-dimensional range profile sequence of buffer memory in buffer, obtain the analysis window of one-dimensional range profile, and remove noise;

Step B, one-dimensional range profile sequences within analysis window all in buffer are carried out cross correlation process; Estimating target envelope amount of movement, and add reference signal time delay correspondence position, obtain the actual range of target;

Step C, obtained the velocity estimation value of one-dimensional range profile by the method for a curve;

Step D, velocity estimation value according to one-dimensional range profile, carry out velocity compensation to one-dimensional range profile;

Step e, carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, determining that the umber of pulse being used as velocity compensation meets error requirements.

Optionally, the described one-dimensional range profile sequence to buffer memory in buffer is carried out pre-service and is comprised:

Gray level image is formed according to the one-dimensional range profile course figure of the one-dimensional range profile sequence of institute's buffer memory in buffer;

Utilize image processing method to extract the edge of described gray level image, obtain reference position and the final position of the analysis window of one-dimensional range profile, and carry out the operation of removing noise.

Optionally, following formula is used to carry out cross correlation process to the one-dimensional range profile sequence within analysis window:

$r^{\left(r\right)}\left(n_k\right)=\underset{i=0}{\overset{N_r-1}{\Σ}}{\mathrm{\ξ}}_H^{\′}{{}^{(k-1)}}_i\·{\mathrm{\ξ}}_H^{\′}{{}^{\left(k\right)}}_{i+n_k},k=1,2,\mathrm{...},J_{\max }-1;$

Wherein, r ^(k)(n _k) be cross correlation process result, n _kfor cross correlation process variable, represent ξ ' _h ^(k)carry out n _kindividual range unit moves, J _maxrepresent the size of buffer; ξ ' _h ^(k-1) _irepresent the one-dimensional range profile after normalization.

Optionally, following formula is used to obtain the actual range p of target _k:

p _k＝p _k-1+cn _k/2Δf，k＝1，2，...，J _max-1；

Wherein, p ₀for the target range of initial time, c is the light velocity, and Δ f is radar swept bandwidth.

Optionally, the described velocity estimation value according to one-dimensional range profile, velocity compensation is carried out to one-dimensional range profile and comprises:

According to the velocity estimation value of one-dimensional range profile, translation compensation process is carried out to each one-dimensional range profile;

Make FFT to the one-dimensional range profile after translation compensation process to estimate, obtain the phase differential ΔΦ of adjacent echoes;

Preset the first threshold value Z _*, as ΔΦ > Z _*time, return and perform step B; And as ΔΦ≤Z _*time, perform step e.

Optionally, described step e comprises:

Carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, obtaining the estimate of variance of each one-dimensional range profile;

The estimate of variance of all one-dimensional range profiles is averaged, obtains mean of variance;

Pre-set the second threshold value NN _*, when time, increase buffer size and return execution steps A; When time, terminate whole flow process.

Optionally, variance evaluation is carried out by following formula:

${\mathrm{\σ}}_i^2=\frac{1}{J_{\max }}\underset{j=0}{\overset{J_{\max }-1}{\Σ}}{({\mathrm{\ξ}}_H^{\′}{{}^{\left(j\right)}}_i-\stackrel{\‾}{{\mathrm{\ξ}}_H^{\′}{{}^{\left(j\right)}}_i})}^2,i=0,1,\mathrm{...},N_r-1;$

Wherein, represent that i-th one-dimensional range profile is J in size _maxbuffer in the mean value of scattering strength.

Optionally, mean of variance is obtained by following formula:

$\stackrel{\‾}{{\mathrm{\σ}}^2}=\frac{1}{N_r}\underset{i=0}{\overset{N_r-1}{\Σ}}\left({\mathrm{\σ}}_i^2\right).$

As above visible, in the speed compensation method of the one-dimensional range profile of moving target of the present invention, because the one-dimensional range profile sequence obtained continuously within a preset time interval by radar is cached in buffer, and pre-service is carried out to the one-dimensional range profile sequence of buffer memory in buffer, obtain the analysis window of one-dimensional range profile, and remove noise; Then, one-dimensional range profile sequences within analysis window all in buffer are carried out cross correlation process; Estimating target envelope amount of movement, and add reference signal time delay correspondence position, obtain the actual range of target; The velocity estimation value of one-dimensional range profile is obtained by the method for a curve; According to the velocity estimation value of one-dimensional range profile, velocity compensation is carried out to one-dimensional range profile; Carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, determining that the umber of pulse being used as velocity compensation meets error requirements, therefore can carry out velocity compensation to echoed signal, effectively eliminate the adverse effect that high-speed motion brings.

Accompanying drawing explanation

Fig. 1 is the schematic flow sheet of the speed compensation method of the one-dimensional range profile of moving target in the embodiment of the present invention.

Fig. 2 is the speed compensation method effect schematic diagram of the one-dimensional range profile of moving target in the embodiment of the present invention.

Embodiment

For making object of the present invention, technical scheme and advantage clearly understand, to develop simultaneously embodiment referring to accompanying drawing, the present invention is described in more detail.

In the inventive solutions, the target echo of broadband signal is one-dimensional range profile, is therefore difficult to, as arrowband situation, follow the tracks of by range gate.So, in the inventive solutions, " can slide relevant " with adjacent twice return (i.e. one-dimensional range profile) and follow the tracks of." sliding relevant " of twice return, be similar to " auto-correlation " said in reception, difference is that auto-correlation only uses echo itself, done time delay to slide, with time delay slippage for variable, calculate the related coefficient of two, represent the position at target place with the peak value of related coefficient, and the speed of estimating target.

Motion compensation is the gordian technique in ISAR imaging technique, generally includes envelope alignment and phase compensation.In the inventive solutions, according to the experience of the envelope alignment in ISAR imaging technique, envelope alignment usually adopts to slide and is correlated with, and maximum as alignment criterion using related coefficient, its alignment error is relevant with the size in the time interval (i.e. visual angle change) of two echoes (i.e. one-dimensional range profile).For the situation that adjacent echoes visual angle change is minimum, Delay Estima-tion error is very little, can be more much smaller than ± 1/4 Range resolution unit.But this refers to the situation that signal to noise ratio (S/N ratio) is very high.When to Long Range Target Tracking, signal to noise ratio (S/N ratio) is much lower, and Delay Estima-tion error now mainly carrys out the impact of self noise, instead of due to visual angle change and cause Range Profile change, signal to noise ratio (S/N ratio) is lower, then evaluated error is also larger, is about ± 1/2 Range resolution unit.The error of the envelope alignment caused due to noise is random completely, there is not correlativity, on average can reduce it by multiple pulse.

Extraterrestrial target in the time of submicrosecond level, radial velocity substantially constant, its radial acceleration is negligible, now, the range difference of adjacent two echo time delays should be constant, thus multiecho one after the other can be done adjacent slip and be correlated with, obtain a series of delay inequality, be then averaging.If each time noise is statistical iteration, the Delay Estima-tion error that it causes is also uncorrelated, and average with N time, the variance of error can be reduced to 1/N.If the time is longer, except the error that noise causes, the Delay Estima-tion error that visual angle change causes also will account for certain ratio, and the error that it causes has correlativity, and average effect is less better.

Present embodiments provide a kind of speed compensation method of one-dimensional range profile of moving target, thus effectively can eliminate the impact that high-speed motion brings.

Fig. 1 is the schematic flow sheet of the speed compensation method of the one-dimensional range profile of moving target in the embodiment of the present invention.As shown in Figure 1, the speed compensation method of the one-dimensional range profile of the moving target in the embodiment of the present invention can comprise step as described below:

Step 101, the one-dimensional range profile sequence obtained continuously within a preset time interval by radar is cached in buffer, and carries out pre-service to the one-dimensional range profile sequence of buffer memory in buffer, obtains the analysis window of one-dimensional range profile, and removes noise.

In the inventive solutions, radar can obtain multiple one-dimensional range profile within a preset time interval continuously, and wherein, single one-dimensional range profile can be expressed as:

ξ＝{ξ ₁，ξ ₂，…，ξ _N}(1)

The multiple one-dimensional range profiles obtained by above-mentioned radar form an one-dimensional range profile sequence and are cached in buffer, then in buffer, one-dimensional range profile sequence can be expressed as:

ξ ^(k)＝{ξ ₁ ^(k)，ξ ₂ ^(k)，…，ξ _N ^(k)}，k＝0，1，…，J _max-1(2)

Wherein, J _maxrepresent the size (i.e. the maximum number of storable one-dimensional range profile) of buffer.

In addition, in buffer, one-dimensional range profile sequence can also be expressed as:

ξ(t _k)＝{ξ ₁(t _k)，ξ ₂(t _k)，…，ξ _N(t _k)}，t _k＝t ₀+k·dt(3)

Wherein, t ₀for initial time, dt is the time interval, k=0,1 ..., J _max-1.

After one-dimensional range profile sequence is cached in buffer, pre-service can be carried out to the one-dimensional range profile sequence of buffer memory in buffer, to obtain the analysis window of one-dimensional range profile, and remove noise.

In the inventive solutions, multiple specific implementation can be used to carry out above-mentioned pre-service.Below, by for a kind of specific implementation wherein, technical scheme of the present invention is described in detail.

Such as, in a preferred embodiment of the present invention, the described one-dimensional range profile sequence to buffer memory in buffer is carried out pre-service and can be comprised:

Gray level image is formed according to the one-dimensional range profile course figure of the one-dimensional range profile sequence of institute's buffer memory in buffer;

Utilize image processing method to extract the edge of described gray level image, obtain reference position and the final position of the analysis window of one-dimensional range profile, and carry out the operation of removing noise.

Wherein, if the reference position of the analysis window of one-dimensional range profile and final position are respectively N _sand N _e, then the analysis window one of this one-dimensional range profile has N _r=(N _e-N _s+ 1) individual range unit.

After the operation carrying out removal noise, the one-dimensional range profile sequence after pre-service can be expressed as:

ξ′ ^(k)＝{ξ′ _i ^(k)}，i＝0，1，…，N _r-1，k＝0，1，…，J _max-1(4)

In the inventive solutions, can also be normalized each one-dimensional range profile in the one-dimensional range profile sequence after above-mentioned pre-service further, the one-dimensional range profile ξ ' after normalization can be obtained _h ^(k) _i:

${\mathrm{\ξ}}_H^{\′}{{}^{\left(k\right)}}_i=\frac{{\mathrm{\ξ}}_i^{\′\left(k\right)}}{{{\mathrm{\ξ}}_{\max }^{\′}}^{\left(k\right)}},i=0,1,\mathrm{...},N_r-1---\left(5\right)$

Wherein, ξ ' _i ^(k)represent the scattering strength of each range unit in one-dimensional range profile; ξ ' _max ^(k)represent the maximal value of all range unit scattering strengths in each one-dimensional range profile.

Step 102, carries out cross correlation process by one-dimensional range profile sequences within analysis window all in buffer; Estimating target envelope amount of movement, and add reference signal time delay correspondence position, obtain the actual range of target.

Such as, in a preferred embodiment of the present invention, formula as described below can be used to carry out cross correlation process to the one-dimensional range profile sequence within analysis window:

$r^{\left(k\right)}\left(n_k\right)=\underset{i=0}{\overset{N_r-1}{\Σ}}{\mathrm{\ξ}}_H^{\′}{{}^{(k-1)}}_i\·{\mathrm{\ξ}}_H^{\′}{{}^{\left(k\right)}}_{i+n_k},k=1,2,\mathrm{...},J_{\max }-1---\left(6\right)$

Wherein, r ^(k)(n _k) be cross correlation process result, n _kfor cross correlation process variable, represent ξ ' _h ^(k)carry out n _kindividual range unit moves, J _maxrepresent the size of buffer.

Suppose adjacent two envelopes (one-dimensional range profile namely after normalization) ξ ' _h ^(k-1)with ξ ' _h ^(k)between side-play amount be Δ r _k, then n is worked as _k=Δ r _ktime, r ^(k)(n _k) reach maximal value.

In addition, in a preferred embodiment of the present invention, formula as described below can be used to obtain the actual range p of target _k:

p _k＝p _k-1+cn _k/2Δf，k＝1，2，...，J _max-1(7)

Wherein, p ₀for the target range of initial time, c is the light velocity, and Δ f is radar swept bandwidth.

Step 103, obtains the velocity estimation value of one-dimensional range profile by the method for a curve.

In the inventive solutions, can hypothetical target distance p be a curve about time independent variable t, can be expressed as:

p＝V _Tt+p ₀+e(8)

Wherein, e is observational error, V _tfor the radial velocity of target, the variances sigma of e ²=E [e ²], if having k observation to this target, k is the number of times observed, and observation equation is respectively:

$\begin{array}{c}{p}_1=V_T{t}_1+p_0+e_1\\ p_2=V_T{t}_2+p_0+e_2\\ .\\ .\\ .\\ p_k=V_T{t}_k+p_0+e_n\end{array}---\left(9\right)$

Write above-mentioned observation equation as vector form: P=[p ₁, p ₂... p _k] ^t, e=[e ₁, e ₂..., e _k] ^t, V=[V _t, p ₀] ^t, $T={\left[\begin{array}{cccc}{t}_1& t_2& ...& t_k\\ 1& 1& ...& 1\end{array}\right]}^T,$ Then above-mentioned observation equation can be expressed as:

P＝TV+e(10)

The cost function of the least-squares estimation of V is:

$J\left(\hat{V}\right)={\[P-T\hat{V}\]}^H\[P-T\hat{V}\]---\left(11\right)$

The estimated value of V cost function should be made to reach minimum, namely meet condition as described below:

$\frac{\∂J\left(\hat{V}\right)}{\∂\hat{V}}=\frac{\∂}{\∂\hat{V}}({\[P-T\hat{V}\]}^H\[P-T\hat{V}\])=-2T^H\[P-T\hat{V}\]=0---\left(12\right)$

Therefore, the estimated value of V (i.e. the velocity estimation value of one-dimensional range profile) is:

$\hat{V}={\left(T^{H}T\right)}^{-1}{T}^{H}P---\left(13\right)$

Step 104, according to the velocity estimation value of one-dimensional range profile, carries out velocity compensation to one-dimensional range profile.

In the inventive solutions, multiple specific implementation can be used to carry out above-mentioned step 104.Below, by for a kind of specific implementation wherein, technical scheme of the present invention is described in detail.

Such as, in a preferred embodiment of the present invention, the described velocity estimation value according to one-dimensional range profile, carrying out velocity compensation to one-dimensional range profile can comprise:

According to the velocity estimation value of one-dimensional range profile, translation compensation process is carried out to each one-dimensional range profile;

After translation compensation process, the one-dimensional range profile marshalling substantially of each adjacent echoes, therefore can make FFT to one-dimensional range profile after translation compensation process again and estimate, obtain the phase differential ΔΦ of adjacent echoes;

Preset the first threshold value Z _*, as ΔΦ > Z _*time, return and perform step 102; And as ΔΦ≤Z _*time, judge that velocity compensation terminates, perform follow-up step 105.

Step 105, carries out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, determines that the umber of pulse being used as velocity compensation meets error requirements.

In the inventive solutions, multiple specific implementation can be used to carry out above-mentioned step 104.Below, by for a kind of specific implementation wherein, technical scheme of the present invention is described in detail.

Such as, in a preferred embodiment of the present invention, described step 105 can comprise:.

Carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, obtaining the estimate of variance of each one-dimensional range profile;

Such as, variance evaluation can be carried out by formula as described below:

${\mathrm{\σ}}_i^2=\frac{1}{J_{\max }}\underset{j=0}{\overset{J_{\max }-1}{\Σ}}{({\mathrm{\ξ}}_H^{\′}{{}^{\left(j\right)}}_i-\stackrel{\‾}{{\mathrm{\ξ}}_H^{\′}{{}^{\left(j\right)}}_i})}^2,i=0,1,\mathrm{...},N_r-1---\left(14\right)$

Wherein, represent that i-th one-dimensional range profile is J in size _maxbuffer in the mean value of scattering strength.

Then, then the estimate of variance of all one-dimensional range profiles is averaged, obtain mean of variance.Such as, mean of variance can be obtained by formula as described below:

$\stackrel{\‾}{{\mathrm{\σ}}^2}=\frac{1}{N_r}\underset{i=0}{\overset{N_r-1}{\Σ}}\left({\mathrm{\σ}}_i^2\right)---\left(15\right)$

Finally, the second threshold value NN is pre-set _*, when time, then increase buffer size and return execution step 101; And work as time, then determine that the umber of pulse being used as velocity compensation meets error requirements, terminate whole flow process, thus the group velocity completing one-dimensional range profile is estimated and compensation process.

In the inventive solutions, the technique effect of the speed compensation method of the one-dimensional range profile of the moving target provided in the present invention can be provided by the method for simulation analysis.

Such as, in the inventive solutions, simulation analysis can be carried out to point target when radar parameter is known.

For Ka band radar, if centre frequency is 37.5GHz, target radial speed is the maximum measuring speed V in its maximum measuring speed scope _t=12km/s, actual radial velocity is generally all less than this maximum measuring speed, target range radar 300km.When pulse width is Tp=50 μ s, by the translation making Range Profile produce 36m, be equivalent to 480 high-resolution range units, after estimating target velocity, Range Profile translation can be corrected.

Fig. 2 is the speed compensation method effect schematic diagram of the one-dimensional range profile of moving target in the embodiment of the present invention.As shown in Figure 2, in this simulation analysis, employ radar parameter as described below, the one-dimensional range profile of the scattering point that the one-dimensional distance shown in left figure of Fig. 2 similarly is the relative velocity of extraterrestrial target and radar when being 8000m/s.The one-dimensional range profile be through after a velocity compensation shown in the right figure of Fig. 2.According to the contrast of two figure, after carrying out velocity compensation to original one-dimensional range profile, the translation of one-dimensional range profile and broadening are obtained for obvious correction, thus effectively eliminate the adverse effect that high-speed motion brings.

In summary, in the speed compensation method of the one-dimensional range profile of moving target of the present invention, because the one-dimensional range profile sequence obtained continuously within a preset time interval by radar is cached in buffer, and pre-service is carried out to the one-dimensional range profile sequence of buffer memory in buffer, obtain the analysis window of one-dimensional range profile, and remove noise; Then, one-dimensional range profile sequences within analysis window all in buffer are carried out cross correlation process; Estimating target envelope amount of movement, and add reference signal time delay correspondence position, obtain the actual range of target; The velocity estimation value of one-dimensional range profile is obtained by the method for a curve; According to the velocity estimation value of one-dimensional range profile, velocity compensation is carried out to one-dimensional range profile; Carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, determining that the umber of pulse being used as velocity compensation meets error requirements, therefore can carry out velocity compensation to echoed signal, effectively eliminate the adverse effect that high-speed motion brings.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within the scope of protection of the invention.

Claims (8)

1. a speed compensation method for the one-dimensional range profile of moving target, is characterized in that, the method comprises:

Steps A, the one-dimensional range profile sequence obtained continuously within a preset time interval by radar are cached in buffer, and carry out pre-service to the one-dimensional range profile sequence of buffer memory in buffer, obtain the analysis window of one-dimensional range profile, and remove noise;

Step B, one-dimensional range profile sequences within analysis window all in buffer are carried out cross correlation process; Estimating target envelope amount of movement, and add reference signal time delay correspondence position, obtain the actual range of target;

Step C, obtained the velocity estimation value of one-dimensional range profile by the method for a curve;

Step D, velocity estimation value according to one-dimensional range profile, carry out velocity compensation to one-dimensional range profile;

Step e, carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, determining that the umber of pulse being used as velocity compensation meets error requirements.

2. method according to claim 1, is characterized in that, the described one-dimensional range profile sequence to buffer memory in buffer is carried out pre-service and comprised:

Gray level image is formed according to the one-dimensional range profile course figure of the one-dimensional range profile sequence of institute's buffer memory in buffer;

Utilize image processing method to extract the edge of described gray level image, obtain reference position and the final position of the analysis window of one-dimensional range profile, and carry out the operation of removing noise.

3. method according to claim 2, is characterized in that, uses following formula to carry out cross correlation process to the one-dimensional range profile sequence within analysis window:

001"/>

Wherein, r ^(k)(n _k) be cross correlation process result, n _kfor cross correlation process variable, represent right carry out n _kindividual range unit moves, J _maxrepresent the size of buffer; represent the one-dimensional range profile after normalization.

4. method according to claim 3, is characterized in that, uses following formula to obtain the actual range p of target _k:

p _k＝p _k-1+cn _k/2Δf，k＝1，2，...，J _max-1；

Wherein, p ₀for the target range of initial time, c is the light velocity, and Δ f is radar swept bandwidth.

5. method according to claim 4, is characterized in that, the described velocity estimation value according to one-dimensional range profile, carries out velocity compensation comprise one-dimensional range profile:

According to the velocity estimation value of one-dimensional range profile, translation compensation process is carried out to each one-dimensional range profile;

Make FFT to the one-dimensional range profile after translation compensation process to estimate, obtain the phase difference Ф of adjacent echoes;

Preset the first threshold value Z _*, as Δ Ф > Z _*time, return and perform step B; And as Δ Ф≤Z _*time, perform step e.

6. method according to claim 5, is characterized in that, described step e comprises:

Carrying out variance evaluation to carrying out each one-dimensional range profile after velocity compensation, obtaining the estimate of variance of each one-dimensional range profile;

The estimate of variance of all one-dimensional range profiles is averaged, obtains mean of variance;

Pre-set the second threshold value NN _*, when time, increase buffer size and return execution steps A; When time, terminate whole flow process.

7. method according to claim 6, is characterized in that, carries out variance evaluation by following formula:

002"/>

Wherein, represent that i-th one-dimensional range profile is J in size _maxbuffer in the mean value of scattering strength.

8. method according to claim 7, is characterized in that, obtains mean of variance by following formula:

003"/>