CN103885044A - Method for suppressing clutter and noise of narrow-band radar echoes based on CLEAN algorithm - Google Patents

Abstract

The invention belongs to the technical field of clutter suppression and noise suppression of narrow-band radar echoes, and particularly relates to a method for suppressing clutter and noise of the narrow-band radar echoes based on the CLEAN algorithm. The method for suppressing the clutter and the noise of the narrow-band radar echoes based on the CLEAN algorithm comprises the following steps that echo data xN are received through narrow-band radar; in the echo data xN, a target echo sample and a clutter and noise echo sample are selected, so that the clutter frequency point position in the clutter and noise echo sample is obtained; the clutter frequency band position in the clutter and noise echo sample is obtained; according to the clutter frequency point position in the clutter and noise echo sample and the clutter frequency band position in the clutter and noise echo sample, the frequency position range needing to be removed through the CLEAN algorithm is determined; according to the frequency position range needing to be removed through the CLEAN algorithm, corresponding harmonic components in the target echo sample are filtered out through the CLEAN algorithm, and the echo data after clutter suppression are obtained; noise suppression is conducted on the echo data after clutter suppression is conducted.

Description

A kind of assorted inhibition method of making an uproar of Narrow-band Radar echo based on CLEAN algorithm

Technical field

The invention belongs to Narrow-band Radar echo clutter and suppress and noise reduction techniques field, particularly a kind of assorted inhibition method of making an uproar of Narrow-band Radar echo based on CLEAN algorithm, can realize the inhibition of land clutter, meteorological clutter and noise in Narrow-band Radar echo.

Background technology

CLEAN algorithm is a kind of removing the deconvolution technology of signal secondary lobe also being removed in the lump in set specific frequency signal, is widely used improving aspect picture quality.The CLEAN algorithm of radar signal processing field is a kind of signal processing method that can remove accurately single-frequency components, but it requires knownly will remove the frequency point information that component is corresponding.

At radar signal processing field, existing clutter suppression method mainly contains moving-target and shows (MTI), generalized matched filter (GMF) and CLEAN algorithm.MTI method, can be on the doppler spectral structure generation impact of target owing to having nonlinear amplitude modulation characteristics.Someone proposes to utilize the method for broad match filtering to solve existing moving target indication technique suppresses the non-linear modulation of stage to the existence of target doppler spectral problem at clutter.But because the method need to be estimated clutter sample autocorrelation matrix, operand and computational complexity are large.Somebody proposes to utilize CLEAN algorithm to suppress to have the zero-frequency land clutter of certain bandwidth, has well retained Doppler's structure of echo signal in clutter reduction, has obtained good clutter and has suppressed result.But, if this traditional clutter based on CLEAN algorithm suppresses the zero-frequency land clutter that means will accurately suppress to have certain bandwidth,, must need the prior imformation of clutter bandwidth, only obtain in the situation of these prior imformations, clutter component in echoed signal effectively could be removed.

In practical application, radar return not only comprises land clutter, especially, the detection identification of aerial target is accompanied by this non-zero-frequency clutter with certain Doppler's translation and bandwidth of the meteorological clutter being produced by sexual intercourse conventionally.Also do not study how utilizing CLEAN algorithm to remove non-zero-frequency clutter at present.Meanwhile, inevitably all can contain noise component in radar return, classic method does not realize denoising in removing clutter.

Summary of the invention

The object of the invention is to propose a kind of assorted inhibition method of making an uproar of Narrow-band Radar echo based on CLEAN algorithm.The present invention effectively searches for clutter frequency and clutter interval under complex clutter background, and utilize CLEAN algorithm to remove clutter, utilize the noise power estimating simultaneously in search clutter frequency and the interval process of clutter, in conjunction with the feature of CLEAN algorithm, realize squelch.

For realizing above-mentioned technical purpose, the present invention adopts following technical scheme to be achieved.

A kind of assorted inhibition method of making an uproar of Narrow-band Radar echo based on CLEAN algorithm comprises the following steps:

S1: utilize Narrow-band Radar to receive echo data x _n; At echo data x _nin, select target echo sample s _nwith the assorted echo samples c that makes an uproar _n;

S2: obtain the clutter frequency position that impurity removal is made an uproar in echo samples;

S3: obtain the clutter frequency band position that impurity removal is made an uproar in echo samples;

S4: according to described assorted clutter frequency position and assorted clutter frequency band position of making an uproar in echo samples of making an uproar in echo samples, determine the frequency location scope that CLEAN algorithm will be removed;

S5: the frequency location scope that will remove according to described CLEAN algorithm, utilize CLEAN algorithm filtering target echo sample s _nthe harmonic components of middle correspondence, obtains clutter and suppresses back echo data;

S6: clutter is suppressed to back echo data and carry out squelch.

Feature of the present invention and further improvement are:

In step S1, receiving echo data x _nafterwards, first determine target place range unit, target place range unit is expressed as to M ₁individual range unit, and by M ₁the echo data of individual range unit is as target echo sample s _n; Then choose and M ₁individual range unit at a distance of the echo data of 5-8 range unit as the echo samples c that makes an uproar that mixes _n.

Described step S2 specifically comprises the following steps:

S21: to the assorted echo samples c that makes an uproar _nmake Fourier transform, obtain the assorted echo samples frequency point data C that makes an uproar _n, to C _ncarry out the slide window processing that window length is 1 for L step-length, the data that the length that each sliding window is obtained is L are averaged, and utilize a sequence P that length is N-L+1 of all averages compositions _n-L+1, N represents the data length of the assorted echo samples of making an uproar;

S22: to P _n-L+1all averages do ascending order arrange, obtain ascending sequence

S23: by described ascending sequence in before the individual average of N-2L ' at P _n-L+1in position be recorded as set L _oc, L ' is maximum meteorological clutter Doppler drift frequency number;

S24: in step S23

in before the individual average of N-2L ' at P _n-L+1in position, from the assorted echo samples frequency point data C that makes an uproar _nin find the frequency point data of correspondence position, C _nin the frequency point data of the correspondence position that finds be expressed as C _l; By C _lin the amplitude of all frequency point data set up noise Doppler amplitude sample D; Then draw noise gate according to noise Doppler amplitude sample D;

S25: screening C _nmiddle amplitude is greater than the frequency point data of noise gate, C _nin the frequency point data that filters out at C _nin position be clutter frequency position, described clutter frequency positional representation set Q ₁.

In step S3, first screen P _n-L+1middle amplitude is greater than the frequency point data of noise gate, P _n-L+1in the frequency point data that filters out at P _n-L+1in position be clutter frequency band position, described clutter frequency band positional representation for set Q ₂.

Described step S4 specifically comprises the following steps: first definition set Q ₃=Q ₁∩ Q ₂, wherein, ∩ represents set intersection; To gather Q ₄be defined as: by being positioned at interval [q _min, q _max] between consecution natural number form set, q _minrepresent Q ₃minimum value, q _maxq ₃maximal value; Then, definition set Q ₅=Q ₁-Q ₃, the frequency location Range Representation that CLEAN algorithm will be removed is set Q, Q=Q ₄∪ Q ₅; Wherein, Q ₁-Q ₃represent: set Q ₁with set Q ₃subtract computing, ∪ represents set also.

Described step S5 specifically comprises the following steps:

S51: filtering number of times variable i 1 is set, i1=1,2 ...; In the time of i1=1, i1 time domain data is target echo sample s _n, then, execution step S52;

S52: i1 time domain data carried out to Fourier transform, obtain and i1 the frequency domain data X that time domain data is corresponding _i1;

S53: for X _i1, within the scope of the frequency location that will remove at CLEAN algorithm, filter out the frequency domain data X of maximum amplitude _{i1, max}; Draw X _{i1, max}amplitude phase theta _i1, and X _{i1, max}corresponding Doppler frequency f _{i1, clutter}; Then draw i1 the harmonic components s that needs filtering _{i1, clutter}, $s_{i1,\mathrm{clutter}}=({\hat{U}}_{i1}/K)\exp (j2\mathrm{\π}{f}_{i1,\mathrm{clutter}}t+j{\mathrm{\θ}}_{i1}),$ Wherein t represents the time, and K is echo data x _ncorresponding pulse accumulation number;

S54: to corresponding i1 the harmonic components of i1 time domain data filtering, obtain i1+1 frequency domain data X _i1+1;

S55: for X _i1+1, within the scope of the frequency location that will remove at CLEAN algorithm, if filtered out all frequency point data, i1+1 time domain data is that clutter suppresses back echo data; Otherwise, make i1 value add 1, be then back to step S52.

Described step S6 specifically comprises the following steps: the first time domain average power of estimating noise; Utilize CLEAN algorithm to suppress back echo data to described clutter and carry out squelch.

In step S6, utilize CLEAN algorithm to carry out squelch to described clutter inhibition back echo data and comprise the following steps:

S61: filtering number of times variable i 2 is set, i2=1,2 ...; In the time of i2=1, i2 time domain data is that clutter suppresses back echo data, then, and execution step S62;

S62: i2 time domain data carried out to Fourier transform, obtain and i2 the frequency domain data X that time domain data is corresponding _i2;

S63: for X _i2, search and i2 the frequency domain data X that time domain data is corresponding _i2in peak value frequency position; Draw X _i2in amplitude corresponding to peak value frequency position

in phase theta corresponding to peak value frequency position _i2, and X _i2in Doppler frequency f corresponding to peak value frequency position _{i2, harmonic}; Then draw i2 the harmonic components s that needs filtering _{i2, harmonic},

$s_{i2,\mathrm{harmonic}}=({\hat{U}}_{i2}/K)\exp (j2\mathrm{\π}{f}_{i2,\mathrm{harmonic}}t+j{\mathrm{\θ}}_{i2})$

Wherein, t represents the time, and K is echo data x _ncorresponding Narrow-band Radar pulse accumulation number;

S64: to i2 i2 the harmonic components that time domain data filtering is corresponding, obtain i2+1 time domain data;

S65: if i2+1 power corresponding to time domain data is less than the time domain average power of noise, i2+1 time domain data is squelch result; Otherwise, make i2 value add 1, be then back to step S62.

Beneficial effect of the present invention is: the present invention has utilized the assorted echo samples of making an uproar that has correlativity in radar data matrix with target data in distance dimension, and utilize this assorted echo samples of making an uproar to estimate the distributed intelligence of clutter at Doppler domain, in real time and adaptive for CLEAN algorithm provides clutter bandwidth location information accurately, thus the clutter that CLEAN algorithm is extended to non-zero-frequency is suppressed to problem.Utilize the characteristic of CLEAN algorithm to realize squelch simultaneously.

Brief description of the drawings

Fig. 1 is the process flow diagram of the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm of the present invention;

Fig. 2 is the clutter frequency position view that adopts step S2 of the present invention to draw in emulation experiment;

Fig. 3 is result schematic diagram after the sliding window of Narrow-band Radar echo frequency domain that adopts step S2 of the present invention to draw in emulation experiment;

Fig. 4 is the clutter frequency band position view that adopts step S3 of the present invention to draw in emulation experiment;

Fig. 5 is the frequency location scope schematic diagram that adopts CLEAN algorithm that the present invention draws to remove in emulation experiment;

Fig. 6 is the comparison diagram that the reconstructed error of two kinds of signals drawing of emulation experiment changes with signal to noise ratio (S/N ratio).

Embodiment

Below in conjunction with accompanying drawing, the invention will be further described:

With reference to Fig. 1, it is the process flow diagram of the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm of the present invention.For example, the embodiment of the present invention is for the inhibition of making an uproar of mixing of Aircraft Targets echo, and the bandwidth approximate range of Narrow-band Radar described in experimentation of the present invention is that 1MHz is to 3MHz.The assorted inhibition method of making an uproar of the above-mentioned Narrow-band Radar echo based on CLEAN algorithm comprises the following steps:

S1: utilize Narrow-band Radar to receive echo data x _n, the echo data x receiving _ncomprise: the echo data of echo data to a M range unit of the 1st range unit; At echo data x _nin, choose the impurity removal echo samples c that makes an uproar _n, and select target echo sample s _n.N represents the data length (being also the data length of the assorted echo samples of making an uproar) of echo data.Below step S1 is specifically described:

In step S1, receiving echo data x _nafterwards, first determine target place range unit, target place range unit is expressed as to M ₁individual range unit, and by M ₁the echo data of individual range unit is as target echo sample s _n; Then, near target range unit, choose the echo data that only contains clutter and noise as the assorted echo samples of making an uproar, for example, choose and M ₁individual range unit at a distance of the echo data of 5-8 range unit as the echo samples c that makes an uproar that mixes _n.

S2: obtain the clutter frequency position that impurity removal is made an uproar in echo samples.Be described as follows:

Described step S2 specifically comprises the following steps:

S21: determine that sliding window window is long, establishing radar emission signal wavelength is λ, and pulse repetition rate is prf, and sexual intercourse moves fast standard deviation δ _w, δ _winterval be [1.8m/s 4m/s].Minimum meteorological clutter Doppler drift frequency is counted L and is:

$L=\frac{2{\mathrm{\δ}}_{w\min }\×N}{\mathrm{prf}\×\mathrm{\λ}},$

Maximum meteorological clutter Doppler drift frequency is counted L ' and is:

$L^{\′}=\frac{2{\mathrm{\δ}}_{w\max }\×N}{\mathrm{prf}\×\mathrm{\λ}},$

δ _wminfor δ _winterval minimum value, i.e. δ _wmin=1.8, δ _wmaxfor δ _winterval maximal value, i.e. δ _wmax=4.It is long that L is definite sliding window window.For example, when radar emission wavelength λ=0.03, repetition is prf=3KHz, when data length N=512.Calculate

represent to round downwards,

expression rounds up, and at this moment gets L=20 long as sliding window window.

Then to the assorted echo samples c that makes an uproar _nmake Fourier transform, obtain the assorted echo samples frequency point data C that makes an uproar _n, to C _ncarry out the slide window processing that window length is 1 for L step-length, the data that the length that each sliding window is obtained is L are averaged, and utilize a sequence P that length is N-L+1 of all averages compositions _n-L+1, N represents the data length of the assorted echo samples of making an uproar.P _n-L+1in each element be: C _nslide the average of the data that obtain after window at every turn.

S22: to P _n-L+1all averages do ascending order arrange, obtain ascending sequence

S23: by above-mentioned ascending sequence

in before the individual average of N-2L ' at P _n-L+1in position be recorded as set L _oc.

S24: in step S23

in before the individual average of N-2L ' at P _n-L+1in position, from the assorted echo samples frequency point data C that makes an uproar _nin find the frequency point data of correspondence position, C _nin the frequency point data of the correspondence position that finds be expressed as C _l; By C _lin the amplitude of all frequency point data set up noise Doppler amplitude sample D; Then draw noise gate according to noise Doppler amplitude sample D.

Particularly, establishing noise is steady normal process, and noise amplitude distributes through derivation Rayleigh distributed.Normal process remains normal state stably after via linear transformation, to there being also Rayleigh distributed of noise doppler spectral amplitude.The probability distributing density function of noise doppler spectral amplitude is:

$f\left(A\right)=\frac{A}{{{\mathrm{\σ}}_f}^2}\exp (-\frac{A^2}{2{{\mathrm{\σ}}_f}^2})$

Wherein, subscript f (A) represents the probability distribution function of noise Doppler amplitude, utilizes noise Doppler amplitude sample D, utilizes maximum Likelihood to estimate σ _f, A is stochastic variable, the rayleigh distributed fiducial interval that the confidence level of A is 0.9999 is: [0,4.2919 σ _f].

S25: screening C _nmiddle amplitude is greater than the frequency point data of above-mentioned noise gate, C _nin the frequency point data that filters out at C _nin position be clutter frequency position, described clutter frequency positional representation set Q ₁.

S3: obtain the clutter frequency band position that impurity removal is made an uproar in echo samples.

Particularly, first screen P _n-L+1middle amplitude is greater than the frequency point data of above-mentioned noise gate, P _n-L+1in the frequency point data that filters out at P _n-L+1in position be clutter frequency band position, described clutter frequency band positional representation for set Q ₂.

S4: according to above-mentioned assorted clutter frequency position and assorted clutter frequency band position of making an uproar in echo samples of making an uproar in echo samples, determine the frequency location scope that CLEAN algorithm will be removed.

Specifically, definition set Q first ₃=Q ₁∩ Q ₂, wherein, ∩ represents set intersection; To gather Q ₄be defined as: by being positioned at interval [q _min, q _max] between consecution natural number form set, q _minrepresent Q ₃minimum value, q _maxq ₃maximal value; Then, definition set Q ₅=Q ₁-Q ₃, the frequency location Range Representation that CLEAN algorithm will be removed is set Q, Q=Q ₄∪ Q ₅; Wherein, ∪ represents set also, Q ₁-Q ₃represent: set Q ₁with set Q ₃subtract computing, from set Q ₁in remove set Q ₃in element, obtain new set, be designated as Q ₅.

S5: the frequency location scope that will remove according to described CLEAN algorithm, utilize CLEAN algorithm filtering target echo sample s _nthe harmonic components of middle correspondence, obtains clutter and suppresses back echo data.

Step S5 specifically comprises the following steps:

S51: filtering number of times variable i 1 is set, i1=1,2 ...; In the time of i1=1, i1 time domain data is target echo sample s _n, then, execution step S52;

S52: i1 time domain data carried out to Fourier transform, obtain and i1 the frequency domain data X that time domain data is corresponding _i1;

S53: for X _i1, within the scope of the frequency location that will remove at CLEAN algorithm, filter out the frequency domain data X of maximum amplitude _{i1, max}; Draw X _{i1, max}amplitude phase theta _i1, and X _{i1, max}corresponding Doppler frequency f _{i1, clutter}; Then draw i1 the harmonic components s that needs filtering _{i1, clutter}, $s_{i1,\mathrm{clutter}}=({\hat{U}}_{i1}/K)\exp (j2\mathrm{\π}{f}_{i1,\mathrm{clutter}}t+j{\mathrm{\θ}}_{i1}),$ Wherein t represents the time, and K is echo data x _ncorresponding pulse accumulation number;

S54: to corresponding i1 the harmonic components of i1 time domain data filtering, obtain i1+1 frequency domain data X _i1+1;

S55: for X _i1+1, within the scope of the frequency location that will remove at CLEAN algorithm, if filtered out all frequency point data, i1+1 time domain data is that clutter suppresses back echo data

otherwise, make i1 value add 1, be then back to step S52.

S6: clutter is suppressed to back echo data and carry out squelch.

Step S6 specifically comprises the following steps: the first time domain average power of estimating noise; Utilize CLEAN algorithm to suppress back echo data to described clutter and carry out squelch.

The time domain average power of estimating noise comprises the following steps:

By noise Doppler amplitude sample D in step S2, calculate the frequency domain average power ζ of noise ²,

${\mathrm{\ζ}}^2=\frac{{\left|\right|\left(\mathrm{abs}\left(D\right)\right)\left|\right|}^2}{\mathrm{\η}}$

Wherein, abs (D) represents the mould of noise Doppler amplitude sample D, || || represent that asking 2-norm, η is the length of Doppler's amplitude sample D.

Then according to frequency domain average power ζ ², application Paasche Wa Er theorem obtains the time domain average power ε of noise ²:

${\mathrm{\ϵ}}^2=\frac{{\mathrm{\ζ}}^2}{K}$

Wherein, K is echo data x _ncorresponding pulse accumulation number.

Utilizing CLEAN algorithm to carry out squelch to described clutter inhibition back echo data comprises the following steps:

S61: filtering number of times variable i 2 is set, i2=1,2 ...; In the time of i2=1, i2 time domain data is that clutter suppresses back echo data

then, execution step S62;

S62: i2 time domain data carried out to Fourier transform, obtain and i2 the frequency domain data X that time domain data is corresponding _i2;

S63: for X _i2, search and i2 the frequency domain data X that time domain data is corresponding _i2in peak value frequency position; Draw X _i2in amplitude corresponding to peak value frequency position

in phase place corresponding to peak value frequency position

and X _i2in Doppler frequency f corresponding to peak value frequency position _{i2, harmonic}; Then draw i2 the harmonic components s that needs filtering _{i2, harmonic},

$s_{i2,\mathrm{harmonic}}=({\hat{U}}_{i2}/K)\exp (j2\mathrm{\π}{f}_{i2,\mathrm{harmonic}}t+j{\mathrm{\θ}}_{i2})$

Wherein, t represents the time, and K is echo data x _ncorresponding Narrow-band Radar pulse accumulation number;

S64: to i2 i2 the harmonic components that time domain data filtering is corresponding, obtain i2+1 time domain data;

S65: if i2+1 power corresponding to time domain data is less than the time domain average power of noise, i2+1 time domain data is squelch result; Otherwise, make i2 value add 1, be then back to step S62.

Effect of the present invention can be verified by following emulation experiment:

1) experiment scene:

First carry out data declaration:

Clutter database: by the clutter echo data of enrolling in actual environment, form after squelch;

Simulate signal: the three class aircraft echo data that emulation generates, simulate signal does not comprise clutter and noise;

Contain assorted signal: each is to be added and to be formed by the noise signal extracting immediately in the echo signal of emulation and the clutter database of actual measurement containing noise signal, but will ensure signal fuselage doppler spectral and not aliasing of clutter doppler spectral.Signal to noise ratio (simulate signal and noise signal power ratio) is set to 5dB or 10dB at random;

The assorted echo samples of making an uproar: be added and form by noise signal and noise signal in experiment; Each assorted echo samples of making an uproar is under the jurisdiction of one containing assorted signal (with consistent containing clutter scope in assorted signal), signal to noise ratio with add the clutter containing assorted signal consistent, add emulation to generate zero-mean white complex gaussian noise to form the clutter sample of different signal noise ratio level according to corresponding simulate signal energy.Near the echo of range unit can select target place range unit in actual conditions, only containing clutter and noise component.

Contain assorted signals and associated noises: add the zero-mean white complex gaussian noise of emulation to form the signal that noise contains clutter that contains of different signal noise ratio level to the above-mentioned synthetic assorted signal that contains according to corresponding simulate signal energy;

2) experiment content:

Estimate clutter frequency content:

Estimate containing clutter frequency positional information in assorted signal: from the assorted echo samples of making an uproar, find assorted a make an uproar echo corresponding with containing assorted signals and associated noises, after processing according to step S2 of the present invention, obtain Fig. 2.With reference to Fig. 2, it is the clutter frequency position view that adopts step S2 of the present invention to draw in emulation experiment.In Fig. 2, dotted line represents Doppler's spectral line of clutter; Solid line represents Doppler's spectral line of noise.

Estimate containing clutter frequency band positional information in assorted signal: from the assorted echo samples of making an uproar, find assorted a make an uproar echo corresponding with containing assorted signals and associated noises, after processing according to step S2 of the present invention, obtain Fig. 3; With reference to Fig. 3, it is result schematic diagram after the sliding window of Narrow-band Radar echo frequency domain that adopts step S2 of the present invention to draw in emulation experiment.In Fig. 3, represent higher than the curve of noise gate the clutter spectrum estimating.With reference to Fig. 4, it is the clutter frequency band position view that adopts step S3 of the present invention to draw in emulation experiment.In Fig. 4, dotted line represents Doppler's spectral line of clutter, the position of correspondence in Fig. 4 in the clutter spectrum width estimating in Fig. 3; Solid line represents Doppler's spectral line of noise.

Estimate containing region CLEAN frequency positional information in assorted signal: in conjunction with above two steps, after using step S4 of the present invention, step S5 to process, obtain Fig. 5.With reference to Fig. 5, it is the frequency location scope schematic diagram that adopts CLEAN algorithm that the present invention draws to remove in emulation experiment.In Fig. 5, dotted line represents the frequency location scope that need to use CLEAN algorithm to remove.

Reconstructed error comparison:

By the assorted signals and associated noises that contains under different signal to noise ratio (S/N ratio)s, after processing by clutter of the present invention, noise suppressing method respectively, the reconstructed error of calculating and simulate signal; The reconstructed error of simulate signal is defined as:

Wherein, ∑ () represents to sue for peace, || represent to carry out signed magnitude arithmetic(al).

By the assorted signals and associated noises that contains under different signal to noise ratio (S/N ratio)s, do not use any clutter, noise suppressing method processing, calculate the reconstructed error with simulate signal; The reconstructed error containing assorted signals and associated noises that does not carry out impurity elimination denoising is defined as:

With reference to Fig. 6, the comparison diagram that the reconstructed error of two kinds of signals that draw for emulation experiment changes with signal to noise ratio (S/N ratio).In Fig. 6, the line of band " * " represents the reconstructed error curve that the present invention draws, the line of band " o " represents assorted signals and associated noises reconstructed error curve map.

Interpretation:

As can be seen from Figure 6, in whole reconstructed error curve map, be all less than containing assorted signals and associated noises through the signal reconstruction error of processing of the present invention; Under high s/n ratio, noise effect can be ignored substantially, but clutter is appointed right existence, suppresses link because the present invention exists clutter, therefore reconstructed error is still less than containing assorted signals and associated noises under high s/n ratio condition.

Obviously, those skilled in the art can carry out various changes and modification and not depart from the spirit and scope of the present invention the present invention.Like this, if these amendments of the present invention and within modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention is also intended to comprise these changes and modification interior.

Claims (8)

1. the assorted inhibition method of making an uproar of the Narrow-band Radar echo based on CLEAN algorithm, is characterized in that, comprises the following steps:

S1: utilize Narrow-band Radar to receive echo data x _n; At echo data x _nin, select target echo sample s _nwith the assorted echo samples c that makes an uproar _n;

S2: obtain the clutter frequency position that impurity removal is made an uproar in echo samples;

S3: obtain the clutter frequency band position that impurity removal is made an uproar in echo samples;

S4: according to described assorted clutter frequency position and assorted clutter frequency band position of making an uproar in echo samples of making an uproar in echo samples, determine the frequency location scope that CLEAN algorithm will be removed;

S5: the frequency location scope that will remove according to described CLEAN algorithm, utilize CLEAN algorithm filtering target echo sample s _nthe harmonic components of middle correspondence, obtains clutter and suppresses back echo data;

S6: clutter is suppressed to back echo data and carry out squelch.

2. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 1, is characterized in that, in step S1, is receiving echo data x _nafterwards, first determine target place range unit, target place range unit is expressed as to M ₁individual range unit, and by M ₁the echo data of individual range unit is as target echo sample s _n; Then choose and M ₁individual range unit at a distance of the echo data of 5-8 range unit as the echo samples c that makes an uproar that mixes _n.

3. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 1, is characterized in that, described step S2 specifically comprises the following steps:

S21: to the assorted echo samples c that makes an uproar _nmake Fourier transform, obtain the assorted echo samples frequency point data C that makes an uproar _n, to C _ncarry out the slide window processing that window length is 1 for L step-length, the data that the length that each sliding window is obtained is L are averaged, and utilize a sequence P that length is N-L+1 of all averages compositions _n-L+1, N represents the data length of the assorted echo samples of making an uproar;

S22: to P _n-L+1all averages do ascending order arrange, obtain ascending sequence

S23: by described ascending sequence

in before the individual average of N-2L ' at P _n-L+1in position be recorded as set L _oc, L ' is maximum meteorological clutter Doppler drift frequency number;

S24: in step S23 in before the individual average of N-2L ' at P _n-L+1in position, from the assorted echo samples frequency point data C that makes an uproar _nin find the frequency point data of correspondence position, C _nin the frequency point data of the correspondence position that finds be expressed as C _l; By C _lin the amplitude of all frequency point data set up noise Doppler amplitude sample D; Then draw noise gate according to noise Doppler amplitude sample D;

S25: screening C _nmiddle amplitude is greater than the frequency point data of noise gate, C _nin the frequency point data that filters out at C _nin position be clutter frequency position, described clutter frequency positional representation set Q ₁.

4. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 3, is characterized in that, in step S3, first screens P _n-L+1middle amplitude is greater than the frequency point data of noise gate, P _n-L+1in the frequency point data that filters out at P _n-L+1in position be clutter frequency band position, described clutter frequency band positional representation for set Q ₂.

5. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 4, is characterized in that, described step S4 specifically comprises the following steps: first definition set Q ₃=Q ₁∩ Q ₂, wherein, ∩ represents set intersection; To gather Q ₄be defined as: by being positioned at interval [q _min, q _max] between consecution natural number form set, q _minrepresent Q ₃minimum value, q _maxq ₃maximal value; Then, definition set Q ₅=Q ₁-Q ₃, the frequency location Range Representation that CLEAN algorithm will be removed is set Q, Q=Q ₄∪ Q ₅; Wherein, Q ₁-Q ₃represent: set Q ₁with set Q ₃subtract computing, ∪ represents set also.

6. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 1, is characterized in that, described step S5 specifically comprises the following steps:

S51: filtering number of times variable i 1 is set, i1=1,2 ...; In the time of i1=1, i1 time domain data is target echo sample s _n, then, execution step S52;

S52: i1 time domain data carried out to Fourier transform, obtain and i1 the frequency domain data X that time domain data is corresponding _i1;

S53: for X _i1, within the scope of the frequency location that will remove at CLEAN algorithm, filter out the frequency domain data X of maximum amplitude _{i1, max}; Draw X _{i1, max}amplitude

phase theta _i1, and X _{i1, max}corresponding Doppler frequency f _{i1, clutter}; Then draw i1 the harmonic components s that needs filtering _{i1, clutter}, $s_{i1,\mathrm{clutter}}=({\hat{U}}_{i1}/K)\exp (j2\mathrm{\π}{f}_{i1,\mathrm{clutter}}t+j{\mathrm{\θ}}_{i1}),$ Wherein t represents the time, and K is echo data x _ncorresponding pulse accumulation number;

S54: to corresponding i1 the harmonic components of i1 time domain data filtering, obtain i1+1 frequency domain data X _i1+1;

S55: for X _i1+1, within the scope of the frequency location that will remove at CLEAN algorithm, if filtered out all frequency point data, i1+1 time domain data is that clutter suppresses back echo data; Otherwise, make i1 value add 1, be then back to step S52.

7. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 1, is characterized in that, described step S6 specifically comprises the following steps: the first time domain average power of estimating noise; Utilize CLEAN algorithm to suppress back echo data to described clutter and carry out squelch.

8. the assorted inhibition method of making an uproar of a kind of Narrow-band Radar echo based on CLEAN algorithm as claimed in claim 7, is characterized in that, in step S6, utilizes CLEAN algorithm to carry out squelch to described clutter inhibition back echo data and comprises the following steps:

S61: filtering number of times variable i 2 is set, i2=1,2 ...; In the time of i2=1, i2 time domain data is that clutter suppresses back echo data, then, and execution step S62;

S62: i2 time domain data carried out to Fourier transform, obtain and i2 the frequency domain data X that time domain data is corresponding _i2;

S63: for X _i2, search and i2 the frequency domain data X that time domain data is corresponding _i2in peak value frequency position; Draw X _i2in amplitude corresponding to peak value frequency position in phase theta corresponding to peak value frequency position _i2, and X _i2in Doppler frequency f corresponding to peak value frequency position _{i2, harmonic}; Then draw i2 the harmonic components s that needs filtering _{i2, harmonic},

$s_{i2,\mathrm{harmonic}}=({\hat{U}}_{i2}/K)\exp (j2\mathrm{\π}{f}_{i2,\mathrm{harmonic}}t+j{\mathrm{\θ}}_{i2})$

Wherein, t represents the time, and K is echo data x _ncorresponding Narrow-band Radar pulse accumulation number;

S64: to i2 i2 the harmonic components that time domain data filtering is corresponding, obtain i2+1 time domain data;

S65: if i2+1 power corresponding to time domain data is less than the time domain average power of noise, i2+1 time domain data is squelch result; Otherwise, make i2 value add 1, be then back to step S62.

Images