CN107607917B - Sea clutter suppression method based on two-stage Doppler correlation discrimination - Google Patents

Sea clutter suppression method based on two-stage Doppler correlation discrimination Download PDF

Info

Publication number
CN107607917B
CN107607917B CN201710660474.2A CN201710660474A CN107607917B CN 107607917 B CN107607917 B CN 107607917B CN 201710660474 A CN201710660474 A CN 201710660474A CN 107607917 B CN107607917 B CN 107607917B
Authority
CN
China
Prior art keywords
echo signal
doppler
dimensional
matrix
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201710660474.2A
Other languages
Chinese (zh)
Other versions
CN107607917A (en
Inventor
宋万杰
张婷婷
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xidian University
Original Assignee
Xidian University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xidian University filed Critical Xidian University
Priority to CN201710660474.2A priority Critical patent/CN107607917B/en
Publication of CN107607917A publication Critical patent/CN107607917A/en
Application granted granted Critical
Publication of CN107607917B publication Critical patent/CN107607917B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Radar Systems Or Details Thereof (AREA)

Abstract

Hair brushThe invention discloses a sea clutter suppression method based on two-stage Doppler correlation discrimination, which mainly adopts the steps of obtaining an M × N-dimensional time-distance two-dimensional echo signal and obtaining an M subjected to zero-frequency FIR filtering1× N-dimensional echo signal matrix S, M1M-M +1, M representing the maximum order of the zero-frequency FIR filter; according to M after zero-frequency FIR filtering1× N dimension echo signal matrix S, obtaining M2×N1Dimensional Doppler frequency value matrix
Figure DDA0001370439140000011
Wherein M is2＝M1‑1，N1N-4, according to M2×N1Dimensional Doppler frequency value matrix
Figure DDA0001370439140000012
To obtain M3×N2A dimensional binary matrix; wherein M is3＝M2‑1，N2＝N1-M '+1, M' represents the number of range cells contained in the set range cell window, in terms of M3×N2A dimensional binary matrix, which obtains a time-distance two-dimensional Doppler value matrix f after two-stage Doppler correlation discrimination and reset, and the size of the matrix is M3×N2Maintaining; and obtaining a sea clutter suppression result based on the two-stage Doppler correlation discrimination according to the time-distance two-dimensional Doppler value matrix f subjected to the two-stage Doppler correlation discrimination and reset.

Description

Sea clutter suppression method based on two-stage Doppler correlation discrimination
Technical Field
The invention belongs to the technical field of radar signal processing, and particularly relates to a sea clutter suppression method based on two-stage Doppler correlation discrimination, which is suitable for practical engineering application.
Background
The technology of radar inhibition of sea clutter is an important branch in the field of radar signal processing, and target detection under the background of sea clutter is always a difficult problem; due to the reflection action of the sea surface, the echo signal received by the radar contains a large amount of sea clutter energy, so that a target is easy to submerge; and when the high-resolution radar irradiates a rough sea surface at a low ground wiping angle, a radar echo signal even has a 'sea clutter spike effect', and a false target is easy to generate. Different from simple ground clutter, sea clutter presents obvious non-stationarity and non-Gaussian property along with the change of a plurality of factors such as a radar polarization mode, radar resolution, an antenna visual angle, sea conditions, wind direction and the like.
Although people fit the sea clutter by adopting non-Gaussian sea clutter models such as lognormal distribution, Weibull distribution and K distribution to realize target detection, due to various factors of radar and sea conditions, the method for researching the prior clutter statistical characteristic according to the actually-measured sea clutter data and establishing a specific sea clutter model to realize target detection is still poor in effect and has no universality; in traditional radar signal processing, clutter is generally suppressed through frequency domain filtering modes such as moving target display (MTI) and Moving Target Detection (MTD), and target detection is achieved; however, since the doppler spectrum of the sea clutter is generally wide, the measurement of the center of the clutter spectrum and the estimation of the spectral width thereof become difficult.
Disclosure of Invention
Aiming at the defects in the prior art, the method is different from the traditional radar signal processing design self-adaptive filter mode, the invention aims to provide a sea clutter suppression method based on two-stage Doppler correlation discrimination.
The realization idea of the invention is as follows: in order to effectively distinguish a target area and a clutter area, Doppler time correlation judgment and Doppler space correlation judgment are sequentially carried out on time-distance two-dimensional Doppler frequency obtained by alternate frequency measurement, an area meeting the correlation requirement is screened out, namely the target area, and other areas with low correlation are regarded as sea clutter areas. Through two-stage Doppler correlation discrimination binarization processing, all Doppler values of a target area are reset to be 1 value, and all Doppler values of a clutter area are reset to be 0 value; and finally, according to the Doppler value after binarization, the target signal is reserved, and clutter signals are all normalized into an average noise level, so that effective sea clutter suppression is realized.
In order to achieve the above purpose, the present invention is realized by the following specific technical scheme.
A sea clutter suppression method based on two-stage Doppler correlation discrimination comprises the following steps:
step 1, obtaining M × N-dimensional time-distance two-dimensional echo signals and obtaining M after zero-frequency FIR filtering1× N-dimensional echo signal matrix S, M1M-M +1, M representing the maximum order of the zero-frequency FIR filter, M, M1M and N are positive integers greater than 0 respectively;
step 2, according to the M after zero frequency FIR filtering1× N dimension echo signal matrix S, obtaining M2×N1Dimensional Doppler frequency value matrix
Figure BDA0001370439120000021
Wherein M is2＝M1-1，N1＝N-4，M2、N1Are respectively positive integers greater than 0;
step 3, according to M2×N1Dimensional Doppler frequency value matrix
Figure BDA0001370439120000022
To obtain M3×N2A dimensional binary matrix; wherein M is3＝M2-1，N2＝N1-M '+1, M' represents the number of range cells contained in the set range cell window, M3、N2And m' are positive integers greater than 0;
step 4, according to M3×N2A dimensional binary matrix, and a time-distance two-dimensional Doppler value matrix f with the size of M after two-stage Doppler correlation discrimination processing is obtained3×N2Maintaining;
and 5, obtaining a sea clutter suppression result based on the two-stage Doppler correlation discrimination according to the time-distance two-dimensional Doppler value matrix f subjected to the two-stage Doppler correlation discrimination processing.
The invention has the beneficial effects that:
1) the method solves the problem that the traditional radar signal processing is difficult to filter in the frequency domain and calculate the center and the spectrum width of the sea clutter spectrum, and provides a two-stage Doppler correlation discrimination mode different from the traditional radar signal frequency domain filtering processing for suppressing the sea clutter. And distinguishing a target area and a clutter area according to the Doppler time-space correlation difference of the target and the clutter, thereby inhibiting the sea clutter.
2) The method does not depend on a specific sea clutter model when the discrimination processing for inhibiting the sea clutter is carried out, and the method has universality. The sea clutter suppression method has a very obvious effect of suppressing the sea clutter, and compared with the traditional filtering mode, the sea clutter suppression method can effectively reduce the noise-to-noise ratio.
Drawings
The invention is described in further detail below with reference to the following description of the drawings and the detailed description.
FIG. 1 is a flow chart of a sea clutter suppression method based on two-stage Doppler correlation discrimination according to the present invention;
FIG. 2(a) is a Doppler correlation coefficient diagram between two adjacent pulses of the same range cell of the target A obtained by the method of the present invention;
FIG. 2(B) is a diagram of Doppler correlation coefficients between two adjacent pulses of the same range cell of the target B obtained by the method of the present invention;
FIG. 3(a) is a diagram of the 103 th PRI raw sea clutter data after pulse compression;
FIG. 3(b) is a diagram of the 103 th PRI sea clutter data obtained by the method of the present invention.
Detailed Description
Referring to fig. 1, it is a flow chart of a sea clutter suppression method based on two-stage doppler correlation discrimination according to the present invention; the sea clutter suppression method based on the two-stage Doppler correlation discrimination comprises the following steps:
step 1, obtaining M × N-dimensional time-distance two-dimensional echo signals and obtaining M after zero-frequency FIR filtering1× N-dimensional echo signal matrix S, M1M-M +1, M representing the maximum order of the zero-frequency FIR filter, M, M1N are positive integers larger than 0 respectively, and FIR is finite long single-bit impulse response.
The substep of step 1 is:
1a) determining radar, setting target in radar detection range, and repeating pulse period TrReceiving a distance dimension continuous pulse echo signal, performing pulse compression processing on the received distance dimension continuous pulse echo signal, and then performing pulse echo signal rearrangement according to a time dimension to obtain a time-distance two-dimensional echo signal, wherein the time dimension of the time-distance two-dimensional echo signal comprises M pulse echo signals, the distance dimension of the time-distance two-dimensional echo signal means that each pulse echo signal comprises N distance units, and M, N is a positive integer greater than 0.
1b) Recording the sequence of m pulse echo signal pulses starting with the ith pulse echo signal and the echo signal at the jth distance unit as XijAnd starting m pulse echo signal pulses of the ith pulse echo signal and an echo signal sequence X at the jth distance unitijRemoving the influence of ground clutter on subsequent Doppler discrimination processing through an m-order zero-frequency FIR filter to obtain an ith pulse echo signal and an echo signal S at a jth distance unit after zero-frequency FIR filteringijThe calculation expression is as follows:
Sij＝Xij·WT
wherein, XijRepresenting m continuous pulse echo signal pulses starting as the ith pulse echo signal and the echo signal sequence at the jth distance unit, wherein the m continuous pulse echo signal pulses starting as the ith pulse echo signal are from the ith pulse echo signal to the (i + m-1) th pulse echo signal, and Xij＝[xij,x(i+1)j,...,x(i+m-1)j]，xijRepresenting the i-th pulse echo signal, the echo signal at the j-th range cell, x(i+1)jRepresents the (i + 1) th pulse echo signal, the echo signal at the jth range cell, x(i+m-1)jRepresents the (i + m-1) th pulse echo signal and the echo signal at the jth range bin, W represents the m-order FIR filter weight coefficient sequence, and W ═ W1,w2,...,wh,...,wm]，whRepresenting the weight coefficient of the h-th order FIR filter, h ∈ {1,2, …, M }, M representing the maximum order of the zero-frequency FIR filter, and M being a positive integer greater than 0, wherein M is 5, i ∈ {1,2, …, M-1}，j∈{1,2,…,N}，M1The initial values of M-M +1, i and j are each 1.
i＝M1When it is started as M1M pulse echo signal pulses of the pulse echo signal, the echo signal sequence at the jth distance unit is
Figure BDA0001370439120000041
Denotes the M th1A pulse echo signal, an echo signal at a jth range bin,
Figure BDA0001370439120000042
denotes the M th1+1 pulse echo signal, echo signal at jth range bin, xMjRepresenting the mth pulse echo signal, the echo signal at the jth range bin.
1c) Let j have constant value and let i have values from 1 to M1And repeatedly executing 1b), and further respectively obtaining the 1 st pulse echo signal after zero-frequency FIR filtering and the echo signal S at the jth distance unit1jTo Mth after zero frequency FIR filtering1Pulse echo signal, echo signal at jth range unit
Figure BDA0001370439120000043
Recording as the echo signal S at the jth distance unit after zero-frequency FIR filteringj。
1d) Respectively taking the value of j from 1 to N, setting the value of i to 1, and repeatedly executing 1b) and 1c) to further respectively obtain echo signals S at the 1 st distance unit after zero-frequency FIR filtering1Echo signal S to the Nth distance unit after zero frequency FIR filteringNIs recorded as M after zero frequency FIR filtering1× N-dimensional echo signal matrix S.
Step 2, M after zero frequency FIR filtering1× N dimension echo signal matrix S carries out alternate frequency measurement to obtain corresponding time-range two-dimensional doppler frequency values, wherein the interval frequency measurement is used to calculate the doppler frequency of each location point by using the phase difference between two adjacent pulses.
2.1 calculating M after zero-frequency FIR Filtering1× N-dimensional echo signal matrix S, i 'th pulse echo signal, j' th Doppler frequency value at distance unit
Figure BDA0001370439120000044
The calculation expression is as follows:
Figure BDA0001370439120000045
wherein, TrIn the form of a pulse repetition period,
Figure BDA0001370439120000046
represents the correlation function between the ith ' pulse echo signal and the ith ' +1 pulse echo signal at the jth ' range bin,
Figure BDA0001370439120000047
i'∈{1,2,…,M2}，M2＝M1-1，j'∈{1,2,…,N1}，N1＝N-4，k∈{j',j'+1,…,j'+4}，Si'krepresents the ith' pulse echo signal after zero frequency FIR filtering, the echo signal at the kth distance unit, S(i'+1)kThe signals represent the i ' +1 pulse echo signal after zero frequency FIR filtering and the echo signal at the k distance unit, and the initial values of i ' and j ' are 1 respectively.
When i ═ M2And j ═ N1When the temperature of the water is higher than the set temperature,
Figure BDA0001370439120000051
Figure BDA0001370439120000052
representing the Mth after zero-frequency FIR filtering2Pulse echo signals, echo signals at the kth range unit,
Figure BDA0001370439120000053
representing the Mth after zero-frequency FIR filtering1Pulse echo signal, echo signal at kth range bin.
Because the Doppler directions of the same target among several continuous pulses are consistent, a time average correlation function of continuous sliding of two adjacent pulse sequences of the same sampling point is used for replacing a statistical average correlation function; tg-1() Representing an arctangent function with a value range of { -pi, pi }; re represents the operation of the real part, Im represents the operation of the imaginary part; estimated in this way
Figure BDA0001370439120000054
Range (-f)r,+fr)，frIndicating the pulse repetition frequency.
2.2 make j 'constant and make i' take values from 1 to M respectively2And repeatedly executing the steps for 2.1 to further obtain M subjected to zero-frequency FIR filtering respectively1× N-dimensional echo signal matrix S, the value of Doppler frequency at the 1 st pulse echo signal and the jth' th range cell
Figure BDA0001370439120000055
To M after zero frequency FIR filtering1× Mth in N-dimensional echo signal matrix S2Doppler frequency value at jth distance cell of pulse echo signal
Figure BDA0001370439120000056
Is recorded as M after zero frequency FIR filtering1M at jth' range cell in × N-dimensional echo signal matrix S2A Doppler frequency value
Figure BDA0001370439120000057
2.3 let j' take values from 1 to N, respectively1Setting the value of i' to 1, and repeatedly executing 2.1 and 2.2 to obtain M after zero-frequency FIR filtering respectively1M at the 1 st range cell in the × N-dimensional echo signal matrix S2A Doppler frequency value
Figure BDA0001370439120000058
To M after zero frequency FIR filtering1× Nth echo signal matrix S1M at distance unit2A Doppler frequency value
Figure BDA0001370439120000059
Is marked as M2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000510
The M is2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000511
In which contains M2×N1A Doppler frequency value.
Step 3, for M2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000512
Performing Doppler time correlation discrimination reset to obtain M3×N2A dimensional binary matrix; namely to M2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000513
Respectively carrying out Doppler binarization processing on each point: and setting the Doppler value of the position point with Doppler time correlation as a value 1 and setting the Doppler value of the position point without Doppler time correlation as a value 0 according to the Doppler time correlation threshold.
The Doppler time correlation discrimination processing procedure is as follows:
3.1 setting a distance unit window, wherein the distance unit window comprises m 'distance units, and m' in the embodiment takes an empirical value of 5; calculating the sliding of the time-distance two-dimensional Doppler frequency value, and obtaining a two-dimensional Doppler time correlation coefficient which is used for expressing the Doppler correlation between two adjacent pulses of each position point; setting the Doppler frequency of m' range bins in a pulse sequence to be
Figure BDA00013704391200000514
Representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure BDA00013704391200000515
The value of the doppler frequency at the 1 st range cell,
Figure BDA0001370439120000061
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure BDA0001370439120000062
The value of the doppler frequency at the 2 nd range cell,
Figure BDA0001370439120000063
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure BDA0001370439120000064
The pulse echo signal, the value of the doppler frequency at the m' th range bin,
Figure BDA0001370439120000065
M3＝M2-1，
Figure BDA0001370439120000066
N2＝N1-m′+1，
Figure BDA0001370439120000067
is set to an initial value of 1,
Figure BDA0001370439120000068
is 0.
3.2 calculation with
Figure BDA0001370439120000069
M' range cells starting from one range cell, the first
Figure BDA00013704391200000610
Doppler frequency sequence of pulse echo signals and
Figure BDA00013704391200000611
correlation coefficient of Doppler frequency sequence of pulse echo signals
Figure BDA00013704391200000612
The expression is as follows:
Figure BDA00013704391200000613
wherein the content of the first and second substances,
Figure BDA00013704391200000614
is expressed as
Figure BDA00013704391200000615
M' range cells starting from one range cell, the first
Figure BDA00013704391200000616
A sequence of doppler frequencies of the individual pulse echo signals,
Figure BDA00013704391200000617
g∈{0,1,…,m'-1}，
Figure BDA00013704391200000618
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure BDA00013704391200000619
A pulse echo signal of
Figure BDA00013704391200000620
The value of the doppler frequency at each range cell,
Figure BDA00013704391200000621
is expressed as
Figure BDA00013704391200000622
M' range cells starting from one range cell, the first
Figure BDA00013704391200000623
A sequence of doppler frequencies of the individual pulse echo signals,
Figure BDA00013704391200000624
Figure BDA00013704391200000625
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure BDA00013704391200000626
A pulse echo signal of
Figure BDA00013704391200000627
The value of the doppler frequency at each range cell,
Figure BDA00013704391200000628
is expressed as
Figure BDA00013704391200000629
M' range cells starting from one range cell, the first
Figure BDA00013704391200000630
Doppler frequency sequence of pulse echo signals
Figure BDA00013704391200000631
And in the first place
Figure BDA00013704391200000632
M' range cells starting from one range cell, the first
Figure BDA00013704391200000633
Doppler frequency sequence of pulse echo signals
Figure BDA00013704391200000634
The covariance of (a) of (b),
Figure BDA00013704391200000635
is expressed as
Figure BDA00013704391200000636
M' range cells starting from one range cell, the first
Figure BDA00013704391200000637
Doppler frequency sequence of pulse echo signals
Figure BDA00013704391200000638
The variance of (a) is determined,
Figure BDA00013704391200000639
is expressed as
Figure BDA00013704391200000640
M' range cells starting from one range cell, the first
Figure BDA00013704391200000641
Doppler frequency sequence of pulse echo signals
Figure BDA00013704391200000642
The variance of (c).
When in use
Figure BDA00013704391200000643
And is
Figure BDA00013704391200000644
Is obtained by the Nth2M' range units starting from one range unit, Mth3Doppler frequency sequence of +1 pulse echo signals
Figure BDA00013704391200000645
Figure BDA0001370439120000071
g∈{0,1,…,m'-1}，
Figure BDA0001370439120000072
Representing M after zero-frequency FIR filtering1× Mth in N-dimensional echo signal matrix S2Pulse echo signal, Nth2+ g doppler frequency values at range bins.
3.3 order
Figure BDA0001370439120000073
Is unchanged, and is
Figure BDA0001370439120000074
Are respectively 1 to M3Repeating the execution of 3.2 to obtain the first step
Figure BDA0001370439120000075
Doppler frequency sequence correlation coefficient of m' range cells, 1 st pulse echo signal and 2 nd pulse echo signal starting from one range cell
Figure BDA0001370439120000076
To by
Figure BDA0001370439120000077
M' range units starting from one range unit, Mth3Pulse echo signal and Mth3Doppler frequency sequence correlation coefficient of +1 pulse echo signal
Figure BDA0001370439120000078
Is marked as M2×N1Dimensional Doppler frequency value matrix
Figure BDA0001370439120000079
To middle
Figure BDA00013704391200000710
M at distance unit3Coefficient of correlation
Figure BDA00013704391200000711
3.4 order
Figure BDA00013704391200000712
Respectively take 0 to N2-1, mixing
Figure BDA00013704391200000713
Is set to 1, the execution is repeated for 3.2 and 3.3, and M is obtained2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000714
M at the 1 st distance unit3Individual correlation coefficient ρ1To M2×N1Dimensional Doppler frequency value matrix
Figure BDA00013704391200000715
Middle N2M at distance unit3Coefficient of correlation
Figure BDA00013704391200000716
Is marked as M3×N2A matrix of dimensional correlation coefficients ρ, M3×N2The dimensional correlation coefficient matrix rho contains M3×N2A correlation coefficient of M3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA00013704391200000717
A pulse echo signal of
Figure BDA00013704391200000718
The correlation coefficient at each distance unit is recorded as
Figure BDA00013704391200000719
Figure BDA00013704391200000720
M3＝M2-1，
Figure BDA00013704391200000721
N2＝N1-m′+1，
Figure BDA00013704391200000722
And
Figure BDA00013704391200000723
are each 1.
3.5 pairs of M3×N2And (3) performing Doppler time correlation judgment on the dimensional correlation coefficient matrix rho: setting correlation coefficient threshold value
Figure BDA00013704391200000724
Figure BDA00013704391200000725
In this example
Figure BDA00013704391200000726
The value was 0.98.
3.6 if M3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA00013704391200000727
A pulse echo signal of
Figure BDA00013704391200000728
Correlation coefficient at distance unit
Figure BDA00013704391200000729
Greater than or equal to the correlation coefficient threshold value
Figure BDA00013704391200000730
Then M3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA00013704391200000731
A pulse echo signal of
Figure BDA00013704391200000732
Correlation coefficient at distance unit
Figure BDA00013704391200000733
Has Doppler time correlation at the position point, and M is calculated3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA00013704391200000734
A pulse echo signal of
Figure BDA00013704391200000735
Correlation coefficient at distance unit
Figure BDA00013704391200000736
Is set to 1 as M3×N2The j 'th element value in the intermediate matrix is maintained, and the initial value of j' is added with 1, j 'to be 1, j' ∈ {1,2, …, M3×N2}。
If M is3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA0001370439120000081
A pulse echo signal of
Figure BDA0001370439120000082
Correlation coefficient at distance unit
Figure BDA0001370439120000083
Less than a correlation coefficient threshold value
Figure BDA0001370439120000084
Then M3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA0001370439120000085
A pulse echo signal of
Figure BDA0001370439120000086
Correlation coefficient at distance unit
Figure BDA0001370439120000087
Has no Doppler time correlation at the position point and M is calculated3×N2The first in the dimensional correlation coefficient matrix rho
Figure BDA0001370439120000088
A pulse echo signal of
Figure BDA0001370439120000089
Correlation coefficient at distance unit
Figure BDA00013704391200000810
Is set to 0 as M3×N2The j 'th element value in the intermediate matrix is maintained, and the initial value of j' is added with 1, j 'to be 1, j' ∈ {1,2, …, M3×N2}。
Specifically, the obtained M3×N2The size of each correlation coefficient in the dimensional correlation coefficient matrix can be used for judging M3×N2The Doppler time correlation between each correlation coefficient position point in the dimensional correlation coefficient matrix and two adjacent pulses of the distance unit is estimated by a large number of repeated experiments to obtain the Doppler time correlation coefficient threshold range, then the current correlation coefficient threshold is set to be a certain fixed value by the experiment, and when M is equal to M, the current correlation coefficient threshold is set to be a certain fixed value3×N2When each correlation coefficient in the dimensional correlation coefficient matrix is greater than or equal to the threshold fixed value, M is considered3×N2And the position point at the position corresponding to the correlation coefficient in the dimensional correlation coefficient matrix has Doppler time correlation, otherwise, the position point is considered to have no Doppler time correlation.
According to M3×N2Judging the Doppler correlation of the position points at any correlation coefficient in the dimensional correlation coefficient matrix, and carrying out binarization processing on the Doppler values of all the position points: that is, the doppler value of each position point having doppler time correlation is set to 1 value, and the doppler value of each position point having no correlation is set to 0 value.
3.7 order
Figure BDA00013704391200000811
Is unchanged, and is
Figure BDA00013704391200000812
Are respectively 1 to M3Repeating the execution for 3.6 to obtain M3×N2Value of the 1 st element to M in the dimension intermediate matrix3×N2M in the dimension intermediate matrix3Value of element, denoted as M3×N2Dimension of the intermediate matrix
Figure BDA00013704391200000813
M at distance unit3The value of each element.
3.8 order
Figure BDA00013704391200000814
Respectively taking 1 to N2Will be
Figure BDA00013704391200000815
Is set to 1, the execution is repeated for 3.6 and 3.7, and M is obtained3×N2M at the 1 st distance cell in the dimension intermediate matrix3Value of element to M3×N2Nth in dimension intermediate matrix2M at distance unit3Value of element, denoted as M3×N2Dimensional binary matrix, M3×N2Correspondence in the dimensional binary matrix includes M3×N2A Doppler value of M3×N2Dimension two-value matrix
Figure BDA00013704391200000816
A pulse echo signal of
Figure BDA00013704391200000817
The corresponding Doppler value at each range bin is recorded as
Figure BDA00013704391200000818
Step 4, for M3×N2Performing Doppler space correlation discrimination reset on the dimensional binary matrix to obtain P stripsAnd continuously simulating a main lobe sequence of the target so as to determine a clutter region and a target region in the time-distance two-dimensional echo signal. And resetting the Doppler value of each point in the clutter area to be 0 value, and resetting the Doppler value of each point in the target area to be 1 value.
The Doppler space correlation discrimination processing procedure is as follows:
4.1, performing sliding search on the two-dimensional Doppler value subjected to Doppler time correlation discrimination binarization processing to find out all 'continuous quasi-target main lobe sequences'; due to the correlation difference of the Doppler frequencies of the target and the sea clutter, after the Doppler time correlation discrimination binarization processing, the Doppler values of most of the sea clutter areas are already set to be 0 values, and the Doppler values of the target areas are mostly set to be 1 values; however, due to the mutation of the sea clutter, the clutter region after being judged by Doppler time correlation still has 1-value Doppler which is partially scattered, and for the continuous pulse main lobe sequence of the target region, each main lobe is basically accompanied with at least one 1-value Doppler side lobe; according to the area block correlation characteristic, a continuous target pulse main lobe, called as a continuous quasi-target main lobe sequence, can be searched by 1-value Doppler side lobes of continuous pulses, and a rectangular area block which is formed by taking the sequence position as the center is considered to have Doppler spatial correlation.
For convenience of explanation, the specific search process is as follows:
initialization, setting a target window length, wherein the target window length is the width of u pulse echo signals, u is a positive integer greater than 0, and the value of u is usually 7 or 8, and setting b ∈ {0,1, …, u } and the initial value of b to be 0.
Setting M3×N2The ith pulse echo signal and the Doppler values at 3 continuous distance units in the dimensional binary matrix are respectively
Figure BDA0001370439120000091
Wherein
Figure BDA0001370439120000092
Represents M3×N2Corresponding Doppler values of ith pulse echo signal and jth distance unit in dimensional binary matrix，
Figure BDA0001370439120000093
Represents M3×N2The corresponding Doppler value at the ith pulse echo signal and the jth' -1 distance unit in the dimensional binary matrix,
Figure BDA0001370439120000094
represents M3×N2When the ith pulse echo signal in the dimensional binary matrix, the corresponding Doppler value at the jth '″ +1 distance unit and j' ″ -1 ═ 0
Figure BDA0001370439120000095
j”'+1>N2Time of flight
Figure BDA0001370439120000096
i”∈{1,2,…,M3}，M3＝M2-1，j”'∈{1,2,…,N2}，N2＝N1The initial values of-m '+1, i "and j'" are 1, respectively.
4.2 calculating Doppler F at j 'th distance unit when the number of pulse echo signals in the target window length is i' + bi”+b,j”'The expression is as follows:
Figure BDA0001370439120000097
wherein the content of the first and second substances,
Figure BDA0001370439120000098
represents M3×N2The ith '+ b pulse echo signal, the jth' ″ -1 Doppler value at the range bin in the dimensional binary matrix,
Figure BDA0001370439120000099
represents M3×N2The ith '+ b pulse echo signal, the jth' ″ Doppler value at the range bin in the dimensional binary matrix,
Figure BDA00013704391200000910
represents M3×N2The i '+ b pulse echo signal, the j' +1 doppler value at the range bin in the dimensional binary matrix.
4.3 Doppler F at j 'th distance unit if number of pulse echo signals in target window length is i' + bi”+b,j”'Not equal to 0, then
Figure BDA00013704391200000911
Not all 0, i.e. at least one Doppler value of 1, then M3×N2In the dimension binary matrix, the ith '+ b pulse echo signal and the jth' ″ effective target main lobe where the distance unit is located are executed for 4.4; otherwise, abandon M3×N2And (3) the ith ' + b pulse echo signal and the jth ' ″ distance unit in the dimensional binary matrix, adding 1 to the value of j ' ″ and returning to 4.2.
4.4 let b take 0 to u respectively, and repeat 4.2 and 4.3 until obtaining the Doppler F at the j 'th distance unit when the number of pulse echo signals in the target window length is i' +0i”+0,j”'Doppler F at the j 'th distance unit when the number of pulse echo signals in the length of a target window is i' + ui”+u,j”'And if so, if satisfied
Fi”+0,j”'·Fi”+1,j”'·...·Fi”+u,j”'≠0
I.e. Fi”+0,j”',Fi”+1,j”',...,Fi”+u,j”'When all the signals are not 0, the Doppler F at the j 'th distance unit when the number of pulse echo signals in the target window length is i' +0i”+0,j”'Doppler F at the j 'th distance unit when the number of pulse echo signals in the length of a target window is i' + ui”+u,j”'The effective target main lobe which is positioned in u +1 continuous pulse echo signals is recorded as the p-th continuous quasi-target main lobe sequence, the initial value of p is 1, the value of p is added with 1, the value of i 'is not changed, the value of j' is added with 1, and 4.2 is returned; otherwise, let the value of i "be constant and let the value of j'" add 1, returning 4.2.
If j'>N2Then add 1 to the value of i ", set the value of j'" to 1, and return to 4.2.
Until the P continuous quasi-target main lobe sequence is obtained, P<M3×N2And recording the 1 st to the P th continuous quasi-target main lobe sequences obtained at the moment as M3×N2P continuous quasi-target main lobe sequences in the dimension binary matrix; if P is equal to 0, it is indicated that no target exists in the radar detection range, the targets are clutter areas, no Doppler space correlation exists, the target areas are marked as clutter areas in the time-distance two-dimensional echo signals, and 4.7 is executed; if 0<P<M3×N2P is initialized to 1 and 4.5 is performed.
4.5 according to M3×N2P continuous quasi-target main lobe sequences in the dimensional binary matrix determine P target areas in the time-distance two-dimensional echo signals and clutter areas in the time-distance two-dimensional echo signals.
4.5.1 respectively using the p continuous quasi-target main lobe sequence as the center to expand 2 distance units or 2 pulse echo signals to the periphery to form a rectangular area block, and obtaining the p continuous quasi-target main lobe sequence protection area.
4.5.2 let P take 1 to P respectively, and repeat execution of 4.5.1 until the P-th continuous quasi-target main lobe sequence protection region is obtained, at this time, the 1-st continuous quasi-target main lobe sequence protection region to the P-th continuous quasi-target main lobe sequence protection region are obtained, and each continuous quasi-target main lobe sequence protection region is (u +4) × 5 dimensional two-dimensional matrix respectively.
4.5.3 recording the corresponding regions from the 1 st continuous quasi-target main lobe sequence protection region to the P th continuous quasi-target main lobe sequence protection region in the time-distance two-dimensional echo signal as P target regions in the time-distance two-dimensional echo signal, wherein the P target regions in the time-distance two-dimensional echo signal have Doppler spatial correlation; and recording the rest areas except the P target areas in the time-distance two-dimensional echo signal as clutter areas in the time-distance two-dimensional echo signal, wherein the clutter areas in the time-distance two-dimensional echo signal have no Doppler spatial correlation.
4.6 separately setting the Doppler value at each position point in P target areas in the time-distance two-dimensional echo signalSetting the Doppler value at each position point in a clutter region in a time-distance two-dimensional echo signal as 0 respectively, recording the Doppler value as a time-distance two-dimensional Doppler value matrix f after two-stage Doppler correlation discrimination processing, wherein the size of the time-distance two-dimensional Doppler value matrix f is M3×N2And (5) maintaining.
4.7 setting the Doppler value at each position point in the clutter region in the time-distance two-dimensional echo signal as 0 respectively, and recording as a time-distance two-dimensional Doppler value matrix f after two-stage Doppler correlation discrimination processing, wherein the size of the time-distance two-dimensional Doppler value matrix f is M3×N2And (5) maintaining.
And 5, obtaining a sea clutter suppression result based on the two-stage Doppler correlation discrimination according to the time-distance two-dimensional Doppler value matrix f subjected to the two-stage Doppler correlation discrimination processing.
Specifically, the time-distance two-dimensional doppler value matrix f after the two-stage doppler correlation discrimination processing is discriminated and detected, and the substeps are as follows:
5.1 recording the Doppler values of the ith 'pulse echo signal and the jth' distance unit in the time-distance two-dimensional Doppler value matrix f after the two-stage Doppler correlation judgment processing as fi”'j””，i”'∈{1,2,…,M3}，M3＝M2-1，j””∈{1,2,…,N2}，N2＝N1The initial values of-m '+1, i' "and j" "are 1, respectively.
5.2 if the time-distance two-dimensional Doppler value matrix f is subjected to two-stage Doppler correlation judgment processing, the ith 'pulse echo signal and the jth' Doppler value f at the distance unit are obtainedi”'j”” Is 1, then M after zero frequency FIR filtering1The ith 'pulse echo signal and the jth' echo signal in the × N-dimensional echo signal matrix S are kept unchanged and are recorded as M after zero-frequency FIR filtering1× N-dimensional echo signal matrix S, i 'th pulse echo signal, j' th reset echo signal at distance unit.
If the ith' ″ in the time-distance two-dimensional Doppler value matrix f after the two-stage Doppler correlation discrimination processingPulse echo signal, Doppler value f at jth' distance uniti”'j””If 0, let M after zero-frequency FIR filtering1The phase of the ith 'pulse echo signal and the phase of the echo signal at the jth' distance unit in the × N-dimensional echo signal matrix S are unchanged, the amplitude is normalized to be the average noise level, and the average noise level is recorded as M after the zero-frequency FIR filtering1× N-dimensional echo signal matrix S, i 'th pulse echo signal, j' th reset echo signal at distance unit.
5.3 make j "" constant, and make i' "take values from 1 to M, respectively3And repeatedly executing the steps of 5.2 to further obtain M subjected to zero-frequency FIR filtering respectively1× N-dimensional echo signal matrix S from pulse echo signal at 1 st position, reset signal at jth' distance unit to M after zero-frequency FIR filtering1× Mth in N-dimensional echo signal matrix S3The pulse echo signal and the reset echo signal at the jth distance unit are recorded as M after zero frequency FIR filtering1M in × N-dimensional echo signal matrix S3Pulse echo signal, j "" reset echo signal at distance unit.
5.4 setting the value of i 'to 1, and respectively taking the values of j' from 1 to N2Repeatedly executing 5.2 and 5.3 to respectively obtain M3Pulse echo signal, reset echo signal at 1 st range unit to M3Pulse echo signal, Nth2Reset echo signal at one range cell, denoted M3Pulse echo signal, N2Reset echo signal at range unit
Figure BDA0001370439120000121
M after zero frequency FIR filtering1Removing M from × N-dimensional echo signal matrix S3Pulse echo signal, N2Reset echo signal at range unit
Figure BDA0001370439120000122
All echo signals except for
Figure BDA0001370439120000123
Further obtaining a sea clutter suppression result based on two-stage Doppler correlation discrimination
Figure BDA0001370439120000124
Figure BDA0001370439120000125
Sea clutter suppression result based on two-stage Doppler correlation discrimination
Figure BDA0001370439120000126
Target information is reserved, and sea clutter is effectively suppressed.
And the average noise level is the average level value of the residual noise signal after P target areas and clutter are removed from the time-distance two-dimensional echo signal.
The effect of the present invention is further verified and explained below with the combination of simulation experiments.
Simulation conditions
The computer operating system of the experiment is 64-bit Windows 7SP1, and the judgment and the inhibition of the sea clutter based on the two-stage Doppler correlation are realized by using Matlab2014b as a simulation platform. Actual measurement sea clutter data collected by a remote early warning radar in a certain city of east coast of China is used as an experimental data sample in the experiment, and two simulated air targets are artificially added: target a and target B.
(II) simulation content
Firstly, filtering ground clutter of actually measured sea clutter data after original pulse pressure by a zero-frequency FIR filter, and carrying out alternate frequency measurement on the ground clutter data to obtain time-distance two-dimensional Doppler frequency; then, sliding the two-dimensional Doppler frequency to obtain the Doppler time correlation coefficient of each position point, obtaining the Doppler correlation coefficient graph between two adjacent pulses of the same distance unit in the graph 2, and performing binarization on the Doppler value of each position point according to the time correlation coefficient threshold judgment to obtain the binarization Doppler value (part) after the time correlation judgment in the table 1; then, sliding the binary Doppler value, and judging and searching out all 'continuous quasi-target mainlobe sequences' by Doppler space correlation so as to determine a target area and a clutter area; and finally, reserving the target area signals according to the two-dimensional Doppler value after the two-stage correlation judgment, and normalizing the clutter area signals into an average noise level to obtain a front-back comparison graph (103 PRI) of the two-stage Doppler correlation judgment for sea clutter suppression in the graph 3.
FIG. 2(a) is a diagram of Doppler correlation coefficients between two adjacent pulses of the same range bin of a target A obtained by the method of the present invention, and FIG. 2(B) is a diagram of Doppler correlation coefficients between two adjacent pulses of the same range bin of a target B obtained by the method of the present invention; the horizontal coordinates represent distance units, and the vertical coordinates represent Doppler time correlation coefficients; the larger the correlation coefficient is, the higher the Doppler time correlation between the two adjacent pulses of the distance unit and the position point is. As can be seen from fig. 2(a) and 2(B), the doppler correlation coefficients near the object a (548 range cell) and the object B (2023 range cell) are relatively high, while the doppler correlation coefficient at the clutter outside the object range cell is significantly lower.
Fig. 3(a) and 3(b) are pre-and post-103 PRI comparison diagrams for suppressing sea clutter processing by using the two-stage doppler correlation determination technique of the present invention, wherein the abscissa is a distance unit, and the ordinate is a sea clutter signal amplitude (decibel), fig. 3(a) is a 103PRI original sea clutter data diagram after pulse compression processing, and fig. 3(b) is a 103PRI sea clutter data diagram obtained by using the method of the present invention; comparing fig. 3(a) and fig. 3(b), it is found that the sea clutter can be effectively suppressed by using the present invention while preserving the target information.
(III) Table description
Table 1 shows the partial doppler value data of the 100 th and 106 th pulse of the time-distance two-dimensional doppler frequency after the doppler time correlation determination binarization. The A target area is near the 550 th distance unit, and the sea clutter area is near the 890 th distance unit. Looking at table 1, it can be seen that the 550 th range cell 100 th-106 pulses in the target region have at least one 1-valued doppler side lobe in each of the 7 consecutive pulse main lobes, which is referred to herein as the 7 consecutive pulse "pseudo-target main lobe sequence". While the 1-value Doppler points near the 890 th range unit clutter area are obviously scattered and have no spatial correlation.
TABLE 1 binarization Doppler values after Doppler time correlation discrimination
Figure BDA0001370439120000131
Table 2 shows the signal-to-noise ratio and the noise-to-noise ratio after sea clutter suppression in the conventional two-stage filtering process and the two-stage doppler correlation discrimination process of the present invention. In the present example, two simulated nulled targets, in which target A and target B are artificially added.
Table 2 shows the SNR and the SNR of the sea clutter data after sea clutter suppression by the two methods
Figure BDA0001370439120000132
Figure BDA0001370439120000141
As can be seen from the table data, the traditional two-stage AMTI has a certain loss on the target signal-to-noise ratio after cancellation, namely the original 51dB is changed into 43dB, and although the sea clutter is suppressed to a certain degree, the noise-to-noise ratio is still high and is 26.78 dB. The signal-to-noise ratio of the target after the two-stage Doppler correlation discrimination processing is basically not lost and is still about 51dB, the sea clutter suppression effect is obvious, and the noise-to-noise ratio is reduced to 5.69 dB. Therefore, the effectiveness of the sea clutter suppression method by the two-stage Doppler correlation discrimination technology is shown.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (2)

1. A sea clutter suppression method based on two-stage Doppler correlation discrimination is characterized by comprising the following steps:
step 1, obtaining M × N-dimensional time-distance two-dimensional echo signals and obtaining zero-passing frequencyFIR filtered M1× N-dimensional echo signal matrix S, M1M-M +1, M representing the maximum order of the zero-frequency FIR filter, M, M1M and N are positive integers greater than 0 respectively;
the zero-frequency FIR filtered M1× N dimension echo signal matrix S, the process is:
1a) determining radar, setting target in radar detection range, and repeating pulse period TrReceiving a distance dimension continuous pulse echo signal, performing pulse compression processing on the received distance dimension continuous pulse echo signal, and then performing pulse echo signal rearrangement according to a time dimension to obtain a time-distance two-dimensional echo signal, wherein the time dimension of the time-distance two-dimensional echo signal comprises M pulse echo signals, the distance dimension of the time-distance two-dimensional echo signal means that each pulse echo signal comprises N distance units, and M, N is a positive integer greater than 0;
1b) recording the sequence of m pulse echo signal pulses starting with the ith pulse echo signal and the echo signal at the jth distance unit as XijAnd starting m pulse echo signal pulses of the ith pulse echo signal and an echo signal sequence X at the jth distance unitijRemoving the influence of ground clutter on subsequent Doppler discrimination processing through an m-order zero-frequency FIR filter to obtain an ith pulse echo signal and an echo signal S at a jth distance unit after zero-frequency FIR filteringijThe calculation expression is as follows:
Sij＝Xij·WT
wherein, XijRepresenting m continuous pulse echo signal pulses starting as the ith pulse echo signal and the echo signal sequence at the jth distance unit, wherein the m continuous pulse echo signal pulses starting as the ith pulse echo signal are from the ith pulse echo signal to the (i + m-1) th pulse echo signal, and Xij＝[xij,x(i+1)j,...,x(i+m-1)j]，xijRepresenting the i-th pulse echo signal, the echo signal at the j-th range cell, x(i+1)jRepresenting the i +1 th pulse echoSignal, echo signal at jth range cell, x(i+m-1)jRepresents the (i + m-1) th pulse echo signal and the echo signal at the jth range bin, W represents the m-order FIR filter weight coefficient sequence, and W ═ W1,w2,...,wh,...,wm]，whDenotes the h order FIR filter weight coefficients, h ∈ {1,2, …, M }, M denotes the maximum order of the zero frequency FIR filter, and M is a positive integer greater than 0, i ∈ {1,2, …, M1}，j∈{1,2,…,N}，M1The initial values of M-M +1, i and j are 1 respectively;
i＝M1when it is started as M1M pulse echo signal pulses of the pulse echo signal, the echo signal sequence at the jth distance unit is
Figure FDA0002476080390000011
Figure FDA0002476080390000012
Figure FDA0002476080390000013
Denotes the M th1A pulse echo signal, an echo signal at a jth range bin,
Figure FDA0002476080390000014
denotes the M th1+1 pulse echo signal, echo signal at jth range bin, xMjRepresenting the Mth pulse echo signal and the echo signal at the jth distance unit;
1c) let j have constant value and let i have values from 1 to M1And repeatedly executing 1b), and further respectively obtaining the 1 st pulse echo signal after zero-frequency FIR filtering and the echo signal S at the jth distance unit1jTo Mth after zero frequency FIR filtering1Pulse echo signal, echo signal at jth range unit
Figure FDA0002476080390000021
Is recorded as the jth distance after zero-frequency FIR filteringEcho signal S at the off-cellj；
1d) Respectively taking the value of j from 1 to N, setting the value of i to 1, and repeatedly executing 1b) and 1c) to further respectively obtain echo signals S at the 1 st distance unit after zero-frequency FIR filtering1Echo signal S to the Nth distance unit after zero frequency FIR filteringNIs recorded as M after zero frequency FIR filtering1× N-dimensional echo signal matrix S;
step 2, according to the M after zero frequency FIR filtering1× N dimension echo signal matrix S, obtaining M2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000022
Wherein M is2＝M1-1，N1＝N-4，M2、N1Are respectively positive integers greater than 0;
the M is2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000023
Is obtained according to a frequency measurement method at intervals, and the obtaining process comprises the following steps:
2.1 calculating M after zero-frequency FIR Filtering1× N-dimensional echo signal matrix S, i 'th pulse echo signal, j' th Doppler frequency value at distance unit
Figure FDA0002476080390000024
The calculation expression is as follows:
Figure FDA0002476080390000025
wherein, TrIn the form of a pulse repetition period,
Figure FDA0002476080390000026
represents the correlation function between the ith ' pulse echo signal and the ith ' +1 pulse echo signal at the jth ' range bin,
Figure FDA0002476080390000027
i'∈{1,2,…,M2}，M2＝M1-1，j'∈{1,2,…,N1}，N1＝N-4，k∈{j',j'+1,…,j'+4}，Si'krepresents the ith' pulse echo signal after zero frequency FIR filtering, the echo signal at the kth distance unit, S(i'+1)kIndicating an i ' +1 pulse echo signal and an echo signal at a k distance unit after zero-frequency FIR filtering, wherein the initial values of i ' and j ' are respectively 1; tg-1() Representing an arctangent function with a value range of { -pi, pi }; re represents real part operation, Im represents imaginary part operation;
2.2 make j 'constant and make i' take values from 1 to M respectively2And repeatedly executing the steps for 2.1 to further obtain M subjected to zero-frequency FIR filtering respectively1× N-dimensional echo signal matrix S, the value of Doppler frequency at the 1 st pulse echo signal and the jth' th range cell
Figure FDA0002476080390000028
To M after zero frequency FIR filtering1× Mth in N-dimensional echo signal matrix S2Doppler frequency value at jth distance cell of pulse echo signal
Figure FDA0002476080390000029
Is recorded as M after zero frequency FIR filtering1M at jth' range cell in × N-dimensional echo signal matrix S2A Doppler frequency value
Figure FDA0002476080390000031
2.3 let j' take values from 1 to N, respectively1Setting the value of i' to 1, and repeatedly executing 2.1 and 2.2 to obtain M after zero-frequency FIR filtering respectively1M at the 1 st range cell in the × N-dimensional echo signal matrix S2A Doppler frequency value
Figure FDA0002476080390000032
To pass through zero frequency FIR filtered M1× Nth echo signal matrix S1M at distance unit2A Doppler frequency value
Figure FDA0002476080390000033
Is marked as M2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000034
The M is2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000035
In which contains M2×N1A Doppler frequency value;
step 3, according to M2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000036
To obtain M3×N2A dimensional binary matrix; wherein M is3＝M2-1，N2＝N1-M '+1, M' represents the number of range cells contained in the set range cell window, M3、N2And m' are positive integers greater than 0;
the M is3×N2Dimensional binary matrix, is to M2×N1Dimensional Doppler frequency value matrix
Figure FDA0002476080390000037
The Doppler time correlation discrimination binarization is obtained by the following steps:
3.1 setting a distance cell window, and setting the distance cell window to comprise m' distance cells,
Figure FDA0002476080390000038
M3＝M2-1，
Figure FDA0002476080390000039
N2＝N1-m′+1，
Figure FDA00024760803900000310
is set to an initial value of 1,
Figure FDA00024760803900000311
is 0;
3.2 calculation with
Figure FDA00024760803900000312
M' range cells starting from one range cell, the first
Figure FDA00024760803900000313
Doppler frequency sequence of pulse echo signals and
Figure FDA00024760803900000314
correlation coefficient of Doppler frequency sequence of pulse echo signals
Figure FDA00024760803900000315
The expression is as follows:
Figure FDA00024760803900000316
wherein the content of the first and second substances,
Figure FDA00024760803900000317
is expressed as
Figure FDA00024760803900000318
M' range cells starting from one range cell, the first
Figure FDA00024760803900000319
A sequence of doppler frequencies of the individual pulse echo signals,
Figure FDA00024760803900000320
g∈{0,1,…,m'-1}，
Figure FDA00024760803900000321
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure FDA00024760803900000322
A pulse echo signal of
Figure FDA00024760803900000323
The value of the doppler frequency at each range cell,
Figure FDA00024760803900000324
is expressed as
Figure FDA00024760803900000325
M' range cells starting from one range cell, the first
Figure FDA00024760803900000326
A sequence of doppler frequencies of the individual pulse echo signals,
Figure FDA00024760803900000327
Figure FDA00024760803900000328
representing M after zero-frequency FIR filtering1× N dimension echo signal matrix S
Figure FDA00024760803900000329
A pulse echo signal of
Figure FDA00024760803900000330
The value of the doppler frequency at each range cell,
Figure FDA0002476080390000041
is expressed as
Figure FDA0002476080390000042
M' range cells starting from one range cell, the first
Figure FDA0002476080390000043
Doppler frequency sequence of pulse echo signals
Figure FDA0002476080390000044
And in the first place
Figure FDA0002476080390000045
M' range cells starting from one range cell, the first
Figure FDA0002476080390000046
Doppler frequency sequence of pulse echo signals
Figure FDA0002476080390000047
The covariance of (a) of (b),
Figure FDA0002476080390000048
is expressed as
Figure FDA0002476080390000049
M' range cells starting from one range cell, the first
Figure FDA00024760803900000410
Doppler frequency sequence of pulse echo signals
Figure FDA00024760803900000411
The variance of (a) is determined,
Figure FDA00024760803900000412
is expressed as
Figure FDA00024760803900000413
M' range cells starting from one range cell, the first
Figure FDA00024760803900000414
Doppler frequency sequence of pulse echo signals
Figure FDA00024760803900000415
The variance of (a);
3.3 order
Figure FDA00024760803900000416
Is unchanged, and is
Figure FDA00024760803900000417
Are respectively 1 to M3Repeating the execution of 3.2 to obtain the first step
Figure FDA00024760803900000418
Doppler frequency sequence correlation coefficient of m' range cells, 1 st pulse echo signal and 2 nd pulse echo signal starting from one range cell
Figure FDA00024760803900000419
To by
Figure FDA00024760803900000420
M' range units starting from one range unit, Mth3Pulse echo signal and Mth3Doppler frequency sequence correlation coefficient of +1 pulse echo signal
Figure FDA00024760803900000421
Is marked as M2×N1Dimensional Doppler frequency value matrix
Figure FDA00024760803900000422
To middle
Figure FDA00024760803900000423
M at distance unit3Coefficient of correlation
Figure FDA00024760803900000424
3.4 order
Figure FDA00024760803900000425
Respectively take 0 to N2-1, mixing
Figure FDA00024760803900000426
Is set to 1, the execution is repeated for 3.2 and 3.3, and M is obtained2×N1Dimensional Doppler frequency value matrix
Figure FDA00024760803900000427
M at the 1 st distance unit3Individual correlation coefficient ρ1To M2×N1Dimensional Doppler frequency value matrix
Figure FDA00024760803900000428
Middle N2M at distance unit3Coefficient of correlation
Figure FDA00024760803900000429
Is marked as M3×N2A matrix of dimensional correlation coefficients ρ, M3×N2The dimensional correlation coefficient matrix rho contains M3×N2A correlation coefficient of M3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA00024760803900000430
A pulse echo signal of
Figure FDA00024760803900000431
The correlation coefficient at each distance unit is recorded as
Figure FDA00024760803900000432
M3＝M2-1，
Figure FDA00024760803900000433
N2＝N1-m′+1，
Figure FDA00024760803900000434
And
Figure FDA00024760803900000435
the initial values of (a) and (b) are respectively 1;
3.5 pairs of M3×N2And (3) performing Doppler time correlation judgment on the dimensional correlation coefficient matrix rho: setting correlation coefficient threshold value
Figure FDA00024760803900000436
3.6 if M3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA00024760803900000437
A pulse echo signal of
Figure FDA00024760803900000438
Correlation coefficient at distance unit
Figure FDA00024760803900000439
Greater than or equal to the correlation coefficient threshold value
Figure FDA00024760803900000440
Then M3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA00024760803900000441
A pulse echo signal of
Figure FDA00024760803900000442
Correlation coefficient at distance unit
Figure FDA00024760803900000443
Has Doppler time correlation at the position point, and M is calculated3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA00024760803900000444
A pulse echo signal of
Figure FDA00024760803900000445
Correlation coefficient at distance unit
Figure FDA00024760803900000446
Is set to 1 as M3×N2The j 'th element value in the intermediate matrix is maintained, and the initial value of j' is added with 1, j 'to be 1, j' ∈ {1,2, …, M3×N2}；
If M is3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA0002476080390000051
A pulse echo signal of
Figure FDA0002476080390000052
Correlation coefficient at distance unit
Figure FDA0002476080390000053
Less than a correlation coefficient threshold value
Figure FDA0002476080390000054
Then M3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA0002476080390000055
A pulse echo signal of
Figure FDA0002476080390000056
Correlation coefficient at distance unit
Figure FDA0002476080390000057
Has no Doppler time correlation at the position point and M is calculated3×N2The first in the dimensional correlation coefficient matrix rho
Figure FDA0002476080390000058
A pulse echo signal of
Figure FDA0002476080390000059
Correlation coefficient at distance unit
Figure FDA00024760803900000510
Is set to 0 as M3×N2The j 'th element value in the intermediate matrix is maintained, and the initial value of the added value 1, j' of j 'is 1, j' ∈ {1,2, …, M3×N2}；
3.7 order
Figure FDA00024760803900000511
Is unchanged, and is
Figure FDA00024760803900000512
Are respectively 1 to M3Repeating the execution for 3.6 to obtain M3×N2Value of the 1 st element to M in the dimension intermediate matrix3×N2M in the dimension intermediate matrix3Value of element, denoted as M3×N2Dimension of the intermediate matrix
Figure FDA00024760803900000513
M at distance unit3A value of an element;
3.8 order
Figure FDA00024760803900000514
Respectively taking 1 to N2Will be
Figure FDA00024760803900000515
Is set to 1, the execution is repeated for 3.6 and 3.7, and M is obtained3×N2M at the 1 st distance cell in the dimension intermediate matrix3Value of element to M3×N2Nth in dimension intermediate matrix2M at distance unit3Value of element, denoted as M3×N2Dimensional binary matrix, M3×N2Correspondence in the dimensional binary matrix includes M3×N2A Doppler value;
step 4, according to M3×N2A dimensional binary matrix, and a time-distance two-dimensional Doppler value matrix f with the size of M after two-stage Doppler correlation discrimination processing is obtained3×N2Maintaining;
the time-distance two-dimensional Doppler value matrix f after the two-stage Doppler correlation discrimination processing is a pair of M3×N2The dimensional binary matrix is obtained by Doppler space correlation discrimination and reset, and the process is as follows:
4.1 initialization, namely setting a target window length, wherein the target window length is u pulse echo signals, and u is a positive integer greater than 0, and setting the initial value of b ∈ {0,1, …, u }, and the initial value of b is 0;
setting M3×N2The ith pulse echo signal and the Doppler values at 3 continuous distance units in the dimensional binary matrix are respectively
Figure FDA00024760803900000516
Wherein
Figure FDA00024760803900000517
Represents M3×N2The ith pulse echo signal and the jth' corresponding Doppler value at the distance unit in the dimensional binary matrix,
Figure FDA00024760803900000518
represents M3×N2The ith 'pulse echo signal, the jth' ″ -1 in the dimensional binary matrixThe corresponding doppler value at the range cell,
Figure FDA00024760803900000519
represents M3×N2When the ith pulse echo signal in the dimensional binary matrix, the corresponding Doppler value at the jth '″ +1 distance unit and j' ″ -1 ═ 0
Figure FDA0002476080390000061
j”'+1>N2Time of flight
Figure FDA0002476080390000062
i”∈{1,2,…,M3}，M3＝M2-1，j”'∈{1,2,…,N2}，N2＝N1The initial values of-m '+1, i "and j'" are respectively 1;
4.2 calculating Doppler F at j 'th distance unit when the number of pulse echo signals in the target window length is i' + bi”+b,j”'The expression is as follows:
Figure FDA0002476080390000063
wherein the content of the first and second substances,
Figure FDA0002476080390000064
represents M3×N2The ith '+ b pulse echo signal, the jth' ″ -1 Doppler value at the range bin in the dimensional binary matrix,
Figure FDA0002476080390000065
represents M3×N2The ith '+ b pulse echo signal, the jth' ″ Doppler value at the range bin in the dimensional binary matrix,
Figure FDA0002476080390000066
represents M3×N2The pulse echo signal of the ith '+ b and the Doppler value at the distance unit of the jth' +1 in the dimensional binary matrix;
4.3 Doppler F at j 'th distance unit if number of pulse echo signals in target window length is i' + bi”+b,j”'Not equal to 0, then
Figure FDA0002476080390000067
Not all 0, i.e. at least one Doppler value of 1, then M3×N2In the dimension binary matrix, the ith '+ b pulse echo signal and the jth' ″ effective target main lobe where the distance unit is located are executed for 4.4; otherwise, abandon M3×N2In the dimension binary matrix, the ith ' + b pulse echo signal and the jth ' ″ distance unit add 1 to the value of j ' ″ and return to 4.2;
4.4 let b take 0 to u respectively, and repeat 4.2 and 4.3 until obtaining the Doppler F at the j 'th distance unit when the number of pulse echo signals in the target window length is i' +0i”+0,j”'Doppler F at the j 'th distance unit when the number of pulse echo signals in the length of a target window is i' + ui”+u,j”'And if so, if satisfied
Fi”+0,j”'·Fi”+1,j”'·...·Fi”+u,j”'≠0
I.e. Fi”+0,j”',Fi”+1,j”',...,Fi”+u,j”'When all the signals are not 0, the Doppler F at the j 'th distance unit when the number of pulse echo signals in the target window length is i' +0i”+0,j”'Doppler F at the j 'th distance unit when the number of pulse echo signals in the length of a target window is i' + ui”+u,j”'The effective target main lobe which is positioned in u +1 continuous pulse echo signals is recorded as the p-th continuous quasi-target main lobe sequence, the initial value of p is 1, the value of p is added with 1, the value of i 'is not changed, the value of j' is added with 1, and 4.2 is returned; otherwise, the value of i 'is not changed, the value of j' is added with 1, and 4.2 is returned;
if j'>N2Adding 1 to the value of i ', setting the value of j' to 1, and returning to 4.2;
until the P continuous quasi-target main lobe sequence is obtained, P<M3×N2And simulating the 1 st continuous image obtained at this timeThe sequence from the target main lobe sequence to the P-th continuous quasi-target main lobe sequence is marked as M3×N2P continuous quasi-target main lobe sequences in the dimension binary matrix; if P is equal to 0, no target exists in the radar detection range, the targets are clutter areas, the clutter areas are recorded as clutter areas in the time-distance two-dimensional echo signals, and 4.7 is executed; if 0<P<M3×N2Initializing p to 1, executing 4.5;
4.5 according to M3×N2Determining P target areas in the time-distance two-dimensional echo signals and clutter areas in the time-distance two-dimensional echo signals by using P continuous quasi-target main lobe sequences in the dimensional binary matrix;
4.5.1 respectively expanding 2 distance units or 2 pulse echo signals to the periphery by taking the p continuous quasi-target main lobe sequence as a center to form a rectangular area block to obtain a p continuous quasi-target main lobe sequence protection area;
4.5.2, enabling P to respectively take 1 to P, and repeatedly executing 4.5.1 until a P continuous quasi-target main lobe sequence protection area is obtained, at the moment, obtaining a 1 st continuous quasi-target main lobe sequence protection area to a P continuous quasi-target main lobe sequence protection area, wherein each continuous quasi-target main lobe sequence protection area is respectively a (u +4) × 5 dimensional two-dimensional matrix;
4.5.3 recording the corresponding regions from the 1 st continuous quasi-target main lobe sequence protection region to the P th continuous quasi-target main lobe sequence protection region in the time-distance two-dimensional echo signal as P target regions in the time-distance two-dimensional echo signal, and recording the rest regions except the P target regions in the time-distance two-dimensional echo signal as clutter regions in the time-distance two-dimensional echo signal;
4.6 setting the Doppler value at each position point in the P target areas in the time-distance two-dimensional echo signal to be 1, setting the Doppler value at each position point in the clutter area in the time-distance two-dimensional echo signal to be 0, and recording the Doppler values as a time-distance two-dimensional Doppler value matrix f after two-stage Doppler correlation judgment processing, wherein the size of the time-distance two-dimensional Doppler value matrix f is M3×N2Maintaining;
4.7 Doppler values at each location point in the clutter region in the time-distance two-dimensional echo signalRespectively set as 0, and recorded as a time-distance two-dimensional Doppler value matrix f after two-stage Doppler correlation discrimination processing, with the size of M3×N2Maintaining;
and 5, obtaining a sea clutter suppression result based on the two-stage Doppler correlation discrimination according to the time-distance two-dimensional Doppler value matrix f subjected to the two-stage Doppler correlation discrimination processing.
2. The method for suppressing sea clutter based on two-stage doppler correlation decision as claimed in claim 1, wherein the sub-step of step 5 is:
5.1 recording the Doppler values of the ith 'pulse echo signal and the jth' distance unit in the time-distance two-dimensional Doppler value matrix f after the two-stage Doppler correlation judgment processing as fi”'j””，
i”'∈{1,2,…,M3}，M3＝M2-1，j””∈{1,2,…,N2}，N2＝N1-m '+1, i' "and j" "have initial values of 1, respectively;
5.2 if the time-distance two-dimensional Doppler value matrix f is subjected to two-stage Doppler correlation judgment processing, the ith 'pulse echo signal and the jth' Doppler value f at the distance unit are obtainedi”'j””Is 1, then M after zero frequency FIR filtering1The ith 'pulse echo signal and the jth' echo signal in the × N-dimensional echo signal matrix S are kept unchanged and are recorded as M after zero-frequency FIR filtering1× N dimension echo signal matrix S the ith '″ pulse echo signal, the jth' ″ reset echo signal at the distance unit;
if the ith 'pulse echo signal and the jth' Doppler value f at the distance unit in the time-distance two-dimensional Doppler value matrix f are subjected to the two-stage Doppler correlation judgment processingi”'j””If 0, let M after zero-frequency FIR filtering1The phase of the ith 'pulse echo signal and the phase of the echo signal at the jth' distance unit in the × N-dimensional echo signal matrix S are unchanged, the amplitude is normalized to be the average noise level, and the average noise level is recordedFor M after zero-frequency FIR filtering1× N dimension echo signal matrix S the ith '″ pulse echo signal, the jth' ″ reset echo signal at the distance unit;
5.3 make j "" constant, and make i' "take values from 1 to M, respectively3And repeatedly executing the steps of 5.2 to further obtain M subjected to zero-frequency FIR filtering respectively1× N-dimensional echo signal matrix S from pulse echo signal at 1 st position, reset signal at jth' distance unit to M after zero-frequency FIR filtering1× Mth in N-dimensional echo signal matrix S3The pulse echo signal and the reset echo signal at the jth distance unit are recorded as M after zero frequency FIR filtering1M in × N-dimensional echo signal matrix S3Pulse echo signals, reset echo signals at jth "" distance units;
5.4 setting the value of i 'to 1, and respectively taking the values of j' from 1 to N2Repeatedly executing 5.2 and 5.3 to respectively obtain M3Pulse echo signal, reset echo signal at 1 st range unit to M3Pulse echo signal, Nth2Reset echo signal at one range cell, denoted M3Pulse echo signal, N2Reset echo signal at range unit
Figure FDA0002476080390000081
M after zero frequency FIR filtering1Removing M from × N-dimensional echo signal matrix S3Pulse echo signal, N2Reset echo signal at range unit
Figure FDA0002476080390000082
All echo signals except for
Figure FDA0002476080390000083
Further obtaining a sea clutter suppression result based on two-stage Doppler correlation discrimination
Figure FDA0002476080390000084
Figure FDA0002476080390000085
And the average noise level is the average level value of the residual noise signal after P target areas and clutter are removed from the time-distance two-dimensional echo signal.
CN201710660474.2A 2017-08-04 2017-08-04 Sea clutter suppression method based on two-stage Doppler correlation discrimination Active CN107607917B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710660474.2A CN107607917B (en) 2017-08-04 2017-08-04 Sea clutter suppression method based on two-stage Doppler correlation discrimination

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710660474.2A CN107607917B (en) 2017-08-04 2017-08-04 Sea clutter suppression method based on two-stage Doppler correlation discrimination

Publications (2)

Publication Number Publication Date
CN107607917A CN107607917A (en) 2018-01-19
CN107607917B true CN107607917B (en) 2020-07-03

Family

ID=61064826

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710660474.2A Active CN107607917B (en) 2017-08-04 2017-08-04 Sea clutter suppression method based on two-stage Doppler correlation discrimination

Country Status (1)

Country Link
CN (1) CN107607917B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104914429A (en) * 2015-05-19 2015-09-16 西安电子科技大学 Target indication radar system capable of adaptively selecting waveform according to target distance
CN105162556A (en) * 2015-08-19 2015-12-16 南京邮电大学 Large-scale MIMO system channel feedback method based on spatial- temporal correlation

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104007423B (en) * 2014-05-27 2016-09-14 电子科技大学 Folded Clutter in Skywave Radars ocean clutter cancellation method based on chaos sequence prediction
CN104155632A (en) * 2014-07-18 2014-11-19 南京航空航天大学 Improved subspace sea clutter suppression method based on local correlation
CN106443633A (en) * 2016-12-01 2017-02-22 哈尔滨工业大学 Shipborne high frequency ground wave radar sea clutter time domain suppression method

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104914429A (en) * 2015-05-19 2015-09-16 西安电子科技大学 Target indication radar system capable of adaptively selecting waveform according to target distance
CN105162556A (en) * 2015-08-19 2015-12-16 南京邮电大学 Large-scale MIMO system channel feedback method based on spatial- temporal correlation

Also Published As

Publication number Publication date
CN107607917A (en) 2018-01-19

Similar Documents

Publication Publication Date Title
CN104569948B (en) Sub-band adaptive GLRT LTD detection methods under sea clutter background
Yang et al. Dense people counting using IR-UWB radar with a hybrid feature extraction method
Rosenberg Sea-spike detection in high grazing angle X-band sea-clutter
CN106125053B (en) Pulse Doppler radar polarization anti jamming method
Duk et al. Target detection in sea-clutter using stationary wavelet transforms
CN113534120B (en) Multi-target constant false alarm rate detection method based on deep neural network
CN111722199A (en) Radar signal detection method based on convolutional neural network
AU2008327744A1 (en) Real-time robust method for determining the trajectory of one or more cetaceans by means of passive acoustics, using a laptop computer
Carrer et al. Automatic enhancement and detection of layering in radar sounder data based on a local scale hidden Markov model and the Viterbi algorithm
CN111580091B (en) Sea surface weak target detection method based on AR spectrum singular intensity function
CN107167802A (en) A kind of breath signal detection algorithm based on ULTRA-WIDEBAND RADAR
CN109633633A (en) A kind of life signal Enhancement Method based on segmentation classification enhancing processing
CN108872961B (en) Radar weak target detection method based on low threshold
Hu et al. A new way to model nonstationary sea clutter
Kerbaa et al. Model selection of sea clutter using cross validation method
Cotter et al. Classification of broadband target spectra in the mesopelagic using physics-informed machine learning
CN112255607A (en) Sea clutter suppression method
CN114545387A (en) High-altitude parabolic detection and discrimination method based on millimeter wave radar data fitting
CN113608193A (en) Radar multi-target distance and speed estimation method based on UNet
CN107607917B (en) Sea clutter suppression method based on two-stage Doppler correlation discrimination
CN106569190B (en) Sea-surface target detection device and its detection method under a kind of high sea situation
CN106324576B (en) Radar target adaptive matched filter method based on autoregression model
CN112162269B (en) Sea clutter suppression and target detection method based on singular spectrum decomposition
CN105548986B (en) Object detection method based on prewhitening median ratio detector under sea clutter background
CN113759335A (en) Ultra-wideband echo denoising and energy detection method based on time reversal

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant