CN104569950A - Sea-surface target clustering method based on relative power and phase linearity - Google Patents

Abstract

The invention belongs to the technical field of radar target detection, and particularly relates to a sea-surface target clustering method based on relative power and phase linearity. The method includes the steps that target detection is conducted on a three-dimensional data array X received by radar, and a distance-orientation matrix D is obtained; a real phase sequence of M pulses of all distance-orientation resolution cells of the data array X is computed; the phase linearity and power value of all the distance-orientation resolution cells of which the element is 1 in the matrix D are computed; the elements '1' in the matrix D are divided into different communication regions according to eight-neighborhood connectivity; the mean Ej of the power value of all the distance-orientation resolution cells in the jth communication region and the mean Lj of the reciprocal of the phase linearity; the sum zj, r of the relative power and the reciprocal of the phase linearity of the rth distance-orientation resolution cell in the jth communication region is computed; the distance-orientation resolution cell corresponding to the maximum value of zj, r serves as the target position clustered by the jth communication region.

Description

Based on the sea-surface target condensing method of relative power and phase linearity

Technical field

The invention belongs to technical field of radar target acquisition, particularly based on the sea-surface target condensing method of relative power and phase linearity, for the cohesion of object detection results and the extraction of size supplementary under the sea clutter background of airborne or carrier-borne scanning radar.

Background technology

To the many employings of the airborne or shipborne radar low repetition beam scanning mode of operation that maritime patrol is looked, each ripple position residence time is short, umber of pulse is few.Even if under low resolution pattern (distance and bearing), a lot of sea-surface target (as large vessel and islands and reefs etc.) is leaked due to size is large or the strong echo of target causes beam side lobe and is made target show as range-azimuth distributed object.That is, the backward energy of these radar targets can be diffused among contiguous ripple position or range unit, causes and all may will detect target in neighbor distance unit and ripple position.Simply each position can not be detected in practical application that the range-azimuth of target is used as an independent target and is processed, need to condense object element, the exact position of reporting objectives and auxiliary dimension information (echo occupies the number of resolution element).

Current most of condensing methods condense according to the proximity principle on resolution element.This condensing method is accurate not on target location accuracy, and the middle-size and small-size naval vessel near islands and reefs, greatly ship is easily lost in coacervation process.

Summary of the invention

The object of the invention is to the sea-surface target condensing method proposed based on relative power and phase linearity, realize the cohesion of airborne or carrier-borne scanning radar object detection results under sea clutter background and the extraction of size supplementary, improve the accuracy of object detection results.

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

Sea-surface target condensing method based on relative power and phase linearity comprises the following steps:

Step 1, utilize radar emission continuous print pulse signal, radar is utilized to receive echo three-dimensional data array X, echo data array X is the array of a M × K × W dimension, wherein, M represents the pile-up pulse number of echo data array X, and K represents the range unit number of echo data array X, and W represents the ripple figure place of echo data array X; In echo data array X, the element representation of a m pulse kth range unit w ripple position is X (m, k, w), m=1,2 ..., M; K=1,2 ..., K; W=1,2 ..., W;

Step 2, carries out target detection to echo data array X, draws range-azimuth matrix D, each element of range-azimuth matrix D is 0 or 1, range-azimuth matrix D is a size is the matrix of K × W, and range-azimuth matrix D row k is D (k, w) to the element representation that w arranges; If the element D (k that range-azimuth matrix D row k arranges w, w) be 0, then illustrate that kth range unit w ripple position of radar does not have target, if the element D (k that range-azimuth matrix D row k arranges w, w) be 1, then illustrate that there is target kth range unit w ripple position of radar;

Step 3, the two-dimensional position coordinate that any one range unit and any one ripple position form is designated as a range-azimuth resolution element, then the range-azimuth resolution element of radar has K × W, draws the true phase sequence q of radar i-th range-azimuth resolution element M pulse _i, i=1,2 ..., K × W;

Step 4, show that in range-azimuth matrix D, element is the phase linearity of each range-azimuth resolution element of 1;

Step 5, show that in range-azimuth matrix D, element is the performance number of each range-azimuth resolution element of 1;

Step 6, the element being 1 by numerical value in range-azimuth matrix D is divided into multiple connected region according to eight neighborhood connectedness, and the number dividing the connected region drawn is expressed as N;

Step 7, calculates the average E of the performance number of a jth connected region all range-azimuths resolution element _j, and the average L of inverse of linear phase degree of a jth connected region all range-azimuths resolution element _j, j=1,2 ..., N;

Step 8, draw the relative power of a jth connected region r range-azimuth resolution element and opposite linear phase place degree inverse and z _j,r, z _j,r=e _r/ E _j+ 1/ (L _jδ _r), e _rrepresent the performance number of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, δ _rrepresent the linear phase degree of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, r=1,2 ..., S _j, S _jrepresent the number of the range-azimuth resolution element of a jth connected region in range-azimuth matrix D; By the relative power of a jth connected region each range-azimuth resolution element and opposite linear phase place degree inverse and maximal value be designated as: a jth connected region statistic maximal value; Using a jth target location that the range-azimuth resolution element that connected region statistic maximal value is corresponding is condensed into as a jth connected region.

Beneficial effect of the present invention is: in the present invention, combine the relative power strength that make use of and adjacency-azimuth discrimination unit detected and the resolution element phase of echo linearity (the radial motion information representation under low Doppler's resolution condition) determines subject goal position, realize the cohesion to Extended target under sea clutter background, noise jamming can be effectively reduced, distinguish adjacent target, improve accuracy destination number being detected.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the sea-surface target condensing method based on relative power and phase linearity of the present invention.

Embodiment

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

Current most of condensing methods condense according to the proximity principle on resolution element.This condensing method is accurate not on target location accuracy, and the middle-size and small-size naval vessel near islands and reefs, greatly ship is easily lost in coacervation process.Its main cause of losing is that these middle-size and small-size naval vessels need to rely on doppler information to distinguish with contiguous islands and reefs and large ship.In the present invention, we combine the joint objective agglomerative algorithm of relative power strength and the resolution element phase of echo linearity (the radial motion information representation under low Doppler's resolution condition) that make use of and adjacency-azimuth discrimination unit detected.This algorithm improves in nearly step of the simple condensing method based on phase linearity of application in early stage.

With reference to Fig. 1, it is the process flow diagram of the sea-surface target condensing method based on relative power and phase linearity of the present invention.Should comprise the following steps based on the sea-surface target condensing method of relative power and phase linearity:

Step 1, utilize radar emission continuous print pulse signal, radar is utilized to receive echo three-dimensional data array X (range-azimuth-pulse), echo data array X is the array of a M × K × W dimension, wherein, M represents the pile-up pulse number of echo data array X, and K represents the range unit number of echo data array X, and W represents the ripple figure place of echo data array X;

In the embodiment of the present invention, in echo data array X, the element representation of a m pulse kth range unit w ripple position is X (m, k, w), m=1,2 ..., M; K=1,2 ..., K; W=1,2 ..., W.

In the embodiment of the present invention, the two-dimensional position coordinate that arbitrary range unit of radar and arbitrary ripple position form is designated as a range-azimuth resolution element of radar.

Step 2, according to radar target detection method, target detection is carried out to echo data array X, draw range-azimuth matrix D, each element of range-azimuth matrix D is 0 or 1, range-azimuth matrix D is a size is the matrix of K × W, range-azimuth matrix D row k is D (k, w) to the element representation that w arranges; If the element D (k that range-azimuth matrix D row k arranges w, w) be 0, then illustrate that kth range unit w ripple position of radar does not have target, if the element D (k that range-azimuth matrix D row k arranges w, w) be 1, then illustrate that there is target kth range unit w ripple position of radar.

In step 2, radar target detection method is existing radar target detection method, is not described in detail in this.

Step 3, the two-dimensional position coordinate that any one range unit and any one ripple position form is designated as a range-azimuth resolution element, then the range-azimuth resolution element of radar has K × W, draws the true phase sequence q of radar i-th range-azimuth resolution element M pulse _i, i=1,2 ..., K × W.

Step 3 specifically comprises following sub-step:

(3.1) the two-dimensional position coordinate that any one range unit and any one ripple position form is designated as a range-azimuth resolution element, then the range-azimuth resolution element of radar has K × W, calculates winding phase place WrapPhase (m) of radar i-th range-azimuth resolution element m pulse, WrapPhase (m)=arg (X (m, k, w)), m=1,2 ..., M, i=1,2 ..., K × W.

(3.2) to true phase initialize, make true phase UnWrapPhase (the 1)=WrapPhase (1) of radar i-th range-azimuth resolution element the 1st pulse, make the phase difference (2) of the true phase of the winding phase place of radar i-th range-azimuth resolution element the 2nd pulse and the 1st pulse of radar i-th range-azimuth resolution element be:

Δ(2)＝WrapPhase(2)-UnWrapPhase(1)。

(3.3) if abs (Δ (2))≤π, then make UnWrapPhase (2)=UnWrapPhase (1)+Δ (2), then skip to step (3.5), abs () expression takes absolute value; If abs (Δ (2)) > π, then skip to sub-step (3.4).

(3.4) if Δ (2) > π, then making the value of Δ (2) from subtracting 2 π, being then back to sub-step (3.3); If Δ (2) <-π, then making the value of Δ (2) from increasing 2 π, being then back to sub-step (3.3).

(3.5) if the pile-up pulse number M of echo data array X is 2, then skip to sub-step (3.12), if the pile-up pulse number M of echo data array X is greater than (2), then skip to sub-step (3.6).

(3.6) n=2 is made, 3 ..., M; Work as n=3,4 ..., during M, make the phase difference (n) of the true phase of the winding phase place of radar i-th range-azimuth resolution element n-th pulse and (n-1)th pulse of radar i-th range-azimuth resolution element be:

Δ(n)＝WrapPhase(n)-UnWrapPhase(n-1)。

(3.7) if abs (Δ (n))≤π, then sub-step (3.8) is skipped to; If abs (Δ (n)) > is π, then skip to sub-step (3.10).

(3.8) radar echo signal contains noise, when phase differential is close to π, phase ambiguity may occur, and needs to do further judgement.Now judge Δ (n) and Δ (2) whether contrary sign (positive number, a negative), if Δ (n) and Δ (2) jack per line, then skip to sub-step (3.11); If Δ (n) and Δ (2) contrary sign, then skip to sub-step (3.9).

(3.9) make λ=1, if Δ (n) > π is-λ, then makes the value of Δ (n) from subtracting 2 π, then skipping to sub-step (3.11); If Δ (n) <-π is+λ, then makes the value of Δ (n) from increasing 2 π, then skipping to sub-step (3.11);

(3.10) if Δ (n) > is π, then making the value of Δ (n) from subtracting 2 π, being then back to sub-step (3.7); If Δ (n) <-is π, then makes the value of Δ (n) from increasing 2 π, being then back to sub-step (3.7).

(3.11) UnWrapPhase (n) is drawn,

UnWrapPhase(n)＝UnWrapPhase(n-1)+Δ(n)；

Then sub-step (3.12) is skipped to.

(3.12) the true phase sequence q of radar i-th range-azimuth resolution element M pulse is drawn _i, the true phase sequence q of radar i-th range-azimuth resolution element M pulse _ifor:

q _i＝[UnWrapPhase(1),UnWrapPhase(2),…,UnWrapPhase(M)]。

Step 4, show that in range-azimuth matrix D, element is the phase linearity of each range-azimuth resolution element of 1.

Specifically, if the element of i-th range-azimuth resolution element is 1 in range-azimuth matrix D, then the linear phase degree δ of radar i-th range-azimuth resolution element _ifor:

${\mathrm{\&delta;}}_i=\frac{1}{M-1}\underset{n=2}{\overset{M}{\mathrm{\&Sigma;}}}{(\mathrm{\&Delta;}\left(n\right)-\stackrel{\&OverBar;}{k})}^2$

$\stackrel{\&OverBar;}{k}=\frac{1}{M-1}\underset{n=2}{\overset{M}{\mathrm{\&Sigma;}}}\mathrm{\&Delta;}\left(n\right)$

Wherein, n=2,3 ... M, Δ (n)=WrapPhase (n)-UnWrapPhase (n-1), i=1,2 ..., K × W, WrapPhase (n) represents the winding phase place of radar i-th range-azimuth resolution element n-th pulse, and UnWrapPhase (n-1) represents the true phase (after unwrapping phase place) of radar i-th range-azimuth resolution element (n-1)th pulse; represent the mean value of the phase differential of radar i-th range-azimuth resolution element, δ _irepresent the standard deviation of the phase differential of i-th range-azimuth resolution element in radar.

Step 5, show that in range-azimuth matrix D, element is the performance number of each range-azimuth resolution element of 1.

Specifically, if the element of i-th range-azimuth resolution element is 1 in range-azimuth matrix D, then the performance number e of radar i-th range-azimuth resolution element _ifor:

$e_i=\frac{1}{M}\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{X}_{m,i}^2;$

Wherein, i=1,2 ..., K × W, m=1,2 ..., M, X _m,irepresent the element of i-th range-azimuth resolution element m pulse in echo data array X.

Step 6, the element being 1 by numerical value in range-azimuth matrix D is divided into multiple connected region according to eight neighborhood connectedness, and the number dividing the connected region drawn is expressed as N.

The concrete sub-step of step 6 is:

(6.1) choose in range-azimuth matrix D any one element be the range-azimuth resolution element of 1 as initial resolution element, draw the neighborhood resolution element of initial resolution element.In range-azimuth matrix D, the neighborhood resolution element of each resolution element comprises the range-azimuth resolution element of the adjacent range-azimuth resolution element of the level of corresponding resolution element, the vertically adjacent range-azimuth resolution element of corresponding resolution element and the monoclinic phase neighbour of corresponding resolution element; Can find out, when on four angles that range-azimuth matrix D middle distance-azimuth discrimination unit is positioned at range-azimuth matrix D, the neighborhood resolution element of corresponding resolution element has 3; When resolution element in range-azimuth matrix D is positioned at the non-boundary of range-azimuth matrix D, the neighborhood resolution element of corresponding resolution element has 8, and now the usual neighborhood resolution element by corresponding resolution element is designated as the eight neighborhood of corresponding resolution element.

(6.2) judge whether the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element (namely whether there is target) that element is 1; If there is not the range-azimuth resolution element (i.e. there is not target in the neighborhood resolution element of initial resolution element) that element is 1 in the neighborhood resolution element of initial resolution element, then using initial resolution element as a connected region, skip to sub-step (6.5); If the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that at least one element is 1, then skip to sub-step (6.3);

(6.3) according to step 3, the mean value k of the phase differential of initial resolution element is calculated ₀, Δ ₀n () represents the end value of the phase differential of the winding phase place of the n-th pulse of initial resolution element and the true phase of unit's (n-1)th pulse of the initial resolution element of radar in the radar drawn according to step 3.

According to step 3, calculating each element in the neighborhood resolution element of initial resolution element is the mean value of the phase differential of the range-azimuth resolution element of 1, the mean value being the phase differential of the range-azimuth resolution element of 1 by any one element in the neighborhood resolution element of initial resolution element is expressed as k' wherein, Δ ' (n) represent the end value of the phase differential of the true phase of (n-1)th pulse of respective distances-azimuth discrimination unit in the winding phase place of the n-th pulse of respective distances-azimuth discrimination unit in the neighborhood resolution element of the initial resolution element of radar drawn according to step 3 and the neighborhood resolution element of the initial resolution element of radar.

If abs is (atan (k ₀)-atan (k'))>=β, then the element of respective distances-azimuth discrimination unit in the neighborhood resolution element of initial resolution element is got 0 (making this element be 0), wherein, atan () represents negate tangent, abs () expression takes absolute value, β is the positive number of a setting, and such as, β is taken as if abs is (atan (k ₀)-atan (k')) < β, then respective distances-azimuth discrimination unit and initial resolution element in the neighborhood resolution element of initial resolution element are divided into same connected region, in all range-azimuths resolution element of range-azimuth matrix D, remove the range-azimuth resolution element being divided into connected region, obtain the unallocated range-azimuth resolution element to connected region of range-azimuth matrix D; Using respective distances-azimuth discrimination unit in the neighborhood resolution element of resolution element initial in range-azimuth matrix D as initial resolution element, unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, draw the neighborhood resolution element of initial resolution element, skip to sub-step (6.4).

In sub-step (6.3), draw the process of the neighborhood resolution element of initial resolution element and in sub-step (6.1), show that the process of the neighborhood resolution element of initial resolution element is identical, no longer repeats at this.

(6.4) judge whether the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element (namely whether there is target) that element is 1; If the neighborhood resolution element of initial resolution element does not exist the range-azimuth resolution element (i.e. there is not target in the neighborhood resolution element of initial resolution element) that element is 1, then skip to sub-step (6.5); If the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that at least one element is 1, be then back to sub-step (6.3);

(6.5) unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, judge whether that there is element is that the range-azimuth resolution element of 1 is as initial resolution element, if there is no element is that the range-azimuth resolution element of 1 is as initial resolution element, the process then dividing connected region is complete, and record is divided into each connected region range-azimuth resolution element number; Otherwise, if there is element is that the range-azimuth resolution element of 1 is as initial resolution element, unallocated in the range-azimuth resolution element of connected region then in range-azimuth matrix D, choosing any one element is that the range-azimuth resolution element of 1 is as initial resolution element, unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, draw the neighborhood resolution element of initial resolution element, be back to sub-step (6.2).The number dividing the connected region drawn is expressed as N.

Step 7, calculates the average E of the performance number of a jth connected region all range-azimuths resolution element _j, and the average L of inverse of linear phase degree of a jth connected region all range-azimuths resolution element _j, j=1,2 ..., N.

Particularly, in step 7, the average E of the performance number of a jth connected region all range-azimuths resolution element _jfor:

$E_j=\frac{1}{S_j}\underset{r=1}{\overset{S_j}{\mathrm{\&Sigma;}}}{e}_r$

Wherein, e _rrepresent the performance number of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, r=1,2 ..., S _j, S _jrepresent the number of the range-azimuth resolution element of a jth connected region in range-azimuth matrix D.

The average L of the inverse of the linear phase degree of a jth connected region all range-azimuths resolution element _jfor:

$L_j=\frac{1}{S_j}\underset{r=1}{\overset{S_j}{\mathrm{\&Sigma;}}}\frac{1}{{\mathrm{\&delta;}}_r}$

Wherein, δ _rrepresent the linear phase degree of a jth connected region r range-azimuth resolution element in range-azimuth matrix D.

Step 8, draw the relative power of a jth connected region r range-azimuth resolution element and opposite linear phase place degree inverse and z _j,r, z _j,r=e _r/ E _j+ 1/ (L _jδ _r); By the relative power of a jth connected region each range-azimuth resolution element and opposite linear phase place degree inverse and maximal value be designated as: a jth connected region statistic maximal value; Using a jth target location that the range-azimuth resolution element that connected region statistic maximal value is corresponding is condensed into as a jth connected region, store the range-azimuth resolution element number being divided into a jth connected region in the target location that a jth connected region is condensed into.

Based on step 1 to step 8, achieve a kind of sea-surface target condensing method based on relative power and phase linearity, realize the extraction of airborne or carrier-borne scanning radar to the cohesion of Extended target under sea clutter background and size supplementary, improve accuracy destination number being detected.

Below in conjunction with emulation experiment, effect of the present invention is described further.

1) simulation parameter

Employ Observed sea clutter and testing result thereof in emulation experiment, done three groups of experiments, range unit number is 1000, and ripple figure place is 100, and umber of pulse is 6.

2) emulation experiment content

By relatively assessing sea-surface target cohesion result in experiment.The umber of pulse N=6 of each resolution element, forms 100000 resolution elements, adopts the present invention's (relative power and phase linearity method), power method and phase linearity method to condense testing result respectively.Wherein, the process of phase linearity method is: first perform step 1 of the present invention to step 6 (step 5 can be omitted), can obtain the linear phase degree of each target in each connected region; Then the body position of the minimum target of each connected region internal linear phase place degree as target after the cohesion of this connected region is chosen.

With reference to table 1, it is the sea-surface target cohesion comparative result that three experiments adopt several method to draw to measured data.As can be seen from Table 1, the present invention is different from the cohesion result that power method and phase linearity method draw respectively.Often in group experiment Preliminary detection to go out the number number of target all identical, often in group experiment, the result of three kinds of method cohesions is all identical, but, it is all variant that power method and cohesion result of the present invention often organize experiment, phase linearity method and the present invention condense result often to organize experiment all variant, illustrate that power and phase linearity all impact sea-surface target cohesion.Only can neglect the doppler information of target according to power, real target energy maximum can be reacted but cannot ensure that it is best congealing point, and real adjacent target cannot be distinguished, as the middle-size and small-size naval vessel near islands and reefs, large ship; Only then thoroughly have ignored according to linear phase degree the fact that real target somewhere has strong energy echo, so should the competitive relation of comprehensive two kinds of factors, meet the minimum and maximum principle of energy of linear phase degree to determine sea-surface target main body position.The experimental result of measured data shows, the present invention has considered power and linear phase degree, has more accurate sea-surface target cohesion result.

Table 1

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

Claims (6)

1., based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, comprise the following steps:

Step 1, utilize radar emission continuous print pulse signal, radar is utilized to receive echo three-dimensional data array X, echo data array X is the array of a M × K × W dimension, wherein, M represents the pile-up pulse number of echo data array X, and K represents the range unit number of echo data array X, and W represents the ripple figure place of echo data array X; In echo data array X, the element representation of a m pulse kth range unit w ripple position is X (m, k, w), m=1,2 ..., M; K=1,2 ..., K; W=1,2 ..., W;

Step 2, carries out target detection to echo data array X, draws range-azimuth matrix D, each element of range-azimuth matrix D is 0 or 1, range-azimuth matrix D is a size is the matrix of K × W, and range-azimuth matrix D row k is D (k, w) to the element representation that w arranges; If the element D (k that range-azimuth matrix D row k arranges w, w) be 0, then illustrate that kth range unit w ripple position of radar does not have target, if the element D (k that range-azimuth matrix D row k arranges w, w) be 1, then illustrate that there is target kth range unit w ripple position of radar;

Step 3, the two-dimensional position coordinate that any one range unit and any one ripple position form is designated as a range-azimuth resolution element, then the range-azimuth resolution element of radar has K × W, draws the true phase sequence q of radar i-th range-azimuth resolution element M pulse _i, i=1,2 ..., K × W;

Step 4, show that in range-azimuth matrix D, element is the phase linearity of each range-azimuth resolution element of 1;

Step 5, show that in range-azimuth matrix D, element is the performance number of each range-azimuth resolution element of 1;

Step 6, the element being 1 by numerical value in range-azimuth matrix D is divided into multiple connected region according to eight neighborhood connectedness, and the number dividing the connected region drawn is expressed as N;

Step 7, calculates the average E of the performance number of a jth connected region all range-azimuths resolution element _j, and the average L of inverse of linear phase degree of a jth connected region all range-azimuths resolution element _j, j=1,2 ..., N;

Step 8, draw the relative power of a jth connected region r range-azimuth resolution element and opposite linear phase place degree inverse and z _j,r, z _j,r=e _r/ E _j+ 1/ (L _jδ _r), e _rrepresent the performance number of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, δ _rrepresent the linear phase degree of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, r=1,2 ..., S _j, S _jrepresent the number of the range-azimuth resolution element of a jth connected region in range-azimuth matrix D; By the relative power of a jth connected region each range-azimuth resolution element and opposite linear phase place degree inverse and maximal value be designated as: a jth connected region statistic maximal value; Using a jth target location that the range-azimuth resolution element that connected region statistic maximal value is corresponding is condensed into as a jth connected region.

2., as claimed in claim 1 based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, described step 3 specifically comprises following sub-step:

(3.1) the two-dimensional position coordinate that any one range unit and any one ripple position form is designated as a range-azimuth resolution element, then the range-azimuth resolution element of radar has K × W, calculates winding phase place WrapPhase (m) of radar i-th range-azimuth resolution element m pulse, WrapPhase (m)=arg (X (m, k, w)), m=1,2 ..., M, i=1,2 ..., K × W; Phase place is asked in arg () expression;

(3.2) make true phase UnWrapPhase (the 1)=WrapPhase (1) of radar i-th range-azimuth resolution element the 1st pulse, make the phase difference (2) of the true phase of the winding phase place of radar i-th range-azimuth resolution element the 2nd pulse and the 1st pulse of radar i-th range-azimuth resolution element be:

Δ(2)＝WrapPhase(2)-UnWrapPhase(1)；

(3.3) if abs (Δ (2))≤π, then make UnWrapPhase (2)=UnWrapPhase (1)+Δ (2), then skip to step (3.5), abs () expression takes absolute value; If abs (Δ (2)) > π, then skip to sub-step (3.4);

(3.4) if Δ (2) > π, then making the value of Δ (2) from subtracting 2 π, being then back to sub-step (3.3); If Δ (2) <-π, then making the value of Δ (2) from increasing 2 π, being then back to sub-step (3.3);

(3.5) if the pile-up pulse number M of echo data array X is 2, then skip to sub-step (3.12), if the pile-up pulse number M of echo data array X is greater than (2), then skip to sub-step (3.6);

(3.6) n=2 is made, 3 ..., M; Work as n=3,4 ..., during M, make the phase difference (n) of the true phase of the winding phase place of radar i-th range-azimuth resolution element n-th pulse and (n-1)th pulse of radar i-th range-azimuth resolution element be:

Δ(n)＝WrapPhase(n)-UnWrapPhase(n-1)；

(3.7) if abs (Δ (n))≤π, then sub-step (3.8) is skipped to; If abs (Δ (n)) > is π, then skip to sub-step (3.10);

(3.8) if Δ (n) and Δ (2) jack per line, then sub-step (3.11) is skipped to; If Δ (n) and Δ (2) contrary sign, then skip to sub-step (3.9);

(3.9) make λ=1, if Δ (n) > π is-λ, then makes the value of Δ (n) from subtracting 2 π, then skipping to sub-step (3.11); If Δ (n) <-π is+λ, then makes the value of Δ (n) from increasing 2 π, then skipping to sub-step (3.11);

(3.10) if Δ (n) > is π, then making the value of Δ (n) from subtracting 2 π, being then back to sub-step (3.7); If Δ (n) <-is π, then makes the value of Δ (n) from increasing 2 π, being then back to sub-step (3.7);

(3.11) UnWrapPhase (n) is drawn,

UnWrapPhase(n)＝UnWrapPhase(n-1)+Δ(n)；

Then sub-step (3.12) is skipped to;

(3.12) the true phase sequence q of radar i-th range-azimuth resolution element M pulse is drawn _i, the true phase sequence q of radar i-th range-azimuth resolution element M pulse _ifor:

q _i＝[UnWrapPhase(1),UnWrapPhase(2),…,UnWrapPhase(M)]。

3. as claimed in claim 1 based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, in step 4, if the element of i-th range-azimuth resolution element is 1 in range-azimuth matrix D, then the linear phase degree δ of radar i-th range-azimuth resolution element _ifor:

${\mathrm{\&delta;}}_i=\frac{1}{M-1}\underset{n=2}{\overset{M}{\mathrm{\&Sigma;}}}{(\mathrm{\&Delta;}\left(n\right)-\stackrel{\&OverBar;}{k})}^2$

$\stackrel{\&OverBar;}{k}=\frac{1}{M-1}\underset{n=2}{\overset{M}{\mathrm{\&Sigma;}}}\mathrm{\&Delta;}\left(n\right)$

Wherein, n=2,3, ... M, Δ (n)=WrapPhase (n)-UnWrapPhase (n-1), WrapPhase (n) represents the winding phase place of radar i-th range-azimuth resolution element n-th pulse, and UnWrapPhase (n-1) represents the true phase of radar i-th range-azimuth resolution element (n-1)th pulse.

4. as claimed in claim 1 based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, in steps of 5, if the element of i-th range-azimuth resolution element is 1 in range-azimuth matrix D, then the performance number e of radar i-th range-azimuth resolution element _ifor:

$e_i=\frac{1}{M}\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{X}_{m,i}^2;$

Wherein, X _m,irepresent the element of i-th range-azimuth resolution element m pulse in echo data array X.

5., as claimed in claim 2 based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, the concrete sub-step of described step 6 is:

(6.1) choose in range-azimuth matrix D any one element be the range-azimuth resolution element of 1 as initial resolution element, draw the neighborhood resolution element of initial resolution element;

(6.2) judge whether the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that element is 1; If there is not the range-azimuth resolution element that element is 1 in the neighborhood resolution element of initial resolution element, then using initial resolution element as a connected region, skip to sub-step (6.5); If the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that at least one element is 1, then skip to sub-step (6.3);

(6.3) the mean value k of the phase differential of initial resolution element is calculated ₀, Δ ₀n () represents the end value of the phase differential of the winding phase place of the n-th pulse of initial resolution element and the true phase of unit's (n-1)th pulse of the initial resolution element of radar in the radar drawn according to step 3;

Calculating each element in the neighborhood resolution element of initial resolution element is the mean value of the phase differential of the range-azimuth resolution element of 1, the mean value being the phase differential of the range-azimuth resolution element of 1 by any one element in the neighborhood resolution element of initial resolution element is expressed as k' wherein, Δ ' (n) represent the end value of the phase differential of the true phase of (n-1)th pulse of respective distances-azimuth discrimination unit in the winding phase place of the n-th pulse of respective distances-azimuth discrimination unit in the neighborhood resolution element of the initial resolution element of radar drawn according to step 3 and the neighborhood resolution element of the initial resolution element of radar;

If abs is (a tan (k ₀)-a tan (k'))>=β, then the element of respective distances-azimuth discrimination unit in the neighborhood resolution element of initial resolution element is got 0, wherein, a tan () represents negate tangent, abs () expression takes absolute value, and β is the positive number of a setting; If abs is (a tan (k ₀)-a tan (k')) < β, then respective distances-azimuth discrimination unit and initial resolution element in the neighborhood resolution element of initial resolution element are divided into same connected region, in all range-azimuths resolution element of range-azimuth matrix D, remove the range-azimuth resolution element being divided into connected region, obtain the unallocated range-azimuth resolution element to connected region of range-azimuth matrix D; Using respective distances-azimuth discrimination unit in the neighborhood resolution element of resolution element initial in range-azimuth matrix D as initial resolution element, unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, draw the neighborhood resolution element of initial resolution element, skip to sub-step (6.4);

(6.4) judge whether the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that element is 1; If the neighborhood resolution element of initial resolution element does not exist the range-azimuth resolution element that element is 1, then skip to sub-step (6.5); If the neighborhood resolution element of initial resolution element exists the range-azimuth resolution element that at least one element is 1, be then back to sub-step (6.3);

(6.5) unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, judge whether that there is element is that the range-azimuth resolution element of 1 is as initial resolution element, if there is no element is that the range-azimuth resolution element of 1 is as initial resolution element, the process then dividing connected region is complete, and record is divided into the range-azimuth resolution element number of each connected region, the number dividing the connected region drawn is expressed as N; Otherwise, if there is element is that the range-azimuth resolution element of 1 is as initial resolution element, unallocated in the range-azimuth resolution element of connected region then in range-azimuth matrix D, choosing any one element is that the range-azimuth resolution element of 1 is as initial resolution element, unallocated in the range-azimuth resolution element of connected region in range-azimuth matrix D, draw the neighborhood resolution element of initial resolution element, be back to sub-step (6.2).

6., as claimed in claim 1 based on the sea-surface target condensing method of relative power and phase linearity, it is characterized in that, in step 7, the average E of the performance number of a jth connected region all range-azimuths resolution element _jfor:

$E_j=\frac{1}{S_j}\underset{r=1}{\overset{S_j}{\mathrm{\&Sigma;}}}{e}_r$

Wherein, e _rrepresent the performance number of a jth connected region r range-azimuth resolution element in range-azimuth matrix D, r=1,2 ..., S _j, S _jrepresent the number of the range-azimuth resolution element of a jth connected region in range-azimuth matrix D;

The average L of the inverse of the linear phase degree of a jth connected region all range-azimuths resolution element _jfor:

$L_j=\frac{1}{S_j}\underset{r=1}{\overset{S_j}{\mathrm{\&Sigma;}}}\frac{1}{{\mathrm{\&delta;}}_r}$

Wherein, δ _rrepresent the linear phase degree of a jth connected region r range-azimuth resolution element in range-azimuth matrix D.