CN104166127B - A kind of multi-beam interleaved projections and the high ski-runnings assuming parabola Hough transform jump formula objective accumulation detection method more - Google Patents

Abstract

The invention discloses a kind of wave beam interleaved projections and the cunnings assuming parabola Hough transform to jump formula object detection method more, first radar pitching wave beam is divided into the clusters of beams of staggered superposition, the corresponding treatment channel of each clusters of beams, the parabola Hough transform realization cunning being utilized respectively multiple speed hypothesis in each treatment channel jumps formula maneuvering target signal energy by track non-inherent accumulation, then extract signal energy in all clusters of beams treatment channel and accumulate maximum passage as target place passage, and the measurement that this passage many hypothesis parabola model detected is weighted merging as final testing result。The present invention passes through multi-beam interleaved projections, reduces clutter density, improves detection probability, many hypothesis parabola Hough transform TBD technology are compared with general parabola Hough transform simultaneously, parameter space is become 2 dimensions from 5 dimensions, greatly reduces amount of calculation and complexity, is suitable for engineer applied。

Description

A kind of multi-beam interleaved projections and the high ski-runnings assuming parabola Hough transform jump formula objective accumulation detection method more

One, technical field

The present invention is under the jurisdiction of radar target tracking field, it is adaptable to solves the hypersonic cunning of near space and jumps the signal integration detection problem of formula maneuvering target。

Two, background technology

Near space is the spatial domain of span sea level 20～l00km, near space hypersonic aircraft has the features such as speed is fast, mobility good, stealth is strong, in two hours, whole world either objective can be carried out quick precision strike, for instance the X-51 aircraft etc. of the U.S.。Near space hypersonic aircraft generally adopts Sanger trajectory or its derivative trajectory of classics, such trajectory be mainly characterized by aircraft " squander " near space like vertical bounce formula flight, ultimate principle is: the space empty air tightness that 1) liftoff ball is near is big, high-speed aircraft can obtain very lift, thus making it have very big upwards acceleration, the atmosphere that quilt " ejection " air is dense as squandering；2) aircraft is low away from earth surface rear space atmospheric density, and lift is ignored, and aircraft will be again introduced into atmosphere as parabolic motion by action of gravity；3) the intermittent reinforcing of electromotor overcomes the impact of air drag, thus punching press repeatedly, sliding jumping, completes the effect of province's energy, far firing range。

The hypersonic target of near space is generally of certain stealth, so that ground-based radar is faced with the detecting and tracking problem of stealthy maneuvering target。And existing maneuvering target tracking technology is substantially for high s/n ratio situation, seldom relate to the tracking of stealthy maneuvering target, it is impossible to meet the tracking needs of the sliding more motor-driven Stealthy Target of formula of near space。Near space target flight speed is fast, mobility strong, largely limit the time of radar correlative accumulation, thus, adopting non-inherent accumulation is one of effective way, TBD technology based on Hough transform is conventional non-inherent accumulation method, its ultimate principle is to pass through Hough transform, it is the straight line of parameter space by data space point transformation, by judging whether a plurality of straight line of parameter space intersects at a point, judge they are whether data space is positioned on straight line track, thus realizing the integration detection to Stealthy Target。The signal integration detection of the rectilinear motion target that traditional Hough transform is primarily adapted for use on two dimensional surface, and near space target is three-dimensional cunning jumps formula maneuvering target, making radar be faced with three-dimensional " stealthy+high motor-driven " objective accumulation detection difficult problem, traditional Hough transform method can't adapt to the integration detection requirement of the hypersonic highly maneuvering target of near space。

Three, summary of the invention

Problem to be solved by this invention is exactly, and is similar to parabolical feature for the sliding motor-driven track of the section of jumping of the hypersonic target of near space, it is provided that many hypothesis parabola Hough transform integration detection method of the sliding formula maneuvering target that jumps of a kind of near space。

This invention address that described technical problem, adopt technical scheme steps as follows:

1. many hypothesis parabola Hough transform integration detection method of the sliding formula maneuvering target that jumps of near space, it is characterised in that comprise the following steps:

(1) the pitching wave beam of radar is divided into the mutually overlapping clusters of beams of multiple adjacent beams composition, the corresponding treatment channel of each clusters of beams。

(2), in each clusters of beams treatment channel, the measurement of radar detection is handled as follows:

A) three-dimensional measuring that clusters of beams is detected carries out pretreatment, gives up the signal energy measurement lower than a certain relatively low threshold, and remaining three-dimensional measuring is respectively perpendicular and projects to ground level, obtains the pseudo-measurement of the two dimension in horizontal plane；

B) the two dimension puppet measurement obtained is carried out straight line Hough transform, detect the pseudo-straight line measuring space, and three-dimensional measurement theed obtain in this straight line place vertical plane according to this straight line collects [Z₁,...,Z_n]；

C) three-dimensional measuring [Z that will be formed₁,...,Z_n] Coordinate Conversion to the straight line detected in b) step for x-axis, [Z₁,...,Z_n] the x-y-z rectangular coordinate system of any point to be zero, vertical direction be z-axis in measurement, obtain new measurement collection [Z'₁,...,Z'_n]；

D) multiple hypothesis speed V of peak horizontal direction in the sliding section of jumping of the hypersonic target of near space are chosen₁, V₂,...,V_i, at the measurement collection [Z' that c) step obtains₁,...,Z'_n] in the x of each measurement, z coordinate carry out the parallel processing of multiple hypothesis passage, each passage assumes V with a certain speed of the hypersonic target of near space respectively_iIt is parabola known quantity, parabola apex coordinate [x with gravity acceleration g_P,z_P] for the parabola Hough transform of parameter, in the two-dimensional parameter space of parabola apex coordinate, carry out signal energy accumulation, find out accumulation energy value peak, and find out measurement collection corresponding to this maximum [Z "₁,...,Z″_n] as the testing result of this passage；

(3) find out in the treatment channel that all clusters of beams are corresponding through too much assuming that passage that signal energy that parabola Hough transform obtains is maximum, it is believed that this passage is the passage that target exists；

(4) measurement that multiple hypothesis Hough transform in passage corresponding to target place clusters of beams are detected is weighted fusion detection, merges the target measurement measured as finally detecting。

Concrete, in described step (2) d) particularly as follows:

1) according to measuring collection [Z'₁,...,Z'_n] maximum magnitude that all measuring points are distributed in x-z-plane, x-z coordinate plane is divided into N × M the square shaped cells lattice that the length of side is Δ r, each square shaped cells center of a lattice x coordinate respectively B (i), i=1,2,3 ..., N, z coordinate is C (j) respectively, j=1,2,3 ..., M, define element initialize be entirely 0 N × Metzler matrix A (i, j), i=1,2,3, ..., N, j=1,2,3 ..., M, for depositing the energy accumulation value of parameter space；

2) the multiple hypothesis speed V of the hypersonic target of near space is chosen₁, V₂,...,V_i, constructing the parallel processing of multiple hypothesis passage, each assumes that the processing procedure of passage is following (with goal hypothesis speed V_iFor example):

A1) apex coordinate [x is made_P,z_PB]=[(i)], C (j)], make i change from 1 to N, j changes from 1 to M, and measurement is collected [Z'₁,...,...Z'_k...,Z'_n] in x, z coordinate [x_k,z_k] bring following formula one by one into

$|z_p-(z_k+\frac{g}{{{2V}_i}^2}{(x_p-x_k)}^2)|<\mathrm{\&Delta;r}$

Wherein, V_iRepresenting the target velocity (for constant) supposed, g represents acceleration of gravity constant。If the parabola apex coordinate [x that a certain group of i, j are corresponding_P,z_P] meeting above formula, then the parameter space following formula that i, j are corresponding carries out energy accumulation

A (i, j)=A (i, j)+E_k

Wherein E_kRepresent and measure Z'_kSignal energy；

B1) collection [Z' is measured₁,...,...Z'_k...,Z'_n] all through a1) after step processes, find out maximum in accumulation matrix A, and judge the matrix ranks label that this maximum is corresponding, be calculated as i_max,j_max, and according to ranks label i_max,j_maxObtain parabolical apex coordinate [B (i_max),C(j_max))]；

C1) by all measurement point set [Z'₁,...,Z'_n] x, z coordinate [x_k,z_k] bring following formula respectively into and judge

$|C\left(j_{\max }\right)-(z_k+\frac{g}{{{2V}_i}^2}{(B\left(i_{\max }\right)-x_k)}^2)|<\mathrm{\&Delta;r}$

Meet the measurement of above formula extract composition potential track [Z "₁,...,Z″_n]。

Concrete, described step (4) particularly as follows:

1) the energy accumulation matrix A that first will obtain after assuming parabola Hough transform more_m, m=1,2 ... (p represents model of ellipse number, A to p_mRepresent energy accumulation matrix corresponding to m-th model of ellipse) in Energy maximum value A_m(i_max,j_max) compare with a certain predetermined threshold E_th, if A_m(i_max,j_max) less than predetermined threshold, then it is assumed that this Energy maximum value A_m(i_max,j_max) corresponding measurement is unlikely to be target, is no longer participate in remaining step；

2) remaining measurement is concentrated according to markers, and utilize following formula

$d_{\max }=\arg \underset{m=\mathrm{1,2},..p^{\&prime;}}{\max }\left[A_m(i_{\max },j_{\max })\right]$

Obtain the pattern number d that multiple speed assumes that in parabola model, accumulation energy is maximum_max, wherein p' represents through b1) and remaining parabola model number after step。Then numbering d is extracted_maxCorresponding each moment parabolical measures collection Z_max{t_i, i=1,2 .., (wherein l represents for accumulating radar frame number, t l_iRepresent the markers of the i-th frame data)；

3) the measurement Z of acquisition is calculated respectively_max{t_iAnd except model d_maxDistance between the same time scale measurement out of other model inspection in addition, if their distance is less than a certain predetermined threshold R_th, then it is assumed that the two measures and can merge, and fusion process is: Schilling t_iMoment merges flight path z_f{t_i}=Z_max{t_i, then utilize following formula to be fused to these 2 a bit

$z_f\left\{t_i\right\}=z_f\left\{t_i\right\}\&CenterDot;a_1+z_m^j\left\{t_i\right\}\&CenterDot;(1-a_1)$

Wherein,Represent the coordinate that the jth that m-th model of ellipse Hough transform detects measures, z_f{t_iFor t_iMoment merges the coordinate measured, a₁Representing weight, it is sized to

$a_1=\frac{A_{d_{\max }}(i_{\max },j_{\max })}{A_m(i_{\max },j_{\max })+A_{d_{\max }}(i_{\max },j_{\max })}$

In formula, A_m(i_max,j_max) represent energy accumulation matrix A_mIn maximum,Represent energy accumulation matrixIn maximum；

4) find out measurement of each moment number measurement collection more than 1, take out these measurements measuring concentration and carry out following judgement with the measuring point of each moment only one of which measurement one by one

R(Z₁,Z₂) < V_min·(t₂-t₁)

R(Z₁,Z₂) > V_max·(t₂-t₁)

Wherein, Z₁,Z₂Representing the coordinate for two measuring points judged, R () represents the Euclidean distance between asking two to measure, V_min, V_maxRepresent that near space hypersonic aircraft is minimum and most probable velocity, t respectively₂,t₁Representing two respectively and measure corresponding markers, if upper two formulas are all set up, then it is assumed that this measurement in this moment is clutter, is rejected, remaining measuring point is as final object detection results output。

The invention has the beneficial effects as follows, improve the ability of discovery of the hypersonic target of near space, radar pitching wave beam is divided into the clusters of beams of staggered superposition, the corresponding treatment channel of each clusters of beams, the parabola Hough transform realization cunning being utilized respectively multiple speed hypothesis in each treatment channel jumps formula maneuvering target signal energy by track non-inherent accumulation, then extract signal energy in all clusters of beams treatment channel and accumulate maximum passage as target place passage, and the measurement that this passage many hypothesis parabola model detected is weighted merging as final testing result。The present invention passes through multi-beam interleaved projections, reduces clutter density, improves detection probability, many hypothesis parabola Hough transform TBD technology are compared with general parabola Hough transform simultaneously, parameter space is become 2 dimensions from 5 dimensions, greatly reduces amount of calculation and complexity, is suitable for engineer applied。

Four, accompanying drawing explanation

Accompanying drawing 1 is the method step flow chart of the present invention；

When accompanying drawing 2 is to carry out emulation experiment by the present invention, near space punching sliding jumps formula aerial vehicle trajectory figure, and solid line represents target entirety flight course trajectory diagram, and solid line thickened portion represents the one section of cunning section of jumping track of intercepting；

7 moment that accompanying drawing 3 intercepts when being carry out emulation experiment by the present invention in the track of accompanying drawing 2 measure, and the interval time of measurement is 2 seconds；

The three-dimensional amount mapping that when accompanying drawing 4 is to carry out emulation experiment by the present invention, clusters of beams 2-4 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

The three-dimensional amount mapping that when accompanying drawing 5 is to carry out emulation experiment by the present invention, clusters of beams 3-5 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

The three-dimensional amount mapping that when accompanying drawing 6 is to carry out emulation experiment by the present invention, clusters of beams 4-6 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

The three-dimensional amount mapping that when accompanying drawing 7 is to carry out emulation experiment by the present invention, clusters of beams 5-7 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

The three-dimensional amount mapping that when accompanying drawing 8 is to carry out emulation experiment by the present invention, clusters of beams 6-8 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

The three-dimensional amount mapping that when accompanying drawing 9 is to carry out emulation experiment by the present invention, clusters of beams 7-8 detects, wherein zero represents measuring value, ● represent the actual measurements position of near space target；

When accompanying drawing 10 is to carry out emulation experiment by the present invention in the passage that target place clusters of beams is corresponding, parabola model 1 straight line (straight line is parabolical special case) is utilized to carry out energy accumulation figure during Hough transform, x-y plane is the parameter space on parabola summit, and z-axis represents the energy accumulation value of parameter space；

When accompanying drawing 11 is to carry out emulation experiment by the present invention in the clusters of beams respective channel of target place, utilizing parabola model 2 hypothetical target horizontal velocity when parabola summit is 3000 meter per seconds²Time corresponding parabola, carry out energy accumulation figure during Hough transform, x-y plane is the parameter space on parabola summit, and z-axis represents the energy accumulation value of parameter space；

When accompanying drawing 12 is to carry out emulation experiment by the present invention in the clusters of beams respective channel of target place, utilizing parabola model 2 hypothetical target horizontal velocity when parabola summit is 3500 meter per seconds²Time corresponding parabola, carry out energy accumulation figure during Hough transform, x-y plane is the parameter space on parabola summit, and z-axis represents the energy accumulation value of parameter space；

Five, specific implementation method

With reference to the accompanying drawings the invention is described in further details

With reference to accompanying drawing 1, the present invention comprises the following steps that

Step one: the pitching wave beam of radar is divided into the mutually overlapping clusters of beams of multiple adjacent beams composition, such as, pitching beam angle is 1 degree, pitching scope is 2～8 degree, have 8 ripple positions, number from 1 to 8 number consecutivelies, so radar completes 5 ripple positions of search needs of pitching, if carrying out staggered division with 2 wave beams for one bunch, then clusters of beams can be { 2-4,3-5,4-6,5-7,6-8} (wherein 2～9 represent wave beam numbering respectively), then by the treatment channel measuring its correspondence of feeding in each clusters of beams；

Step 2: in each clusters of beams treatment channel, the measurement of radar detection is handled as follows:

1. thresholding relatively low with one by one for the signal energy of three dimensions measurement (distance, orientation, pitching, signal energy) of radar in the observation area of clusters of beams is compared, energy is measured upright projection to horizontal plane more than the three dimensions of this thresholding, forming the pseudo-measurement of two dimension under rectangular coordinate system in horizontal plane, conversion formula is as follows:

$\left\{\begin{array}{c}{r}^{\&prime;}=r\cos \left(\mathrm{\&epsiv;}\right)\\ {\mathrm{\&theta;}}^{\&prime;}=\mathrm{\&theta;}\end{array}\right.$

Wherein, r, θ, ε represent the distance of radar three-dimensional spatial measurement, orientation, pitching, r', θ ' represent conversion after two dimension pseudo-measure distance, orientation；

2. pseudo-for two dimension in plane measurement is carried out straight line Hough transform, detects the pseudo-straight line measuring space, concretely comprise the following steps:

A1) observation area in r'-θ ' plane being divided into N × M resolution cell, and assume that resolution cell distance is for Δ r, orientation is Δ θ, so each resolution cell midpoint, observation area distance respectively B (i), i=1,2,3 ..., N, orientation is C (j) respectively, j=1, and 2,3 ..., M, and define a N × M matrix A (i, j), i=1,2,3 ..., N, j=1,2,3 ..., M, it is 0 entirely that each element of matrix initializes, for depositing the energy accumulation value of parameter space；

A2) two dimension polar coordinate puppet is measured polar coordinate and utilize equation below

$\left\{\begin{array}{c}x=r^{\&prime;}\cos \left({\mathrm{\&theta;}}^{\&prime;}\right)\\ y=r^{\&prime;}\sin \left({\mathrm{\&theta;}}^{\&prime;}\right)\end{array}\right.$

Be converted to x, y-coordinate, take out respectively under rectangular coordinate system in order and measure, and measurement is carried out Coordinate Conversion, the lower left corner of coordinate translation to observation area is obtained new measurement coordinate [x_k,y_k], for the measurement coordinate that each is new, bring point coordinates in resolution cell [r, θ]=[B (i), C (j)] (i changes from 1 to N, and j changes from 1 to M) into following formula one by one and compare

|r_i-[x_k·cos(θ_j)+y_k·sin(θ_j)] | < Δ r

If above formula is set up, then parameter space utilizes equation below to accumulate

A (i, j)=A (i, j)+E_k

Wherein E_kRepresent and measure [x_k,y_k] signal energy；

A3) maximum in energy accumulation matrix A is extracted, it is thus achieved that the i that this maximum is corresponding_max,j_max, and according to i_max,j_maxObtain the parameter of straight line

A4) by all radar two dimension puppet measuring point coordinate [x_k,y_k] bring following formula one by one into

$|r_{i_{\max }}-[x_k\&CenterDot;\cos \left({\mathrm{\&theta;}}_{j_{\max }}\right)+y_k\&CenterDot;\sin \left({\mathrm{\&theta;}}_{j_{\max }}\right)\left]\right|<\mathrm{\&Delta;r}$

Find out the three-dimensional measuring collection [Z meeting above formula₁,...,Z_n]。

3. three-dimensional measuring [the Z that will be formed in step 3₁,...,Z_n] Coordinate Conversion is to straight lineFor x-axis, [Z₁,...,Z_n] the x-y-z rectangular coordinate system of any point to be initial point, vertical direction be z-axis in measurement, obtain new measurement collection [Z'₁,...,Z'_n], switch process is:

B1) by three-dimensional measuring [Z₁,...,Z_n] it being converted into rectangular coordinate [x, y, z] by polar coordinate [r, θ, ε], application conversion formula is

$\left\{\begin{array}{c}x=r\cos \left(\mathrm{\&theta;}\right)\cos \left(\mathrm{\&epsiv;}\right)\\ y=r\sin \left(\mathrm{\&theta;}\right)\cos \left(\mathrm{\&epsiv;}\right)\\ z=r\sin \left(\mathrm{\&epsiv;}\right)\end{array}\right.$

B2) undertaken moving to measurement collection [Z by the rectangular coordinate system measurement after conversion₁,...,Z_n] in any one point, it is assumed that move to Z₁Point, Z₁Rectangular coordinate be [x₁,y₂,z₃], then [Z₁,...,Z_n] in arbitrfary point [x, y, z] translation after coordinate be

$\left\{\begin{array}{c}{x}^{\&prime;}=x-x_1\\ y^{\&prime;}=y-y_1\\ z^{\&prime;}=z-z_1\end{array}\right.$

B3) by b2) conversion after measurement coordinate rotate to x-axis and straight lineOverlapping, each measures [Z₁,...,Z_n] coordinate after rotation [x ", y ", z "] converts as follows

$\left\{\begin{array}{c}[x^{\&prime;\&prime;},y^{\&prime;\&prime;}]=[x^{\&prime;},y^{\&prime;}]\&CenterDot;\left[\begin{array}{cc}\cos \left({\mathrm{\&theta;}}_{j_{\max }}\right)& \sin \left({\mathrm{\&theta;}}_{j_{\max }}\right)\\ -\sin \left({\mathrm{\&theta;}}_{j_{\max }}\right)& \cos \left({\mathrm{\&theta;}}_{j_{\max }}\right)\end{array}\right]\\ z^{\&prime;\&prime;}=z^{\&prime;}\end{array}\right.$

All after conversion, obtain new measurement collection [Z'₁,...,Z'_n]。

4. new measurement collection is carried out many hypothesis parabola Hough transform, comprises the following steps that

C1) according to measuring collection [Z'₁,...,Z'_n] maximum magnitude that all measuring points are distributed in x-z-plane, x-z coordinate plane is divided into N × M the square shaped cells lattice that the length of side is Δ r, each square shaped cells center of a lattice x coordinate respectively B (i), i=1,2,3 ..., N, z coordinate is C (j) respectively, j=1,2,3 ..., M, define element initialize be entirely 0 N × Metzler matrix A (i, j), i=1,2,3, ..., N, j=1,2,3 ..., M, for depositing the energy accumulation value of parameter space；

C2) the multiple hypothesis speed V of the hypersonic target of near space is chosen₁, V₂,...,V_i, constructing the parallel processing of multiple hypothesis passage, each assumes that the processing procedure of passage is following (with goal hypothesis speed V_iFor example):

C3) apex coordinate [x is made_P,z_PB]=[(i)], C (j)], make i change from 1 to N, j changes from 1 to M, and measurement is collected [Z'₁,...,...Z'_k...,Z'_n] in x, z coordinate [x_k,z_k] bring following formula one by one into

$|z_p-(z_k+\frac{g}{{{2V}_i}^2}{(x_p-x_k)}^2)|<\mathrm{\&Delta;r}$

Wherein, V_iRepresenting the target velocity (for constant) supposed, g represents acceleration of gravity constant。If the parabola apex coordinate [x that a certain group of i, j are corresponding_P,z_P] meeting above formula, then the parameter space following formula that i, j are corresponding carries out energy accumulation

A (i, j)=A (i, j)+E_k

Wherein E_kRepresent and measure Z'_kSignal energy；

C4) collection [Z' is measured₁,...,...Z'_k...,Z'_n] all through c3) after step processes, find out maximum in accumulation matrix A, and judge the matrix ranks label that this maximum is corresponding, be calculated as i_max,j_max, and according to ranks label i_max,j_maxObtain parabolical apex coordinate [B (i_max),C(j_max))]；

C5) by all measurement point set [Z'₁,...,Z'_n] x, z coordinate [x_k,z_k] bring following formula respectively into and judge

$|C\left(j_{\max }\right)-(z_k+\frac{g}{{{2V}_i}^2}{(B\left(i_{\max }\right)-x_k)}^2)|<\mathrm{\&Delta;r}$

Meet the measurement of above formula extract composition potential track [Z "₁,...,Z″_n]。

Step 3: find out through too much assuming that passage that signal energy that parabola Hough transform obtains is maximum in the treatment channel that all clusters of beams are corresponding, it is believed that this passage is the passage that target exists；

Step 4: the many hypothesis parabola Hough transform testing result Weighted Fusion in the clusters of beams at target place, specific as follows

1. the energy accumulation matrix A that first will obtain after assuming parabola Hough transform more_m, m=1,2 ... (p represents model of ellipse number, A to p_mRepresent energy accumulation matrix corresponding to m-th model of ellipse) in Energy maximum value A_m(i_max,j_max) compare with a certain predetermined threshold E_th, if A_m(i_max,j_max) less than predetermined threshold, then it is assumed that this Energy maximum value A_m(i_max,j_max) corresponding measurement is unlikely to be target, is no longer participate in remaining step；

2. remaining measurement is concentrated according to markers, and utilize following formula

$d_{\max }=\arg \underset{m=\mathrm{1,2},..p^{\&prime;}}{\max }\left[A_m(i_{\max },j_{\max })\right]$

Obtain the pattern number d that multiple speed assumes that in parabola model, accumulation energy is maximum_max, wherein p' represents through b1) and remaining parabola model number after step。Then numbering d is extracted_maxCorresponding each moment parabolical measures collection Z_max{t_i, i=1,2 .., (wherein l represents for accumulating radar frame number, t l_iRepresent the markers of the i-th frame data)；

3. calculate the measurement Z of acquisition respectively_max{t_iAnd except model d_maxDistance between the same time scale measurement out of other model inspection in addition, if their distance is less than a certain predetermined threshold R_th, then it is assumed that the two measures and can merge, and fusion process is: Schilling t_iMoment merges flight path z_f{t_i}=Z_max{t_i, then utilize following formula to be fused to these 2 a bit

$z_f\left\{t_i\right\}=z_f\left\{t_i\right\}\&CenterDot;a_1+z_m^j\left\{t_i\right\}\&CenterDot;(1-a_1)$

Wherein,Represent the coordinate that the jth that m-th model of ellipse Hough transform detects measures, z_f{t_iFor t_iMoment merges the coordinate measured, a₁Representing weight, it is sized to

$a_1=\frac{A_{d_{\max }}(i_{\max },j_{\max })}{A_m(i_{\max },j_{\max })+A_{d_{\max }}(i_{\max },j_{\max })}$

In formula, A_m(i_max,j_max) represent energy accumulation matrix A_mIn maximum,Represent energy accumulation matrixIn maximum；

4. find out measurement of each moment number measurement collection more than 1, take out these measurements measuring concentration and carry out following judgement with the measuring point of each moment only one of which measurement one by one

R(Z₁,Z₂) < V_min·(t₂-t₁)

R(Z₁,Z₂) > V_max·(t₂-t₁)

Wherein, Z₁,Z₂Representing the coordinate for two measuring points judged, R () represents the Euclidean distance between asking two to measure, V_min, V_maxRepresent that near space hypersonic aircraft is minimum and most probable velocity, t respectively₂,t₁Representing two respectively and measure corresponding markers, if upper two formulas are all set up, then it is assumed that this measurement in this moment is clutter, is rejected, remaining measuring point is as final object detection results output。

The effect of the present invention can be further illustrated by following emulation experiment:

Simulated environment: being provided with the initial velocity magnitude of near space hypersonic aircraft is 3000 meter per seconds, course angle is 270 degree (counterclockwise for just), and the angle of pitch is 10 degree, initial position co-ordinates is [0,300000 meters, 70000 meters], quality is 1000 kilograms, assume in aircraft flight process by four power effects, i.e. gravity, thrust, lift, resistance, wherein thrust is mainly used in overcoming resistance, afterburning in an intermittent fashion in the aircraft punching press stage, the direction of power is contrary with resistance。The computing formula being subject to lift in aircraft flight process is

L=0.5 C_lρSV²

ρ=ρ₀e^-Bh

Wherein, lift coefficient C_l=1, aircraft front face area S=1 rice², V represents the velocity magnitude of aircraft, and ρ represents atmospheric density, the atmospheric density ρ on earth's surface₀(equal to 0.0034 lb/ft³), B is that constant (is equal toFoot^-1), h represents aircraft terrain clearance (unit foot), and lift is vertical with aircraft speed direction and upwards。Resistance calculation formulae is

D=0.5 C_dρSV²

Wherein, resistance coefficient C_d=0.48, ρ, S, V meaning is identical with lift formula。Above simulation parameter is utilized to generate three dimensions " the squander formula " flight path (as shown in Figure 2) of near space target, 2 seconds cunning sections of jumping at this track of interval take a little, take 7 points (as shown in Figure 3) altogether, with these 7 points for target actual position (as shown in Figure 4) produce radar measurement and add clutter (with radar resolution for distance 100 meters, 1 degree of orientation, signal to noise ratio be 6dB condition random produce clutter) constitute dim target detection scene to verify the effectiveness of the inventive method。

Simulation result and analysis: by the inventive method, after wave beam is carried out staggered sub-clustering, the measurement figure obtained is such as shown in accompanying drawing 4 to accompanying drawing 8, the measurement of target drops in some clusters of beams (clusters of beams 6-8), it can be seen that greatly reduce clutter number during Hough transform after projection by wave beam sub-clustering；After multiple clusters of beams passages are carried out many hypothesis parabola Hough transform, by extracting the maximum of energy accumulation, obtain target and be present in clusters of beams 6-8, and extract the measurement that 3 the parabola model Hough transform (such as accompanying drawing 9-Figure 11) in clusters of beams 6-8 treatment channel detect, it is weighted fusion treatment by the measurement that 3 model inspection are gone out, obtain final testing result (as shown in Figure 12), can be seen that, measurement corresponding to target detected by correct, it was demonstrated that the effectiveness of algorithm。

Claims (3)

1. a wave beam interleaved projections and the cunnings assuming parabola Hough transform jump formula object detection method more, it is characterised in that comprise the following steps:

(1) the pitching wave beam of radar is divided into the mutually overlapping clusters of beams of multiple adjacent beams composition, the corresponding treatment channel of each clusters of beams；

(2), in each clusters of beams treatment channel, the measurement of radar detection is handled as follows:

A) three-dimensional measuring that clusters of beams is detected carries out pretreatment, gives up the signal energy measurement lower than a certain relatively low threshold, and remaining three-dimensional measuring is respectively perpendicular and projects to ground level, obtains the pseudo-measurement of the two dimension in horizontal plane；

B) the two dimension puppet measurement obtained is carried out straight line Hough transform, detect the pseudo-straight line measuring space, and three-dimensional measurement theed obtain in this straight line place vertical plane according to this straight line collects [Z₁,...,Z_n]；

C) three-dimensional measuring [Z that will be formed₁,...,Z_n] Coordinate Conversion to the straight line detected in b) step for x-axis, [Z₁,...,Z_n] the x-y-z rectangular coordinate system of any point to be zero, vertical direction be z-axis in measurement, obtain new measurement collection [Z'₁,...,Z'_n]；

D) multiple hypothesis speed V of peak horizontal direction in the sliding section of jumping of the hypersonic target of near space are chosen₁, V₂,...,V_i, at the measurement collection [Z' that c) step obtains₁,...,Z'_n] in the x of each measurement, z coordinate carry out the parallel processing of multiple hypothesis passage, each passage assumes V with a certain speed of the hypersonic target of near space respectively_iIt is parabola known quantity, parabola apex coordinate [x with gravity acceleration g_P,z_P] for the parabola Hough transform of parameter, in the two-dimensional parameter space of parabola apex coordinate, carry out signal energy accumulation, find out accumulation energy value peak, and find out measurement collection corresponding to this peak [Z "₁,...,Z”_n] as the testing result of this passage；

(3) find out in the treatment channel that all clusters of beams are corresponding through too much assuming that passage that signal energy that parabola Hough transform obtains is maximum, it is believed that this passage is the passage that target exists；

(4) measurement that multiple hypothesis Hough transform in passage corresponding to target place clusters of beams are detected is weighted fusion detection, merges the target measurement measured as finally detecting。

2. a kind of wave beam interleaved projections according to claim 1 and the cunnings assuming parabola Hough transform jump formula object detection method more, it is characterised in that the d in described step (2)) particularly as follows:

1) according to measuring collection [Z'₁,...,Z'_n] maximum magnitude that all measuring points are distributed in x-z-plane, x-z coordinate plane is divided into N × M the square shaped cells lattice that the length of side is Δ r, each square shaped cells center of a lattice x coordinate respectively B (i), i=1,2,3, ..., N, z coordinate is C (j) respectively, j=1,2,3 ..., M, define element initialize be entirely 0 N × Metzler matrix A_m(i, j), i=1,2,3 ..., N, j=1,2,3 ..., M, m=1,2 ... p, p represent model of ellipse number, A_mRepresent the energy accumulation matrix that m-th model of ellipse is corresponding, for depositing the energy accumulation value of parameter space；

2) multiple hypothesis speed V of peak horizontal direction in the sliding section of jumping of the hypersonic target of near space are chosen₁, V₂,...,V_i, constructing the parallel processing of multiple hypothesis passage, each assumes that the processing procedure of passage is as follows:

A1) apex coordinate [x is made_P,z_PB]=[(i), C (j)], make i change from 1 to N, j changes from 1 to M, and measurement is collected [Z'₁,...,...Z'_k...,Z'_n] in x, z coordinate [x_k,z_k] bring following formula one by one into

Wherein, V_iRepresenting the target velocity supposed, for constant, g represents acceleration of gravity constant, if the parabola apex coordinate [x that a certain group of i, j are corresponding_P,z_P] meeting above formula, then the parameter space following formula that i, j are corresponding carries out energy accumulation

A_m(i, j)=A_m(i,j)+E_k

Wherein E_kRepresent and measure Z'_kSignal energy；

B1) collection [Z' is measured₁,...,...Z'_k...,Z'_n] all through a1) and step process after, find out accumulation matrix A_mMiddle maximum, and judge the matrix ranks label that this maximum is corresponding, it is calculated as i_max,j_max, and according to ranks label i_max,j_maxObtain parabolical apex coordinate [B (i_max),C(j_max)]；

C1) by all measurement point set [Z'₁,...,Z'_n] x, z coordinate [x_k,z_k] bring following formula respectively into and judge

Meet the measurement of above formula extract composition potential track [Z "₁,...,Z”_n]。

3. a kind of wave beam interleaved projections according to claim 1 and the cunnings assuming parabola Hough transform jump formula object detection method more, it is characterised in that described step (4) particularly as follows:

1) the energy accumulation matrix A that first will obtain after assuming parabola Hough transform more_mMiddle Energy maximum value A_m(i_max,j_max) compare with a certain predetermined threshold E_th, if A_m(i_max,j_max) less than predetermined threshold, then it is assumed that this Energy maximum value A_m(i_max,j_max) corresponding measurement is unlikely to be target, is no longer participate in remaining step；

2) remaining measurement is concentrated according to markers, and utilize following formula

Obtain the pattern number d that multiple speed assumes that in parabola model, accumulation energy is maximum_max, wherein p' represents through b1) and remaining parabola model number after step, then extracts numbering d_maxCorresponding each moment parabolical measures collection Z_max{t_i, i=1,2 .., l, wherein l represents for accumulating radar frame number, t_iRepresent the markers of the i-th frame data；

3) the measurement Z of acquisition is calculated respectively_max{t_iAnd except model d_maxDistance between the same time scale measurement out of other model inspection in addition, if their distance is less than a certain predetermined threshold R_th, then it is assumed that the two measures and can merge, and fusion process is: Schilling t_iMoment merges flight path z_f{t_i}=Z_max{t_i, then utilize following formula to be fused to these 2 a bit

Wherein,Represent the coordinate that the jth that m-th model of ellipse Hough transform detects measures, z_f{t_iFor t_iMoment merges the coordinate measured, a₁Representing weight, it is sized to

In formula, A_m(i_max,j_max) represent energy accumulation matrix A_mIn maximum,Represent energy accumulation matrixIn maximum；

4) find out measurement of each moment number measurement collection more than 1, take out these measurements measuring concentration and carry out following judgement with the measuring point of each moment only one of which measurement one by one

R(Z₁,Z₂)<V_min·(t₂-t₁)

R(Z₁,Z₂)>V_max·(t₂-t₁)

Wherein, Z₁,Z₂Representing the coordinate for two measuring points judged, R () represents the Euclidean distance between asking two to measure, V_min, V_maxRepresent that the hypersonic target of near space is minimum and most probable velocity, t respectively₂,t₁Representing two respectively and measure corresponding markers, if upper two formulas are all set up, then it is assumed that this measurement in this moment is clutter, is rejected, remaining measuring point is as final object detection results output。