CN103885051B - Based on the method for parameter estimation of the simple scattering point cone target of time-frequency imaging - Google Patents

Abstract

The invention belongs to signal processing technology field, relate to radar imaging technology, disclose a kind of method for parameter estimation of the simple scattering point cone target based on time-frequency imaging.Should carry out in azimuth dimension the instantaneous RD image that time frequency analysis obtains cone target by pulse pressure echo of adjusting the distance; Barycenter displacement alignment is carried out to make up the out of true of motion compensation in early stage to instantaneous RD image; The echo cycle is estimated by calculating instantaneous picture related coefficient to the instantaneous RD image of center of gravity translational alignment; Then cone asymmetry parameter is estimated according to echo time-frequency distributions; The gyro frequency of cone is estimated in conjunction with echo cycle and cone asymmetry parameter; Finally estimate cone height and facies basialis pyramidis radius according to the position of scattering point in instantaneous RD image.The present invention has simple to operate, and make full use of the positional information of scattering center in image area, estimated accuracy is high, is not subject to the advantages such as the restriction of concrete model simultaneously.

Description

Based on the method for parameter estimation of the simple scattering point cone target of time-frequency imaging

Technical field

The invention belongs to signal processing technology field, relate to radar imaging technology, be specifically related to a kind of method for parameter estimation of the simple scattering point cone target based on time-frequency imaging.

Background technology

While the high-speed flight aloft of space cone target, be also attended by the micromotions such as spin to keep the stability of self.Space cone target is often with the parts such as empennage, groove, and its echo scattering properties can regard as the echoing characteristics of simple scattering point.In wideband radar imaging field, the echoed signal of the space cone after motion-compensated can be compressed in different distance frequencies, thus can carry out high-resolution imaging to target.High-resolution imaging and the parameter estimation of space cone target are significant for Ballistic Missile Targets identification.

A kind of method for estimating rotating speed of target of inverse synthetic aperture radar (ISAR) is disclosed in Tsing-Hua University's patented claim " method for estimating rotating speed of target based on the inverse synthetic aperture radar (ISAR) analyzed during sky " (number of patent application: 201010209955.X, publication number: CN102121990A).The method is one group of basis function at wavenumber domain spatial configuration, is used for representing echo data, and analysis when sky carries out to target scattering point, according to the scattering point locus of extracting and spatial position change rate information, fit object rotating speed.The deficiency that the method exists is, extracting scattering point phase coefficient, but not making full use of image area scattering point positional information by analyzing during wavenumber domain empty; The method only have estimated the rotating speed of target, can not estimate the physical dimension of target.

Xian Electronics Science and Technology University's patented claim " smooth procession cone parameter estimation method based on high-resolution ISAR imaging " (number of patent application: disclose a kind of high-resolution imaging method for smooth precession cone target 201310386501.3).The method, by echo translational compensation and Range Walk Correction, adopts the three-dimensional imaging of match search method realization to cone target of amplitude and phase place, and utilizes the positional information of scattering point to estimate size and the fine motion information of cone.This algorithm carries out imaging and parameter estimation based on the Precession model of smooth cone, because algorithm is subject to the restriction of concrete model, makes the parameter estimating error when model mismatch larger.

Summary of the invention

The object of the invention is to overcome above-mentioned the deficiencies in the prior art, a kind of method for parameter estimation of the simple scattering point cone target based on time-frequency imaging is proposed, adopt the method for instantaneous picture related coefficient to estimate the swing circle of cone target, and estimate the size of cone on this basis.The method overcome tradition empty time analytical approach in can not make full use of the positional information of scattering point in image area and estimate the defect of cone size, compensate for the deficiency that parameter estimating error that the model mismatch for smooth cone model brings is larger simultaneously.

Realizing basic ideas of the present invention is: by adjusting the distance, pulse pressure echo carries out in azimuth dimension the instantaneous RD image that time frequency analysis obtains cone target; Barycenter displacement alignment is carried out to make up the out of true of motion compensation in early stage to instantaneous RD image; The echo cycle is estimated by calculating instantaneous picture related coefficient to the instantaneous RD image of center of gravity translational alignment; Then cone asymmetry parameter is estimated according to echo time-frequency distributions; The gyro frequency of cone is estimated in conjunction with echo cycle and cone asymmetry parameter; Finally estimate cone height and facies basialis pyramidis radius according to the position of scattering point in instantaneous RD image.

The method for parameter estimation of a kind of simple scattering point cone target based on time-frequency imaging of the present invention, is characterized in that, comprise the following steps:

Step 1, obtain the distance pulse pressure echo of the simple scattering point cone target of inverse synthetic aperture radar (ISAR) admission, this distance pulse pressure echo is distance verses time 2-D data;

Step 2, the distance verses time 2-D data of pulse pressure of adjusting the distance echo does Short Time Fourier Transform along slow time orientation, obtains distance verses time-Doppler's three-dimensional data, extracts the instantaneous RD image of cone target from distance verses time-Doppler's three-dimensional data;

Step 3, by the instantaneous RD image reform translational alignment of cone target to its geometric center, obtains the instantaneous RD image after translational alignment;

Step 4, to the instantaneous RD image after translational alignment, utilizes circulation auto-correlation and circular AMDF Copula, calculates the instantaneous picture related coefficient of the instantaneous RD image after translational alignment;

Wherein, circulation auto-correlation and circular AMDF Copula following formula represent:

$F\left(k\right)=\frac{\underset{n=1}{\overset{L}{\mathrm{\Σ}}}\left(\mathrm{\Σ}\right|I_i^{\′}(n,m)I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)\left|\right)}{\underset{i=1}{\overset{L}{\mathrm{\Σ}}}\left(\mathrm{\Σ}\right|I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)-I_i^{\′}(n,m)\left|\right)},k=\mathrm{1,2},\·\·\·L/2$

Wherein, F (k) is instantaneous picture related coefficient, and k is time variable, and L represents the number of instantaneous RD image, I _i' (n, m) represents t after translational alignment _ithe instantaneous RD image in moment, n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and mod (i+k, L) represents i+k L remainder;

Step 5, according to the instantaneous picture related coefficient obtained, therefrom takes out moment corresponding to instantaneous picture related coefficient maximal value place, as the echo cycle T of distance pulse pressure echo ₀estimated value;

Step 6, the time-frequency distributions curve according to distance pulse pressure echo estimates cone asymmetry parameter, and its concrete sub-step is as follows:

(1) from distance verses time-Doppler's three-dimensional data, take out the time m-doppler data that any distance frequency is corresponding, also just obtain the time-frequency distributions curve of the distance pulse pressure echo of this distance frequency;

(2) time-frequency distributions of distance pulse pressure echo is utilized to estimate the asymmetry parameter of cone, wherein, in the time-frequency distributions line of distance pulse pressure echo, consider that simple scattering point there will be eclipse phenomena, the time-frequency distributions curve of scattering point disconnects, but the time-frequency distributions curve within the scope of the SEE time of scattering point in a swing circle is continuous print; Set certain scattering point as with reference to scattering point, be greater than semi-cone angle at the angle of radar line of sight and cone axis of symmetry, and the supplementary angle being less than semi-cone angle be in situation, for continuous time-frequency distributions Curves corresponding to reference scattering point time range in:

If a () does not have other time-frequency distributions curves, but be carved with other scattering points when the head and the tail with reference to continuous time-frequency distributions curve corresponding to scattering point when occurring, the asymmetry parameter N of cone ₀be 2;

If b () has two other time-frequency distributions curves to exist, and when scattering point corresponding to these two curves can not occur at any one time simultaneously, the asymmetry parameter N of cone ₀be 3;

If c () has two other time-frequency distributions curves to exist, and the moment that scattering point corresponding to these two curves is only 0 when the Doppler with reference to scattering point occurs simultaneously, the asymmetry parameter N of cone ₀be 4;

If d () has four other time-frequency distributions curves to exist, so asymmetry parameter N of cone ₀be 5;

Step 7, the echo cycle of Binding distance pulse pressure echo and cone asymmetry parameter, estimate cone gyro frequency:

$f_0=\frac{1}{N_0{T}_0}$

Wherein, f ₀cone gyro frequency, N ₀the asymmetry parameter of cone, T ₀it is the echo cycle of distance pulse pressure echo;

Step 8, estimate cone height and facies basialis pyramidis radius, its concrete sub-step is as follows:

(1), on the instantaneous RD image before translational alignment or after translational alignment, utilize the distance frequency at two scattering point places of distance dimension lie farthest away to estimate the length of cone, and estimate according to the following formula:

Wherein, the height of cone that what l represented is, c represents propagation velocity of electromagnetic wave, N _rrepresent vertex of a cone scattering point and the maximum difference of boring the distance frequency arrived corresponding to end scattering point in instantaneous RD image, B represents the bandwidth of radar emission linear FM signal, represent the radar line of sight angle of pitch;

(2), on the instantaneous RD image before translational alignment or after translational alignment, utilize the localizer unit at two of azimuth dimension lie farthest away scattering point places to estimate facies basialis pyramidis radius, and estimate according to the following formula:

$d=\frac{\mathrm{\λ}{N}_a}{8\mathrm{\πf}{N}_g}\mathrm{PRF}$

Wherein, what d represented is facies basialis pyramidis radius, and λ represents electromagnetic wavelength, N _arepresent the maximum difference of the scattering point place localizer unit of facies basialis pyramidis, f represents gyro frequency, N _grepresent that the window of time window function g (τ-t) is long, PRF indicating impulse repetition frequency;

Step 9, exports the parameter estimation result of simple scattering point cone target: instantaneous picture related coefficient, cone gyro frequency, cone height and facies basialis pyramidis radius.

Wherein, feature of the present invention and further improvement are:

The concrete sub-step of a, step 2 is:

(1) for distance pulse pressure echo, do time frequency analysis according to the following formula along slow time orientation, obtain distance verses time-Doppler's three-dimensional data:

$Q(r,f,t)={\∫}_{-\∞}^{\∞}\left[x(r,\mathrm{\τ})g^{*}(\mathrm{\τ}-t)\right]e^{-j2\mathrm{\πf\τ}}\mathrm{d\τ}$

Wherein, Q (r, f, t) signal x (r is represented, τ) in the result of the Short Time Fourier Transform of t, r represents distance frequency, and f represents Doppler frequency, t represents the slow time, x (r, τ) represents the echo sequence of distance frequency r, and τ is integration variable, g (τ-t) represents the time window function near t, N _gfor the window of time window function g (τ-t) is long;

(2) from distance verses time-Doppler's three-dimensional data, then take out the distance-Doppler data of any time according to the following formula, obtain the instantaneous RD image in one group of this moment:

I _i(r,f)=Q(r,f,t=t _i)

Wherein, I _i(r, f) represents t _ithe instantaneous RD image in moment, r represents distance frequency, and f represents Doppler, and Q represents distance verses time-Doppler's three-dimensional data, and t represents the slow time, t _iit is a certain instantaneous moment.

The concrete sub-step of b, step 3 is:

(1) according to the following formula instantaneous RD image is tieed up along distance, asks its distance dimension center of gravity:

$D_i(n,m)=\frac{{\mathrm{\Σ}}_{n=1}^N\left(n{I}_i(n,m)\right)}{{\mathrm{\Σ}}_{n=1}^N{I}_i(n,m)}$

Wherein, D _i(n, m) represents t _imoment is instantaneous RD image distance dimension center of gravity, and n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and N represents the distance frequency number of instantaneous RD image, I _i(n, m) represents t _ithe instantaneous RD image in moment;

(2) according to the following formula to instantaneous RD image distance dimension center of gravity D _i(n, m), along azimuth dimension, asks the two-dimentional center of gravity of instantaneous RD image:

$P_i(n,m)=\frac{{\mathrm{\Σ}}_{m=1}^M\left(n{D}_i(n,m)\right)}{{\mathrm{\Σ}}_{m=1}^M{D}_i(n,m)}$

Wherein, P _i(n, m) represents t _imoment is the two-dimentional center of gravity of instantaneous RD image, D _i(n, m) represents t _imoment is instantaneous RD image distance dimension center of gravity, and n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and M represents the localizer unit number of instantaneous RD image;

(3) the two-dimentional center of gravity position translation of the instantaneous RD image of trying to achieve is snapped to the geometric center position of instantaneous RD image, make t before translational alignment _ithe instantaneous RD image I in moment _i(r, f) is transformed to t after translational alignment _ithe instantaneous RD image I in moment _i' (r, f).

Compared with prior art, the present invention has outstanding substantive distinguishing features and significant progress.In prior art, based on the method for estimating rotating speed of target of the inverse synthetic aperture radar (ISAR) analyzed during sky, scattering point phase coefficient is extracted by analyzing during wavenumber domain empty, but image area scattering point positional information is not made full use of, only have estimated the rotating speed of target, can not estimate the physical dimension of target; Based on the smooth procession cone parameter estimation method of high-resolution ISAR imaging, its algorithm is subject to the restriction of concrete model, makes the parameter estimating error when model mismatch larger.But, the present invention adopts the method for time frequency analysis to obtain the instantaneous RD image of cone target, and utilize instantaneous picture related coefficient to estimate cone gyro frequency, overcome existing wavenumber domain empty time analytical approach can not make full use of the deficiency of the positional information of scattering point in image area, and have estimated cone height and facies basialis pyramidis radius; Simultaneously the present invention adopts the imparametrization method of time frequency analysis to come imaging and parameter estimation, and overcome the defect that evaluated error under existing concrete model mismatch condition is large, the observation for instantaneous RD image has the advantage be easily understood.

Accompanying drawing explanation

Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.

Fig. 1 is the process flow diagram of the method for parameter estimation of the simple scattering point cone target based on time-frequency imaging of the present invention;

Fig. 2 is that wherein, transverse axis is the time, and the longitudinal axis is Copula range value based on the echo cycle that circulation auto-correlation and circular AMDF Copula are estimated in the present invention;

Fig. 3 is the time-frequency distributions curve map of echo of the present invention, and transverse axis is the time, and the longitudinal axis is Doppler frequency;

Fig. 4 (a), Fig. 4 (b), Fig. 4 (c), Fig. 4 (d), Fig. 4 (e), Fig. 4 (f), Fig. 4 (g), Fig. 4 (h), Fig. 4 are (i) one group of instantaneous RD image of the present invention, wherein, every instantaneous RD image header is moment value, in figure, transverse axis is localizer unit number, and the longitudinal axis is distance frequency number.

Fig. 5 is the simple scattering point cone object module figure that emulation experiment of the present invention uses, and wherein, 4 symmetrical buses of this simple scattering point cone target surface embedded in 8 little balls, every bar bus have two little balls.

Embodiment

With reference to Fig. 1, the method for parameter estimation of the simple scattering point cone target based on time-frequency imaging of the present invention, detailed step is as follows:

Step 1, obtain the distance pulse pressure echo of the single scattering point cone target of inverse synthetic aperture radar (ISAR) admission, this distance pulse pressure echo is distance verses time 2-D data.

Inverse synthetic aperture radar (ISAR) ISAR launches linear FM signal, and receives the echoed signal of single scattering point cone target, and does distance pulse pressure along fast time orientation to echoed signal, obtains distance pulse pressure echo.

Step 2, the distance verses time 2-D data of pulse pressure of adjusting the distance echo does Short Time Fourier Transform along slow time orientation, obtains distance verses time-Doppler's three-dimensional data, extracts the instantaneous RD image of cone target from distance verses time-Doppler's three-dimensional data.

The concrete sub-step of this step is:

(1) for distance pulse pressure echo, do time frequency analysis according to the following formula along slow time orientation, obtain distance verses time-Doppler's three-dimensional data:

$Q(r,f,t)={\∫}_{-\∞}^{\∞}\left[x(r,\mathrm{\τ})g^{*}(\mathrm{\τ}-t)\right]e^{-j2\mathrm{\πf\τ}}\mathrm{d\τ}$

Wherein, Q (r, f, t) signal x (r is represented, τ) in the result of the Short Time Fourier Transform of t, r represents distance frequency, and f represents Doppler frequency, t represents the slow time, x (r, τ) represents the echo sequence of distance frequency r, and τ is integration variable, g (τ-t) represents the time window function near t, N _gfor the window of time window function g (τ-t) is long.

(2) from distance verses time-Doppler's three-dimensional data, then take out the distance-Doppler data of any time according to the following formula, obtain the instantaneous RD image in one group of this moment:

I _i(r,f)=Q(r,f,t=t _i)

Wherein, I _i(r, f) represents t _ithe instantaneous RD image in moment, r represents distance frequency, and f represents Doppler, and Q represents distance verses time-Doppler's three-dimensional data, and t represents the slow time, t _iit is a certain instantaneous moment.

Step 3, by the instantaneous RD image reform translational alignment of cone target to its cluster center, obtains the instantaneous RD image after translational alignment.

Its concrete sub-step is as follows:

(1) according to the following formula instantaneous RD image is tieed up along distance, asks its distance dimension center of gravity:

$D_i(n,m)=\frac{{\mathrm{\Σ}}_{n=1}^N\left(n{I}_i(n,m)\right)}{{\mathrm{\Σ}}_{n=1}^N{I}_i(n,m)}$

Wherein, D _i(n, m) represents t _imoment is instantaneous RD image distance dimension center of gravity, and n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and N represents the distance frequency number of instantaneous RD image, I _i(n, m) represents t _ithe instantaneous RD image in moment;

(2) according to the following formula to instantaneous RD image distance dimension center of gravity D _i(n, m), along azimuth dimension, asks the two-dimentional center of gravity of instantaneous RD image:

$P_i(n,m)=\frac{{\mathrm{\Σ}}_{m=1}^M\left(n{D}_i(n,m)\right)}{{\mathrm{\Σ}}_{m=1}^M{D}_i(n,m)}$

Wherein, P _i(n, m) represents t _imoment is the two-dimentional center of gravity of instantaneous RD image, D _i(n, m) represents t _imoment is instantaneous RD image distance dimension center of gravity, and n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and M represents the localizer unit number of instantaneous RD image;

(3) the two-dimentional center of gravity position translation of the instantaneous RD image of trying to achieve is snapped to the geometric center position of instantaneous RD image, make t before translational alignment _ithe instantaneous RD image I in moment _i(r, f) is transformed to t after translational alignment _ithe instantaneous RD image I in moment _i' (r, f).

Step 4, to the instantaneous RD image after translational alignment, utilizes circulation auto-correlation and circular AMDF Copula, calculates the instantaneous picture related coefficient of the instantaneous RD image after translational alignment.

Here circulation auto-correlation and circular AMDF Copula can represent with following formula:

$F\left(k\right)=\frac{\underset{n=1}{\overset{L}{\mathrm{\Σ}}}\left(\mathrm{\Σ}\right|I_i^{\′}(n,m)I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)\left|\right)}{\underset{i=1}{\overset{L}{\mathrm{\Σ}}}\left(\mathrm{\Σ}\right|I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)-I_i^{\′}(n,m)\left|\right)},k=\mathrm{1,2},\·\·\·L/2$

Wherein, F (k) is instantaneous picture related coefficient, and k is time variable, and L represents the number of instantaneous RD image, I _i' (n, m) represents t after translational alignment _ithe instantaneous RD image in moment, n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and mod (i+k, L) represents i+k L remainder.

Step 5, according to the instantaneous R image correlation coefficient obtained, therefrom takes out moment corresponding to instantaneous picture related coefficient maximal value place, as the echo cycle T of distance pulse pressure echo ₀estimated value.

Step 6, the time-frequency distributions curve according to distance pulse pressure echo estimates cone asymmetry parameter, and its concrete sub-step is as follows:

(1) from distance verses time-Doppler's three-dimensional data, take out the time m-doppler data that any distance frequency is corresponding, also just obtain the time-frequency distributions curve of the distance pulse pressure echo of this distance frequency;

(2) time-frequency distributions of distance pulse pressure echo is utilized to estimate the asymmetry parameter of cone, wherein, in the time-frequency distributions line of distance pulse pressure echo, consider that simple scattering point there will be eclipse phenomena, the time-frequency distributions curve of scattering point disconnects, but the time-frequency distributions curve within the scope of the SEE time of scattering point in a swing circle is continuous print; Set certain scattering point as with reference to scattering point, be greater than semi-cone angle at the angle of radar line of sight and cone axis of symmetry, and the supplementary angle being less than semi-cone angle be in situation, for continuous time-frequency distributions Curves corresponding to reference scattering point time range in:

If a () does not have other time-frequency distributions curves, but be carved with other scattering points when the head and the tail with reference to continuous time-frequency distributions curve corresponding to scattering point when occurring, the asymmetry parameter N of cone ₀be 2;

If b () has two other time-frequency distributions curves to exist, and when scattering point corresponding to these two curves can not occur at any one time simultaneously, the asymmetry parameter N of cone ₀be 3;

If c () has two other time-frequency distributions curves to exist, and the moment that scattering point corresponding to these two curves is only 0 when the Doppler with reference to scattering point occurs simultaneously, the asymmetry parameter N of cone ₀be 4;

If d () has four other time-frequency distributions curves to exist, so asymmetry parameter N of cone ₀be 5.

Generally, the space cone target asymmetry parameter 2 ~ 5 of aerial high-speed flight, lower than 2 or do not had practical significance more than 5, the present invention is not considered.

Step 7, the echo cycle of Binding distance pulse pressure echo and cone asymmetry parameter, estimate cone gyro frequency:

$f_0=\frac{1}{N_0{T}_0}$

Wherein, f ₀cone gyro frequency, N ₀the asymmetry parameter of cone, T ₀it is the echo cycle of distance pulse pressure echo.

Step 8, estimate cone height and facies basialis pyramidis radius, its concrete sub-step is as follows:

(1), on the instantaneous RD image before translational alignment, utilize the distance frequency at two scattering point places of distance dimension lie farthest away to estimate the length of cone, and estimate according to the following formula:

Wherein, the height of cone that what l represented is, c represents propagation velocity of electromagnetic wave, N _rrepresent vertex of a cone scattering point and the maximum difference of boring the distance frequency arrived corresponding to end scattering point in instantaneous RD image, B represents the bandwidth of radar emission linear FM signal, represent the radar line of sight angle of pitch;

(2) on instantaneous RD image, utilize the localizer unit at two of azimuth dimension lie farthest away scattering point places to estimate facies basialis pyramidis radius, and estimate according to the following formula before translational alignment:

$d=\frac{\mathrm{\λ}{N}_a}{8\mathrm{\πf}{N}_g}\mathrm{PRF}$

Wherein, what d represented is facies basialis pyramidis radius, and λ represents electromagnetic wavelength, N _arepresent the maximum difference of the scattering point place localizer unit of facies basialis pyramidis, f represents gyro frequency, N _grepresent that the window of time window function g (τ-t) is long, PRF indicating impulse repetition frequency.

In this step, the instantaneous RD image before translational alignment is adopted to carry out the estimation of cone height and facies basialis pyramidis radius, also the instantaneous RD image after translational alignment can be used to carry out the estimation of cone height and facies basialis pyramidis radius, the estimated result of the two is identical, what its reason was that algorithm for estimating adopts is mathematic interpolation, and translational alignment transfer pair mathematic interpolation result does not impact.

Step 9, exports the parameter estimation result of simple scattering point cone target: instantaneous picture related coefficient, cone gyro frequency, cone height and facies basialis pyramidis radius.

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

(1) simulated conditions:

The present invention's research be the cone model of simple scattering point, concrete emulation experiment adopts the model as Fig. 5, and the height of cone is 1m, and facies basialis pyramidis radius is 0.2m, and cone is rotationally symmetric body, and in cone, symmetry embedded in 8 little balls, and crown radius is 0.02m.The radar return data of research institute are produced by certain electromagnetic simulation software, launch linear FM signal, and electromagnetic wave adopts horizontal polarization mode incident, horizontal polarization mode receives, the electromagnetic carrier frequency of radar emission is 10GHz(X wave band), bandwidth is 2GHz, and frequency number is 51.

Emulation experiment is for single cone target, and emulation experiment optimum configurations is as follows: pulse repetition rate is 1200Hz, and integration time is 1s, and initial time radar line of sight position angle and the angle of pitch are 50.6 ° and 37.5 ° respectively, and spin frequency is 3Hz.

(2) content is emulated:

According to above-mentioned simulated conditions, test in MATLAB7.0 software, according to method provided by the invention, by adjusting the distance, pulse pressure echo carries out in azimuth dimension the instantaneous RD image that time frequency analysis obtains cone target; Barycenter displacement alignment is carried out to make up the out of true of motion compensation in early stage to instantaneous RD image; By calculating instantaneous picture related coefficient to the instantaneous RD image of center of gravity translational alignment, estimate the echo cycle; Then cone asymmetry parameter is estimated according to echo time-frequency distributions; The gyro frequency of cone is estimated in conjunction with echo cycle and cone asymmetry parameter; Cone height and facies basialis pyramidis radius is estimated according to the position of scattering point in instantaneous RD image.

Fig. 2 be obtain based on the method for parameter estimation of the simple scattering point cone target of time-frequency imaging instantaneous picture related coefficient is calculated to the instantaneous RD image of center of gravity translational alignment, can estimate that the echo cycle is 0.083s.

Fig. 3 is the time-frequency distributions curve map of the cone target obtained based on the method for parameter estimation of the simple scattering point cone target of time-frequency imaging.In Fig. 3, for in the time range with reference to continuous curve place corresponding to scattering point, also have two curves, and the moment that scattering point corresponding to curve can only be 0 in the Doppler frequency with reference to scattering point occurs simultaneously, therefore, can estimate that the asymmetry parameter of cone is 4, its result meets the cone model of Fig. 5.

Fig. 4 is the one group of instantaneous RD image obtained based on the method for parameter estimation of the simple scattering point cone target of time-frequency imaging, 25 and 35, N of two distance frequencies of the longitudinal lie farthest away of scattering point _r=35-25=10, the cone height that can estimate is l=0.95m, 37 and 95, N of two localizer units of the horizontal lie farthest away of scattering point _a=95-37=58, the facies basialis pyramidis radius that can estimate is d=0.216m.

Visible employing the inventive method all has very high precision for the estimated result of the gyro frequency of cone target, cone height and facies basialis pyramidis radius, and demonstrates validity of the present invention.

The emulation of the cone model of the simple scattering point under other asymmetry parameter, the precision that its estimated result is very high too, repeats no more herein.

Claims (3)

1., based on a method for parameter estimation for the simple scattering point cone target of time-frequency imaging, it is characterized in that, comprise the following steps:

Step 1, obtain the distance pulse pressure echo of the simple scattering point cone target of inverse synthetic aperture radar (ISAR) admission, this distance pulse pressure echo is distance verses time 2-D data;

Step 2, the distance verses time 2-D data of pulse pressure of adjusting the distance echo does Short Time Fourier Transform along slow time orientation, obtains distance verses time-Doppler's three-dimensional data, extracts the instantaneous RD image of cone target from distance verses time-Doppler's three-dimensional data;

Step 3, by the instantaneous RD image reform translational alignment of cone target to its geometric center, obtains the instantaneous RD image after translational alignment;

Step 4, to the instantaneous RD image after translational alignment, utilizes circulation auto-correlation and circular AMDF Copula, calculates the instantaneous picture related coefficient of the instantaneous RD image after translational alignment;

Wherein, circulation auto-correlation and circular AMDF Copula following formula represent:

$F\left(k\right)=\frac{\underset{i=1}{\overset{L}{\Σ}}(\Σ|I_i^{\′}(n,m)I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)\left|\right)}{\underset{i=1}{\overset{L}{\Σ}}(\Σ|I_{\mathrm{mod}(i+k,L)}^{\′}(n,m)-I_i^{\′}(n,m)\left|\right)},k=1,2,\mathrm{...}L/2$

Wherein, F (k) is instantaneous picture related coefficient, and k is time variable, and L represents the number of instantaneous RD image, I ' _i(n, m) represents t after translational alignment _ithe instantaneous RD image in moment, n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and mod (i+k, L) represents i+k L remainder;

Step 5, according to the instantaneous picture related coefficient obtained, therefrom takes out moment corresponding to instantaneous picture related coefficient maximal value place, as the echo cycle T of distance pulse pressure echo ₀estimated value;

Step 6, the time-frequency distributions curve according to distance pulse pressure echo estimates cone asymmetry parameter, and its concrete sub-step is as follows:

(1) from distance verses time-Doppler's three-dimensional data, take out the time m-doppler data that any distance frequency is corresponding, also just obtain the time-frequency distributions curve of the distance pulse pressure echo of this distance frequency;

(2) time-frequency distributions of distance pulse pressure echo is utilized to estimate the asymmetry parameter of cone, wherein, in the time-frequency distributions curve of distance pulse pressure echo, consider that simple scattering point there will be eclipse phenomena, the time-frequency distributions curve of scattering point disconnects, but the time-frequency distributions curve within the scope of the SEE time of scattering point in a swing circle is continuous print; Set certain scattering point as with reference to scattering point, be greater than semi-cone angle at the angle of radar line of sight and cone axis of symmetry, and under being less than the supplementary angle situation of semi-cone angle, for continuous time-frequency distributions Curves corresponding to reference scattering point time range in:

If a () does not have other time-frequency distributions curves, but be carved with other scattering points when the head and the tail with reference to continuous time-frequency distributions curve corresponding to scattering point when occurring, the asymmetry parameter N of cone ₀be 2;

If b () has two other time-frequency distributions curves to exist, and when scattering point corresponding to these two curves can not occur at any time simultaneously, the asymmetry parameter N of cone ₀be 3;

If c () has two other time-frequency distributions curves to exist, and the moment that scattering point corresponding to these two curves is only 0 when the Doppler with reference to scattering point occurs simultaneously, the asymmetry parameter N of cone ₀be 4;

If d () has four other time-frequency distributions curves to exist, so asymmetry parameter N of cone ₀be 5;

Step 7, the echo cycle of Binding distance pulse pressure echo and cone asymmetry parameter, estimate cone gyro frequency:

$f_0=\frac{1}{N_0{T}_0}$

Wherein, f ₀cone gyro frequency, N ₀the asymmetry parameter of cone, T ₀it is the echo cycle of distance pulse pressure echo;

Step 8, estimate cone height and facies basialis pyramidis radius, its concrete sub-step is as follows:

(1), on the instantaneous RD image before translational alignment or after translational alignment, utilize the distance frequency at two scattering point places of distance dimension lie farthest away to estimate the length of cone, and estimate according to the following formula:

Wherein, the height of cone that what l represented is, c represents propagation velocity of electromagnetic wave, N _rrepresent vertex of a cone scattering point and the maximum difference of boring the distance frequency arrived corresponding to end scattering point in instantaneous RD image, B represents the bandwidth of radar emission linear FM signal, represent the radar line of sight angle of pitch;

(2), on the instantaneous RD image before translational alignment or after translational alignment, utilize the localizer unit at two of azimuth dimension lie farthest away scattering point places to estimate facies basialis pyramidis radius, and estimate according to the following formula:

$d=\frac{{\mathrm{\λN}}_a}{8{\mathrm{\πf}}_0{N}_g}PRF$

Wherein, what d represented is facies basialis pyramidis radius, and λ represents electromagnetic wavelength, N _arepresent the maximum difference of the scattering point place localizer unit of facies basialis pyramidis, f ₀represent cone gyro frequency, N _grepresent that the window of time window function g (τ-t) is long, PRF indicating impulse repetition frequency;

Step 9, exports the parameter estimation result of simple scattering point cone target: instantaneous picture related coefficient, cone gyro frequency, cone height and facies basialis pyramidis radius.

2. the method for parameter estimation of the simple scattering point cone target based on time-frequency imaging according to claim 1, it is characterized in that, the concrete sub-step of step 2 is:

(1) for distance pulse pressure echo, do time frequency analysis according to the following formula along slow time orientation, obtain distance verses time-Doppler's three-dimensional data:

$Q(r,f,t)={\∫}_{-\mathrm{\∞}}^{\mathrm{\∞}}\[x(r,\mathrm{\τ})g^{*}(\mathrm{\τ}-t)\]e^{-j2\mathrm{\π}f\mathrm{\τ}}d\mathrm{\τ}$

Wherein, Q (r, f, t) signal x (r is represented, τ) in the result of the Short Time Fourier Transform of t, r represents distance frequency, and f represents Doppler frequency, t represents the slow time, x (r, τ) represents the echo sequence of distance frequency r, and τ is integration variable, g (τ-t) represents the time window function near t, N _gfor the window of time window function g (τ-t) is long;

(2) from distance verses time-Doppler's three-dimensional data, then take out the distance-Doppler data of any time according to the following formula, obtain the instantaneous RD image in one group of this moment:

I _i(r,f)＝Q(r,f,t＝t _i)

Wherein, I _i(r, f) represents t _ithe instantaneous RD image in moment, r represents distance frequency, and f represents Doppler frequency, and Q represents distance verses time-Doppler's three-dimensional data, and t represents the slow time, t _iit is a certain instantaneous moment.

3. the method for parameter estimation of the simple scattering point cone target based on time-frequency imaging according to claim 1, it is characterized in that, the concrete sub-step of step 3 is:

(1) according to the following formula instantaneous RD image is tieed up along distance, asks its distance dimension center of gravity:

$D_i(n,m)=\frac{{\mathrm{\Σ}}_{n=1}^N\left({\mathrm{nI}}_i\right(n,m))}{{\mathrm{\Σ}}_{n=1}^N{I}_i(n,m)}$

Wherein, D _i(n, m) represents t _imoment instantaneous RD image distance dimension center of gravity, n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and N represents the distance frequency number of instantaneous RD image, I _i(n, m) represents t _ithe instantaneous RD image in moment;

(2) according to the following formula to instantaneous RD image distance dimension center of gravity D _i(n, m), along azimuth dimension, asks the two-dimentional center of gravity of instantaneous RD image:

$P_i(n,m)=\frac{{\mathrm{\Σ}}_{m=1}^M\left({\mathrm{nD}}_i\right(n,m))}{{\mathrm{\Σ}}_{m=1}^M{D}_i(n,m)}$

Wherein, P _i(n, m) represents t _ithe two-dimentional center of gravity of moment instantaneous RD image, D _i(n, m) represents t _imoment instantaneous RD image distance dimension center of gravity, n represents the distance dimension index of instantaneous RD image, and m represents the azimuth dimension index of instantaneous RD image, and M represents the localizer unit number of instantaneous RD image;

(3) the two-dimentional center of gravity position translation of the instantaneous RD image of trying to achieve is snapped to the geometric center position of instantaneous RD image, make t before translational alignment _ithe instantaneous RD image I in moment _i(r, f) is transformed to t after translational alignment _ithe instantaneous RD image I in moment _i' (r, f).