CN101846741A - Inverse synthetic aperture radar imaging data segment selecting method - Google Patents

Abstract

The invention relates to an inverse synthetic aperture radar imaging data segment selecting method. In the method, image contrast is taken as a standard of evaluating image focusing degree, and the optimal imaging data segment selecting method can be simply expressed by a mathematical expression (topt, delta topt)=arg[max(IC)], wherein the topt is the central time of the optimal imaging data segment; the delta topt is the length of the optimal imaging data segment; and the IC is the image contrast; the key for determining the optimal data segment is to calculate the topt and the delta topt from the expression; and the method comprises the following four steps of: 1, generating an ISAR image sequence; 2, calculating image sequence contrast; 3, estimating the central time topt of the optimal data segment; and 4, estimating the length delta topt of the optimal data segment. Images with high resolution and good focusing can be generated according to the optimal data segment.

Description

A kind of inverse synthetic aperture radar imaging data segment selecting method

Technical field

The present invention relates to a kind ofly, belong to the radar imagery technical field based on the maximized inverse synthetic aperture radar (ISAR) imaging data segment selecting method of picture contrast.

Background technology

Radar imagery not only the resolution height, penetrability is good, antijamming capability is strong, and can be round-the-clock, round-the-clock work can remedy the deficiency of visible light and infrared imagery technique.ISAR is another the radar imagery technology that grows up on SAR (synthetic-aperture radar) basis, and purpose is to solve the imaging problem of moving target.In the practical application, ISAR is except fixing on the ground, also can be contained on the motion platforms such as aircraft, naval vessel,, in strategic defensive, anti-satellite, tactical weapon and radar astronomy, significant application value be arranged all the motive target imaging on aerial, sea or ground.

Distance-Doppler (R-D) formation method is still the classical way of ISAR imaging, and the ultimate principle of R-D imaging can illustrate with mount model.Fig. 1 is the mount model imaging schematic diagram, and hypothetical target is positioned on the rotatable platform, and target is wound perpendicular to the x-y plane with uniform angular velocity and rotated.Distance between radar and the target rotation center O is r _a, target is ω around the angular velocity of O point rotation.Generally speaking, the distance between radar and the target is greater than the physical dimension (r of target _ar ₀).So, at constantly slow time t _m, certain a bit (r on the target ₀, θ ₀) to the distance of radar be:

r(t _m)＝[r ₀ ²+r _a ²-2r ₀r _a?cos(θ ₀+ωt _m+π/2)] ^1/2

(1)

≈r _a+x ₀sin(ωt _m)+y ₀cos(ωt _m)

Wherein, (x ₀, y ₀) expression point target A initial imaging time is at the position coordinates of vertical and horizontal.

The Doppler frequency of echoed signal is by the radial motion speed v of scattering point on the target with respect to radar _rSo decision is Doppler frequency f _dCan be write as:

$f_d=\frac{2v_r}{\mathrm{\&lambda;}}=\frac{2}{\mathrm{\&lambda;}}\frac{\mathrm{dr}\left(t_m\right)}{d{t}_m}=\frac{2x_0\mathrm{\&omega;}}{\mathrm{\&lambda;}}\cos \left(\mathrm{\&omega;}{t}_m\right)-\frac{2{\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="HSA00000118371600022.png,HSA00000118371600031.png"\; state="\{\{state\}\}">y</figure-callout>}}_0\mathrm{\&omega;}}{\mathrm{\&lambda;}}\sin \left(\mathrm{\&omega;}{t}_m\right)---\left(2\right)$

Suppose a synthetic aperture in the time angle θ=ω t that target turns over _mVery little, then formula (1) and (2) can be approximately:

r(t _m)＝r _a+y ₀ (3)

$f_d=\frac{2x_0\mathrm{\&omega;}}{\mathrm{\&lambda;}}---\left(4\right)$

This shows, utilize distance time-delay and the Doppler frequency of analyzing echoed signal, just can determine scattering point (x ₀, y ₀) the position.Equidistant plane is one group of parallel plane perpendicular to radar line of sight (RLOS) direction, and etc. the Doppler frequency plane parallel in the plane that forms by target rotating shaft and RLOS, as shown in Figure 1.Formula (4) is an approximate expression of deriving under the condition of target rotation angle little (3 °-5 °), and the signal processing method that can utilize distance-Doppler's two dimension to divide in this case carries out imaging.But when θ was too big when the echo data time is longer, formula (4) just had been false, and Doppler frequency no longer is approximately constant, and therefore simple service range-Doppler's method imaging meeting causes image blurring.

In addition, can be based on the ISAR formation method of time frequency analysis to the long echo imaging, this method is the time-frequency distributions that obtains echo by time frequency analysis, is stitched together by the time frequency analysis result each range unit, obtains target image sequence.But can't determine best the imaging moment and optimal imaging data segment, length by the method.

In order to remedy the deficiency of above-mentioned formation method, the present invention has designed a kind of based on the maximized optimal imaging data segment selecting method of contrast, can realize choosing the optimal imaging data segment automatically from the long echo data, is used to generate the ISAR image of focusing.

Summary of the invention

The problem that solves:

The object of the present invention is to provide a kind of inverse synthetic aperture radar imaging data segment selecting method, is a kind of based on the maximized optimal imaging data segment selecting method of contrast, realizes obtaining the resolution height from the long echo data, focuses on good ISAR image.

Technical scheme:

This method is with the standard of picture contrast as the assessing image focus degree, and the system of selection of optimal imaging data segment can be shown by simple table with mathematic(al) representation:

(t _opt，Δt _opt)＝arg[max(IC)] (5)

T wherein _OptBe the center moment of optimal imaging data segment, Δ t _OptBe the length of optimal imaging data segment, IC is a picture contrast.As seen the key of determining the optimum data section solves t exactly from (5) formula _OptWith Δ t _Opt, the solution that provides is divided into following four steps:

Step 1 generates the ISAR image sequence

To the echoed signal after the motion compensation, a given initial data length Δ t _InWith the sliding step of data segment, echo data according to primary data section and sliding step segmentation, is carried out bidimensional Fast Fourier Transform (FFT) (2DFT) imaging to each echo data section, obtain a series of ISAR image sequences.

Step 2 sequence of computed images contrast

The intensity of each image in the image sequence that obtains in the calculation procedure one respectively.Definition image intensity expression formula is:

$I(\mathrm{\&eta;},v;t,\mathrm{\&Delta;t})={\left|2\mathrm{DFT}\left(s(\hat{t},t_m)\right)\right|}^2---\left(6\right)$

Wherein

Be certain section echoed signal, (η υ) is the coordinate of image, (t, Δ t) be the data segment chosen the center constantly and length.

The contrast of computed image again, the definition of picture contrast (IC) is:

$\mathrm{IC}(t,\mathrm{\&Delta;t})=\frac{\sqrt{m\left\{\right[I(\mathrm{\&eta;},v;t,\mathrm{\&Delta;t})-m{\left[I(\mathrm{\&eta;},v;t,\mathrm{\&Delta;t})\right]]}^2}}{m\left[I(\mathrm{\&eta;},v;t,\mathrm{\&Delta;t})\right]}---\left(7\right)$

Wherein, m represents average, I presentation video intensity.Contrast is the normalization standard variance of image intensity.When image focusing, the intensity of scattering point is big, so use the criterion of contrast as image focusing.

Step 3 is estimated the center moment t of optimum data section _Opt

The contrast of searching image sequence (IC) is about the maximal value of t, and the center of data segment of getting IC maximal value correspondence is constantly as the center of optimum data section t constantly _Opt

Step 4 is estimated the length Δ t of optimum data section _Opt

Determine the center moment t of optimum data section _OptAfter, the length Δ t of conversion primary data section calculates the IC of the data segment imaging of different length, finds the maximal value of IC about Δ t, and the data length of getting IC maximal value correspondence is as optimum data length Δ t _Opt, detailed process is:

(3) with t _OptBe the center, data length is increased 2n (representative value of n is 4,5,6) on the basis of initial length, and calculate the IC of the image that generates thus.If IC has increased, data length is increased 2n again, no longer increase up to IC.

(4) data length last in (1) is increased by 2 (n-k) more one by one, (k=1 ..., n), and calculate IC, obtain maximal value up to IC.

If Δ t _InWhen increasing 2n for the first time, IC has reduced, then with Δ t _InReduce 2n gradually, no longer increase up to IC.Then, make data length reduce 2 (n-k) one by one, (k=1 ..., n), up to obtaining the IC maximal value.

Beneficial effect:

The present invention is a kind of inverse synthetic aperture radar imaging data segment selecting method, can realize choosing automatically the optimum data section center moment and the data segment, length that are used for imaging from long echo data, thereby avoid directly using the R-D imaging to cause image blur.The ISAR image resolution ratio height that obtains by the method, focus on, can provide high-quality ISAR image for follow-up radar target recognition and detection.

Description of drawings

Fig. 1 is the mount model imaging schematic diagram.

Fig. 2 is the process flow diagram of the inventive method.

Fig. 3 is a Doppler frequency change curve in time.

Fig. 4 is the image sequence contrast figure of three kinds of different signal-noise ratio signals.

Fig. 5 is that the lateral separation of three kinds of data lengths is as map of magnitudes.

Embodiment

In conjunction with the accompanying drawings, describe embodiments of the present invention in detail by emulation.Fig. 2 is the implementing procedure figure of the inventive method.

Be the validity of checking this method, at first the echo data of a scattering point of emulation is specifically chosen process with this echo data explanation optimal imaging data segment.Radar parameter is in the artificial echo: wavelength X=0.03m, bandwidth B=100MHz, imaging accumulated time T=0.5s, pulse repetition rate prf=4000Hz.The hypothetical target rotational speed ω=π rad/s that remains unchanged, the scattering point initial coordinate is (x ₀, y ₀)=(10,0), radar and the distance R=20000m of target's center.

By Doppler frequency expression formula (3) the Doppler frequency prior imformation of this echo as can be known:

$f_d=\frac{2x_0\mathrm{\&omega;}}{\mathrm{\&lambda;}}\cos \left(\mathrm{\&omega;}{t}_m\right)-\frac{2{\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="HSA00000118371600022.png,HSA00000118371600031.png"\; state="\{\{state\}\}">y</figure-callout>}}_0\mathrm{\&omega;}}{\mathrm{\&lambda;}}\sin \left(\mathrm{\&omega;}{t}_m\right)=\frac{2\&times;10\&times;\mathrm{\&pi;}}{0.03}\cos \left(\mathrm{\&pi;}{t}_m\right)$

Fig. 3 is the Doppler frequency change curve in time of this echo, even visible target rotational speed is a constant, becomes when Doppler frequency is still, thereby causes the R-D image blur.Doppler frequency changes slowly constantly at t=0 as can be seen from Figure 3, so optimum data section center constantly in theory should be at the t=0 place.

Adopt method provided by the invention to estimate the center moment and the length of optimal imaging data segment below:

Step 1, supposition primary data length are 256 sampled point (Δ t _In=0.064s), sliding step is 32 sampled points (0.008s).Respectively echo data being added signal to noise ratio (S/N ratio) is the white Gaussian noise of 10dB and 0dB.The target echo signal expression formula is:

$s_r(\hat{t},t_m)=a_r\left(\hat{t}-\frac{2r\left(t_m\right)}{c}\right)a_a\left(t_m\right)\exp \left(\mathrm{j\&pi;\&gamma;}{\left(\hat{t}-\frac{2r\left(t_m\right)}{c}\right)}^2\right)\exp (-j\frac{4\mathrm{\&pi;}}{\mathrm{\&lambda;}}r\left(t_m\right))+n(\hat{t},t_m)---\left(8\right)$

Wherein

Represent the fast time, λ is the wavelength that transmits, and γ is the frequency modulation rate,

Be noise.

According to Δ t _InTo with sliding step

Segmentation and imaging obtain comprising the image sequence of 55 ISAR subimages.

Step 2, use image intensity expression formula

The intensity of sequence of computed images is designated as I respectively _k(η, υ; t _k, Δ t _In), (k=1,2 ..., 55).Wherein

Be certain data segment echoed signal, (η υ) is the coordinate of image, and (t, Δ t) is the center moment and the length of this data segment.

Re-use picture contrast (IC) expression formula

The contrast of subimage in the sequence of computed images is designated as IC (t respectively _k, Δ t _In), (k=1,2 ..., 55).In the IC expression formula, m represents average, I presentation video intensity.Figure 4 shows that the contrast result of calculation of image sequence correspondence.

Step 3, search out contrast maximal value moment corresponding in the image sequence, t=0 constantly.T=0 is exactly the center moment t of optimum data section constantly _OptAs seen from Figure 4, noise only can influence the amplitude of picture contrast, does not influence the peak of contrast, the position of image sequence maximum-contrast all the time constantly at t=0, to provide prior imformation consistent for Doppler frequency among estimated result and Fig. 3.

The length Δ t of step 4, estimation optimum data section _OptBe divided into for two steps:

(1) with t _OptBe the center, with data length Δ t at Δ t _InIncrease 2n (n=4 in this example) on the basis, and calculate the contrast IC (t of the image that generates thus _Opt, Δ t).If IC is (t _Opt, Δ t)＞IC (t _Opt, Δ t _In), then Δ t continues to increase 2n, no longer increases up to IC.

(2) Δ t is increased by 2 (n-k) more one by one, (k=1 ..., n), and calculate IC, obtain maximal value up to IC.

When if Δ t increases 2n for the first time in (1), IC (t _Opt, Δ t _In)＞IC (t _Opt, Δ t), then Δ t is reduced 2n gradually, no longer increase up to IC.Then, make Δ t reduce 2 (n-k) one by one, (k=1 ..., n), obtain maximal value up to IC.

The data length Δ t of IC maximal value correspondence is exactly optimum data segment length Δ t _OptThe optimum data length that estimates in this example is 264 sampled points, i.e. Δ t _Opt=0.066s.Fig. 5 has provided optimum length (264 sampled points), long data length (444 sampled points) and the lateral separation of short three kinds of data lengths of data length (66 sampled points) as map of magnitudes.As seen long data length is owing to Doppler's the very high secondary lobe of Shi Bianyou, and the lateral resolution of short data length is low, and the optimum data section had both guaranteed that secondary lobe was low, had higher lateral resolution again.

In sum, the optimum data section center that the present invention estimates constantly is positioned at Doppler frequency and changes the slowly time period, and the optimum length of data segment has resolution ratio height, advantage that secondary lobe is low. Therefore, the optimum data section of using the present invention to select is carried out imaging, can generate the ISAR image that focusing effect is good, resolution ratio is high.

Claims (2)

1. inverse synthetic aperture radar imaging data segment selecting method, this method is with the standard of picture contrast as the assessing image focus degree, and the system of selection of optimal imaging data segment can be shown by simple table with mathematic(al) representation:

(t _opt，Δt _opt)＝arg[max(IC)] (5)

T wherein _OptBe the center moment of optimal imaging data segment, Δ t _OptBe the length of optimal imaging data segment, IC is a picture contrast; The key of determining the optimum data section solves t exactly from (5) formula _OptWith Δ t _Opt, specifically be divided into following four steps:

Step 1 generates the ISAR image sequence

To the echoed signal after the motion compensation, a given initial data length Δ t _InWith the sliding step of data segment, echo data according to primary data section and sliding step segmentation, is carried out bidimensional Fast Fourier Transform (FFT) imaging to each echo data section, obtain a series of ISAR image sequences;

Step 2 sequence of computed images contrast

The intensity of each image in the image sequence that obtains in the calculation procedure one respectively; Definition image intensity expression formula is:

$I(\mathrm{\&eta;},\mathrm{\&upsi;};t,\mathrm{\&Delta;t})={\left|2\mathrm{DFT}\left(s(\hat{t},t_m)\right)\right|}^2---\left(6\right)$

Wherein Be certain section echoed signal, (η υ) is the coordinate of image, (t, Δ t) be the data segment chosen the center constantly and length;

The contrast of computed image again, the definition of IC is:

$\mathrm{IC}(t,\mathrm{\&Delta;t})=\frac{\sqrt{m\left\{{[I(\mathrm{\&eta;},\mathrm{\&upsi;};t,\mathrm{\&Delta;t})-m[I(\mathrm{\&eta;},\mathrm{\&upsi;};t,\mathrm{\&Delta;t})\left]\right]}^2\right\}}}{m\left[I(\mathrm{\&eta;},\mathrm{\&upsi;};t,\mathrm{\&Delta;t})\right]}---\left(7\right)$

Wherein, m represents average, I presentation video intensity; Contrast is the normalization standard variance of image intensity; When image focusing, the intensity of scattering point is big, so use the criterion of contrast as image focusing;

Step 3 is estimated the center moment t of optimum data section _Opt

The contrast of searching image sequence is about the maximal value of t, and the center of data segment of getting IC maximal value correspondence is constantly as the center of optimum data section t constantly _Opt

Step 4 is estimated the length Δ t of optimum data section _Opt

Determine the center moment t of optimum data section _OptAfter, the length Δ t of conversion primary data section calculates the IC of the data segment imaging of different length, finds the maximal value of IC about Δ t, and the data length of getting IC maximal value correspondence is as optimum data length Δ t _Opt

2. a kind of inverse synthetic aperture radar imaging data segment selecting method according to claim 1 is characterized in that: the detailed process of described step 4 is:

(1) with t _OptBe the center, data length is increased 2n (representative value of n is 4,5,6) on the basis of initial length, and calculate the IC of the image that generates thus; If IC has increased, data length is increased 2n again, no longer increase up to IC;

(2) data length last in the said process (1) is increased by 2 (n-k) more one by one, (k=1 ..., n), and calculate IC, obtain maximal value up to IC;

If Δ t _InWhen increasing 2n for the first time, IC has reduced, then with Δ t _InReduce 2n gradually, no longer increase up to IC; Then, make data length reduce 2 (n-k) one by one, (k=1 ..., n), up to obtaining the IC maximal value.

Images