CN102508244B - Ground moving target detection and parameter estimation method - Google Patents

Abstract

The invention provides a ground moving target detection and parameter estimation method. The technical scheme realizes the moving target detection and two-dimensional velocity estimation by combining with an interference image sequence with multiple sub-bands and sub-apertures on the basis of a three-way ultra-wide bandwidth (UWB) synthetic aperture radar (SAR) image. The specific steps of the method are as follows: 1, generating a sub-band sub-aperture image sequence; 2, generating a sub-band sub-aperture interference image sequence; 3, detecting the moving target; and 4, estimating the two-dimensional velocity of the moving target. By using the ground moving target indication and parameter estimation method provided by the invention, the moving target and a static target can be distinguished on an imaging position and a phase position; the radial velocity of the moving target can be estimated clearly; and the azimuth speed and distance speed can also be estimated.

Description

A kind of ground moving object detects and method for parameter estimation

Technical field

The invention belongs to the SAR signal processing technical field, relating to a kind of ground moving object detects and method for parameter estimation, especially utilizing triple channel UWB SAR (Ultra Wide-band Synthetic Aperture Radar, Research on UWB SAR) image to carry out ground moving object detects and method for parameter estimation.

Background technology

UWB SAR system has the foliage penetration ability, can obtain simultaneously two-dimentional high-resolution imaging radar.Utilize hyperchannel UWB SAR system to realize GMTI (Ground Moving Target Indication, ground moving object indication) function can be scouted the static and moving target of in a big way open smooth or leafage covering area, supervision and imaging.Utilize motion different from the phase place of static target at UWB SAR image, carry out moving object detection, owing to obtained the long relevant processing time, stronger to the detectability of microinching target.These class methods change detection, along the rail interference etc.Since the phase delta phi of moving target by distance to speed v _yDetermine, there is the deficiency of two aspects in these class methods: (1) is because Δ _φTake 2 π as the cycle, even so v _yThe Δ non-vanishing, that it is corresponding _φAlso may be zero, produce the blind speed phenomenon when detecting target; In addition, different v _yCorresponding Δ _φMay be identical, produce distance during estimated parameter to the velocity ambiguity problem; (2) because phase delta _φOnly and v _yRelevant, so these class methods can only detect with estimated distance to the non-vanishing moving target of speed, and can't detect distance to speed be zero and the orientation to the non-vanishing moving target of speed, also can't estimate to speed the orientation of target.

Above blind speed and velocity ambiguity problem, usually adopt multifrequency or many baselines mode to solve, the multifrequency scheme needs a plurality of signal emitting-sources, and many baseline schema need four receiving antennas to form three not redundant baselines at least, then need to suppress more hyperchannel if consider clutter.In addition, Canadian RADARSAT-2 system has proposed the antenna operating mode of many triggerings, has realized the effect of many baselines with two passages by increasing virtual phase center.The wavelength of UWB SAR system is longer, require long baseline, be subjected to the restriction of antenna size and platform size, at four receiving antennas of carrier aircraft device and to form virtual base very difficult, too much number of antennas has increased hardware and storage burden simultaneously, adopts multifrequency or many trigger methods also to improve the complicacy of system.

Summary of the invention

The objective of the invention is the image based on triple channel UWB SAR, unite many subbands multiple sub-apertures interference image sequence, realize the estimation of moving object detection and two-dimension speed.

The sub-subaperture image of different subbands has different subband center frequency f _mWith sub-aperture view angle theta _nOn each the sub-aperture of width of cloth subband interference image, the interference image of target has following rule:

(1) about image space: the image space of moving target interference image in each the sub-aperture of width of cloth subband interference image with θ _nChange continuously, and and f _mIrrelevant.The image space of static target interference image all is constant in each the sub-aperture of width of cloth subband interference image.

(2) about the interference image amplitude: when the RCS of moving target changes when very little with visual angle and centre frequency, the amplitude of its interference image is approximate constant in each the sub-aperture of width of cloth subband interference image.If but the amplitude of moving target interference image in certain the sub-aperture of width of cloth subband interference image is zero, the blind speed phenomenon namely occurred.The interference image amplitude perseverance of static target in each the sub-aperture of width of cloth subband interference image is zero.

(3) about interferometric phase: moving target is θ without fuzzy interferometric phase _nContinuous function.The interferometric phase of static target is always zero.

Technical scheme of the present invention is that a kind of ground moving object detects and method for parameter estimation, it is characterized in that, may further comprise the steps:

The first step generates the sub-subaperture image sequence of subband

At first the UWB SAR image of three passages is done two-dimensional Fourier transform to wavenumber domain (k _x-k _r), adopt again bandpass filter respectively the UWB SAR image of each passage of transforming to wavenumber domain to be carried out digital filtering, the passband of bandpass filter is provided by following formula:

$\left.\begin{array}{c}|\sqrt{{\mathrm{<figure-callout\; id="2"\; label="k\; x"\; filenames="HDA0000106254760000031.png"\; state="\{\{state\}\}">k</figure-callout>}}_x^{}+{(k_r+k_c)}^2}-\frac{{4\mathrm{\&pi;f}}_m}{c}|\&le;\frac{4\mathrm{\&pi;}\&CenterDot;\mathrm{\&Delta;f}}{c}\\ \tan ({\mathrm{\&theta;}}_n-\frac{\mathrm{\&Delta;\&Theta;}}{2})\&le;\frac{k_x}{k_r+k_c}\&le;\tan ({\mathrm{\&theta;}}_n+\frac{\mathrm{\&Delta;\&Theta;}}{2})\end{array}\right\}$ (formula one)

Wherein c is propagation velocity of electromagnetic wave, k _cThe carrier frequency wave number of UWB SAR system, f _m(m=1,2..., M) is subband center frequency, θ _n(n=1,2..., N) is visual angle, sub-aperture, and Δ f is the subband bandwidth, and Δ Θ is accumulation angle, sub-aperture.M divides the subband number that obtains with the bandwidth of UWB SAR.N divides the sub-aperture number that obtains with the aperture of UWB SAR.The value of M, N, Δ f and Δ Θ is determined according to the actual requirements.

Note is generated by the UWB SAR image of the individual passage of l (l=1,2,3), subband center frequency is f _m, visual angle, sub-aperture is θ _nThe sub-subaperture image of subband be $S_{\mathrm{mn}}^{\left(l\right)}(m=\mathrm{1,2}...,M,n=\mathrm{1,2}...,N).$

Will With Carry out registration again clutter suppress to obtain image

Will

With

Carry out registration again clutter suppress to obtain image

Second step generates the sub-aperture of subband interference image sequence

Utilize following formula to calculate subband aperture interference image sequence I _Mn(m=1,2..., M, n=1,2..., N):

$I_{\mathrm{mn}}=S_{\mathrm{mn}}^{\left(A\right)}\&CenterDot;[S_{\mathrm{mn}}^{\left(B\right)}{]}^{*}$ (formula two)

Wherein * represents to get conjugation.

The 3rd step was detected moving target

Antithetical phrase belt aperture interference image sequence I _Mn(m=1,2..., M, n=1,2..., N) carries out following two operations:

Operation one arranges Phase Threshold and amplitude threshold, to every width of cloth I _MnCarry out CFAR (Constant FalseAlarm Rate, CFAR) and detect, Phase Threshold and amplitude threshold are determined by concrete applied environment.

If operation two is not at any width of cloth I _MnIn detect moving target, then the technical program stops.If at least one width of cloth I _MnIn can detect moving target, to detect the maximum width of cloth I of moving target number _Mn(be designated as I as the reference image _Ref), if two width of cloth and above I are arranged _MnDetected moving target number is maximum and identical, selects the highest width of cloth of signal to noise ratio (S/N ratio) as the reference image I _RefTo I _RefIn detected each moving target, carry out one by one following joint detection process:

With I _RefCentered by, and with this moving target at I _RefIn the position be reference point, on other sub-aperture of all subbands interference image, in the distance range of the twice resolution element centered by reference point the search whether operate in addition a detected moving target: if do not search moving target, change I over to _RefThe joint detection process of middle next moving target; If interference image searches moving target in the sub-aperture of at least one width of cloth subband, then the range value with the moving target that searches in the afore-mentioned distance scope in the interference image of the sub-aperture of all subbands extracts, consist of a new set A, total M * N range value in the set A.The sub-subaperture image of subband that does not search moving target is made as zero corresponding to the range value in the set A.The standard deviation of range value among the set of computations A, with amplitude in itself and the set A peaked half relatively: if standard deviation greater than amplitude among the A peaked half, then this target not united detect, no longer process this target, change I over to _RefThe joint-detection of next moving target; If standard deviation be not more than among the A amplitude peaked half, then finish the joint-detection of this moving target, and this moving target put into detect goal set.

The 4th step was estimated the two-dimension speed of moving target

To each moving target in the detection goal set of the 3rd step, estimate its radial velocity and two-dimension speed according to following steps one by one.

(1) step, the radial velocity of estimation moving target.

Extract the phase sequence Δ φ of moving target _Mn(m=1,2..., M, n=1,2..., N) estimates that according to projection theorem this moving target is at different θ _nRadial velocity under (n=1,2..., N)

Value realized by following operation:

Many subbands estimation model is set up in operation ():

(formula three) be v wherein _a, B,

Represent respectively corresponding carrier aircraft speed, base length and the antenna incident angle constantly of UWB SAR image data acquiring, k _m(m=1,2 ..., M) be

Relative f _mAmbiguity period, be integer to be found the solution,

With respect to independent variable k _mDependent variable, thereby be designated as

To each m, give k _mRespectively 7 integers of assignment-3 in 3 scopes calculate 7 by formula three

Travel through all m, obtain the M group, 7 every group

Operation (two) is organized at M In, every group optional one Formation contains the set of M element, and calculates the variance of this set.

Operation (three) repetitive operation (two) travels through all possible combination, altogether can obtain 7 ^MThe variance of each set is calculated in individual set.

One of variance minimum is selected in operation (four) from the set of operation (three) gained, the mean value of this set is exactly estimated value

(2) step, the two-dimension speed of estimation moving target.

Utilize the radial velocity of (one) step estimation

Estimate the two-dimension speed of this moving target by formula four, namely the orientation is to speed

With the distance to speed

(formula four)

Wherein, $\mathrm{\&alpha;}=\underset{n=1}{\overset{\mathrm{<figure-callout\; id="2"\; label="N\; cos"\; filenames="HDA0000106254760000031.png"\; state="\{\{state\}\}">N</figure-callout>}}{\mathrm{\&Sigma;}}}\cos {2\mathrm{\&theta;}}_n,$ $\mathrm{\&beta;}=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}\sin {2\mathrm{\&theta;}}_n.$

Adopt the present invention can reach following technique effect:

1. owing to having utilized large bandwidth and the large field angle characteristic of UWB SAR system, can obtain a series of sub-aperture of subband interference images of having indicated different center frequency and visual angle, moving target and static target can be distinguished at image space and phase place.

2. aspect moving object detection, if blind speed appears in certain moving target, be that its amplitude in certain the sub-aperture of width of cloth subband interference image is zero, in the 3rd step of technical solution of the present invention, standard deviation and maximal value by relatively its amplitude set in the interference image sequence of the sub-aperture of subband, still it can be detected, namely eliminate blind speed; If certain moving target has the orientation to speed, because its radial velocity is non-vanishing, and the 3rd step of technical solution of the present invention is to detect the moving target with radial velocity, thereby also it can be detected.

3. aspect the moving target parameter estimation, by technical solution of the present invention many subbands estimation model in the 4th step, can be without the radial velocity of estimating faintly moving target; And then utilizing formula four, the orientation that can estimate moving target is to speed and distance to speed.

Description of drawings

Fig. 1 is realization flow figure of the present invention;

Fig. 2 is the schematic diagram of UWB SAR image spectrum and the sub-subaperture image frequency spectrum of some subbands;

Fig. 3-Fig. 5 utilizes the present invention certain airborne measured data to be carried out the processing example of moving object detection and parameter estimation;

Fig. 3 is first passage f _mThe sub-subaperture image of the subband at each visual angle during for 400MHz;

Fig. 4 is f _mThe sub-aperture of the subband at each visual angle interference image during for 400MHz;

Fig. 5 is the data result of processing example.

Embodiment

Fig. 1 is realization flow figure of the present invention.Whole flow process mainly was divided into for four steps: the sub-subaperture image sequence of first step generation subband, second step generate the sub-aperture of subband interference image sequence, the 3rd step detection moving target, the 4th step estimation moving target radial velocity and two-dimension speed.

The first step will be noted the selection at subband bandwidth, accumulation angle, sub-aperture, subband center frequency, visual angle, sub-aperture in the process that the sub-subaperture image of subband generates.The relation that needs consideration resolution and moving target to defocus.If target velocity is higher, then need lower resolution, subband bandwidth, sub-aperture accumulation angle can be smaller; If target velocity is lower, then need higher resolution, subband bandwidth, sub-aperture accumulation angle can be larger.Subband center frequency and visual angle, sub-aperture must be all in UWB SAR bandwidth and angulars field of view.

The registration of the sub-subaperture image of subband can adopt conventional method, referring to " multiband SAR images registration and integration technology research " (Chang Yulin, Master of engineering academic dissertation, the National University of Defense technology, 2005).Clutter suppresses to adopt DPCA (Displaced Phase Centre Antenna, DPCA) method, referring to " modern DPCA technical research " (Wang Yongliang etc., " electronic letters, vol ",, the 28th volume, the 6th phase in 2000).DPCA is linear operation, does not affect the linear relationship of target phase and speed.

The CFAR in the 3rd step detects referring to " target detection and discriminating in the UWB SAR image " (Fang Xueli, Doctor of engineering academic dissertation, the National University of Defense technology, 2004).

In the 3rd step the moving target range value and in the 4th step the extraction of moving target phase value referring to " based on the binary channels UWB SAR moving object detection of ATI and distance to velocity estimation ", " signal processing " (Chang Yulin, signal is processed, 2008, the 24th volume, the 5th phase).

Fig. 2 is the schematic diagram of the sub-subaperture image frequency spectrum of subband.Large fan ring is UWB SAR image spectrum, and the width on its angle and the longitudinal axis is respectively aperture and the bandwidth of UWB SAR; Little fan ring (oblique line frame) in the large fan ring is the sub-subaperture image frequency spectrum of a width of cloth subband, and the frequency that the large fan ring center of circle is departed from the little fan ring center of circle is f _m, the expression subband center frequency, the angle that the large fan ring center of circle is departed from the little fan ring center of circle is θ _n, representing visual angle, sub-aperture, the angle of little fan ring is sub-aperture accumulation angle Δ Θ, the inside and outside semidiameter of little fan ring is subband bandwidth deltaf f.

Fig. 3-Fig. 5 adopts the present invention to carry out the example that measured data is processed.

Fig. 3 is to first passage, utilizes the f that bandpass filter generates in the first step _mBe 400MHz, each visual angle, the sub-subaperture image sequence of subband that do not have clutter to suppress, transverse axis be the orientation to, the longitudinal axis be oblique distance to, covered the scope of 479 * 999m (orientation * oblique distance).M is taken as 5, N and is taken as 3.θ _nFrom (a) to (e) is respectively-9 ° ,-4.5 °, 0 °, 9 °, 4.5 °.Having comprised 4 moving targets in the scene, is respectively near bright short line segment " 1 ", " 2 " in the sub-subaperture image of each width of cloth subband, " 3 ", " 4 " numeral.

Fig. 4 utilizes the first step of the present invention and second step to process f _mThe sub-aperture of the subband at each visual angle interference image during for 400MHz, visual angle, sub-aperture from (a) to (e) are respectively-9 ° ,-4.5 °, 0 °, 9 °, 4.5 °, (f) are angle value corresponding to different colours.Can observe 4 short line segments that color is darker than surrounding pixel, be 4 moving targets, can be detected at an easy rate.

Fig. 5 utilizes technical scheme of the present invention, to the estimated result of moving target two-dimension speed.Listed 4 moving target orientation among the figure to speed with apart from actual value, estimated value and error to speed, and illustrated each target is at which f _mAnd θ _nThe sub-aperture of subband interference image in blind speed and radial velocity blooming have appearred.

That blind speed occurs is target 2 (f _m=400MHz, θ _nIn the time of=-9 °), target 3 (f _m=450MHz, θ _nIn the time of=4.5 °); That the appearance radial velocity is fuzzy is target 2 ((f _m=400,450MHz, all θ _n), target 3 (f _m=350,400,450MHz, all θ _n), target 4 (f _m=350,400,450MHz, all θ _n).Estimate that at last each target velocity that obtains does not have blind speed, and it is fuzzy to have eliminated radial velocity.

All moving targets all have been detected out, are not subject to the impact of blind speed, and radial velocity is all by without fuzzy estimating, and the absolute error of orientation to speed and distance to the velocity estimation value is no more than 0.31m/s, and relative error is no more than 10%.Illustrate that technical solution of the present invention can overcome blind speed and radial velocity blooming, with the orientation of higher precision estimating target to speed and distance to speed.

Claims (1)

1. a ground moving object detects and method for parameter estimation, it is characterized in that, may further comprise the steps:

The first step, the sub-subaperture image sequence of generation subband;

At first UWB SAR (Ultra Wide-band Synthetic Aperture Radar, the Research on UWB SAR) image of three passages is done two-dimensional Fourier transform to wavenumber domain (k _x-k _r), adopt again bandpass filter respectively the UWB SAR image of each passage of transforming to wavenumber domain to be carried out digital filtering, the passband of bandpass filter is provided by following formula:

$\left.\begin{array}{c}|\sqrt{k_x^2+{(k_r+k_c)}^2}-\frac{4\mathrm{\&pi;}{f}_m}{c}|\&le;\frac{4\mathrm{\&pi;}\&CenterDot;\mathrm{\&Delta;f}}{c}\\ \tan ({\mathrm{\&theta;}}_n-\frac{\mathrm{\&Delta;\&Theta;}}{2})\&le;\frac{k_x}{k_r+k_c}\&le;\tan ({\mathrm{\&theta;}}_n+\frac{\mathrm{\&Delta;\&Theta;}}{2})\end{array}\right\}$ (formula one)

Wherein c is propagation velocity of electromagnetic wave, k _cThe carrier frequency wave number of UWB SAR system, f _mSubband center frequency, m=1,2..., M, θ _nVisual angle, sub-aperture, n=1,2..., N, Δ f are the subband bandwidth, Δ Θ is accumulation angle, sub-aperture; M divides the subband number that obtains with the bandwidth of UWB SAR; N divides the sub-aperture number that obtains with the aperture of UWB SAR; The value of M, N, Δ f and Δ Θ is determined according to the actual requirements;

Note is generated by the UWB SAR image of l passage, subband center frequency is f _m, visual angle, sub-aperture is θ _nThe sub-subaperture image sequence of subband be

L=1,2,3;

Will

With

Carry out registration again clutter suppress to obtain image sequence

Will

With

Carry out registration again clutter suppress to obtain image sequence

Second step, the sub-aperture of generation subband interference image sequence;

Utilize following formula to calculate subband aperture interference image sequence I _Mn:

$I_{\mathrm{mn}}=S_{\mathrm{mn}}^{\left(A\right)}\&CenterDot;{\left[S_{\mathrm{mn}}^{\left(B\right)}\right]}^{*}$ (formula two)

Wherein * represents to get conjugation;

The 3rd step, detection moving target;

Antithetical phrase belt aperture interference image sequence I _MnCarry out following two operations:

Operation one arranges Phase Threshold and amplitude threshold, to every width of cloth I _MnCarry out CFAR (Constant False Alarm Rate, CFAR) and detect, Phase Threshold and amplitude threshold are determined by concrete applied environment;

If operation two is not at any width of cloth I _MnIn detect moving target, then this method stops; If at least one width of cloth I _MnIn can detect moving target, to detect the maximum width of cloth I of moving target number _MnAs the reference image I _RefIf two width of cloth and above I are arranged _MnDetected moving target number is maximum and identical, selects the highest width of cloth of signal to noise ratio (S/N ratio) as the reference image I _RefTo I _RefIn detected each moving target, carry out one by one following joint detection process:

With I _RefCentered by, and with this moving target at I _RefIn the position be reference point, on other sub-aperture of all subbands interference image, in the distance range of the twice resolution element centered by reference point the search whether operate in addition a detected moving target: if do not search moving target, change I over to _RefThe joint detection process of middle next moving target; If interference image searches moving target in the sub-aperture of at least one width of cloth subband, then the range value with the moving target that searches in the afore-mentioned distance scope in the interference image of the sub-aperture of all subbands extracts, consist of a new set A, total M * N range value in the set A; The sub-subaperture image of subband that does not search moving target is made as zero corresponding to the range value in the set A; The standard deviation of range value among the set of computations A, with amplitude in itself and the set A peaked half relatively: if standard deviation greater than amplitude among the A peaked half, then this target not united detect, no longer process this target, change I over to _RefThe joint-detection of next moving target; If standard deviation be not more than among the A amplitude peaked half, then finish the joint-detection of this moving target, and this moving target put into detect goal set;

The two-dimension speed of the 4th step, estimation moving target;

To each moving target in the detection goal set of the 3rd step, estimate its radial velocity and two-dimension speed according to following steps one by one;

(1) step, the radial velocity of estimation moving target;

Extract the phase sequence Δ φ of moving target _Mn, estimate that according to projection theorem this moving target is at different θ _nUnder radial velocity

Value realized by following operation:

Operation (one), set up many subbands estimation model:

(formula three)

V wherein _a, B,

Represent respectively corresponding carrier aircraft speed, base length and the antenna incident angle constantly of UWB SAR image data acquiring, k _mBe

Relative f _mAmbiguity period, be integer to be found the solution,

With respect to independent variable k _mDependent variable; To each m, give k _mRespectively 7 integers of assignment-3 in 3 scopes calculate 7 by formula three

Travel through all m, obtain the M group, 7 every group

Operation (two) is organized at M

In, every group optional one

Formation contains the set of M element, and calculates the variance of this set;

Operation (three), repetitive operation (two) travels through all possible combination, altogether can obtain 7 ^MThe variance of each set is calculated in individual set;

Operation (four), of from the set of operation (three) gained, selecting the variance minimum, the mean value of this set is exactly estimated value

(2) step, the two-dimension speed of estimation moving target;

Utilize the radial velocity of (one) step estimation

Estimate the two-dimension speed of this moving target by formula four, namely the orientation is to speed

With the distance to speed

(formula four)

Wherein, $\mathrm{\&alpha;}=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}\cos {2\mathrm{\&theta;}}_n,$ $\mathrm{\&beta;}=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}\sin {2\mathrm{\&theta;}}_n.$

Images