CN101666879B - Method for improving resolution of linear-array three-dimensional imaging synthetic aperture radars - Google Patents

Abstract

The invention provides a method for improving the resolution of linear-array three-dimensional imaging synthetic aperture radars. The method adopts two linear-array three-dimensional imaging synthetic aperture radars with orthogonal motion tracks for imaging of the same region, and then blends two acquired images by discrete wavelet transform technology so as to obtain a linear-array three-dimensional imaging synthetic aperture radar image with high resolution. The invention has the following advantages of realizing high-resolution imaging of the linear-array three-dimensional imaging synthetic aperture radars by using a relatively short array antenna and solving the problem of relatively low tangential track resolution of images acquired by the linear-array three-dimensional imaging synthetic aperture radars. The method can be widely applied in the fields such as synthetic aperture radar imaging, earth remote sensing and geological mapping.

Description

A kind of method that improves resolution of linear-array three-dimensional imaging synthetic aperture radars

Technical field

The invention belongs to the Radar Technology field, it is particularly related to linear-array three-dimensional imaging synthetic aperture radars (LASAR) technical field of imaging.

Background technology

Linear-array three-dimensional imaging synthetic aperture radars (LASAR) is that linear array antenna is fixed on the platform of motion, with synthetic two-dimensional planar array, and a kind of novel polarization sensitive synthetic aperture radar system that carries out three-dimensional imaging.Linear-array three-dimensional imaging synthetic aperture radars can be realized that present single antenna synthetic-aperture radar is irrealizable three-dimensional ground is carried out to the ability of picture, become the synthetic-aperture radar hot research fields at present.The document of understanding and having delivered according to the inventor, for example: J.Klare, A.Brenner; J.Ender; " A New Airborne Radar for 3D Imaging-Image Formation using the ARTINO Principle-", EUSAR, Dresden; Germany; 2006.BASSEM R.MAHAFZA, MITCH SAJJADI " Three-dimensional SAR imaging using linear array in transverse motion " IEEE transaction on aerospace and electronic system VOL 32, NO.1JANUARY 1996; Owing to receive the restriction of array antenna length; The flight path resolution of cutting of the image that linear-array three-dimensional imaging synthetic aperture radars obtains is generally less than along the flight path directional resolution, in order to improve the resolution of linear-array three-dimensional imaging synthetic aperture radars, and must its corresponding resolution integration technology of research.Understand according to the inventor, about method for merging resolutions of linear array three-dimensional imaging synthetic aperture radars, the technical literature of still not publishing at present.

Summary of the invention

For the flight path resolution of cutting that overcomes the image that has the linear-array three-dimensional imaging synthetic aperture radars acquisition now is generally less than along the problem of flight path directional resolution; The invention provides a kind of method that improves resolution of linear-array three-dimensional imaging synthetic aperture radars, adopt method of the present invention can access the high-resolution linear array three-dimensional imaging synthetic aperture radar image after the orthogonal traces three-dimensional imaging synthetic aperture radar merges.

Describe content of the present invention for ease, at first make following term definition:

Definition 1, linear-array three-dimensional imaging synthetic aperture radars (LASAR)

Linear-array three-dimensional imaging synthetic aperture radars (LASAR) is that linear array antenna is fixed on the platform of motion, with synthetic two-dimensional planar array, and a kind of novel polarization sensitive synthetic aperture radar system that carries out three-dimensional imaging.See document for details: R.Giret, H.Jeuland, P.Enert; " A Study of a 3D-SAR Concept for a Millimeter-Wave Imaging Radar onboard an UAV "; European Radar Conference, 2004, pp 201-204..Owing to receive the restriction of array antenna length, the flight path resolution of cutting of the image that linear-array three-dimensional imaging synthetic aperture radars obtains is generally less than along the flight path directional resolution.

Definition 2, linear-array three-dimensional imaging synthetic aperture radars image

The linear-array three-dimensional imaging synthetic aperture radars image is meant the linear-array three-dimensional imaging synthetic aperture radars data is carried out the data that obtain after the imaging processing, wherein comprised the distribution of diverse location place scattering coefficient in the space.See document for details: J.Klare, A.Brenner, J.Ender, " A New Airborne Radar for 3D Imaging-Image Formation using the ARTINO Principle-", EUSAR, Dresden, Germany, 2006.Can obtain the length and the width of this image by the linear-array three-dimensional imaging synthetic aperture radars image, note is P and Q respectively.

Definition 3, orthogonal traces three-dimensional imaging synthetic aperture radar

The orthogonal traces three-dimensional imaging synthetic aperture radar is meant that two linear-array three-dimensional imaging synthetic aperture radars are along orthogonal movement locus flight; Its trajectory direction is remembered respectively and is done x direction and y direction; The synthetic aperture direction of synthetic-aperture radar SAR-A is along the x direction; The image that is become by synthetic-aperture radar SAR-A has higher resolution in the x direction, and the y direction is the true aperture direction of radar SAR-A, and the resolution of y direction is lower.Synthetic-aperture radar SAR-B is just in time opposite with it.Its fundamental diagram sees accompanying drawing 1 for details, and the course of work of orthogonal traces three-dimensional imaging synthetic aperture radar sees document for details: J.Klare, A.Brenner; J.Ender; " A New Airborne Radar for 3D Imaging-Image Formation using the ARTINO Principle-", EUSAR, Dresden; Germany, 2006.

Definition 4, x (y) direction low resolution linear-array three-dimensional imaging synthetic aperture radars image

Utilize the orthogonal traces three-dimensional imaging synthetic aperture radar can obtain two width of cloth linear-array three-dimensional imaging synthetic aperture radars images; Wherein a width of cloth linear-array three-dimensional imaging synthetic aperture radars image has high y directional resolution and low x directional resolution; Be called x direction low resolution linear-array three-dimensional imaging synthetic aperture radars image, note is G ^xAn other width of cloth linear-array three-dimensional imaging synthetic aperture radars image has high x directional resolution and low y directional resolution, is called y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image, and note is G ^y

Definition 5, one-dimensional discrete wavelet decomposition and reconstruct

For one-dimensional signal f _J+1(n), its quick orthogonal wavelet transformation relation is as follows:

Orthogonal wavelet decomposes fast:

${\tilde{f}}_j\left(n\right)=f_{j+1}\left(n\right)*h\left(2n\right)$

${\tilde{d}}_j\left(n\right)=f_{j+1}\left(n\right)*h\left(2n\right)$

Wherein, * representes convolution, and n representes the discrete signal point.H (2n) and g (2n) expression are carried out carrying out 1/2 sampling after low pass h (n) (high pass g (the n)) filtering to the high-resolution approximation signal.G (n) is the corresponding Hi-pass filter of h (n).

Orthogonal wavelet is rebuild fast:

The proof of above-mentioned formula sees document Stephane Mallat for details, " A Wavelet Tour of Signal processing ", [c] 2nd ed.Chap.VII, Academic press Elsevier Pte Ltd, 2003.The form that quick orthogonal wavelet transformation can be used system chart is referring to accompanying drawing 2.

Definition 6,2-d discrete wavelet decomposition and reconstruction

Piece image f _J+1(n ₁, n ₂) can carry out discrete wavelet through formula once and decompose:

${\tilde{f}}_j(n_1,n_2)=f_{j+1}(n_1,n_2)*\left[h\left(2n_1\right)h\left(2n_2\right)\right]$

${\tilde{d}}_j^1(n_1,n_2)=f_{j+1}(n_1,n_2)*\left[h\left(2n_1\right)g\left(2n_2\right)\right]$

${\tilde{d}}_j^2(n_1,n_2)=f_{j+1}(n_1,n_2)*\left[g\left(2n_1\right)h\left(2n_2\right)\right]$

${\tilde{d}}_j^3(n_1,n_2)=f_{j+1}(n_1,n_2)*\left[g\left(2n_1\right)g\left(2n_2\right)\right]$

Wherein, * representes convolution.n ₁, n ₂The vertical dimension of the horizontal peacekeeping pixel of representing two dimensional image respectively.H (n) is a low-pass filter, and g (n) is corresponding Hi-pass filter.H (2n ₁), h (2n ₂) represent respectively down-sampling behind the vertical dimension of the horizontal peacekeeping of the image LPF, g (2n ₁), g (2n ₂) represent respectively down-sampling after the high-pass filtering of the vertical dimension of the horizontal peacekeeping of image.Decomposing back

is the low frequency coefficient matrix of former figure;

and

is respectively level, the high frequency coefficient matrix of vertical and diagonal.

Utilize the method for above-mentioned coefficient wavelet reconstruction to be:

Wherein:

The flow process of discrete wavelet decomposition and reconstruction is seen accompanying drawing 3.The method of discrete wavelet decomposition and reconstruction sees for details: Stephane Mallat, " A Wavelet Tour of Signal processing ", [c] 2nd ed.Chap.VII, Academic press Elsevier Pte Ltd, 2003, pp.221-314

Definition 7, the Meyer small echo:

The Meyer small echo is to be made up by low-pass filter

to form.Concrete construction method sees for details: Stephane Mallat; " A Wavelet Tour of Signal processing "; [c] 2nd ed.Chap.VII; Academic press Elsevier Pte Ltd, 2003, the expression formula of

is following:

$\hat{h}\left(\mathrm{\ω}\right)=\left\{\begin{array}{cc}\sqrt{2}& \mathrm{\ω}\∈[-\mathrm{\π}/3,\mathrm{\π}/3]\\ 0& \mathrm{\ω}\∈[-\mathrm{\π},-2\mathrm{\π}/3]\∪[-2\mathrm{\π}/3,\mathrm{\π}]\end{array}\right.$

The invention provides a kind of method that improves resolution of linear-array three-dimensional imaging synthetic aperture radars, it comprises following step:

Step 1, acquisition low resolution diameter radar image

Utilizing the orthogonal traces three-dimensional imaging synthetic aperture radar to obtain x direction low resolution linear array three-dimensional imaging closes

Pore-forming aperture radar image G ^xWith y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^y

Step 2, the high resolving power of low-resolution image dimension carried out discrete wavelet decompose:

Adopt the Meyer small echo, utilize one dimension wavelet decomposition formula

Wherein, h (2n) representes after the high-resolution approximation signal is through low-pass filter h (n) filtering it to be carried out the signal that 1/2 sampling obtains, With

Be the coefficient of dissociation under the corresponding Wavelet Orthogonal base, f _J+1(n) be discrete representation form under resolution j+1; At first to y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^yCarry out x direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix

With horizontal direction high frequency coefficient matrix

Then to x direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^xCarry out y direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix

With vertical direction high frequency coefficient matrix

Wherein, the low resolution linear-array three-dimensional imaging synthetic aperture radars image of y direction and x direction obtains the low frequency coefficient matrix through the one-dimensional discrete wavelet decomposition

Be identical, n ₁, n ₂The vertical dimension of the horizontal peacekeeping pixel of representing two dimensional image respectively; n ₁, n ₂Span depend on the size of two dimensional image;

Step 3, utilize the wavelet coefficient that obtains in the step 2 to carry out the zero padding of discrete wavelet difference:

The use of low-frequency coefficient matrix

horizontal high-frequency coefficient matrix

and vertical high frequency coefficients matrix

according to the formula:

Calculated frequency coefficient matrix

corresponding coefficient matrix

horizontal high-frequency coefficient matrix

corresponding coefficient matrix

vertical high-frequency coefficient matrix

corresponding coefficient matrix

Step 4, wavelet reconstruction obtain fused images:

Obtained in Step 3 using the low frequency coefficient matrix

corresponding coefficient matrix horizontal high-frequency coefficient matrix

corresponding coefficient matrix

vertical high-frequency coefficient matrix corresponding coefficient matrix

By two-dimensional wavelet reconstruction formula:

Calculate the high-resolution linear array three-dimensional imaging synthetic aperture radar image after resolution merges Wherein * representes convolution, n ₁, n ₂The pixel of representing the vertical dimension of horizontal peacekeeping direction respectively, h (n ₁), g (n ₁) expression horizontal dimension direction low pass and high-pass filtering, h (n ₂), g (n ₂) vertical low pass and the high-pass filtering of tieing up direction of expression.

Through aforesaid operations, can obtain the high-resolution linear array three-dimensional imaging synthetic aperture radar image after the orthogonal traces three-dimensional imaging synthetic aperture radar merges.

Innovative point of the present invention is the lower problem of flight path resolution of cutting to the image of linear-array three-dimensional imaging synthetic aperture radars acquisition; Adopt the linear-array three-dimensional imaging synthetic aperture radars of two movement locus quadratures that the same area is carried out to picture; Adopt the wavelet transform technology then; Two width of cloth images that obtain are merged, thereby obtain high-resolution linear-array three-dimensional imaging synthetic aperture radars image.The invention solves the lower problem of flight path resolution of cutting of the image of linear-array three-dimensional imaging synthetic aperture radars acquisition.

The invention has the advantages that and utilize short array antenna to realize the linear-array three-dimensional imaging synthetic aperture radars high-resolution imaging.The present invention can be applied to fields such as synthetic aperture radar image-forming, earth remote sensing.

Description of drawings

Fig. 1 is an orthogonal traces three-dimensional imaging synthetic aperture radar fundamental diagram

Wherein synthetic-aperture radar SAR-A and synthetic-aperture radar SAR-B represent two linear-array three-dimensional imaging synthetic aperture radars of orthogonal motion respectively; Synthetic-aperture radar SAR-A and synthetic-aperture radar SAR-B move along x and y direction respectively.

Fig. 2 is the process flow diagram of one-dimensional signal wavelet decomposition and reconstruct.Wherein (a) is exploded view, (b) is restructuring graph.Wherein, " ↓ 2 ", expression 1/2 sampling " ↑ 2 " is illustrated in odd bits zero padding .h and representes low-pass filter, and g representes Hi-pass filter.a _I+1The one-dimensional signal that expression is decomposed, a _iThe low frequency coefficient that the back produces, d are decomposed in expression _iThe high frequency coefficient that the back produces is decomposed in expression.

Fig. 3 is the process flow diagram of two dimensional image wavelet decomposition and reconstruct.Wherein (a) is exploded view, (b) is restructuring graph.Wherein, " ↓ 2 " expression 1/2 sampling, " ↑ 2 " are illustrated in the odd bits zero padding..H representes low-pass filter, and g representes Hi-pass filter.a _I+1The two dimensional image that expression is decomposed, a _iThe low frequency coefficient matrix that the back produces, d are decomposed in expression ¹ _iThe horizontal high frequency coefficient matrix that the back produces, d are decomposed in expression ² _iThe vertical high frequency matrix of coefficients that the back produces, d are decomposed in expression ³ _iThe diagonal high frequency coefficient matrix that the back produces is decomposed in expression.

The linear-array three-dimensional imaging synthetic aperture radars original image of Fig. 4 for adopting in the embodiment

Black rectangle 1,2,3,4 is represented four scattering points that can distinguish mutually respectively among the figure.

X direction low the differentiate linear-array three-dimensional imaging synthetic aperture radars image of Fig. 5 for adopting in the embodiment

Black rectangle 5,6 is represented bluring owing to the low x directional image that causes of x directional resolution respectively among the figure.

Y direction low the differentiate linear-array three-dimensional imaging synthetic aperture radars image of Fig. 6 for adopting in the embodiment

Black rectangle 7,8 is represented bluring owing to the low y directional image that causes of y directional resolution respectively among the figure.

Fig. 7 is that the image 3 that adopts the present invention to provide method to obtain merges back linear-array three-dimensional imaging synthetic aperture radars image with image 4

Rectangle 9,10,11,12 representes to adopt four scattering points that can distinguish each other that obtain behind the present invention respectively.Can find out that from Fig. 4,5,6,7 method for merging resolutions of linear array three-dimensional imaging synthetic aperture radars that the present invention proposes can improve the resolution of linear-array three-dimensional imaging synthetic aperture radars image.

Fig. 8 is the inventive method process flow diagram

Embodiment

The present invention mainly adopts the method for emulation experiment to verify, institute in steps, conclusion all on MATLAB7.0 checking correct.The practical implementation step is following:

Step 1, the high resolving power of low-resolution image dimension carried out discrete wavelet decompose:

Use the Meyer wavelet basis, to x direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^xCarry out y direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix Right with the horizontal direction high frequency coefficient

To y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^yCarry out x direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix

Right with the vertical direction high frequency coefficient ${\tilde{d}}_j^2({\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="H2008100459746E00012.png"\; state="\{\{state\}\}">n</figure-callout>}}_1,n_2).$

Step 2, wavelet reconstruction obtain fused images:

Carry out discrete wavelet reconstruct, the image after obtaining merging

with obtaining matrix of coefficients

in the step 1

Can find out that through the specific embodiment of the invention method for merging resolutions of linear array three-dimensional imaging synthetic aperture radars provided by the present invention can merge two width of cloth images, thereby obtain high-resolution linear-array three-dimensional imaging synthetic aperture radars image.

Claims (1)

1. method for merging resolutions of linear array three-dimensional imaging synthetic aperture radars based on wavelet transform is characterized in that it may further comprise the steps:

Step 1, acquisition low resolution diameter radar image

Utilize the orthogonal traces three-dimensional imaging synthetic aperture radar to obtain x direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^xWith y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^y

Step 2, the high resolving power of low-resolution image dimension carried out discrete wavelet decompose:

Adopt the Meyer small echo, utilize one dimension wavelet decomposition formula

Wherein, h (2n) representes after the high-resolution approximation signal is through low-pass filter h (n) filtering it to be carried out the signal that 1/2 sampling obtains,

With

Be the coefficient of dissociation under the corresponding Wavelet Orthogonal base, f _J+1(n) be discrete representation form under resolution j+1; At first to y direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^yCarry out x direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix With horizontal direction high frequency coefficient matrix

Then to x direction low resolution linear-array three-dimensional imaging synthetic aperture radars image G ^xCarry out y direction one-dimensional discrete wavelet decomposition, obtain the low frequency coefficient matrix

With vertical direction high frequency coefficient matrix

Wherein, the low resolution linear-array three-dimensional imaging synthetic aperture radars image of y direction and x direction obtains the low frequency coefficient matrix through the one-dimensional discrete wavelet decomposition

Be identical, n ₁, n ₂The vertical dimension of the horizontal peacekeeping pixel of representing two dimensional image respectively; n ₁, n ₂Span depend on the size of two dimensional image;

Step 3, utilize the wavelet coefficient that obtains in the step 2 to carry out the zero padding of discrete wavelet difference:

The use of low-frequency coefficient matrix?

horizontal high-frequency coefficient matrix?

and vertical high frequency coefficients matrix?

according to the formula:

Calculated frequency coefficient matrix?

corresponding coefficient matrix?

horizontal high-frequency coefficient matrix?

corresponding coefficient matrix?

vertical high-frequency coefficient matrix?

corresponding coefficient matrix?

Step 4, wavelet reconstruction obtain fused images:

Obtained in Step 3 using the low frequency coefficient matrix?

corresponding coefficient matrix?

horizontal high-frequency coefficient matrix?

corresponding coefficient matrix?

vertical high-frequency coefficient matrix?

corresponding coefficient matrix?

By two-dimensional wavelet reconstruction formula:

Calculate the high-resolution linear array three-dimensional imaging synthetic aperture radar image after resolution merges

Wherein * representes convolution, n ₁, n ₂The pixel of representing the vertical dimension of horizontal peacekeeping direction respectively, h (n ₁), g (n ₁) expression horizontal dimension direction low pass and high-pass filtering, h (n ₂), g (n ₂) vertical low pass and the high-pass filtering of tieing up direction of expression.

Images