CN104008270A - Multichannel reconstruction method and device for periodical non-uniform sampling SAR (synthetic aperture radar) signals - Google Patents

Abstract

The invention discloses a multichannel reconstruction method for periodical non-uniform sampling SAR (synthetic aperture radar) signals. The method includes: periodically changing pulse repetition frequency of a radar transmitter to acquire a periodical non-uniform sampling SAR echo signal, constructing a constraint underdetermined system of linear equations of a reconstruction filter according to channel features of the periodical non-uniform sampling SAR echo signal to acquire signal power after multichannel reconstruction, combining the underdetermined system of linear equations and minimum conditions of the signal power to acquire the optimal reconstruction filter, and using the optimal reconstruction filter to complete multichannel reconstruction.

Description

A kind of hyperchannel method for reconstructing and device of periodically nonuniform sampling SAR signal

Technical field

The recovery problem of non-homogeneous synthetic-aperture radar of the cycle that the present invention relates to (SAR, Synthetic Aperture Radar) signal, relates in particular to a kind of hyperchannel method for reconstructing and device of periodically nonuniform sampling SAR signal.

Background technology

Development along with New System new ideas SAR, orientation becomes a subject matter of many New System SAR to the periodically nonuniform sampling problem of SAR signal, be prevalent in discrete phases Centers multibeam SAR, multi-platform many bases SAR and prompt change in pulse repetition rate SAR.In order to recover uniform sampling signal, a kind of effective method is to adopt hyperchannel reconstruction algorithm.

Yet due to the impact of system noise and antenna sidelobe, hyperchannel is rebuild and can be amplified noise power and azimuth ambiguity signal power.And nonuniform sampling is more serious, it is more that noise power and azimuth ambiguity power are just exaggerated, will be larger on the impact of the performance of radar system and radar image.

Summary of the invention

For solving the problem of existing existence, the present invention mainly provides a kind of hyperchannel method for reconstructing and device of periodically nonuniform sampling SAR signal, make in recovering antenna main lobe in signal, minimise interference signal, improves the performance of radar system and the quality of radar image.

Technical scheme of the present invention is achieved in that

The hyperchannel method for reconstructing that the invention provides a kind of periodically nonuniform sampling SAR signal, the method comprises:

The pulse repetition rate of the change radar transmitter in cycle, obtains orientation to the SAR echoed signal of periodically nonuniform sampling;

What according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, build constraint reestablishing wave filter owes to determine system of linear equations;

Obtain the signal power after hyperchannel is rebuild;

The condition that minimizes of owing to determine system of linear equations and described signal power described in associating, obtains optimum reconstruction filter, utilizes the reconstruction filter of described optimum to complete hyperchannel reconstruction.

In such scheme, described acquisition orientation to the SAR echoed signal of periodically nonuniform sampling is: due to the change in cycle radar transmitter pulse repetition rate, the echoed signal obtaining in orientation to being the SAR echoed signal of periodically nonuniform sampling, and a plurality of nonuniform sampling points in one-period are regarded as from different passages, the variation characteristic of pulse repetition rate determines the multi-channel feature of each passage.

In such scheme, the described multi-channel feature according to the SAR echoed signal of described periodically nonuniform sampling build constraint reestablishing wave filter owe determine system of linear equations and be: according to the multi-channel feature of the SAR echoed signal of periodically nonuniform sampling, determine prefilter matrix, then according to described prefilter matrix, build constraint reestablishing wave filter owe to determine system of linear equations.

In such scheme, described prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; N represents N nonuniform sampling point in one-period, and non-uniformly sampled signals derives from N passage; M represents the frequency band number of processing; H _i(f) represent the characteristic filtering device of i passage in N passage, for:

H _i(f)＝exp{-j2πf·t _i}

Wherein, t _irepresent that i passage is with respect to the time delay of reference channel; J represents imaginary unit; F represents orientation frequency, and f ∈ I ₁, I _jthe frequency range that represents j frequency band, for:

$I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M.$

In such scheme, described in owe to determine system of linear equations and be:

H _M×N·P _j＝A，j＝1，2，…，M

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition.

In such scheme, the signal power after described acquisition hyperchannel is rebuild is: determine the covariance matrix of passage, according to described covariance matrix, obtain the signal power after hyperchannel is rebuild.

In such scheme, the covariance matrix of described passage is:

R _S＝E{S ^H(f)S(f)}

Wherein, E{} represents to ask expectation; . ^hrepresent to ask the conjugate transpose of vector or matrix; Vector S represents the vector that the signal by each passage forms, and described vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)]

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage.

In such scheme, the signal power after described hyperchannel is rebuild is: E (| SP _j| ²)={ P _j ^hr _sp _j.

In such scheme, the described underdetermined system of equations of described associating and signal power minimize condition, obtain optimum reconstruction filter and are: combine the described underdetermined system of equations and signal power and minimize condition and obtain a linear restriction minimum power (LCMP), for:

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.$

Utilize method of Lagrange multipliers to solve described LCMP, obtain optimum reconstruction filter and be: P _j=(R _s ^-1) H ^h(HR _s ^-1h ^h) ^-1a.

In such scheme, the optimum reconstruction filter of described utilization completes hyperchannel and is redeveloped into: the row vector S that the signal of each passage is formed is multiplied by the reconstruction filter column vector P of described optimum _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, obtain the uniform sampling signal after rebuilding.

The present invention also provides a kind of hyperchannel reconstructing device of periodically nonuniform sampling synthetic-aperture radar SAR signal, and this hyperchannel reconstructing device comprises: signal acquisition module, equation build module, signal power determination module and rebuild module; Wherein,

Signal acquisition module, for the pulse repetition rate of the change radar transmitter in cycle, obtains orientation to the SAR echoed signal of periodically nonuniform sampling;

Equation builds module, for what build constraint reestablishing wave filter according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, owes to determine system of linear equations;

Signal power determination module, for obtaining the signal power after reconstruction;

Rebuild module, for owing to determine the condition that minimizes of system of linear equations and described signal power described in combining, obtain optimum reconstruction filter, utilize described reconstruction filter to complete hyperchannel and rebuild.

In such scheme, described equation builds module, specifically for determining prefilter matrix H according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling _{m * N}, then according to described prefilter matrix H _{m * N}what build constraint reestablishing wave filter owes to determine system of linear equations.

In such scheme, described prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; H _i(f) represent the characteristic filtering device of i passage in N passage, for:

H _i(f)＝exp{-j2πf·t _i}

Wherein, t _irepresent that i passage is with respect to the time delay of reference channel; J represents imaginary unit; F represents orientation frequency, and f ∈ I ₁, I _jthe frequency range that represents j frequency band, for:

$I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M.$

In such scheme, described in owe to determine system of linear equations and be:

H _M×N·P _j＝A，j＝1，2，…，M

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition; N represents N nonuniform sampling point in one-period, and non-uniformly sampled signals derives from N passage; M represents the frequency band number of processing.

In such scheme, described signal power determination module, specifically for determining the covariance matrix of passage, then the signal power after being rebuild according to described covariance matrix.

In such scheme, the covariance matrix of described passage is:

R _S＝E{S ^H(f)S(f)}

Wherein, E{} represents to ask expectation; . ^hrepresent to ask the conjugate transpose of vector or matrix; Vector S represents the vector that the signal by each passage forms, and described vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)]

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage.

In such scheme, the signal power after described definite reconstruction is: E (| SP _j| ²)={ P _j ^hr _sp _j.

In such scheme, described reconstruction module, minimizes condition and obtains a linear restriction minimum power (LCMP) specifically for combining the described underdetermined system of equations and signal power:

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.$

And utilize method of Lagrange multipliers to solve described LCMP, obtain optimum reconstruction filter and be: P _j=(R _s ^-1) H ^h(HR _s ^-1h ^h) ^-1a.

In such scheme, described reconstruction module, the concrete reconstruction filter column vector P that is also multiplied by described optimum for the row vector S that the signal of each passage is formed _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, obtain the uniform sampling signal after rebuilding.

Hyperchannel method for reconstructing and the device of a kind of periodically nonuniform sampling SAR signal providing of the present invention, the pulse repetition rate of the change radar transmitter in cycle, obtain orientation to the SAR echoed signal of periodically nonuniform sampling, what according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, build constraint reestablishing wave filter owes to determine system of linear equations, determine the covariance matrix of each passage, obtain the signal power after reconstruction, and owe to determine system of linear equations and signal minimizes condition described in associating, obtain optimum reconstruction filter, utilize the reconstruction filter of described optimum to complete hyperchannel reconstruction, so, not only can recover signal in antenna main lobe, reconstruct the useful signal within the scope of antenna beam 3dB, again can be by the minimum power of azimuth ambiguity signal and noise signal, improve azimuth ambiguity and equivalent backscattering coefficient, thereby improve performance and the radar image quality of radar system.

Accompanying drawing explanation

The schematic flow sheet of the hyperchannel method for reconstructing of the periodically nonuniform sampling SAR signal that Fig. 1 provides for the embodiment of the present invention;

The distribution schematic diagram of the periodically nonuniform sampling signal that Fig. 2 provides for the embodiment of the present invention;

Uniform sampling signal distributions schematic diagram after the reconstruction that Fig. 3 provides for the embodiment of the present invention;

The hyperchannel reconstructing device structural representation of the periodically nonuniform sampling SAR signal that Fig. 4 provides for the embodiment of the present invention;

The empty graph of a relation frequently of single channel signal and hyperchannel reconstruction signal after the time-continuous signal that Fig. 5 provides for the embodiment of the present invention, sampling;

The concrete exemplary plot of the periodically nonuniform sampling signal that Fig. 6 provides for the embodiment of the present invention;

Imaging results figure in the desirable uniform sampling situation that Fig. 7 provides for the embodiment of the present invention;

Fig. 8 rebuilds the imaging results figure after (M=8) for the traditional hyperchannel that the embodiment of the present invention provides;

Fig. 9 rebuilds the imaging results figure after (M=7) for the improved hyperchannel that the embodiment of the present invention provides;

The norm comparison diagram of the reconstruction filter vector when M=7 that Figure 10 provides for the embodiment of the present invention and M=8.

Embodiment

In the embodiment of the present invention, the pulse repetition rate of the change radar transmitter in cycle, obtain orientation to the SAR echoed signal of periodically nonuniform sampling, what according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, build constraint reestablishing wave filter owes to determine system of linear equations, obtain the signal power after rebuilding, and owe to determine the condition that minimizes of system of linear equations and signal power described in associating, and obtain optimum reconstruction filter, utilize described reconstruction filter to complete hyperchannel and rebuild.

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

As shown in Figure 1, this hyperchannel method for reconstructing mainly comprises the following steps the schematic flow sheet of the hyperchannel method for reconstructing of the periodically nonuniform sampling SAR signal that the embodiment of the present invention provides:

Step 101: the pulse repetition rate of the change radar transmitter in cycle, obtains orientation to the SAR echoed signal of periodically nonuniform sampling;

Concrete, due to the change in cycle the pulse repetition rate of radar transmitter, the SAR echoed signal that the orientation that makes to obtain is periodically nonuniform sampling to echoed signal; And a plurality of nonuniform sampling points in one-period are regarded as from different passages, the variation characteristic of described pulse repetition rate has determined the multi-channel feature of each passage;

Radar emission signal is not to continue transmitting, but transmitting has the pulse signal in the time interval, the echoed signal that radar receiver receives is 2D signal, be divided into orientation to distance to, the orientation receiving is discrete signal to signal, between two sampled point, be spaced apart 1/ pulse repetition rate, therefore, in actual applications, the cyclical variation of pulse repetition rate, will cause orientation to the nonuniform sampling of signal, the orientation of acquisition is cycle signal heterogeneous to SAR echoed signal.

Step 102: what build constraint reestablishing wave filter according to the multi-channel feature of the SAR echoed signal of periodically nonuniform sampling owes to determine system of linear equations;

Concrete, according to the multi-channel feature of periodically nonuniform sampling SAR echoed signal, determine prefilter matrix, then according to described prefilter matrix, build constraint reestablishing wave filter owe to determine system of linear equations;

The distribution of periodically nonuniform sampling signal as shown in Figure 2, wherein, t _arepresent the orientation time; T represents one-period; In figure, in one-period T, have N nonuniform sampling point, whole non-uniformly sampled signals can be regarded as and derive from N passage, supposes take that passage 1 is as reference, and i passage is t with respect to the time delay of reference channel _i, in N passage, i channel characteristics wave filter is:

H _i(f)＝exp{-j2πf·t _i} (1)

Wherein, j represents imaginary unit; F represents orientation frequency;

Figure 3 shows that the uniform sampling signal distribution plots after reconstruction, wherein, M represents to process the number of frequency band, wherein T _erepresent equivalent uniform sampling interval; In improved hyperchannel method for reconstructing provided by the invention, M is less than or equal to N;

According to above-mentioned channel characteristics wave filter, can obtain prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; And orientation frequency f ∈ I ₁, I _jthe frequency range that represents j frequency band, for: $I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M;---\left(3\right)$

According to above-mentioned prefilter matrix, constructed owing determined system of linear equations and is:

H _M×N·P _j＝A，j＝1，2，…，M (4)

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition.

Step 103: obtain the signal power after rebuilding;

Concrete, determine the covariance matrix of passage then the signal power after being rebuild according to described covariance matrix;

Determine that first covariance matrix need to determine the vectorial S that the signal by each passage forms, described vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)] (5)

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage; So the covariance matrix consisting of vectorial S is:

R _S＝E{S ^H(f)S(f)} (6)

Wherein, E{} represents to ask expectation; . ^hrepresent to ask the conjugate transpose of vector or matrix;

Signal power after hyperchannel is rebuild is by by vectorial S and the vectorial P of reconstruction filter _jthe acquisition of multiplying each other, for:

E(|SP _j| ²)＝{P _j ^HR _SP _j} (7)

Signal power after described hyperchannel is rebuild comprises: signal power, the interior signal power of secondary lobe and noise power in main lobe; When described reconstruction filter meet (4) formula owe to determine system of linear equations time, in the main lobe in signal power, signal power remains unchanged.

Step 104: owe to determine system of linear equations and described signal power minimal condition described in associating, obtain optimum reconstruction filter, utilize the reconstruction filter of described optimum to complete hyperchannel reconstruction;

Concrete, it is hour required satisfied condition of signal power after the hyperchannel that obtains in step 102 is rebuild that described signal power minimizes condition, and this minimizes condition and represents with following formula:

minimize E(|SP _j| ²)＝min{P _j ^HR _SP _j} (8)

By described in owe to determine system of linear equations (4) and above-mentioned signal power to minimize the system of equations that condition (8) joins together to obtain be linear restriction minimum power (LCMP, Linear Constraint Minimize Power):

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.---\left(9\right)$

Utilize method of Lagrange multipliers to solve above-mentioned LCMP system of equations, obtain optimum reconstruction filter and be:

P _j＝(R _S ^-1)H ^H(HR _S ^-1H ^H) ^-1A (10)

When M=N, described in, P _j=H ^-1a is identical with conventional reconstruction filter;

According to the reconstruction filter of described optimum, complete hyperchannel and rebuild, concrete, the row vector S that the signal of each passage is formed is multiplied by optimum reconstruction filter column vector P _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, the uniform sampling signal after being rebuild.

The embodiment of the present invention also provides a kind of hyperchannel reconstructing device of periodically nonuniform sampling SAR signal, the structure of this hyperchannel reconstructing device as shown in Figure 4, comprising: signal acquisition module 40, equation build module 41, signal power determination module 42 and rebuild module 43; Wherein,

Signal acquisition module 40, for the pulse repetition rate of the change radar transmitter in cycle, obtains orientation to cycle SAR echoed signal heterogeneous;

Due to the change in cycle the pulse repetition rate of radar transmitter, so, the SAR echoed signal that the orientation of acquisition is periodically nonuniform sampling to echoed signal; A plurality of nonuniform sampling points in one-period are regarded as from different passages, and the variation characteristic of described pulse repetition rate has determined the multi-channel feature of each passage;

Equation builds module 41, for what build constraint reestablishing wave filter according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, owes to determine system of linear equations;

Signal power determination module 42, for obtaining the signal power after reconstruction;

Rebuild module 43, for owe shown in combining to determine system of linear equations and shown in the minimal condition of signal power, obtain optimum reconstruction filter, utilize described reconstruction filter to complete hyperchannel and rebuild;

Described equation builds module 41, specifically for determining prefilter matrix H according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling _{m * N}, then according to described prefilter matrix H _{m * N}what build constraint reestablishing wave filter owes to determine system of linear equations;

In periodically nonuniform sampling signal distributions as shown in Figure 2, have N nonuniform sampling point in one-period T, whole non-uniformly sampled signals is regarded as and derived from N passage, suppose take that passage 1 is as reference, i passage is t with respect to the time delay of reference channel _i, in N passage, i channel characteristics wave filter is:

H _i(f)＝exp{-j2πf·t _i} (1)

Wherein, j represents imaginary unit; F represents orientation frequency;

Homogeneous signal distribution plan after reconstruction as shown in Figure 3, wherein, has M to process frequency band in one-period T;

According to shown in channel characteristics wave filter can obtain prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; And orientation frequency f ∈ I ₁, wherein, I _jthe frequency range that represents j frequency band, for: $I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M;---\left(3\right)$

According to above-mentioned prefilter matrix H _{m * N}constructed owe to determine system of linear equations be:

H _M×N·P _j＝A，j＝1，2，…，M (4)

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition;

Signal power determination module 42, specifically for determining the covariance matrix of passage, then the signal power after being rebuild according to described covariance matrix;

The vectorial S that the covariance matrix of passage consists of the signal of passage determines, vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)] (5)

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage;

According to the covariance matrix of the definite passage of vectorial S, be:

R _S＝E{S ^H(f)S(f)} (6)

According to the signal power after the reconstruction of described covariance matrix gained, be:

E(|SP _j| ²)＝{P _j ^HR _SP _j} (7)

Described reconstruction module 43, specifically for owing to determine system of linear equations (4) described in combining and signal power minimizes condition, obtains a linear restriction minimum power problem;

Described signal power minimizes condition and is (7) condition hour, can be expressed as following formula:

minimize E(|SP _j| ²)＝min{P _j ^HR _SP _j} (8)

Will described in owe to determine system of linear equations (4) and signal power and minimize condition (8) and join together, obtain linear restriction minimum power:

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.---\left(9\right)$

Utilize method of Lagrange multipliers to solve above-mentioned system of equations (9), obtain optimum reconstruction filter and be:

P _j＝(R _S ^-1)H ^H(HR _S ^-1H ^H) ^-1A (10)

Described reconstruction module 43, also for being multiplied by the row vector S of the signal composition of each passage the reconstruction filter column vector P of described optimum _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, the uniform sampling signal after just can being rebuild;

Above-mentioned sampling module 40 can be realized by hardware sample circuit, specifically can complete with analog-digital conversion a/d chip;

Above-mentioned equation builds module 41, signal power determination module 42 and rebuilds module 43 and can on the CPU of computing machine, realize, and also can realize by the signal processing chip in other signal handling equipment, as DSP, ARM, FPGA etc.

As shown in Figure 5, wherein θ is instantaneous angle of squint to the empty relation frequently of the single channel signal after time-continuous signal, sampling and hyperchannel reconstruction signal, P _ijfor vectorial P _ji element, B _dfor 3dB doppler bandwidth;

In Fig. 5, transverse axis f represents signal frequency, the sine value that the longitudinal axis is θ, and continuous signal obtains single channel sampled signal after over-sampling, the band limiting of discrete signal after sampling, after sampling, signal is between cycle repetition frequency CRF; The hyperchannel method for reconstructing that single channel signal after sampling is provided through the above embodiment of the present invention is rebuild, and obtains hyperchannel reconstruction signal, B _dbe depicted as the doppler bandwidth of 3dB, within the scope of this, in figure, dark-shaded partly represents useful signal, and light dash area represents undesired signal, and the hyperchannel method for reconstructing that the present invention proposes and the object of device are to make the undesired signal in figure to reach minimum.

Figure 6 shows that the embodiment of the present invention provides a concrete example of periodically nonuniform sampling signal; As shown in Figure 6, one-period T is divided into 8 sections to non-homogeneous signal distributions of cycle, can find out that 8 sampled point right and wrong in cycle T are equally distributed.

For different reconstruction band numbers, it is also different that the reconstruction filter that utilizes (10) to obtain carries out the image of the rear imaging of hyperchannel reconstruction;

For example, the parameters of system is as shown in table 1:

Under the systematic parameter shown in table 1, the imaging results in desirable uniform sampling situation as shown in Figure 7;

The grey level on figure right side represents signal energy, and the background color in left hand view is more deeply felt and shown that signal energy is less; In left-side images, the image A at 0 place, direction position represents useful signal, the A of upper and lower both sides represents undesired signal, the energy of undesired signal is the weighting of all undesired signal energy, as can be seen from Figure, in desirable uniform sampling situation, background color is very dark, general in-58 left and right of grey level, illustrates that the energy of undesired signal is very little;

Periodically nonuniform sampling, traditional hyperchannel is rebuild after (the number M=8 that processes frequency band), and imaging results is as shown in Figure 8;

In this situation, by the grey level of background color, can find out that signal energy is probably in-35 left and right, so the energy of undesired signal obviously wants large many than ideal situation (Fig. 7);

Periodically nonuniform sampling, improved hyperchannel provided by the invention is rebuild after (the number M=7 that processes frequency band), and imaging results is as shown in Figure 9;

In this situation, by the grey level of background color, can find out that signal energy is greatly about-50 left and right, the energy of undesired signal is less;

Comparison diagram 8 and Fig. 9, can find out: the number that reduces to process frequency band, and utilize and the invention provides hyperchannel method for reconstructing, with optimum reconstruction filter, carry out hyperchannel and rebuild the azimuth ambiguity that obtains and signal to noise ratio (S/N ratio) (shown in Fig. 8, what M=8) obtain will get well than conventional hyperchannel method for reconstructing; Because during M=7, the reconstruction filter vector norm of each frequency (undesired signal enlargement factor) is little many during than M=8, the norm value while Figure 10 shows that M=8 and M=7, in figure, || || ₂represent to ask 2-norm, reconstruction filter vector norm when dotted line represents M=8, reconstruction filter vector norm when solid line represents M=7.

The application is that process flow diagram and/or the block scheme of the method, device or the computer program that provide according to embodiment described.Should understand can be in computer program instructions realization flow figure and/or block scheme each flow process and/or the flow process in square frame and process flow diagram and/or block scheme and/or the combination of square frame.Can provide these computer program instructions to the processor of multi-purpose computer, special purpose computer, Embedded Processor or other programmable data processing device to produce a machine, the instruction of carrying out by the processor of computing machine or other programmable data processing device is produced for realizing the device in the function of flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame appointments.

The pulse repetition rate of the change radar transmitter in embodiment of the present invention cycle, obtain orientation to the SAR echoed signal of periodically nonuniform sampling, what according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, build constraint reestablishing wave filter owes to determine system of linear equations, obtain the signal power after hyperchannel is rebuild, described in associating, owe to determine again the condition that minimizes of system of linear equations and described signal power, obtain optimum reconstruction filter, utilize the reconstruction filter of described optimum to complete hyperchannel reconstruction, not only can reconstruct the useful signal within the scope of antenna beam 3dB, and the interfering signal power that formed by azimuth ambiguity and noise of energy minimization, improve azimuth ambiguity and equivalent back scattering system.

The above, be only preferred embodiment of the present invention, is not intended to limit protection scope of the present invention, all any modifications of doing within the spirit and principles in the present invention, is equal to and replaces and improvement etc., within all should being included in protection scope of the present invention.

Claims (19)

1. a hyperchannel method for reconstructing for periodically nonuniform sampling synthetic-aperture radar SAR signal, is characterized in that, described hyperchannel method for reconstructing comprises:

The pulse repetition rate of the change radar transmitter in cycle, obtains orientation to the SAR echoed signal of periodically nonuniform sampling;

What according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, build constraint reestablishing wave filter owes to determine system of linear equations;

Obtain the signal power after hyperchannel is rebuild;

The condition that minimizes of owing to determine system of linear equations and described signal power described in associating, obtains optimum reconstruction filter, utilizes the reconstruction filter of described optimum to complete hyperchannel reconstruction.

2. hyperchannel method for reconstructing according to claim 1, it is characterized in that, described acquisition orientation to the SAR echoed signal of periodically nonuniform sampling is: due to the change in cycle radar transmitter pulse repetition rate, the echoed signal obtaining in orientation to being the SAR echoed signal of periodically nonuniform sampling, and a plurality of nonuniform sampling points in one-period are regarded as from different passages, the variation characteristic of pulse repetition rate determines the multi-channel feature of each passage.

3. hyperchannel method for reconstructing according to claim 1, it is characterized in that, the described multi-channel feature according to the SAR echoed signal of described periodically nonuniform sampling build constraint reestablishing wave filter owe determine system of linear equations and be: according to the multi-channel feature of the SAR echoed signal of periodically nonuniform sampling, determine prefilter matrix, then according to described prefilter matrix, build constraint reestablishing wave filter owe to determine system of linear equations.

4. hyperchannel method for reconstructing according to claim 3, is characterized in that, described prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; N represents N nonuniform sampling point in one-period, and non-uniformly sampled signals derives from N passage; M represents the frequency band number of processing; H _i(f) represent the characteristic filtering device of i passage in N passage, for:

H _i(f)＝exp{-j2πf·t _i}

Wherein, t _irepresent that i passage is with respect to the time delay of reference channel; J represents imaginary unit; F represents orientation frequency, and f ∈ I ₁, I _jthe frequency range that represents j frequency band, for:

$I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M.$

5. hyperchannel method for reconstructing according to claim 4, is characterized in that, described in owe to determine system of linear equations and be:

H _M×N·P _j＝A，j＝1，2，…，M

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition.

6. hyperchannel method for reconstructing according to claim 1, is characterized in that, the signal power after described acquisition hyperchannel is rebuild is: determine the covariance matrix of passage, according to described covariance matrix, obtain the signal power after hyperchannel is rebuild.

7. hyperchannel method for reconstructing according to claim 6, is characterized in that, wherein, the covariance matrix of described passage is:

R _S＝E{S ^H(f)S(f)}

Wherein, E{} represents to ask expectation; . ^hrepresent to ask the conjugate transpose of vector or matrix; Vector S represents the vector that the signal by each passage forms, and described vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)]

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage.

8. hyperchannel method for reconstructing according to claim 7, is characterized in that, the signal power after described hyperchannel is rebuild is: E (| SP _j| ²)={ P _j ^hr _sp _j.

9. according to the hyperchannel method for reconstructing described in claim 1 to 8 any one, it is characterized in that, the described underdetermined system of equations of described associating and signal power minimize condition, obtaining optimum reconstruction filter is: combine the described underdetermined system of equations and signal power and minimize condition and obtain a linear restriction minimum power (LCMP), for:

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.$

Utilize method of Lagrange multipliers to solve described LCMP, obtain optimum reconstruction filter and be: P _j=(R _s ^-1) H ^h(HR _s ^-1h ^h) ^-1a.

10. hyperchannel method for reconstructing according to claim 9, is characterized in that, the optimum reconstruction filter of described utilization completes hyperchannel and is redeveloped into: the row vector S that the signal of each passage is formed is multiplied by the reconstruction filter column vector P of described optimum _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, obtain the uniform sampling signal after rebuilding.

The hyperchannel reconstructing device of 11. 1 kinds of periodically nonuniform sampling synthetic-aperture radar SAR signals, is characterized in that, described hyperchannel reconstructing device comprises: signal acquisition module, equation build module, signal power determination module and rebuild module; Wherein,

Signal acquisition module, for the pulse repetition rate of the change radar transmitter in cycle, obtains orientation to the SAR echoed signal of periodically nonuniform sampling;

Equation builds module, for what build constraint reestablishing wave filter according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling, owes to determine system of linear equations;

Signal power determination module, for obtaining the signal power after reconstruction;

Rebuild module, for owing to determine the condition that minimizes of system of linear equations and described signal power described in combining, obtain optimum reconstruction filter, utilize described reconstruction filter to complete hyperchannel and rebuild.

12. hyperchannel reconstructing devices according to claim 11, is characterized in that, described equation builds module, specifically for determining prefilter matrix H according to the multi-channel feature of the SAR echoed signal of described periodically nonuniform sampling _{m * N}, then according to described prefilter matrix H _{m * N}what build constraint reestablishing wave filter owes to determine system of linear equations.

13. hyperchannel reconstructing devices according to claim 12, is characterized in that, described prefilter matrix H _{m * N}for:

Wherein, CRF indication cycle repetition frequency; H _i(f) represent the characteristic filtering device of i passage in N passage, for:

H _i(f)＝exp{-j2πf·t _i}

Wherein, t _irepresent that i passage is with respect to the time delay of reference channel; J represents imaginary unit; F represents orientation frequency, and f ∈ I ₁, I _jthe frequency range that represents j frequency band, for:

$I_j=(-\frac{M\·\mathrm{CRF}}{2}+(j-1)\mathrm{CRF},-\frac{M\·\mathrm{CRF}}{2}+j\·\mathrm{CRF}),j=1,\·\·\·,M.$

14. hyperchannel reconstructing devices according to claim 12, is characterized in that, described in owe to determine system of linear equations and be:

H _M×N·P _j＝A，j＝1，2，…，M

Wherein, P _jrepresent that j is processed the reconstruction filter vector that frequency band is corresponding; A=[a ₁..., a _j..., a _m] ^t, wherein, a _j=1 represents to retain j processes frequency spectrum corresponding to frequency band, a _k=0 (k=1,2 ..., M, k ≠ j) other M-1 frequency spectrum corresponding to processing frequency band of expression inhibition; N represents N nonuniform sampling point in one-period, and non-uniformly sampled signals derives from N passage; M represents the frequency band number of processing.

15. hyperchannel reconstructing devices according to claim 11, is characterized in that, described signal power determination module, and specifically for determining the covariance matrix of passage, then the signal power after being rebuild according to described covariance matrix.

16. hyperchannel reconstructing devices according to claim 15, is characterized in that, the covariance matrix of described passage is:

R _S＝E{S ^H(f)S(f)}

Wherein, E{} represents to ask expectation; . ^hrepresent to ask the conjugate transpose of vector or matrix; Vector S represents the vector that the signal by each passage forms, and described vectorial S is:

S＝[S ₁(f)，S ₂(f)，...，S _N(f)]

Wherein, S _i(f), i=1,2 ..., N represents the signal of i passage in N passage.

17. hyperchannel reconstructing devices according to claim 16, is characterized in that, the signal power after described definite reconstruction is: E (| SP _j| ²)={ P _j ^hr _sp _j.

18. hyperchannel reconstructing devices according to claim 11, is characterized in that, described reconstruction module minimizes condition and obtains a linear restriction minimum power (LCMP) specifically for combining the described underdetermined system of equations and signal power:

$\left\{\begin{array}{c}\min \mathrm{imizeE}\left({\left|{\mathrm{SP}}_j\right|}^2\right)=\min \left\{{P_j}^H{R}_S{P}_j\right\}\\ \mathrm{subjectto}{H}_{M\×N}\·P_j=A,j=\mathrm{1,2},\·\·\·,M\end{array}\right.$

And utilize method of Lagrange multipliers to solve described LCMP, obtain optimum reconstruction filter and be: P _j=(R _s ^-1) H ^h(HR _s ^-1h ^h) ^-1a.

19. hyperchannel reconstructing devices according to claim 18, is characterized in that, described reconstruction module, the concrete reconstruction filter column vector P that is also multiplied by described optimum for the row vector S that the signal of each passage is formed _jobtain j the frequency spectrum of processing frequency band, M the frequency spectrum of processing frequency band coupled together, obtain the uniform sampling signal after rebuilding.