CN102680973B - Method for removing repeat-pass InSAR (Interferometric Synthetic Aperture Radar) nonsynchronous signals based on burst mode - Google Patents

Abstract

The invention discloses a method for removing repeat-pass InSAR (Interferometric Synthetic Aperture Radar) nonsynchronous signals based on a burst mode. The method is used for interfering removal of nonsynchronous signals in combination with azimuth time domain convolution imaging, and can be applied to interference processing of mode data such as ScanSAR and the like. The method comprises the following steps of: calculating the burst synchronism by using SAR original signal data, additional pass data and the like through a similar InSAR space base line calculating method, and setting a time node for distinguishing synchronous data from nonsynchronous data; and after azimuth time domain convolution imaging, extracting a phase only containing a synchronous signal convolution result for serving as a phase of an output picture element, and extracting an amplitude only containing a convolution result of synchronous signals and nonsynchronous signals for serving as an amplitude of the output picture element, so that amplitude information is ensured while nonsynchronous signals are removed. The method can further be used for processing a group of interference data; and under the condition, a group of synchronisms is required to be calculated and a group of time nodes is required to be set, so that a position of an imaging result corresponds to a group of output picture elements at last.

Description

Repeat track InSAR nonsynchronous signal removal method based on the burst pattern

Technical field

The present invention relates to a kind of interfering synthetic aperture radar and measure (InSAR) data processing method, specifically is a kind of nonsynchronous signal removal method of the repeat track InSAR data based on the burst pattern.

Background technology

Satellite-borne synthetic aperture radar in recent years (SAR) system almost all possesses the mode of operation based on burst, as the ScanSAR pattern, has expanded the spatial observation scope of SAR greatly.The same with conventional band pattern, the data of burst pattern also can be used for interfering handling to obtain DEM or face of land deformation data.

Burst is the set of pulses that SAR launches and receives in flight course, be made up of mutual discontinuous burst one by one based on the data of burst pattern, thereby the orientation of terrain object is discontinuous to Doppler frequency spectrum.If the burst data are interfered processing, it is certain overlapping then to require the Doppler frequency spectrum of major-minor image burst to have, namely require major-minor image when data are obtained certain scan-synchronized will be arranged, this control to satellite orbit and attitude thereof has proposed very high requirement.In fact, more or less all there is certain non-synchronous data in the burst data of obtaining for repeat track SAR, and it also can bring noise to interferogram, so it should be removed to improve the quality of interferogram on the contrary to the not contribution of resolution of interferogram.Present burst nonsynchronous signal removal method mainly contains two kinds:

(1) in raw data, directly non-synchronous data is removed before the imaging.This method has also been removed its corresponding amplitude information when removing asynchronous phase information, be unfavorable for the image registration of back.

(2) before interfering processing, by filtering nonsynchronous signal is removed after the imaging, but the burst fringe region can produce noise because of the part convolution after the filtering.

Summary of the invention

The present invention proposes a kind of new method, is the burst nonsynchronous signal to be removed in imaging process in the orientation, can effectively overcome the problem that above-mentioned two kinds of methods exist.

Technical scheme of the present invention is as follows:

A kind of repeat track InSAR nonsynchronous signal removal method based on the burst pattern comprises the steps:

1) the original SAR signal data of input and subsidiary orbital data thereof etc.;

2) utilize SAR locus and the speed of major-minor image correspondence, calculate the synchronism of major-minor image burst by the temporal information of interative computation and burst;

3) with step 2) in the synchronism that calculates and pre-set threshold relatively, if synchronism is higher than threshold value, so burst is carried out distance to compression with apart from proofreading and correct to migration; Then data are carried out the orientation to compression, simultaneously nonsynchronous signal is removed.Otherwise, abandon interfering and handle.

Described repeat track InSAR nonsynchronous signal removal method based on the burst pattern is characterized in that, in the step 1), can import a pair of SAR data, also can import one group of SAR data; Simultaneously, also need import the subsidiary orbital data of SAR data etc., comprising locus and the speed of SAR.

Described repeat track InSAR nonsynchronous signal removal method based on the burst pattern is characterized in that step 2) in, if handle a pair of SAR data, so only estimate that this is to the synchronism of data; If handle one group of SAR data, to estimate so

Individual synchronism, x presentation video scape number wherein, x〉2.

Described repeat track InSAR nonsynchronous signal removal method based on the burst pattern is characterized in that, in the step 3), process following (as shown in Figure 2, corresponding to the operation in the figure l frame of broken lines) is removed to compression and nonsynchronous signal in concrete orientation:

3.1) at burst timing node is set according to synchronism result of calculation, to distinguish synchronous and asynchronous data;

3.2) burst and orientation are multiplied each other to matched filter, and multiplied result added up, get the phase place that only contains the synchronizing signal accumulation result as the phase place of output pixel, get contain synchronizing signal and nonsynchronous signal accumulation result amplitude as the amplitude of exporting pixel.

Described repeat track InSAR nonsynchronous signal removal method based on the burst pattern is characterized in that step 3.1) in, if processing is a pair of SAR data, this so only is set to the timing node of data; If what handle is one group of SAR data, the right timing node of corresponding SAR data is set respectively then.

Described repeat track InSAR nonsynchronous signal removal method based on the burst pattern, it is characterized in that, step 2) method of the synchronism of the major-minor image burst of described calculating is: for the data line among burst of master image, calculate its corresponding SAR locus (R _M) and speed, it is capable to seek burst corresponding in the auxilliary image by interative computation then; The capable corresponding SAR position (P of burst when auxilliary image _S) on the given velocity reversal of master image with P _MThe corresponding deviation of master image corresponding row stops iteration during less than pre-set threshold; Find out the corresponding burst of major-minor image capable after, according to the zero-time of burst separately, calculate corresponding row time of burst zero-time relatively separately, can determine the burst synchronism of major-minor image.

Advantage of the present invention mainly contains:

With respect to the method for directly removing non-synchronous data in raw data, this method has kept whole amplitude informations, and this will be conducive to the registration between image; With respect to the method for removing nonsynchronous signal after the imaging by filtering, the noise of bringing for the burst fringe region because of the part convolution when this method has been avoided filtering.

Aspect calculated amount, with respect to the method for by filtering nonsynchronous signal being removed before interfering processing after the imaging, when the processing single interference was right, because the time domain convolution algorithm plays a leading role, calculated amount of the present invention was the former several times; But handle interference data in enormous quantities to the time, because filtering plays a leading role, calculated amount of the present invention is on the contrary less than the former.

Description of drawings

Fig. 1 is for adopting the interference processing flow chart based on the burst pattern of the inventive method.

Fig. 2 is the concrete operations synoptic diagram in Fig. 1 frame of broken lines.

Embodiment

Below in conjunction with accompanying drawing the described technical scheme of summary of the invention is elaborated.Fig. 1 is for adopting the interference processing flow chart based on the burst pattern of the inventive method.Fig. 2 is the concrete operations synoptic diagram in Fig. 1 frame of broken lines.

When the original SAR signal data of input, handle equally with conventional interference, need to extract and interfere the parameters of using in the processing, also to calculate information such as doppler centroid simultaneously.

The synchronism of burst can be calculated by the subsidiary orbital data of SAR data, and its process is similar to the calculating of interfering Space Baseline: for the data line among burst of master image, calculate its corresponding SAR locus (P _M) and speed, it is capable to seek burst corresponding in the auxilliary image by interative computation then.The capable corresponding SAR position (P of burst when auxilliary image _S) on the given velocity reversal of master image with P _MDeviation during less than pre-set threshold, stop iteration; Find out the corresponding burst of major-minor image capable after, according to the zero-time of burst separately, calculate corresponding row time of burst zero-time relatively separately, can determine the burst synchronism of major-minor image.

Compare with conventional band pattern signal, the burst signal with its distance to characteristics identical, the difference of the two be embodied in the orientation to.Be higher than the interference data that preestablishes threshold value for synchronism, burst carried out distance compression and distance after the migration correction, the present invention carries out nonsynchronous signal in conjunction with the orientation to imaging processing and removes.

Ignore the orientation to antenna radiation pattern, the orientation of a burst can be described as to signal

$s_a(\mathrm{\η},{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{cn}})=\underset{n=0}{\overset{+\∞}{\mathrm{\Σ}}}{s}_{\mathrm{an}}(\mathrm{\η},{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{cn}})$ $\left(1\right)$

$=\underset{n=0}{\overset{+\∞}{\mathrm{\Σ}}}\mathrm{rect}\left(\frac{\mathrm{\η}-{\mathrm{\η}}_{\mathrm{cn}}}{T_b}\right)\·\exp \{-j\frac{4\mathrm{\π}\sqrt{R_0^2+V_r^2(\mathrm{\η}-{\mathrm{\η}}_{\mathrm{dn}})}}{\mathrm{\λ}}\}$

Wherein, η be the orientation to the time, η _DnBe zero Doppler's moment of point target, η _CnBe the burst center of point target, T _bBe the burst duration, R ₀Be low coverage, λ is radar wavelength, V _rBe effective radar speed, n is point target number, s _An() be the orientation of a single point target to signal, rect () is rectangular function.

The complete signal copy of a point target is

$g_{\mathrm{an}}(\mathrm{\η},{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{dcn}})=h_{\mathrm{an}}^{*}(-\mathrm{\η},{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{dcn}})$

$=\mathrm{rect}\left(\frac{\mathrm{\η}-{\mathrm{\η}}_{\mathrm{dcn}}}{T_a}\right)\·\exp \{j\frac{4\mathrm{\π}[R_c-\sqrt{R_0^2+V_r^2(\mathrm{\η}-{\mathrm{\η}}_{\mathrm{dn}})}]}{\mathrm{\λ}}\}---\left(2\right)$

Wherein, η _DcnFor the corresponding orientation of beam center to the time, R _cBe the oblique distance of beam center correspondence, T _aBe the whole synthetic aperture time, h _An() is corresponding matched filter, and * represents conjugation.In the following formula with R _cThe phase place of bringing is considered as constant term, and this phase term is the phase place of target after the imaging, it is not included in the point target signal herein, and all the other are approximately linear frequency modulation phase place.By the time domain convolution algorithm burst is carried out the orientation to imaging, can finish by following relevant or convolution so

${\∫}_{-\∞}^{+\∞}{s}_{\mathrm{an}}(u,{\mathrm{\η}}_{\mathrm{an}},{\mathrm{\η}}_{\mathrm{cn}})g_{\mathrm{an}}^{*}(u-\mathrm{\η},{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{dcn}})\mathrm{du}$ $\left(3\right)$

$={\∫}_{-\∞}^{+\∞}{s}_{\mathrm{an}}(u,{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{cn}})h_{\mathrm{an}}(\mathrm{\η}-u,{\mathrm{\η}}_{\mathrm{dn}},{\mathrm{\η}}_{\mathrm{dcn}})\mathrm{du}.$

The convolutional calculation result of a point adding up after to be exactly that matched filter and burst are point-to-point multiply each other, adding up is that pointwise is carried out.Suppose the synchronism of known major-minor image, so timing node can be set before adding up, to distinguish synchronously and asynchronous data, cumulative process is from an end of synchrodata then, when cumulative process has reached this timing node, accumulation result leaves among the variable p, adds up then and proceeds, and last accumulation result leaves among the variable m.Only comprised synchronous signal among the p, but its amplitude resolution has reduced; Both comprised synchronizing signal among the m, and also comprised nonsynchronous signal, its amplitude resolution is corresponding to whole burst bandwidth.What final imaging results needed is the phase place of p and the amplitude of m, so the amplitude of p is replaced to the amplitude of m, its advantage is both to have removed asynchronous signal in the interferometric phase, has kept the amplitude of full resolution (corresponding to whole burst bandwidth) again.As can be seen, the extra computation of removing corresponding to nonsynchronous signal in the whole process only is the amplitude replacement operation.

At present, often will handle large batch of interferogram (Interferogram Stack) in the interventional applications, one group of SAR data is namely arranged, each the scape data in the data all might be right with these other data composition interference of organizing in the data, to interfere processing, the present invention also can be used for this situation.At this moment, for other pairing interference data, the nonsynchronous signal of scape data both may be at the head end of burst, also may be at the tail end of burst.For scape SAR data, suppose to have k and l to carry out the removal of nonsynchronous signal to needing respectively from burst head end and tail end to interfering.Before adding up, a k and l timing node is set respectively, it is right to interfering to correspond respectively to above-mentioned k and l.Suppose to begin to add up from the burst head end of correspondence, add up finish after, the part accumulation result can exist in the following vector

P=(p′ ₁，p′ ₂，…，p′ _k，p ₁，p ₂，…，p _l，) （4）

Wherein, (p ' ₁, p ' ₂..., p ' _k) and (p ₁, p ₂..., p _l) correspond respectively to k and l is right to interfering.And complete convolution results leaves among the variable m.Being used for next step so interferes the imaging results of handling to be expressed as

O=(P+A)B （5.1）

Wherein

For a burst, its calculated amount comprises that time domain convolution algorithm and the inferior amplitude of z (k+l) substitute to be calculated, and wherein z is expressed as the orientation of picture back burst to the pixel number.

Claims (5)

1. the repeat track InSAR nonsynchronous signal removal method based on the burst pattern comprises the steps:

1) the original SAR signal data of input and subsidiary orbital data thereof;

2) utilize SAR locus and the speed of major-minor image correspondence, calculate the synchronism of major-minor image burst by the temporal information of interative computation and burst;

3) with step 2) in the synchronism that calculates and pre-set threshold relatively, if synchronism is higher than threshold value, so burst is carried out distance to compression with apart from proofreading and correct to migration; Then data are carried out the orientation to compression, simultaneously nonsynchronous signal is removed; Otherwise, abandon interfering and handle;

In the step 3), concrete orientation is as follows to compression and nonsynchronous signal removal process:

3.1) at burst timing node is set according to synchronism result of calculation, to distinguish synchronous and asynchronous data;

3.2) burst and orientation are multiplied each other to matched filter, and multiplied result added up, get the phase place that only contains the synchronizing signal accumulation result as the phase place of output pixel, get contain synchronizing signal and nonsynchronous signal accumulation result amplitude as the amplitude of exporting pixel.

2. the repeat track InSAR nonsynchronous signal removal method based on the burst pattern as claimed in claim 1 is characterized in that, in the step 1), can import a pair of SAR data, also can import one group of SAR data; Simultaneously, also need import the subsidiary orbital data of SAR data, comprising locus and the speed of SAR.

3. the repeat track InSAR nonsynchronous signal removal method based on the burst pattern as claimed in claim 1 is characterized in that step 2) in, if handle a pair of SAR data, so only estimate that this is to the synchronism of data; If handle one group of SAR data, to estimate so

Individual synchronism, x presentation video scape number wherein, x＞2.

4. the repeat track InSAR nonsynchronous signal removal method based on the burst pattern as claimed in claim 1 is characterized in that step 3.1) in, if processing is a pair of SAR data, this so only is set to the timing node of data; If what handle is one group of SAR data, the right timing node of corresponding SAR data is set respectively then.

5. the repeat track InSAR nonsynchronous signal removal method based on the burst pattern as claimed in claim 1, it is characterized in that, step 2) method of the synchronism of the major-minor image burst of described calculating is: for the data line among burst of master image, calculate its corresponding SAR locus P _MAnd speed, it is capable to seek burst corresponding in the auxilliary image by interative computation then; The capable corresponding SAR position P of burst when auxilliary image _SOn the given velocity reversal of master image with P _MThe corresponding deviation of master image corresponding row stops iteration during less than pre-set threshold; Find out the corresponding burst of major-minor image capable after, according to the zero-time of burst separately, calculate corresponding row time of burst zero-time relatively separately, can determine the burst synchronism of major-minor image.

Images