1. An error estimation and compensation algorithm for satellite-borne multi-channel SAR moving target imaging is characterized by comprising the following steps:

firstly, distance compression is carried out on each channel data of the SAR, then channel amplitude and phase errors are estimated by adopting amplitude comparison and correlation methods and used as input of inverse filtering, and inverse filtering operation is carried out to complete spectrum reconstruction; finally, performing secondary distance compression, migration correction and azimuth compression to obtain the whole image of the SAR echo;

step two, extracting a moving target from the image domain of the whole image obtained in the step one;

step three, detecting whether the false targets on the two sides of each moving target can be observed or not in the whole image, if the false targets can be observed, measuring the amplitude ratio of the current moving target and the false targets on the two sides, wherein the amplitude ratio of the three is expressed by a formula as follows:

wherein PRF is pulse repetition frequency, phase corresponding to complex number alphaIndicating phase error between two channels, i.e.The unknown number is to be solved; d is the channel spacing, v is the satellite velocity;

calculating to obtain the phase error alpha between channels corresponding to the current moving target according to the formula (1), and then executing the step four;

if no false target exists, executing step six;

step four, intercepting data of the area where the current moving target is located in the data after the distance compression in the step one, taking the phase error alpha of the current channel as input, performing inverse filtering imaging again, and performing distance secondary compression, migration correction and azimuth compression to obtain an image of the area;

step five, whether the region where the moving target is located can observe a false target is checked again in the region image, if the false target can be detected, the amplitude ratio is measured, the current channel phase error alpha is obtained by calculation again by using the formula (1) in the step three, and the step four is executed; when the false target corresponding to the current moving target is eliminated, executing a sixth step;

and step six, executing the step three to the step five until the channel phase error estimation and compensation are completed for all the moving targets.

2. The error estimation and compensation algorithm for imaging the satellite-borne multi-channel SAR moving target as claimed in claim 1, wherein in the second step, when the moving target is extracted, a self-focusing operation is performed to compensate for the defocusing caused by the azimuth motion.

3. The error estimation and compensation algorithm for on-board multi-channel SAR moving-target imaging of claim 2, characterized in that the self-focusing operation is implemented by a maximum contrast algorithm.

4. The error estimation and compensation algorithm for satellite-borne multi-channel SAR moving target imaging as claimed in claim 1, characterized in that in the first step, when the whole image of SAR echo is obtained, the distance compression is performed on each channel data of SAR.

5. The error estimation and compensation algorithm for satellite-borne multi-channel SAR moving-target imaging as claimed in claim 4, characterized in that in the step one, after distance compression, channel amplitude and phase errors are estimated by using amplitude comparison and correlation methods, and used as input of inverse filtering, and inverse filtering operation is performed to complete spectrum reconstruction.

6. The error estimation and compensation algorithm for spaceborne multi-channel SAR moving target imaging as claimed in claim 5, wherein in the first step, after the spectrum reconstruction, the distance secondary compression, the migration correction and the azimuth compression are performed to obtain the whole image of the SAR echo.

Background

Synthetic Aperture Radar (SAR) is a full-time and all-weather high-resolution microwave remote sensing imaging radar, and can be installed on flight platforms such as airplanes and satellites. The method has unique advantages in the aspects of environmental monitoring, marine observation, resource exploration, crop estimation, mapping, military affairs and the like, and can play a role which is difficult to play by other remote sensing means.

Compared with the traditional single-channel SAR system, the multi-channel SAR realizes that the PRF is increased on the premise of keeping the range-direction ambiguity not to be reduced by increasing the sampling of the azimuth-direction space, thereby realizing higher azimuth-direction resolution and being an effective realization mode of the high-resolution wide swath SAR. Channel errors have a significant impact on the quality of multi-channel imaging, mainly resulting in the appearance of false targets, which directly affect the viewing of the image. In practice, channel errors are inevitable due to imbalance of hardware, change of posture, target motion and the like. The channel errors comprise amplitude errors and phase errors, the amplitude errors are not changed in a space mode, the traditional amplitude comparison method can be adopted for measurement, and the accuracy can meet the imaging requirement; aiming at channel phase errors, a large number of channel phase error estimation algorithms such as an interference phase method, an orthogonal sub-aperture method (OSM), a Signal Subspace Comparison Method (SSCM), an antenna direction diagram method (APM), an adaptive weighted minimum mean square error method (AWLS) and the like are provided, but the channel phase errors in application scenes of the algorithms can be regarded as fixed values, and the precondition is not satisfied in a plurality of different moving target scenes, so that the algorithms cannot be applied.

A channel error calibration algorithm based on a strong scattering source is proposed in 2017 in the university of Shanghai traffic, and the algorithm can effectively calibrate the average channel error in the whole observation process and can also effectively calibrate the channel error changing along with the direction, so that the direction ambiguity of a strong target is effectively inhibited. The algorithm has the defects that an isolated strong scattering source exists in an image, the azimuth length which can be calibrated by a single strong scattering source is short, and a plurality of calibration sources are respectively calibrated by adopting a sub-aperture method and then spliced to obtain a complete image. In fact, the purpose of this method is to estimate the phase error of the corrected azimuth space-variant, not the channel error correction for the presence of multiple different moving objects, and the algorithm is based on distance-compressed data, where there may be multiple objects overlapping in the range-doppler domain, where it is difficult to distinguish them. Although the authors also propose a preprocessing method, the algorithm further increases the computational load of the processing and the zeroing operation therein may suppress the useful signal.

Disclosure of Invention

The invention aims to overcome the defects of the prior art, and provides an error estimation and compensation algorithm for satellite-borne multi-channel SAR moving target imaging, in order to solve the problem of error estimation of a multi-channel satellite-borne SAR moving target in a complex scene with a plurality of different moving targets.

An error estimation and compensation algorithm for satellite-borne multi-channel SAR moving target imaging comprises the following steps:

firstly, distance compression is carried out on each channel data of the SAR, then channel amplitude and phase errors are estimated by adopting amplitude comparison and correlation methods and used as input of inverse filtering, and inverse filtering operation is carried out to complete spectrum reconstruction; finally, performing secondary distance compression, migration correction and azimuth compression to obtain the whole image of the SAR echo;

step two, extracting a moving target from the image domain of the whole image obtained in the step one;

step three, detecting whether the false targets on the two sides of each moving target can be observed or not in the whole image, if the false targets can be observed, measuring the amplitude ratio of the current moving target and the false targets on the two sides, wherein the amplitude ratio of the three is expressed by a formula as follows:

wherein PRF is pulse repetition frequency, phase corresponding to complex number alphaIndicating phase error between two channels, i.e.The unknown number is to be solved; d is the channel spacing, v is the satellite velocity;

calculating to obtain the phase error alpha between channels corresponding to the current moving target according to the formula (1), and then executing the step four;

if no false target exists, executing step six;

step four, intercepting data of the area where the current moving target is located in the data after the distance compression in the step one, taking the phase error alpha of the current channel as input, performing inverse filtering imaging again, and performing distance secondary compression, migration correction and azimuth compression to obtain an image of the area;

step five, whether the region where the moving target is located can observe a false target is checked again in the region image, if the false target can be detected, the amplitude ratio is measured, the current channel phase error alpha is obtained by calculation again by using the formula (1) in the step three, and the step four is executed; when the false target corresponding to the current moving target is eliminated, executing a sixth step;

and step six, executing the step three to the step five until the channel phase error estimation and compensation are completed for all the moving targets.

Preferably, in the second step, when the moving target is extracted, a self-focusing operation is performed to compensate for defocusing caused by the azimuth motion.

Preferably, the self-focusing operation is implemented by a maximum contrast algorithm.

Preferably, in the first step, when the whole image of the SAR echo is obtained, distance compression is performed on data of each channel of the SAR first.

Preferably, in the first step, after the distance compression is performed, the amplitude and phase errors of the channel are estimated by using an amplitude comparison and correlation method, and the estimated channel amplitude and phase errors are used as the input of inverse filtering to perform inverse filtering operation to complete spectrum reconstruction.

Preferably, in the first step, after the frequency spectrum is reconstructed, distance secondary compression, migration correction and azimuth compression are performed to obtain the whole image of the SAR echo.

The invention has the following beneficial effects:

aiming at the problem of channel error estimation in a complex scene with multiple moving targets, the invention provides an error estimation and compensation algorithm for imaging of satellite-borne multi-channel SAR moving targets, which estimates the intensity ratio of a false target to a real target one by one and reversely deduces channel errors according to the channel errors and the intensity ratio of the false target to the real target, thereby realizing the estimation of all moving target channel errors in the complex scene and finally compensating the channel errors to realize the inhibition of each false target;

in addition, the method also considers the defocusing phenomenon caused by target motion, performs self-focusing processing on the defocusing phenomenon, and has the effect of obtaining the two-channel SAR imaging result under the high-quality complex scene.

Drawings

Fig. 1 is a flow chart of a channel error estimation and compensation algorithm in a two-channel SAR system in a complex scene.

Fig. 2 is a schematic diagram of complex scene motion ship position setting.

Fig. 3 shows a two-channel SAR imaging result in a complex scene.

FIG. 4(a) is a result of imaging a self-focusing forward motion ship; fig. 4(b) is the imaging result of the moving ship after self-focusing.

FIG. 5 is a relationship between phase error of 2-channel SAR channels and amplitude ratio of false target to true target.

FIG. 6(A) is a false target corresponding to target A before iterative compensation of channel error; FIG. 6(B) shows the false target corresponding to target A after the iterative compensation of the channel error of 0.19 rad; FIG. 6(C) shows the false target corresponding to target A after iterative compensation of the channel error 2 π -0.19 rad.

FIG. 7(A) shows a false target corresponding to target C before iterative compensation of channel error; FIG. 7(B) shows the false target corresponding to target C after the channel error is iteratively compensated by 0.39 rad; FIG. 7(C) shows the false target corresponding to target C after the channel error is iteratively compensated for 0.1rad again.

Detailed Description

The following describes an embodiment of the method in detail with reference to the drawings and an embodiment, and a two-channel SAR system is taken as an example to describe a specific embodiment.

According to the error estimation and compensation algorithm for imaging of the satellite-borne multi-channel SAR moving target, disclosed by the invention, the distance compression is firstly carried out on each channel data, and the target is not redispersed in the distance direction after the distance compression is carried out, so that the data volume in subsequent interception can be greatly reduced.

Then, estimating channel amplitude and phase errors by adopting amplitude comparison and correlation methods, which are input of inverse filtering, and performing inverse filtering operation to complete spectrum reconstruction; and then, performing secondary distance compression, migration correction and azimuth compression to obtain the whole image. It should be noted that the phase error of the overall data estimation is equivalent to the average of the phase error of the entire map, which is reasonable for a scene such as a simple scene (e.g. a single land), however, for an image with multiple moving objects, the phase error estimation cannot achieve the channel error estimation and compensation for each moving object, so that a false object corresponding to the moving object still exists in the complex scene image with multiple different moving objects.

Extracting a moving target in an image domain, and performing self-focusing operation to compensate defocusing caused by azimuth motion, wherein a maximum contrast algorithm is adopted in a self-focusing algorithm; detecting whether the false target can be observed or not, if the false target can be observed, measuring the amplitude ratio of the false target and the real target, then reversely deducing the phase error of the channel corresponding to the target according to the amplitude ratio, intercepting the data of the region in the data after distance compression, performing inverse filtering imaging again to obtain an image, checking whether the false target can be observed or not in the region again, and finishing the estimation of the phase error of the channel of the region to eliminate the false target through several times of iteration operation.

Then, the same operation is carried out on the next moving target until the channel phase error estimation and compensation are completed on all the moving targets.

An error estimation and compensation algorithm for satellite-borne multi-channel SAR moving target imaging is shown in a flow chart of the algorithm in figure 1, and the basic implementation process is as follows:

firstly, distance compression is carried out on each channel data, and the target is not redispersed in the distance direction after the distance compression is carried out, so that the data volume intercepted by a subsequent opposite moving target can be greatly reduced; estimating channel amplitude and phase errors by adopting amplitude comparison and a correlation method, and performing inverse filtering operation to complete spectrum reconstruction by using the channel amplitude and phase errors as input of inverse filtering; and performing secondary distance compression, migration correction and azimuth compression to obtain the whole image.

Step two, extracting a moving target from the image domain of the whole image obtained in the step one, and performing self-focusing operation to compensate defocusing caused by azimuth motion, wherein the self-focusing algorithm adopts a maximum contrast algorithm;

step three, for the two-channel SAR, if a channel phase error exists, two false targets exist at the symmetrical position of each real moving target along the azimuth direction, and the image domain false target is far away from the real target PRF²/f_drA pixel, wherein PRF is the pulse repetition frequency, f_drAdjusting the frequency for Doppler; therefore, whether a false target of each moving target can be observed or not is detected in the whole image, if the false target can be observed, target amplitude ratios of the real target A, the false target A _ v1 and the false target A _ v2 are measured, and the amplitude ratios of the real target A, the false target A _ v1 and the false target A _ v2 are expressed by the following formula:

wherein the phase of the complex number alpha corresponds toIndicating phase error between two channels, i.e.The unknown number is to be solved; d is the channel distance, v is the satellite velocity, therefore, the step judges whether a false target corresponding to the moving target exists, if yes, the channel phase error alpha corresponding to the moving target is obtained by calculation according to the formula (1); if no false target exists, step six is performed.

Step four, intercepting the data of the area where the moving target is located in the data after the distance compression in the step one, taking the phase error alpha of the current channel as input, performing inverse filtering imaging again, and performing distance secondary compression, migration correction and azimuth compression to obtain an image of the area;

step five, whether the region where the moving target is located can observe a false target is checked again in the image, if the false target can be detected, the current channel phase error is obtained by measuring and calculating again by using the formula (1) in the step three, and the step four is executed; when the false target corresponding to the current moving target is eliminated, executing a sixth step;

and step six, executing the step three to the step five, and performing the same operation on the next moving target until the channel phase error estimation and compensation are completed on all the moving targets.

It can be seen that the algorithm can realize channel error estimation and compensation corresponding to all moving targets by sequentially processing each moving target in an image domain, so that the algorithm is suitable for complex scenes with a plurality of moving targets.

Example (b):

in order to verify the effectiveness of the two-channel SAR channel error estimation and compensation method in a complex scene with a plurality of different moving targets, the present embodiment adopts surface target data to develop a simulation experiment, the SAR system simulation parameters are consistent with those in table 1, the moving target position is set as shown in fig. 2, and considering that the sea surface ship speed is generally 0-30 knots, the moving target speed setting is shown in table 2. Setting the overall channel error as α 2 · exp (j π/2), and because there are multiple moving targets, then using a channel error estimation algorithm, where the estimation result is α _ esti ═ 2.001 · exp (j · 1.37), the amplitude error is estimated accurately, the phase error estimation result and the overall channel error have a phase error of 0.2rad, and the imaging result is as shown in fig. 3, the channel phase errors of these moving ships are about 0 to 0.6rad, the false targets on the graph after the modulus quantization should be very weak, however, because the actual quantization usually does not use a simple modulus normalization quantization method, but uses more complex image parameters (such as mean and variance) for quantization to achieve a better observable visual effect, so as to achieve the effect of weak targets and also to influence the observation due to the occurrence of false targets, as shown in fig. 3, the circle is a false target. It can be seen that so many false objects can affect the severe image viewing.

TABLE 1 partial System parameters

TABLE 2 Complex scene two-channel SAR motion naval vessel speed settings

Refocusing the moving target, fig. 4 shows the imaging result after the ship self-focusing, where the image is in the azimuth direction in the transverse direction and in the distance direction in the longitudinal direction, it can be seen that the target before the self-focusing has the defocus in the azimuth direction, and the target after the self-focusing has a good focus in the azimuth direction.

The relationship between the channel phase error and the amplitude ratio of the false target and the real target is drawn according to the formula 1, as shown in fig. 5, it can be seen that because the amplitude ratio is a modulus ratio and because of the particularity of the formula (1), one amplitude ratio of the false target and the real target corresponds to two channel phase errors symmetrical by pi, which causes difficulty in deriving the channel phase error from the amplitude ratio of the false target and the real target, and the method for solving the problem is to adopt two possible phase errors to image respectively, the correct phase error can reduce the amplitude of the false target, and the wrong phase error can increase the amplitude of the false target, so that the real channel phase error can be judged.

The parameters and channel error settings at the current moving target are analyzed and table 3 shows the theoretical channel errors for different moving targets. And extracting the moving target with the false target, wherein whether the false target exists in the target can be determined by observing whether the false target exists at the position where the false target appears. Because the actual amplitude ratio of the measured false target to the real target is lower than the theoretical value due to the false target defocusing, the false target defocusing is found to cause amplitude loss of about 6dB through simulation calculation. Firstly, selecting a target A, measuring to obtain the actual amplitude ratio of the target false target to the real target to be-32 dB, compensating the amplitude ratio loss caused by defocusing to be-26 dB, and setting the corresponding channel error to be 0.19rad or 2 pi-0.19 rad. The corresponding channel errors are respectively adopted for spectrum reconstruction and imaging, the obtained results are shown in fig. 6, it can be seen that the false target before the channel error is iteratively compensated is obvious, the false target disappears after the channel error is iteratively compensated by 0.19rad, the channel error compensation is correct, the false target is larger after the channel error is iteratively compensated by 2 pi-0.19 rad, and the channel error compensation is wrong, so that the estimation of the target A channel error is completed, and is 0.19 rad.

TABLE 3 corresponding channel error of each sports ship

Observing object B finds that object B has no false object, that is, B does not need channel error estimation, and in fact the channel error of object B is-0.02 rad, so that false objects caused by small channel errors cannot be observed. And observing and measuring that the actual amplitude ratio of the false target and the real target of the target C is-26 dB, wherein the amplitude ratio is-20 dB after the amplitude ratio is lost due to defocusing compensation, and the corresponding channel error is 0.39rad or 2 pi-0.39 rad. The false target amplitude is greatly reduced after the iterative compensation channel error is 0.39rad but still exists, the actual amplitude ratio of the false target to the real target at the moment of the measurement target C is-38 dB, the amplitude ratio is-32 dB after the amplitude ratio loss caused by defocusing compensation, and the corresponding channel error is 0.1rad or 2 pi-0.1 rad. The false target amplitude disappears after compensating the channel error by 0.1 rad. The target C channel error estimate is thus completed to 0.49 rad. FIG. 7 illustrates a false target corresponding to a target C before and after iterative compensation of a channel error, where initially the channel error corresponding to the target C is large, as shown in FIG. (A), and at this time the false target has a large amplitude; after the channel error is compensated by the first iteration, the false target amplitude is greatly reduced, as shown in a graph (B); the false target amplitude substantially disappears after the second iteration compensates for the channel error, as shown in graph (C). This demonstrates the effectiveness of the iterative image domain channel error estimation and compensation algorithm.

By the operation, all false targets are eliminated after channel error estimation and compensation are carried out on each target with the false target. Table 4 shows the residual channel errors of each moving ship after channel error estimation and compensation, and it can be seen that all target residual channel errors are less than 0.1rad, and such low channel errors ensure that false targets do not appear. And the maximum number of iterative operation times in channel error estimation is 2, and the interception of the data after the distance compression ensures that the data volume is about 1% of the original data volume in the iterative operation, so that the algorithm efficiency is very high.

TABLE 4 residual channel errors of each moving ship after channel error estimation and compensation

Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, it will be apparent to those skilled in the art that various modifications may be made without departing from the principles of the invention and these are considered to fall within the scope of the invention.