CN101639530B - SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation - Google Patents
SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation Download PDFInfo
- Publication number
- CN101639530B CN101639530B CN2009100833457A CN200910083345A CN101639530B CN 101639530 B CN101639530 B CN 101639530B CN 2009100833457 A CN2009100833457 A CN 2009100833457A CN 200910083345 A CN200910083345 A CN 200910083345A CN 101639530 B CN101639530 B CN 101639530B
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- sar
- echo signal
- dimensional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000009466 transformation Effects 0.000 title claims abstract description 34
- 238000007781 pre-processing Methods 0.000 title claims abstract description 22
- 238000003384 imaging method Methods 0.000 claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 22
- 238000005070 sampling Methods 0.000 claims abstract description 10
- 238000001228 spectrum Methods 0.000 claims description 16
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 abstract description 5
- 230000000452 restraining effect Effects 0.000 abstract description 2
- 230000001131 transforming effect Effects 0.000 abstract 2
- 238000003672 processing method Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 5
- 238000012805 post-processing Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000593 degrading effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000011888 foil Substances 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
Images
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides an SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation. The method is the preprocessing of an original echo signal before SAR imaging processing, aiming at de-noising the echo signal before noise mixes into an SAR imaging system to heighten the signal to noise ratio of the SAR imaging system and improve the imaging quality of the SAR image. The method comprises the following steps: (1) obtaining the SAR original echo signal, and obtaining a two-dimensional base band array echo signal for preprocessing by sampling and demodulating; (2) two-dimensional mixed transforming (DFT-DWT) the echo signal; (3) removing an out-band coefficient part and a high-frequency coefficient part of the echo signal, wherein the out-band coefficient part and the high-frequency coefficient part are two-dimensionally mixed transformed; (4) inversely two-dimensionally mixed transforming the signal after coefficient operation; and (5) taking the preprocessed echo signal as an input signal of the SAR imaging processing system, and generating the SAR image. The method is suitable for a preprocessing stage in the SAR imaging processing system, and has better effect in restraining strong interference, detecting weak objects, improving the SAR inmaging quality, etc.
Description
Technical Field
The invention relates to the technical field of Synthetic Aperture Radar (SAR) signal processing, which is used in an SAR imaging system. In particular, the invention relates to a denoising processing method aiming at SAR original echo signals in a preprocessing stage before SAR imaging processing. The method can be applied to the fields of noise elimination of the SAR imaging system, weak target detection, SAR image processing and the like.
Background
Synthetic Aperture Radars (SAR) have attracted extensive research interest for more than half a century due to their advantages of high resolution, high signal-to-noise ratio, and all-weather, and have been used in a variety of fields, such as terrain mapping, ground monitoring, resource exploration, and the like. In military applications in particular, synthetic aperture radar has significant capabilities in both battlefield terrain mapping and target detection, and therefore plays an increasingly important role. In practice, however, the SAR system inevitably suffers from interference from various noise sources, such as thermal noise, ground or ocean clutter, electromagnetic interference, etc., and thus its imaging quality and detection of targets are severely affected.
In recent decades, many SAR image processing methods have emerged to address the problem of SAR image speckle noise, where typical algorithms such as Lee, Kuan, Frost, etc. However, all of these processes can be classified as post-processing methods, which are processing methods that generate images for SAR after SAR imaging processing, and thus are greatly affected by SAR imaging quality. But in some severe cases, the imaging performance of SAR will be severely deteriorated. For example, echo signals reflected by very distant targets will be disturbed by various noise and clutter present in the space and thus become very weak. In addition, for some anti-radar targets (such as stealth aircraft), the echo signal becomes too weak to be detected. In addition, SAR imaging systems suffer from ground or ocean clutter interference, atmospheric cloud occlusion, foil strip interference in electronic warfare, and interference from enemy jammers, among others. In these cases, the target signal may be completely drowned in noise, thereby extremely degrading the signal-to-noise ratio of the SAR image. Since strong noise has been difficult to separate by the hybrid SAR imaging system after SAR image generation, post-processing methods have become more limited in improving SAR imaging quality and target detection.
SAR preprocessing is a processing method for SAR raw echo signals prior to SAR imaging processing. Compared with the SAR generated image, the SAR original echo signal retains more characteristics of the target echo signal. Unlike an irregular noise signal, the target echo signal has some unique characteristics that can be used to separate and cancel the noise. Especially when the interference is very strong, the contribution of the preprocessing is more significant, which can obviously improve the signal-to-noise ratio of the SAR system. Therefore, an effective denoising preprocessing method can be designed, which is very important for improving the performance of the SAR system and improving the effect of post-processing.
The SAR echo signal is a typical two-dimensional space-time signal that can be processed using some two-dimensional signal processing method. Existing two-dimensional signal processing methods, which typically employ a two-dimensional Discrete Fourier Transform (DFT) or a two-dimensional Discrete Wavelet Transform (DWT) to remove noise, do not fully take into account that some signals may have different characteristics in two directions. Although Wavelet Transform (WT) has good local characteristics in the time-frequency domain, it is difficult to remove narrowband interference due to its poor frequency resolution. Unlike DWT, DFT has a good ability to analyze spectral characteristics, and it can be used to remove out-of-band noise, but is not very effective for wideband noise cancellation. Since DFT and DWT have different characteristics, it is sometimes not sufficient to use only one type of transform in both directions when analyzing complex signals. Through analysis, a target echo signal in the SAR echo signals is a band-limited signal in a time domain, and the frequency band of the target echo signal is determined by the radar signal generated by the SAR system; but it is a broadband signal in the spatial domain, and its frequency characteristics are uncertain. Due to the characteristic of the SAR echo signal, the conventional two-dimensional signal processing method is not very effective in extracting the target signal and removing the noise in the SAR echo signal.
Disclosure of Invention
Considering that a target echo signal in the SAR echo signal is a band-limited signal in the time domain, and the frequency band of the target echo signal is determined by a radar signal generated by the SAR system; however, the SAR echo signal denoising preprocessing method is a broadband signal in a space domain, the frequency characteristic of the SAR echo signal is uncertain, the traditional two-dimensional signal processing method is not very effective in the aspects of extracting a target signal and removing noise in the SAR echo signal, and the SAR echo signal denoising preprocessing method based on two-dimensional hybrid transform (DFT-DWT) is provided by the invention.
The invention aims to improve the anti-interference capability of the SAR system, improve the imaging quality of the SAR system and provide a higher-quality denoised image for subsequent SAR image processing and post-processing work such as target detection and identification.
The method of the invention comprises the following steps: before the SAR system is subjected to imaging processing, a denoising pretreatment is firstly carried out on an original echo signal, so that noise and interference signals are filtered once before being mixed into the SAR imaging system, and an original target signal is not damaged.
The invention adopts a two-dimensional hybrid transform (DFT-DWT) method for denoising SAR echo signals. The two-dimensional hybrid transform is a new processing algorithm for analyzing and processing complex two-dimensional signals, and different transforms and inverse transforms can be adopted according to different noise distributions of the signals in the horizontal direction and the vertical direction. The invention combines and uses the discrete Fourier transform and the discrete wavelet transform in two directions of signals and processes SAR echo signals in a mixed transform domain. The invention flexibly applies two transformations to different directions of the SAR echo signal, and solves the problem that the traditional two-dimensional transformation can not process complex noise with different characteristics in different directions of the SAR echo signal.
The present invention employs the following definition of two-dimensional hybrid transform (DFT-DWT).
For a two-dimensional N1×N2S (n) of1，n2) The two-dimensional hybrid transform (DFT-DWT) is realized by the following steps:
(1) for s (n)1，n2) Along a first variable n1One-dimensional discrete Fourier transform of direction
Wherein <math><mrow><msub><mi>W</mi><msub><mi>N</mi><mn>1</mn></msub></msub><mo>=</mo><mi>exp</mi><mrow><mo>(</mo><mo>-</mo><mi>j</mi><mfrac><mrow><mn>2</mn><mi>π</mi></mrow><msub><mi>N</mi><mn>1</mn></msub></mfrac><mo>)</mo></mrow><mo>,</mo></mrow></math> S(k1，n2) Is n1The spectrum in the direction.
(2) For S (k)1，n2) Along a second variable n2One-dimensional discrete wavelet transform with one layer of decomposition is carried out in the direction to obtain the original signal s (n)1，n2) Two-dimensional hybrid transformation of (2), denoted as [ X ]L，XH]The form is as follows:
wherein
And
respectively, orthogonal scale equation and wavelet equation, m being the number of layers of decomposition, XLAnd XHAre respectively S (k)1，n2) The low and high frequency parts of the wavelet decomposition.
In addition, the implementation steps of the corresponding inverse two-dimensional hybrid transformation are as follows:
(1) for [ X ]L，XH]Along n2The direction is subjected to one-dimensional inverse discrete wavelet transform, which has the following form:
(2) for S (k)1，n2) Along n1One-dimensional inverse discrete Fourier transform is carried out on the direction to reconstruct the original signal s (n)1，n2)
The denoising principle of SAR pretreatment in the invention is as follows: the SAR raw echo signal s (t, u) is a two-dimensional space-time signal. Unlike the signals generated by interference and noise, the echo signal of the target is a band-limited signal in the time domain t direction, and the frequency band of the echo signal is determined by the frequency band of the radar signal; but it is a broadband signal in the spatial domain u direction, and its frequency characteristics are uncertain. In addition, since the reflecting surface of the target is generally continuous in space, the time domain spectrum of the echo signal of the target has a large correlation in the spatial domain.
The invention takes the characteristics of the target echo signal into consideration, and adopts two-dimensional mixed transform (DFT-DWT) to ensure that the noise in two directions of space and time in the echo signal can be well separated and removed: in the time domain direction, the one-dimensional Fourier transform is carried out on the echo signal, and the target signal can be separated into a pass band of a time domain frequency spectrum, so that out-of-band noise in the time domain direction can be effectively removed; because the time domain frequency spectrum of the target signal has larger continuity in the spatial domain direction, after the one-dimensional wavelet transform is carried out on the spatial domain direction of the time domain frequency spectrum, the target signal is mainly concentrated on the low-frequency part, so that the noise in the spatial domain direction can be effectively removed after the high-frequency part after the wavelet transform is removed.
The SAR echo signal denoising preprocessing method provided by the invention comprises the following specific implementation steps:
step 1: obtaining a two-dimensional space-time echo signal s (t, u) containing noise received by an SAR system, sampling the time domain t direction to obtain a two-dimensional array echo signal, and determining omega [ omega ] as a frequency band according to the Nyquist theoremc-ω0，ωc+ω0]SAR radar signal (ω)cIs the carrier frequency, omega0Single-sided bandwidth), sampling frequency ωsShould be greater than 2 omega0. Echo signal s (n)1，n2) Demodulation to baseband by sb(n1，n2)＝s(n1，n2)·exp(-jωcn1) Wherein s isb(n1，n2) Obtaining a baseband two-dimensional array echo signal.
Step 2: will sb(n1，n2) As a pretreatmentThe input signal is denoised, and the specific processing steps are as follows:
(a) to sb(n1，n2) Performing two-dimensional hybrid transformation, wherein the transformed result is [ XL，XH]The specific transformation process is described by the above formulas (1) and (2).
(b) Two-dimensional hybrid transformation [ X ] of echo signalsL，XH]Performing coefficient operation: remove [ X ]L，XH]The out-of-band and high-frequency coefficient portions are specifically operated as shown in
(c) For [ X ] after coefficient operationL，XH]Performing inverse two-dimensional mixed transformation, and reconstructing the denoised echo signal s by referring to the formulas (3) and (4)d(n1，n2)。
And step 3: to denoise the preprocessed output signal sb(n1，n2) And carrying out SAR imaging processing to reconstruct an SAR image.
According to one aspect of the invention, a SAR echo signal denoising preprocessing method based on two-dimensional hybrid transformation is provided, which comprises the following steps:
(1) acquiring an SAR target echo signal;
(2) carrying out time domain sampling on the SAR target echo signal;
(3) time domain demodulation is carried out on the result obtained by time domain sampling to obtain SAR baseband echo signal sb(n1，n2)；
(4) For the SAR baseband echo signal sb(n1，n2) Performing two-dimensional hybrid transformation to obtain two-dimensional hybrid transformation [ X ] of SAR target echo signalL，XH]；
(5) Two-dimensional hybrid transformation [ X ] of SAR target echo signalsL，XH]Performing coefficient operation to remove out-of-band and high-frequency coefficient parts to obtain the result [ X ] after the coefficient operationL，XH]d；
(6) For the result [ X ] after coefficient operationL，XH]dInverse two-dimensional hybrid transformation is carried out to obtain a denoised echo signal sd(n1，n2)；
(7) De-noising the echo signal sd(n1，n2) And as an input signal of the SAR imaging processing system, generating an SAR image by utilizing an SAR reconstruction algorithm.
According to an aspect of the present invention, the two-dimensional hybrid transformation in the above step (4) includes the steps of:
(4-1) along SAR base band echo signal sb(n1，n2) The time domain direction of the SAR target echo signal is subjected to one-dimensional discrete Fourier transform to obtain a time domain frequency spectrum S of the SAR target echo signalb(k1，n2)，
Wherein <math><mrow><msub><mi>W</mi><msub><mi>N</mi><mn>1</mn></msub></msub><mo>=</mo><mi>exp</mi><mrow><mo>(</mo><mo>-</mo><mi>j</mi><mfrac><mrow><mn>2</mn><mi>π</mi></mrow><msub><mi>N</mi><mn>1</mn></msub></mfrac><mo>)</mo></mrow><mo>,</mo></mrow></math> Sb(k1，n2) For the time domain n of the baseband echo signal1Frequency spectrum in the direction, Sb(k1，n2) The target signal contained in the signal has a baseband frequency band range of <math><mrow><msub><mi>k</mi><mn>1</mn></msub><mo>∈</mo><mo>[</mo><mo>-</mo><msub><mi>N</mi><mn>1</mn></msub><mfrac><msub><mi>ω</mi><mn>0</mn></msub><msub><mi>ω</mi><mi>s</mi></msub></mfrac><mo>,</mo><msub><mi>N</mi><mn>1</mn></msub><mfrac><msub><mi>ω</mi><mn>0</mn></msub><msub><mi>ω</mi><mi>s</mi></msub></mfrac><mo>]</mo><mo>;</mo></mrow></math>
(4-2) aligning the time domain spectrum S in the spatial domain directionb(k1，n2) Performing one-dimensional wavelet transform, whose wavelet decomposition process is as follows
Wherein
Andrespectively orthogonal scale equation and wavelet equation, m is the number of decomposed layers, and the result [ X ] after two-dimensional mixed transformation is obtained by the formulaL，XH]Wherein X isL(k1，k21) and XH(k1，k21) are each Sb(k1，n2) Along the spatial domain n2Low and high frequency parts after directional wavelet decomposition.
According to an aspect of the present invention, the inverse two-dimensional hybrid transformation in the above step (6) includes the steps of:
(6-1) pairs of [ XL，XH]dEdge k2One-dimensional inverse discrete wavelet transform is carried out in the direction to remove the high-frequency part XH(k1，k21), in the form:
(6-2) to Sd(k1，n2) Edge k1One-dimensional inverse discrete Fourier transform is carried out on the direction to reconstruct a denoised echo signal sd(n1，n2) The form is as follows:
for the purpose of further illustrating the principles and features of the present invention, reference will now be made in detail to the present invention, examples of which are illustrated in the accompanying drawings.
Drawings
FIG. 1 is a flow diagram according to an embodiment of the present invention.
Fig. 2 is a noiseless SAR reconstructed image.
Fig. 3 shows the result of two-dimensional hybrid transformation of the target echo without noise.
Fig. 4 shows the result of two-dimensional hybrid transformation of the echo signal containing noise.
Fig. 5 is the result of a two-dimensional hybrid transform of a noisy echo signal after a coefficient operation.
FIG. 6 is a comparison of echo signals with different SNR after denoising pre-processing according to an embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings.
The satellite-borne or airborne SAR imaging system flies at a constant speed during working, transmits radar signals to the ground at the same time interval, and receives SAR echo signals at the same time. The radar signal emitted by the SAR imaging system has radar signal parameters such as specific waveform, carrier frequency, frequency bandwidth and the like, and the SAR imaging system samples, demodulates and images the echo signal by using the parameters. In practice, the SAR imaging system is subject to strong interference from various noise sources, and these interference noise signals are received by the SAR system together with the target echo signal and subjected to imaging processing, thereby seriously affecting the reconstruction quality of the SAR image.
FIG. 1 is a flow diagram according to an embodiment of the present invention. According to the method provided by the invention, the actual SAR imaging system is simulated, the condition that the SAR system is subjected to strong interference is simulated in the simulation, and Gaussian white noise with a certain noise-signal ratio is added in an echo signal. The SAR echo signal is denoised by utilizing the SAR denoising preprocessing method in the invention.
Referring to the implementation flowchart 1, the specific implementation process of denoising pre-processing proposed by the present invention is as follows:
step 1, obtaining a target echo signal s (t, u), wherein t is a time domain direction, and u is a space domain direction;
step 2, adding Gaussian white noise n (t, u) into the target echo signal;
step 3, obtaining SAR echo signals s containing noisen(t, u) with the model sn(t，u)＝s(t，u)+n(t，u)；
Step 4, according to the frequency spectrum range omega E [ -omega ] of radar signal0+ωc，ω0+ωc](its baseband band is ω e [ - ω [) in0，ω0]，ωcCarrier frequency of radar signal) to SAR echo signal snSampling in the time domain direction of (t, u), and setting the sampling frequency omega according to the Nyquist theorems≥2ω0To obtain N1×N2Two-dimensional array echo signal sn(n1，n2) (wherein N is1And N2Are respectively sn(n1，n2) In the direction and space domain n2Length of (1);
step 5, according to carrier frequency omega of radar signalcFor SAR echo signal sn(n1，n2) Demodulating the time domain direction to obtain an SAR baseband echo signal sb(n1，n2)
Namely, it is
sb(n1，n2)＝sn(n1，n2)·exp(-jωcn1)
Wherein s isb(n1，n2) The signal is a baseband signal after time domain demodulation;
step 6, SAR baseband echo signal sb(n1，n2) Performing two-dimensional hybrid transformation:
(1) along the time domain n1Direction pair sb(n1，n2) By one-dimensional Discrete Fourier Transform (DFT), i.e.
Wherein <math><mrow><msub><mi>W</mi><msub><mi>N</mi><mn>1</mn></msub></msub><mo>=</mo><mi>exp</mi><mrow><mo>(</mo><mo>-</mo><mi>j</mi><mfrac><mrow><mn>2</mn><mi>π</mi></mrow><msub><mi>N</mi><mn>1</mn></msub></mfrac><mo>)</mo></mrow><mo>,</mo></mrow></math> Sb(k1，n2) Time domain n being an echo1The spectrum in the direction. Based on the baseband frequency spectrum range omega E-omega of radar signal0，ω0]It can be known that Sb(k1，n2) The target signal contained in the signal has a baseband frequency band range of <math><mrow><msub><mi>k</mi><mn>1</mn></msub><mo>∈</mo><mo>[</mo><mo>-</mo><msub><mi>N</mi><mn>1</mn></msub><mfrac><msub><mi>ω</mi><mn>0</mn></msub><msub><mi>ω</mi><mi>s</mi></msub></mfrac><mo>,</mo><msub><mi>N</mi><mn>1</mn></msub><mfrac><msub><mi>ω</mi><mn>0</mn></msub><msub><mi>ω</mi><mi>s</mi></msub></mfrac><mo>]</mo><mo>;</mo></mrow></math>
(2) Along the spatial domain n2Direction versus time domain spectrum Sb(k1，n2) One-dimensional Discrete Wavelet Transform (DWT) with one-layer decomposition is performed as follows
WhereinAnd
respectively, an orthogonal scale equation and a wavelet equation, and m is the number of decomposed layers. The result after two-dimensional hybrid transformation [ X ] can be obtained by the formulaL，XH]Wherein X isL(k1，k21) and XH(k1，k21) are each Sb(k1，n2) Along the spatial domain n2Low and high frequency parts after directional wavelet decomposition.
Step 7, the result [ X ] of SAR echo signal after two-dimensional hybrid transformationL，XH]Performing coefficient operation: will [ X ]L，XH]Is set to 0, and the high frequency component X is setHAll coefficients of (a) are set to 0. The specific operation can be represented by the following formula:
the result after the coefficient operation is denoted as [ X ]L，XH]d。
Step 8, operating the result [ X ] of the above coefficientL，XH]dPerforming inverse two-dimensional hybrid transformation:
(1) to [ X ]L，XH]dEdge k2One-dimensional Inverse Discrete Wavelet Transform (IDWT) is performed in the direction due to the elimination of the high frequency part XH(k1，k21), in the form:
(2) to Sd(k1，n2) Edge k1One-dimensional Inverse Discrete Fourier Transform (IDFT) is carried out on the direction to reconstruct the echo signal s after de-noisingd(n1，n2) The form is as follows:
step 9, de-noising the echo signal sd(n1，n2) As an input signal of SAR imaging processing, s is processed by using SAR imaging processing algorithm (such as RD algorithm, CS algorithm, wave number domain imaging algorithm, etc. in the prior art)d(n1，n2) Carrying out imaging processing;
and step 10, obtaining a denoised SAR reconstruction image.
In a specific embodiment according to the present invention, an existing SAR imaging simulation program is used (see in particular: Soumekh M. synthetic Aperture radio Signal Processing with MATLAB Algorithms)[M]New York, USA, John Wiley, 1999), generates SAR target echo signals s (n) with 5 aircraft targets1，n2) In the experiment, the frequency band of the radar signal is set to be omega E [500HZ, 1500HZ]Sampling frequency ω in the time domainsSet to 2000HZ, generated target echo signal s (n)1，n2) Is N1×N2A two-dimensional array signal of 850 × 564. The SAR wave number domain imaging algorithm is used for imaging processing, and a reconstructed SAR image is shown in figure 2.
Without adding noise, SAR target echo signal s (n)1，n2) Demodulation to a baseband signal sb(n1，n2) And for this two-dimensional signal sb(n1，n2) Performing two-dimensional hybrid transformation, and FIG. 3 shows the SAR target echo signal after the two-dimensional hybrid transformation [ X ] of the noise-free SAR target echo signalL，XH]Amplitude 3D plot, where the vertical axis represents [ X ]L，XH]Amplitude value of (d), the horizontal axis being k1Direction and longitudinal axis k2And (4) direction. Wherein, the edge k1The time domain spectrum passband of the direction is k1∈[213，638](ii) a Edge k2In the direction, the first half (k)2E (1, 282)) is a low-frequency part X after one layer of wavelet decompositionLThe second half (k)2E (283, 564)) is the high frequency part XH。
As shown in FIG. 3, the target echo signal is almost totally separated into a low frequency part X after being subjected to two-dimensional hybrid transformationLWithin the passband. Therefore, according to the method, the method for removing the out-of-band and high-frequency parts of the SAR target echo signal after two-dimensional hybrid transformation is reliable, noise in two directions of the two-dimensional signal can be effectively removed, and the SAR target echo signal is not damaged.
In the simulation experiment, the situation that the SAR imaging system is subjected to strong interference is simulated, and the target echo signal s (n)1，n2) Adding white Gaussian noise n (n)1，n2) Noise-to-Signal-Ratio (NS) of the echo Signal containing NoiseThe specific calculation of R) is given by:
wherein, <math><mrow><mi>P</mi><mrow><mo>(</mo><mi>n</mi><mo>)</mo></mrow><mo>=</mo><munderover><mi>Σ</mi><mrow><msub><mi>n</mi><mn>1</mn></msub><mo>=</mo><mn>0</mn></mrow><mrow><msub><mi>N</mi><mn>1</mn></msub><mo>-</mo><mn>1</mn></mrow></munderover><munderover><mi>Σ</mi><mrow><msub><mi>n</mi><mn>2</mn></msub><mo>=</mo><mn>0</mn></mrow><mrow><msub><mi>N</mi><mn>2</mn></msub><mo>-</mo><mn>1</mn></mrow></munderover><mi>n</mi><mrow><mo>(</mo><msub><mi>n</mi><mn>1</mn></msub><mo>,</mo><msub><mi>n</mi><mn>2</mn></msub><mo>)</mo></mrow></mrow></math> the power of all noise signals; <math><mrow><mi>P</mi><mrow><mo>(</mo><mi>s</mi><mo>)</mo></mrow><mo>=</mo><munderover><mi>Σ</mi><mrow><msub><mi>n</mi><mn>1</mn></msub><mo>=</mo><mn>0</mn></mrow><mrow><msub><mi>N</mi><mn>1</mn></msub><mo>-</mo><mn>1</mn></mrow></munderover><munderover><mi>Σ</mi><mrow><msub><mi>n</mi><mn>2</mn></msub><mo>=</mo><mn>0</mn></mrow><mrow><msub><mi>N</mi><mn>2</mn></msub><mo>-</mo><mn>1</mn></mrow></munderover><msub><mi>s</mi><mi>b</mi></msub><mrow><mo>(</mo><msub><mi>n</mi><mn>1</mn></msub><mo>,</mo><msub><mi>n</mi><mn>2</mn></msub><mo>)</mo></mrow></mrow></math> is N1×N2SAR baseband echo signal sb(n1，n2) The power of all target echo signals contained in (a).
The process of operating on the coefficients of the two-dimensional hybrid transform of the SAR echo signal can be referred to fig. 4 and 5 (the meaning and representation of the coordinates of fig. 4 and 5 are the same as fig. 3): FIG. 4 shows a SAR echo signal s containing noisen(n1，n2) Two-dimensional hybrid transformation result of [ X ]L，XH]The noise-to-signal ratio of the white gaussian noise mixed therein is NSR 500. FIG. 5 is [ X ]L，XH]Result after coefficient manipulation [ XL，XH]dOnly retained [ X ] after coefficient manipulationL，XH]Low frequency component X ofLThe coefficient part within the mid-passband.
In order to test the effect of the denoising pretreatment provided by the invention, the imaging effects of 4 groups of SAR images are compared, wherein each group of images are respectively SAR reconstruction images which do not perform denoising pretreatment on echo signals under a certain noise-signal ratio. As shown in fig. 6, the 4 groups of images are SAR reconstructed images with noise-to-signal ratios of NSR 200, 500, 800, and 1000 in sequence, where the left image of each group is an SAR reconstructed image without denoising, and the right image is a corresponding SAR reconstructed image after denoising preprocessing.
As can be seen from the comparison of fig. 6, after the denoising pre-processing provided by the present invention, the image quality after SAR reconstruction is significantly improved, and the target in the image becomes clearer. Especially when the noise-to-signal ratio NSR exceeds 800, the objects in the noisy image almost disappear, but can be re-visualized in the reconstructed image after the denoising pre-processing. Therefore, the denoising preprocessing method provided by the invention has good effects in the aspects of restraining strong interference, improving the signal-to-noise ratio of the SAR image and the like.
Although specific embodiments of the present invention have been described above, it will be understood by those skilled in the art that these specific embodiments are merely illustrative and that various omissions, substitutions and changes in the form of the detail of the methods and systems described above may be made by those skilled in the art without departing from the spirit and scope of the invention. For example, it is within the scope of the present invention to combine the steps of the above-described methods to perform substantially the same function in substantially the same way to achieve substantially the same result. Accordingly, the scope of the invention is to be limited only by the following claims.
Claims (3)
1. A SAR echo signal denoising preprocessing method based on two-dimensional hybrid transformation is characterized by comprising the following steps:
(1) acquiring an SAR target echo signal;
(2) carrying out time domain sampling on the SAR target echo signal;
(3) time domain demodulation is carried out on the result obtained by time domain sampling to obtain SAR two-dimensional baseband echo signal sb(n1，n2)；
(4) For the SAR two-dimensional baseband array echo signal sb(n1，n2) Performing two-dimensional hybrid transformation to obtain two-dimensional hybrid transformation [ X ] of SAR target echo signalL，XH]；
(5) Two-dimensional hybrid transformation [ X ] of SAR target echo signalsL，XH]Performing coefficient operation to remove out-of-band and high-frequency coefficient parts to obtain the result [ X ] after the coefficient operationL，XH]d；
(6) For the result [ X ] after coefficient operationL，XH]dInverse two-dimensional hybrid transformation is carried out to obtain a denoised echo signal sd(n1，n2)；
(7) De-noising the echo signal sd(n1，n2) And as an input signal of the SAR imaging processing system, generating an SAR image by utilizing an SAR reconstruction algorithm.
2. The method of claim 1, wherein the step (4) comprises the steps of:
(4-1) along SAR two-dimensional baseband echo signal sb(n1，n2) The time domain direction of the SAR target echo signal is subjected to one-dimensional discrete Fourier transform to obtain a time domain frequency spectrum S of the SAR target echo signalb(k1，n2)，
Wherein
Sb(k1，n2) Time domain n being an echo1Frequency spectrum in the direction, Sb(k1，n2) The target signal contained in the signal has a baseband frequency band range of
Wherein: n is a radical of1Is s isn(n1，n2) In the direction n1Length of (c), ωsIn order to be able to sample the frequency,
ω0boundary frequencies of the baseband frequency band;
(4-2) aligning the time domain spectrum S in the spatial domain directionb(k1，n2) Performing one-dimensional wavelet transform, whose wavelet decomposition process is as follows
WhereinAnd
respectively orthogonal scale equation and wavelet equation, m is the number of decomposed layers, and the result [ X ] after two-dimensional mixed transformation is obtained by the formulaL，XH]Wherein X isL(k1，k21) and XH(k1，k21) are each Sb(k1，n2) Along the spatial domain n2Low and high frequency parts after directional wavelet decomposition.
3. The method of claim 1, wherein the step (6) comprises the steps of:
(6-1) pairs of [ XL，XH]dEdge k2One-dimensional inverse discrete wavelet transform is carried out in the direction, and the high-frequency part X is removedH(k1，k21), in the form:
(6-2) to Sd(k1，n2) Edge k1One-dimensional inverse discrete Fourier transform is carried out on the direction to reconstruct a denoised echo signal sd(n1，n2) The form is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100833457A CN101639530B (en) | 2009-05-04 | 2009-05-04 | SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100833457A CN101639530B (en) | 2009-05-04 | 2009-05-04 | SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101639530A CN101639530A (en) | 2010-02-03 |
| CN101639530B true CN101639530B (en) | 2011-09-07 |
Family
ID=41614612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100833457A Expired - Fee Related CN101639530B (en) | 2009-05-04 | 2009-05-04 | SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101639530B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105513020B (en) * | 2015-11-30 | 2018-06-19 | 东南大学 | A kind of method for removing ISAR image stripe interferences |
| CN112578383B (en) * | 2020-11-19 | 2023-06-30 | 西安电子科技大学 | High maneuvering platform TOPS SAR imaging method based on extended wave number spectrum reconstruction |
| CN112649808B (en) * | 2020-12-30 | 2023-04-07 | 西安电子科技大学 | Bistatic forward-looking SAR wave number domain imaging method based on shift configuration |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6133989A (en) * | 1993-02-09 | 2000-10-17 | Advanced Scientific Concepts, Inc. | 3D imaging laser radar |
| CN1781032A (en) * | 2003-02-26 | 2006-05-31 | 电信网络信息有限公司 | Improved sodar sounding of the lower atmosphere |
-
2009
- 2009-05-04 CN CN2009100833457A patent/CN101639530B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6133989A (en) * | 1993-02-09 | 2000-10-17 | Advanced Scientific Concepts, Inc. | 3D imaging laser radar |
| CN1781032A (en) * | 2003-02-26 | 2006-05-31 | 电信网络信息有限公司 | Improved sodar sounding of the lower atmosphere |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101639530A (en) | 2010-02-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111142105B (en) | ISAR imaging method for complex moving target | |
| CN112987003B (en) | HFM signal separation method and system in active sonar | |
| CN103454622B (en) | Broadband radar target complex echo de-noising method based on sparseness constraints | |
| CN106199532A (en) | Gpr Signal noise-reduction method based on mixing Fourier wavelet analysis | |
| CN103091665A (en) | Compressed sensing synthetic aperture radar radio frequency interference suppression handling method | |
| CN110632573A (en) | Airborne broadband radar space-time two-dimensional keystone transformation method | |
| CN101639530B (en) | SAR echo signal de-noising preprocessing method based on two-dimensional mixed transformation | |
| Chen et al. | Suppressive interference suppression for airborne SAR using BSS for singular value and eigenvalue decomposition based on information entropy | |
| CN109061626B (en) | Method for detecting low signal-to-noise ratio moving target by step frequency coherent processing | |
| CN101930605B (en) | Synthetic Aperture Radar (SAR) image target extraction method and system based on two-dimensional mixing transform | |
| CN108845318B (en) | Satellite-borne high-resolution wide-range imaging method based on Relax algorithm | |
| CN106199539A (en) | Ground bounce removal method based on prewhitening filter | |
| CN109471084B (en) | Electromagnetic scattering mechanism decomposition method | |
| Kelly et al. | RFI suppression and sparse image formation for UWB SAR | |
| CN109061590B (en) | Radar reference signal extraction method based on blind self-adaptive oblique projection | |
| CN103885044B (en) | A kind of miscellaneous suppressing method of making an uproar of Narrow-band Radar echo based on CLEAN algorithm | |
| CN115220008A (en) | Broadband radar target energy accumulation method based on sub-band processing | |
| Li et al. | Deep learning for interference mitigation in time-frequency maps of fmcw radars | |
| CN111398912B (en) | Synthetic aperture radar interference suppression method based on tensor low-rank approximation | |
| Wang et al. | LFM signal perception based on Wavelet transform and Time-Frequency Technology | |
| CN113589253A (en) | Method for detecting weak echo signal based on wavelet transform algorithm of pseudo time domain | |
| CN112462344A (en) | Method for extracting radar signal intra-pulse characteristic parameters through time-frequency domain transformation processing | |
| CN104076361A (en) | Super-resolution wide-area imaging method for airborne battlefield monitoring radar of unmanned aerial vehicle | |
| Vehmas et al. | Mitigation of RFI in High-Resolution SAR Data–Algorithm Overview and Experimental Demonstration | |
| CN112269179B (en) | Airspace high-resolution detection method for low-noise target |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110907 Termination date: 20160504 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |