SAR noise frequency modulation saturation interference suppression method based on time-frequency characteristics

Abstract

The application discloses a SAR noise FM saturation interference suppression method based on time-frequency characteristics and belongs to the technical field of radar signal processing. The method comprises the following steps: a SAR echo model under NFM saturation interference is established, short-time Fourier transform is performed on the disturbed SAR echo, fundamental wave and high-order harmonic detection is performed, notch filtering is performed in each STFT window, the echo time domain is transformed back, and finally, SAR echo data after interference suppression is obtained. The application can effectively suppress NFM saturation interference, the effectiveness of the method is verified through semi-actual measurement data of Radarsat-1, and compared with a time domain cancellation method, the method has better saturation interference suppression capability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar signal processing, and particularly relates to a SAR noise frequency modulation saturation interference suppression method based on time-frequency characteristics. BACKGROUND

[0002] Synthetic aperture radar (SAR) can use the relative motion between radar and target to obtain high-resolution surface images or data. The all-weather, all-day imaging capability and high-resolution image quality of SAR make it a powerful remote sensing tool. SAR plays an important role in the fields of earth science, resource exploration, military application, environmental monitoring, etc. However, SAR is vulnerable to radio frequency interference (RFI) in the same frequency band. There are many RFI sources that can threaten SAR, such as communication equipment, radars and jammers working in the same band. High-power RFI can drown out target echoes, thus seriously degrading the quality of SAR images. Usually, RFI appears as stripes on the SAR image, which obscures the target of interest. Therefore, it is of great significance to study effective methods to suppress RFI in SAR systems.

[0003] In order to suppress RFI, researchers have conducted research from different angles, including parametric methods, non-parametric methods, semi-parametric methods and machine learning methods. Parametric methods involve establishing a parametric model of interference signals and estimating model parameters to estimate and extract interference. Non-parametric methods use the intensity feature difference between interference and target echo signals to filter out interference. Semi-parametric methods convert the complex signal separation problem into a hyperparameter optimization problem. The above interference suppression methods need to be analyzed specifically for specific environments, which leaves room for performance improvement. Machine learning methods can automatically extract hierarchical features of targets in images and have been successfully applied to the field of computer vision.

[0004] The existing methods can achieve good suppression effect in dealing with various interferences. However, these methods often ignore the saturation interference. The saturation phenomenon caused by the interference is not uncommon in synthetic aperture radar, which is usually caused by the radiation signals of the same frequency radar, jammer and other devices of the ground base station. The power of these radiation sources is very strong, and the interference power of the direct SAR receiving end only suffers from distance square attenuation, which leads to the saturation of the SAR receiving channel and triggers the nonlinear clipping distortion of the receiving channel. This nonlinear distortion causes the distortion of the interference signal in the time domain and the frequency domain. In the time domain, the signal amplitude is truncated, causing amplitude and phase distortion. In the frequency domain, due to the nonlinear characteristics of the saturation distortion, the interference spectrum produces false components. If we ignore the special characteristics of the saturation interference and still use the existing interference analysis and suppression method, it will lead to the mismatch between the signal model and the suppression method, thus worsening the interference suppression effect. Compared with other saturation interference types, the noise frequency modulation saturation interference not only has strong interference radiation power, but also has wide spectrum and small amplitude dynamic range, so there is still a lot of research space for the suppression method of the noise frequency modulation saturation interference. SUMMARY

[0005] The present application provides a SAR noise frequency modulation saturation interference suppression method based on time-frequency characteristics, which is used for suppressing the saturation interference and its high-order harmonics in a low intermediate frequency SAR receiver, and effectively improves the imaging quality of the SAR.

[0006] The technical solution for achieving the object of the present application is as follows: a SAR noise frequency modulation saturation interference suppression method based on time-frequency characteristics, comprising the following steps:

[0007] Step 1: based on the response of the low intermediate frequency SAR receiver, a SAR echo model under the condition of NFM saturation interference is established;

[0008] Step 2: the time domain and frequency domain characteristics of the single frequency and wideband saturation interference are analyzed, the time-frequency domain characteristics of the noise frequency modulation saturation interference are obtained, and the processing method for NFM interference suppression is determined;

[0009] Step 3: saturation interference fundamental wave and high-order harmonic detection, determination of the notch threshold of each short-time Fourier window, and elimination of the fundamental wave and high-order harmonic of the saturation interference exceeding the threshold in each short-time Fourier window;

[0010] Step 4: inverse STFT is performed on the data, the time-frequency domain signal after interference suppression is restored to the time domain, and the echo data after interference suppression is obtained.

[0011] A computer device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the program.

[0012] A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method.

[0013] A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method.

[0014] Compared with the prior art, the present application has the beneficial effects that: the present application utilizes the harmonic characteristics of the saturated interference, accurately detects the fundamental harmonic and high-order harmonic of the NFM saturated interference in the SAR echo, and realizes effective suppression of the NFM saturated interference; the interference suppression is performed in the manner of notch filtering, does not involve the matrix inversion operation of the traditional algorithm, and has higher efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 It is a flow chart of the SAR noise frequency modulation saturated interference suppression based on the time-frequency characteristics under the condition of the low intermediate frequency SAR receiver.

[0016] Figure 2 It is a structure block diagram of the low intermediate frequency SAR receiver.

[0017] Figure 3 It is a schematic diagram of the single frequency saturated interference.

[0018] Figure 4 It is the frequency spectrum before and after the single frequency signal saturation, (a) is the non-saturated frequency spectrum of the single frequency signal, and (b) is the saturated frequency spectrum of the single frequency signal.

[0019] Figure 5 It is the frequency spectrum before and after the LFM signal saturation under the low intermediate frequency receiver, (a) is the non-saturated frequency spectrum of the LFM signal, and (b) is the saturated frequency spectrum of the LFM signal.

[0020] Figure 6 It is the time-frequency domain before and after the LFM signal saturation under the low intermediate frequency receiver, (a) is the non-saturated time-frequency domain diagram of the LFM signal, and (b) is the saturated time-frequency domain diagram of the LFM signal.

[0021] Figure 7 It is the time-frequency domain diagram before and after the NFM signal saturation under the low intermediate frequency receiver, (a) is the non-saturated time-frequency domain diagram of the NFM signal, and (b) is the saturated time-frequency domain diagram of the NFM signal.

[0022] Figure 8 It is the imaging effect diagram of the frequency domain notch method and the method of the present application, (a) is the imaging effect diagram of the method of the present application, and (b) is the imaging effect diagram of the frequency domain notch method.

[0023] Figure 9 It is the local ship target enlarged diagram of the imaging effect diagram of the frequency domain notch method and the method of the present application, (a) is the local enlarged diagram of the method of the present application, and (b) is the local enlarged diagram of the frequency domain notch method. DETAILED DESCRIPTION

[0024] The application belongs to the technical field of radar signal processing, and discloses a SAR noise frequency modulation (NFM) interference suppression method based on time-frequency characteristics of saturated interference signals.

[0025] The application will be further described in detail in terms of specific embodiments.

[0026] Reference Figure 1 The SAR noise frequency modulation saturated interference suppression method based on time-frequency characteristics provided by the application comprises the following steps:

[0027] Step 1: establishing a SAR echo model under the condition of NFM saturated interference based on the response of a low intermediate frequency SAR receiver;

[0028] The structural block diagram of the low intermediate frequency SAR receiver is shown in FIG. 1. Figure 2 The low intermediate frequency receiver converts echo signals to a lower intermediate frequency, and then converts them into digital signals through an analog-digital converter (ADC). The main functions of automatic gain control (AGC) and manual gain control (MGC) are to adjust the gain of the echo and amplify its power to an appropriate level to adapt to the intensity variation of the scene scattering. However, when facing an interference much stronger than the echo signal, this may exceed the dynamic range of AGC or MGC, causing the signal power input to the ADC to exceed its sampling threshold, thereby causing saturation.

[0029] Under the condition of using the low intermediate frequency SAR receiver, the NFM interference signal received by the receiver can be expressed as

[0030]

[0031] where the amplitude A0 and the intermediate frequency f0 are constants, Δf represents the frequency offset amplitude, and m(τ) is a modulated signal controlled by random noise. The randomness of m(τ) leads to random frequency offset. Replace the phase of the NFM interference signal with to obtain the NFM interference signal when the receiver is saturated

[0032]

[0033] Where A j (t) represents the nonlinear change in the amplitude of the NFM interference signal when the receiver becomes saturated. Assume the signal input to the ADC is S. in Then the output signal S out It can be represented as

[0034] S out =A[sat(S in )]

[0035]

[0036] Where A[·] is the quantization model function, and sat(·) is the saturation function / saturation conversion model of the ADC. The dynamic range of the ADC is [-S a ,S a Within this input range, the ADC operates in a linear state. When the input signal amplitude exceeds this dynamic range, the ADC will operate in a non-linear state, and the system will saturate.

[0037] When saturation is not present, the interference signal, after quadrature demodulation, yields the following two baseband signals.

[0038]

[0039] Therefore, the process of the interference signal passing through the ADC, saturating, and then undergoing quadrature demodulation is equivalent to s I (t) and s Q (t) is input into the saturation transition model sat(·), and we can obtain the saturated signal.

[0040]

[0041] Where C is the saturation coefficient, and C < 1. When the amplitude of the received signal exceeds CA... j When (t), it indicates signal saturation.

[0042] Step 2: Analyze the time-domain and frequency-domain characteristics of single-frequency and broadband saturated interference, simulate the time-frequency domain characteristics of noise FM saturated interference, and determine the processing method for NFM interference suppression.

[0043] First, we analyze single-frequency saturation interference. For example... Figure 3 As shown, single-frequency saturation interference It is still a periodic signal, specifically represented as

[0044]

[0045] Where T represents the period, and n represents the count representing the period. Initial phase τ is a constant. cThe time instant at which saturation first occurs on the time axis. Single frequency jamming signal It consists of two parts, one is the periodic modulation of local cosine function signal s1, which can be divided into two single frequency signals and square wave signals s3, the second is a rectangular wave signal s2 with regular periodicity, as shown in the green part of s1, the red part of s2, expressed by formula Figure 3

[0046]

[0047]

[0048] s2 and s3 are expanded into Fourier series, which is

[0049]

[0050] It can be seen that s2 only contains odd harmonic components, and s3 only contains even harmonic components. According to the properties of trigonometric functions, after multiplying s3 with Only contains odd harmonic components. Therefore, it can be concluded that single frequency saturation jamming is composed of multiple high order harmonics, and the frequency of high order harmonic is the odd multiple of the original signal frequency.

[0051] Next, we analyze the wideband saturation jamming in the general form. Wideband saturation jamming cannot be directly compared with periodic single frequency saturation jamming, but it can be approximated by a smooth non-affine function g(·) to approximate the saturation function sat(·)

[0052]

[0053] The Taylor expansion of the function tanh(x) is obtained as

[0054]

[0055] where B 2n is the Bernoulli number. It can be seen that the expansion of tanh(x) is an odd item, and does not contain even items and direct current components. Therefore, in the case of using low intermediate frequency SAR receiver, the unsaturated part of the signal in the received echo and its quadrature demodulation are represented as

[0056]

[0057] By reducing the power of high order terms in the signal and ignoring constant terms, we can get

[0058]

[0059] ​​The results show that when the echo signal saturates, higher-order harmonics are generated, and the frequencies of these harmonics are all odd multiples of the original signal frequency. The intermediate frequencies of the higher-order harmonics also become corresponding odd multiples, and positive and negative frequency values ​​alternate. As the frequency multiple increases, the amplitude of the harmonics gradually decreases.

[0060] Figure 4 The graph shows the spectrum of a single-frequency signal before and after saturation. As can be seen from the graph, the single-frequency signal only contains odd harmonic frequencies. Figure 5 The frequency spectrum of the LFM signal before and after saturation under a low-IF receiver is shown below. Figure 5 The results show the bandwidth of the higher-order harmonics of the broadband saturated interference signal and the correspondence between the intermediate frequency and the fundamental harmonic. The saturated LFM signal becomes a combination of multiple LFM signals. The frequency center and bandwidth of these LFM signals are odd multiples of the original LFM signal, and their amplitudes gradually decrease, consistent with the theoretical analysis. Figure 6 The figure shows the time-frequency domain of the LFM signal before and after saturation in a low-IF receiver. It can be seen that the saturated LFM signal generates multiple LFM harmonics, with the modulation frequencies of these harmonics being odd multiples of the fundamental frequency, alternating between positive and negative values. For example, the modulation frequencies of the 3rd and 5th harmonics are -3 and 5 times the fundamental frequency, respectively. Figure 7 For the time and frequency of the NFM signal before and after saturation under low-IF receiver, from Figure 7 As can be seen from (a), the interference radiated power is strong, the spectrum is wide, and the frequency variation is relatively random. Figure 7 As shown in (b), saturated NFM interference contains significant higher-order harmonics, particularly the 3rd and 5th harmonics. However, the amplitudes of these higher-order harmonics are weaker than the fundamental frequency, and some instantaneous spectra contain numerous spurious signals, making it impossible to determine the harmonic location based on amplitude alone; they can even be ignored. Therefore, the specific method for suppressing NFM interference can be determined as follows: detect the fundamental frequency and higher-order harmonics of the interference separately, determine the notch threshold for each short-time Fourier window, and then perform notch filtering.

[0061] Step 3: Detect the fundamental and higher-order harmonics of saturated NFM interference, determine the notch threshold for each short-time Fourier window, and eliminate the fundamental and higher-order harmonics of saturated interference that exceed the threshold within each short-time Fourier window.

[0062] Step 3.1: Perform fundamental harmonic detection for saturated NFM interference. Perform STFT on the i-th slow-time echo x[n] with interference, and the result is expressed as...

[0063]

[0064] Where m represents the m-th time in the STFT result, and k represents X. i The k-th frequency in N, w[·] represents the window function, and N r N represents the number of points in each STFT time window.fft The number of FFT points representing the time window. After performing STFT, the instantaneous frequency spectrum X i [m, : ] is obtained. Then the interference in the instantaneous frequency spectrum is detected using the Maximum-to-Mean Ratio (MMR) adaptive threshold criterion, which is expressed as

[0065]

[0066] where R represents the root mean square error of MMSE estimation, T wide is the interference detection threshold adaptively obtained, which is set as the sum of the mean and standard deviation of the instantaneous frequency spectrum. When R < R th (R th represents the tolerance of the interference strength, which is set by experience), it means that there is no interference or the interference power is relatively low, and interference suppression is not needed. When R ≥ R th , it means that there is relatively high power interference, and interference suppression is needed. When R ≥ R th , the instantaneous frequency spectrum is compared with the threshold determined by the MMR criterion, and if the frequencies of the instantaneous frequency spectrum exceed the threshold, it is considered that there is interference on these frequencies. The frequencies satisfying these conditions are recorded, thereby obtaining the position set k i of the interference fundamental harmonic, denoted as

[0067]

[0068] Step 3.2, high-order harmonic detection of saturated NFM interference. It is known that in the instantaneous frequency spectrum of saturated interference, the center frequency and bandwidth of the high-order harmonic are still odd multiples of the fundamental harmonic. Therefore, the center frequency and bandwidth of the interference fundamental harmonic X i [m, k i ] can be directly obtained from , thereby deducing the position of the interference high-order harmonic in the instantaneous frequency spectrum and obtaining the position set k of the qth interference harmonic. Then the method of estimating the interference-to-signal ratio (ISR) of the interference harmonic is used to determine whether to apply the notch filter. After obtaining the position of the interference high-order harmonic in the instantaneous frequency spectrum, the SIR of the interference harmonic can be expressed as

[0069]

[0070] where represents the power of the qth interference harmonic, and P s represents the power of the useful signal. ISR (q)> 0 dB represents the power of the qth interference harmonic exceeds the power of the signal. In this case, the interference harmonic needs to be notch filtered in the time-frequency domain. At the same time, the position set of the interference high-order harmonic which seriously affects the useful signal can be obtained, and the set is combined with the position set of the interference fundamental wave, so that the comprehensive position set Ψ which needs to be notch filtered in the instantaneous spectrum is obtained m .

[0071] Step 3.3, saturated NFM interference suppression, eliminating the fundamental wave and high-order harmonic of saturated interference exceeding the threshold in each short-time Fourier window.

[0072] According to step 3.2, the position set Ψ corresponding to the interference fundamental wave and high-order harmonic with larger power can be obtained m . Through the position set Ψ m , a binary mask corresponding to the instantaneous spectrum of the time sample m can be constructed, denoted as

[0073]

[0074] By combining the binary mask corresponding to each time sample, a binary mask matrix Q[m, k] with the same size as X i [m, k] can be obtained. Then, by multiplying Q[m, k] and X i [m, k], the time-frequency spectrum Y which does not contain the interference fundamental wave and high-power harmonic can be obtained, denoted as

[0075]

[0076] Step 4, inverse STFT is performed on the data, and the time-frequency domain signal Y after interference suppression is restored to the time domain to obtain the echo data after interference suppression. This process is denoted as

[0077]

[0078] The above process completes the suppression of saturated NFM interference, and removes the interference fundamental wave and high-order harmonic caused by saturation.

[0079] For the specific implementation method of the present application, the frequency domain notch method and the SAR NFM saturated interference suppression method based on time-frequency characteristics of the present application are used for interference suppression, and the SAR echo data after suppression is imaged, and the imaging result is as shown in Figure 8 .

[0080] The simulation experiment adopts semi-actual SAR data, and the interference suppression result is evaluated through the imaging result. The simulation experiment uses the measured data of Radarsat-1 and manually adds saturated NFM interference. The specific parameters of the Radarsat-1 platform are as follows: carrier frequency: 5.3 GHz, transmitted signal bandwidth: 30.116 MHz, pulse width: 41.75 us, flight platform speed: 7062 m / s, range direction sampling rate: 32.317 MHz, azimuth direction sampling rate: 1256.98 MHz, and azimuth direction frequency modulation rate: 1733 Hz / s. The specific parameters of the NFM interference signal are as follows: frequency offset amplitude 0.63946 MHz, time width 41.75 us, and signal-to-interference ratio: -27.6824 dB.

[0081] Figure 8 In the displayed imaging result, the frequency domain wave-trapping method and the present application can both achieve certain interference suppression effect, Figure 9 A local enlarged view of the above shows that the ship display effect of the present application is clearer, and the interference suppression effect is better. It can be seen that the present application can provide much higher interference suppression effect than the conventional frequency domain wave-trapping method under the same ISR condition.

[0082] The above only describes the preferred embodiments of the present application, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A method for suppressing SAR noise frequency-modulated saturation interference based on time-frequency characteristics, characterized in that, Includes the following steps: Step 1: Based on the low-IF SAR receiver response, establish a SAR echo model under noise frequency modulation (NFM) saturation interference. Step 2: Analyze the time and frequency domain characteristics of single-frequency and broadband saturated interference to obtain the time and frequency domain characteristics of noise FM saturated interference and determine the treatment method for NFM interference suppression. Step 3: Detect saturated interference fundamental and higher harmonics, determine the notch threshold for each short-time Fourier window, and eliminate the fundamental and higher harmonics of saturated interference that exceed the threshold within each short-time Fourier window. Step 4: Perform inverse STFT on the data to restore the time-frequency domain signal after interference suppression to the time domain, and obtain the echo data after interference suppression.

2. The SAR noise frequency-modulated saturation interference suppression method based on time-frequency characteristics according to claim 1, characterized in that, Step 1 is as follows: Using a low-to-intermediate-frequency SAR receiver, the received NFM interference signal is represented as: Where amplitude A0 and intermediate frequency f0 are constants, Δf represents the frequency offset amplitude, and m(τ) is a modulated signal controlled by random noise. The randomness of m(τ) leads to random frequency offset; By replacing the phase of the NFM interference signal, the NFM interference signal when the receiver saturates is obtained. Where A j (t) represents the nonlinear change in the amplitude of the NFM interference signal when the receiver becomes saturated; assuming the signal input to the ADC is S in Then the output signal S out Represented as S out =A[sat(S in )] Where A[·] is the quantization model function, and sat(·) is the saturation function / saturation conversion model of the ADC; the dynamic range of the ADC is [-S a ,S a Within this input range, the ADC operates in a linear state; when the input signal amplitude exceeds this dynamic range, the ADC will operate in a non-linear state, and the system will saturate. When no saturation occurs, the interference signal, after quadrature demodulation, yields the following two baseband signals. Therefore, the process of the interference signal passing through the ADC, saturating, and then undergoing quadrature demodulation is equivalent to s I (t) and s Q (t) is input into the saturation transformation model sat(·) to obtain the saturated signal. and Where C is the saturation coefficient, and C < 1; when the amplitude of the received signal exceeds CA j When (t), it indicates signal saturation.

3. The SAR noise frequency-modulated saturation interference suppression method based on time-frequency characteristics according to claim 2, characterized in that, Step 2: Analyze the time and frequency domain characteristics of single-frequency and broadband saturated interference to obtain the time and frequency domain characteristics of noise FM saturated interference, and determine the processing method for NFM interference suppression, specifically: First, we analyze single-frequency saturation interference; single-frequency saturation interference... It is a periodic signal, specifically represented as Where T represents the period, n represents the count representing the period, and the initial phase. τ is a constant. c The moment when saturation first occurs on the time axis; single-frequency interference signal. It consists of two parts: first, a periodically modulated local sine / cosine function signal s1, which can be decomposed twice into the product of a single-frequency signal and a square wave signal s3; second, a rectangular wave signal s2 with regular periodicity, expressed by the formula as follows: Expanding s2 and s3 into Fourier series, we get s2 contains only odd harmonic components, and s3 contains only even harmonic components; according to the properties of trigonometric functions, s3 and... After multiplication, It contains only odd harmonic components; therefore, it can be concluded that single-frequency saturation interference consists of multiple higher-order harmonics, and the frequencies of the higher-order harmonics are odd multiples of the original signal frequency. The saturation function sat(·) is approximated by smoothing the non-affine function g(·): Performing a Taylor expansion on the function tanh(x), we obtain Among them B 2n It is a Bernoulli number, and the exponents of the expansion of tanh(x) are all odd-numbered, and it does not include even-numbered terms or a DC component; therefore, when using a low-IF SAR receiver, the unsaturated portion of the received echo signal is received. and its orthogonal demodulation They are respectively represented as By raising the higher-order terms of the signal to a power of 1, and ignoring the constant term, we obtain: 。 4. The SAR noise frequency-modulated saturation interference suppression method based on time-frequency characteristics according to claim 3, characterized in that, Step 3: Detection of saturated NFM interference fundamental and higher-order harmonics; determine the notch threshold for each short-time Fourier window; and eliminate the fundamental and higher-order harmonics of saturated interference exceeding the threshold within each short-time Fourier window. Specifically: Step 3.1: Perform fundamental harmonic detection for saturated NFM interference; perform STFT on the i-th slow-time echo x[p] with interference, and the result is expressed as... Where m represents the m-th time in the STFT result, and k represents X. i The k-th frequency in N, w[·] represents the window function, and N r N represents the number of points in each STFT time window. fft This represents the number of FFT points in the time window; after performing STFT, the instantaneous frequency spectrum X corresponding to any time sample m is obtained. i [m,:]; then, interference in the instantaneous frequency spectrum is detected, and an adaptive threshold criterion for the maximum average ratio is used, which is expressed as: Where R represents the root mean square error of the MMSE estimate, and T wide The adaptively obtained interference detection threshold is set as the sum of the mean and standard deviation of the instantaneous frequency spectrum; when R < R th No interference suppression is required, where R th This indicates the tolerance to interference intensity; when R ≥ R th Interference suppression is required when R ≥ R th At that time, the instantaneous frequency spectrum is compared with a threshold determined by the MMR criterion. If the frequency of the instantaneous frequency spectrum exceeds the threshold, interference is considered to exist at these frequencies. The frequencies that meet these conditions are recorded, thereby obtaining the location k of the interfering fundamental harmonic. i The set of, denoted as Step 3.2, perform high-order harmonic detection of saturated NFM interference; it is known that in the instantaneous frequency spectrum of saturated interference, the center frequency and bandwidth of the high-order harmonics are still odd multiples of the fundamental frequency; therefore, from The interference fundamental wave X can be directly obtained in the middle. i [m,k i By determining the center frequency and bandwidth, the positions of the higher-order harmonics of the interference in the instantaneous frequency spectrum can be derived, and the set of positions of the q-th harmonic of the interference can be obtained. Then, the interference-to-signal ratio (ISR) of the interfering harmonics is estimated to determine whether notch filtering should be applied; after obtaining the location of the higher-order interfering harmonics in the instantaneous frequency spectrum, the ISR of the interfering harmonics is expressed as: in Let P represent the power of the q-th interference harmonic, and Pq represent the power of the interference harm s Indicates the power of the useful signal; ISR (q) >0dB indicates that the power of the q-th interference harmonic exceeds the power of the signal. In this case, notch filtering is required for this interference harmonic in the time-frequency domain. Simultaneously, the location set of the higher-order interference harmonics is obtained, and this set is combined with the location set of the fundamental interference wave to obtain the comprehensive location set Ψ for notch filtering in the instantaneous spectrum. m ; Step 3.3, Saturated NFM interference suppression, eliminates the fundamental harmonics and higher-order harmonics of saturated interference exceeding the threshold within each short-time Fourier window; According to step 3.2, the set of positions Ψ corresponding to the high-power fundamental and higher-order harmonic interference is obtained. m ; through the location set Ψ m To construct a binary mask for the instantaneous spectrum of the corresponding time sample m, denoted as: By combining the binary masks corresponding to each time sample, a mask corresponding to X is obtained. i A binary mask matrix Q[m,k] of uniform size [m,k] is then created by comparing Q[m,k] with X. i Multiplying [m,k] yields the time spectrum that does not contain interfering fundamental frequency and high-power harmonics. Represented as 5. The SAR noise frequency-modulated saturation interference suppression method based on time-frequency characteristics according to claim 4, characterized in that, Step 4: Perform inverse STFT on the data to obtain the time-frequency domain signal after interference suppression. Restoring to the time domain Obtain the echo data after interference suppression, specifically 6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of any of the methods described in claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method described in any one of claims 1-5.

8. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-5.

Images