Clutter frequency domain suppression method based on correlation spectrum

Abstract

This invention relates to the field of signal processing technology and provides a ground clutter frequency domain suppression method based on correlation spectrum. Its key feature is the use of phase information from the autocorrelation spectral density of weather radar echo signals to identify and suppress ground clutter signals. The method includes: adaptively selecting an optimal data window function based on the clutter-to-noise ratio of the weather radar echo signal and performing power-preserving processing; calculating the autocorrelation spectral density and power spectral density, and effectively distinguishing clutter from broadband weather signals using the local deviation characteristics of the autocorrelation spectral density relative to the ideal linear phase, thereby performing clutter identification and filtering. Based on the autocorrelation spectral density after filtering out ground clutter, an inverse transform is performed to calculate the autocorrelation function, completing the time-domain estimation of the weather radar detection variables. This invention effectively solves the problem that traditional ground clutter suppression methods cannot accurately identify and eliminate ground clutter mixed in with weather signals.

Description

Technical Field

[0001] This invention relates to the field of signal processing technology, and in particular to a method for suppressing ground clutter in the frequency domain based on correlation spectrum. Background Technology

[0002] Ground clutter consists of non-weather echoes reflected from surface targets such as mountains and buildings. These targets have average Doppler velocities of zero or near zero and generally very small spectral widths. Therefore, when using low elevation angle scanning, ground clutter echoes can mix with weather echoes, hindering the identification of weather phenomena and significantly impacting the performance of weather radar in acquiring radar detection variables. For example, under conditions of strong ground clutter contamination, reflectivity estimates will increase, while velocity and spectral width estimates will approach zero. Without ground clutter removal, the accuracy of basic variables and even derived products will be difficult to guarantee. Based on these considerations, ground clutter must be processed to mitigate its negative impact.

[0003] In existing technologies, there are generally two methods for suppressing ground clutter: time-domain and frequency-domain.

[0004] In the time domain, digital high-pass filters are generally used for processing. IIR elliptic filters, which have advantages such as equal passband ripple and narrow transition band, are typically employed.

[0005] In the frequency domain, the Gaussian Model Adaptive Processing (GMAP) method is generally used. This algorithm assumes that the spectra of clutter signals and weather signals are both Gaussian in shape. It uses this characteristic to identify and filter the spectral components of ground clutter, and then uses a Gaussian model to compensate for and recover weather signals that may exist near the zero velocity location.

[0006] In the time-domain processing methods mentioned above, due to the transient response during the filtering process of the IIR elliptic filter, its clutter suppression performance will be reduced by the transient response, making it difficult to completely filter out ground clutter in the signal. In addition, the filtering performance of the filter is related to the clutter intensity. When the clutter is strong, its attenuation may be insufficient, while when the clutter is weak or there is no clutter, it will cause excessive attenuation. When the frequency components of the weather echo signal and the clutter signal are very close, the filter will also cause a large attenuation of the weather echo signal in the stopband, which will also affect the accuracy of variable estimation in the radar echo.

[0007] In the frequency domain processing methods mentioned above, GMAP also applies filtering to weather signals without clutter, which can affect the accuracy of variable estimation in radar echoes. Summary of the Invention

[0008] This invention provides a ground clutter frequency domain suppression method based on correlation spectrum, which solves the problem that traditional weather radar ground clutter processing methods cannot accurately identify the situation where weather signals and ground clutter signals are mixed, and also have the disadvantage of generally suppressing weather signals near zero velocity at the same time.

[0009] To achieve the above objectives, this invention introduces the autocorrelation spectral density (ASD) in the frequency domain during clutter identification and processing, and simultaneously utilizes the amplitude and phase information in the correlation spectrum, employing the following technical solution: The ground clutter frequency domain suppression method based on correlation spectrum includes the following steps: (1) Signal sampling: Perform staggered pulse sampling or single pulse sampling on the weather signal to obtain the original complex voltage sequence of the received echo signal. The number of pulses in the original complex voltage sequence is denoted as M, and the value of the sequence index m ranges from 0 to M-1. (2) Subsequence reconstruction: Based on the pulse repetition interval mode (single pulse repetition interval or staggered pulse repetition interval) of the radar operation, the uniform sampling subsequence is reconstructed based on the original complex voltage sequence obtained in step (1). In the staggered pulse repetition interval mode, at least three uniform sampling subsequences of length M are obtained, and the sampling period of the uniform sampling subsequence is denoted as the uniform sampling period. (3) Data window processing: Based on the clutter ratio of the uniform sampled subsequence obtained in step (2), select the data window with sidelobe attenuation greater than the clutter ratio and the largest sidelobe level; perform power preservation processing on the selected data window so that the total signal power after weighting the data window is consistent with the original signal power; (4) Spectral density calculation: Based on the data window processed with power preservation in step (3), calculate the discrete Fourier transform of the uniformly sampled subsequence with a time shift of a specific interval; based on the discrete Fourier transform results, calculate the autocorrelation spectral density (ASD) of the uniformly sampled subsequence with a lag of a specific order, and the power spectral density (PSD) of the uniformly sampled subsequence. (5) Clutter identification: Based on the autocorrelation spectral density (ASD) and power spectral density (PSD) obtained in step (4), phase spectrum clutter identification and amplitude spectrum clutter identification are performed respectively; among them, phase spectrum clutter identification utilizes the phase characteristics of autocorrelation spectral density (ASD) with a specific hysteresis order, and amplitude spectrum clutter identification combines the amplitude of autocorrelation spectral density (ASD) with the power of spectral floor noise. (6) Clutter filtering: For the clutter spectrum components identified in step (5), clutter filtering is performed on the amplitude spectrum and the phase spectrum respectively; interpolation compensation is performed on the missing positions after removing clutter spectrum points in the amplitude spectrum, and the corresponding ideal phase line is inserted into the missing positions after removing clutter spectrum points in the phase spectrum. (7) Inverse transformation and variable estimation: Convert the frequency domain signal after filtering out clutter in step (6) into a time domain signal, calculate specific autocorrelation parameters based on the time domain signal, and use the traditional PPP (pulse pair processing) method to estimate radar variables.

[0010] In this specification, in step (2), when the radar is operating in single pulse repetition interval mode, the pulse repetition period in this mode is equal to the uniform sampling period described in step (2); the original complex voltage sequence is directly used as the base sequence to reconstruct l+1 uniform sampling subsequences of length M, where l is the hysteresis order (positive integer). Each uniform sampling subsequence is obtained by sampling samples from different starting indices in the original complex voltage sequence with the pulse repetition period as the sampling interval, and the sample index m of each subsequence is in the range of 0 to M-1.

[0011] In this specification, in step (2), when the radar operates in the staggered pulse repetition interval mode, this mode includes a first repetition interval corresponding to odd pulses and a second repetition interval corresponding to even pulses, and the time ratio of the first repetition interval to the second repetition interval is equal to the staggered ratio ( : ),in , are positive integers and < , and Coprime, and The sum of the two is as follows Based on the original complex voltage sequence, four uniform sampling subsequences of length M are reconstructed. The four subsequences start from the first, second, third and fourth samples of the original complex voltage sequence, respectively, and are all extracted with the sum of the first repetition interval and the second repetition interval as the sampling interval. The sample index m of each subsequence ranges from 0 to M-1.

[0012] In this specification, in step (3), the selected data window is one of the rectangular window, Hanning window, Blackman window, and Nuttall window; the specific method of power preservation processing is to make the sum of the squares of the weighted values ​​of the data window at all sample points equal to the number of pulses M of the uniform sampling subsequence, so as to ensure that no additional power attenuation or gain is introduced after the data window is weighted.

[0013] In this specification, in step (4), the time shift of the specific interval includes 0, / 1, 1+ / The corresponding time interval; the autocorrelation spectral density (ASD) of the specific order includes the autocorrelation spectral density of lag 1 (lag-1ASD), lag... / lag- / ASD), lag / lag- / The calculation of all autocorrelation spectral densities (ASD) is based on the product of the complex conjugate of the discrete Fourier transform result of the corresponding time-shifted sequence and the original result, and is normalized by combining the sampling period and the number of pulses.

[0014] In this specification, the specific method for phase spectrum clutter identification in step (5) is as follows: using the phase characteristics of the lag-1ASD obtained in step (4), the ground clutter signal has a narrow spectral width and a large power gradient. The phase of its lag-1ASD is relatively flat near the spectral peak and has a significant deviation from the ideal linear phase. Based on this deviation characteristic, the clutter spectral components near zero velocity are identified. When the phase of the lag-1ASD is less than the maximum absolute value of the phase of the lag-1ASD obtained from the clutter model established based on the ideal Gaussian spectrum, it is considered that there is a clutter signal.

[0015] In this specification, the specific method for amplitude spectrum clutter identification in step (5) is as follows: first, calculate the spectral noise power of the lag-1 autocorrelation spectral density (lag-1ASD) obtained in step (4); remove the spectral components with amplitudes smaller than the spectral noise power from the clutter spectral components identified in step (5) to obtain the final clutter spectral components, thereby avoiding misjudging weather signals with low noise or power as clutter.

[0016] In this specification, in step (6), the interpolation compensation method for the amplitude spectrum is: linear interpolation is performed on the missing positions after removing clutter spectral points in logarithmic units to ensure the continuity of the amplitude spectrum; the ideal phase line insertion method for the phase spectrum is: the ideal phase line corresponding to each clutter spectral point is determined according to the true velocity characteristics of the weather signal, and the missing phase spectral points are inserted into the corresponding ideal phase line to ensure the rationality of the phase spectrum.

[0017] In this specification, during phase spectrum clutter identification in step (5), for weather radars operating at staggered pulse repetition intervals, in addition to utilizing the phase characteristics of the lag-1 autocorrelation spectral density (lag-1ASD), the lag obtained in step (4) is also combined. / lag- / ASD) and lag / lag- / The phase characteristics of the fractional autocorrelation spectral density (ASD) are analyzed. If the phases of both ASDs are on the ideal phase line corresponding to zero velocity, ground clutter is determined to exist; otherwise, no clutter is determined and no further filtering is performed. A rapid discrimination method is employed, which determines whether the zero-frequency coefficient based on the fractional hysteresis autocorrelation spectral density is greater than... / If the value is greater than this, it is determined that there is no noise and no further filtering operation is performed.

[0018] In this specification, in step (7), the specific autocorrelation parameter includes four parameters: zero-order hysteresis autocorrelation calculated based on the power spectral density (PSD) after clutter removal, hysteresis autocorrelation calculated based on the hysteresis after clutter removal, and so on. / Zero-order hysteresis autocorrelation calculated from power spectral density (PSD), and hysteresis after clutter filtering. / First-order hysteresis autocorrelation calculated using first-order autocorrelation spectral density (ASD), and hysteresis autocorrelation after clutter filtering. / The first-order hysteresis autocorrelation is calculated using the first-order autocorrelation spectral density (ASD). These four autocorrelation parameters can be directly adapted to the traditional PPP (pulse pair processing) process to achieve radar variable estimation without modifying the original signal processing logic.

[0019] In summary, the present invention has at least the following beneficial effects: This invention utilizes phase information absent in the PSD (Physical Detector) to determine the presence of clutter based on the ASD (Aspect-Defined Segment) phase of the current sequence. Sequences without clutter are left unprocessed; for sequences with clutter, the clutter range is automatically determined by combining the ASD amplitude, and the notch width of the filter is dynamically adjusted for more precise filtering. After filtering, linear interpolation is performed on the filtered position to compensate for any potential weather signals that may have been filtered out. The clutter-free ASD is then inversely transformed back to its original input form, adding a ground clutter processing step in a "modular" manner. This allows for parameter estimation of weather radar without altering the existing PPP (Physical Detection and Detection) processing. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the ground clutter frequency domain suppression method based on correlation spectrum involved in this invention.

[0022] Figure 2 This is a schematic diagram of the staggered pulse sequence extraction involved in this invention.

[0023] Figure 3 This is a schematic diagram of the time-domain waveform and frequency-domain response of the Rect, Hanning, BlackMan, and Nuttall windows involved in this invention.

[0024] Figure 4 This is a schematic diagram of ASD aliasing of a uniform sequence to a staggered sequence involved in this invention. Detailed Implementation

[0025] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0026] The following disclosure provides many different implementations or examples for carrying out different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. Furthermore, reference numerals and / or reference letters may be repeated in different examples of the embodiments of the present invention; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0027] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0028] like Figure 1 As shown, this embodiment provides a ground clutter frequency domain suppression method based on correlation spectrum, including the following steps: (1) Signal sampling: Perform staggered pulse sampling or single pulse sampling on the weather signal to obtain the original complex voltage sequence of the received echo signal. The number of pulses in the original complex voltage sequence is denoted as M, and the value of the sequence index m ranges from 0 to M-1. (2) Subsequence reconstruction: According to the pulse repetition interval mode (single pulse repetition interval or staggered pulse repetition interval) of the radar, when the radar is operating in the staggered pulse time interval mode, based on the original complex voltage sequence obtained in step (1), at least three uniform sampling subsequences of length M are reconstructed, and the sampling period of the uniform sampling subsequence is denoted as the uniform sampling period. (3) Data window processing: Based on the clutter ratio of the uniform sampled subsequence obtained in step (2), select the data window with sidelobe attenuation greater than the clutter ratio and the largest sidelobe level; perform power preservation processing on the selected data window so that the total signal power after weighting the data window is consistent with the original signal power; (4) Spectral density calculation: Based on the data window processed with power preservation in step (3), calculate the discrete Fourier transform of the uniformly sampled subsequence with a time shift of a specific interval; based on the discrete Fourier transform results, calculate the autocorrelation spectral density (ASD) of the uniformly sampled subsequence with a lag of a specific order, and the power spectral density (PSD) of the uniformly sampled subsequence. (5) Clutter identification: Based on the autocorrelation spectral density (ASD) and power spectral density (PSD) obtained in step (4), phase spectrum clutter identification and amplitude spectrum clutter identification are performed respectively; among them, phase spectrum clutter identification utilizes the phase characteristics of autocorrelation spectral density (ASD) with a specific hysteresis order, and amplitude spectrum clutter identification combines the amplitude of autocorrelation spectral density (ASD) with the power of spectral floor noise. (6) Clutter filtering: For the clutter spectrum components identified in step (5), clutter filtering is performed on the amplitude spectrum and the phase spectrum respectively; interpolation compensation is performed on the missing positions after removing clutter spectrum points in the amplitude spectrum, and the corresponding ideal phase line is inserted into the missing positions after removing clutter spectrum points in the phase spectrum. (7) Inverse transformation and variable estimation: Convert the frequency domain signal after filtering out clutter in step (6) into a time domain signal, calculate specific autocorrelation parameters based on the time domain signal, and use the traditional PPP (pulse pair processing) method to estimate radar variables.

[0029] In this specification, in step (2), when the radar is operating in single pulse repetition interval mode, the pulse repetition period in this mode is equal to the uniform sampling period described in step (2); the original complex voltage sequence is directly used as the base sequence to reconstruct l+1 uniform sampling subsequences of length M, where l is the hysteresis order (positive integer). Each uniform sampling subsequence is obtained by sampling samples from different starting indices in the original complex voltage sequence with the pulse repetition period as the sampling interval, and the sample index m of each subsequence is in the range of 0 to M-1.

[0030] In some embodiments, in step (2), when the radar operates in the staggered pulse repetition interval mode, this mode includes a first repetition interval corresponding to odd pulses and a second repetition interval corresponding to even pulses, and the time ratio of the first repetition interval to the second repetition interval is equal to the staggered ratio ( : ),in , are positive integers and < , and Coprime, and The sum of the two is as follows Based on the original complex voltage sequence, four uniform sampling subsequences of length M are reconstructed. The four subsequences start from the first, second, third and fourth samples of the original complex voltage sequence, respectively, and are all extracted with the sum of the first repetition interval and the second repetition interval as the sampling interval. The sample index m of each subsequence ranges from 0 to M-1.

[0031] In some embodiments, in step (3), the selected data window is selected from one of the rectangular window, Hanning window, Blackman window, and Nuttall window; the specific method of power preservation processing is to make the sum of the squares of the weighted values ​​of the data window at all sample points equal to the number of pulses M of the uniform sampling subsequence, so as to ensure that no additional power attenuation or gain is introduced after the data window is weighted.

[0032] In some embodiments, in step (4), the time shift of the specific interval includes 0, / 1, 1+ / The corresponding time interval; the autocorrelation spectral density (ASD) of the specific order includes the autocorrelation spectral density of lag 1 (lag-1ASD), lag... / lag- / ASD), lag / lag- / The calculation of all autocorrelation spectral densities (ASD) is based on the product of the complex conjugate of the discrete Fourier transform result of the corresponding time-shifted sequence and the original result, and is normalized by combining the sampling period and the number of pulses.

[0033] In some embodiments, the specific method for phase spectrum clutter identification in step (5) is as follows: Utilizing the phase characteristics of the lag-1ASD obtained in step (4), the ground clutter signal has a narrow spectral width and a large power gradient. The phase of its lag-1ASD exhibits a relatively flat characteristic near the spectral peak, showing a significant deviation from the ideal linear phase. Based on this deviation characteristic, clutter spectral components near zero velocity are identified. Wherein, when the phase of the lag-1ASD is less than the maximum absolute value of the phase of the lag-1ASD obtained from the clutter model established based on the ideal Gaussian spectrum, it is considered that a clutter signal exists.

[0034] In some embodiments, the specific method for amplitude spectrum clutter identification in step (5) is as follows: first, calculate the spectral noise power of the lag-1 autocorrelation spectral density (lag-1ASD) obtained in step (4); remove the spectral components with amplitudes smaller than the spectral noise power from the clutter spectral components identified in step (5) to obtain the final clutter spectral components, thereby avoiding misjudging weather signals with low noise or power as clutter.

[0035] In some embodiments, in step (6), the interpolation compensation method for the amplitude spectrum is: linear interpolation is performed on the missing positions after removing clutter spectral points in logarithmic units to ensure the continuity of the amplitude spectrum; the ideal phase line insertion method for the phase spectrum is: the ideal phase line corresponding to each clutter spectral point is determined according to the true velocity characteristics of the weather signal, and the missing phase spectral points are inserted into the corresponding ideal phase line to ensure the rationality of the phase spectrum.

[0036] In some embodiments, during phase spectrum clutter identification in step (5), for weather radars operating at staggered pulse repetition intervals, in addition to utilizing the phase characteristics of the lag-1 autocorrelation spectral density (lag-1ASD), the lag obtained in step (4) is also combined. / lag- / ASD) and lag / lag- / The phase characteristics of the fractional autocorrelation spectral density (ASD) are analyzed. If the phases of both ASDs are on the ideal phase line corresponding to zero velocity, ground clutter is determined to exist; otherwise, no clutter is determined and no further filtering is performed. A rapid discrimination method is employed, which determines whether the zero-frequency coefficient based on the fractional hysteresis autocorrelation spectral density is greater than... / If the value is greater than this, it is determined that there is no noise and no further filtering operation is performed.

[0037] In some embodiments, in step (7), the specific autocorrelation parameter includes four parameters: zero-order hysteresis autocorrelation calculated based on the power spectral density (PSD) after clutter removal, hysteresis autocorrelation calculated based on the hysteresis after clutter removal, and zero-order hysteresis autocorrelation calculated based on the power spectral density (PSD) after clutter removal. / Zero-order hysteresis autocorrelation calculated from power spectral density (PSD), and hysteresis after clutter filtering. / First-order hysteresis autocorrelation calculated using first-order autocorrelation spectral density (ASD), and hysteresis autocorrelation after clutter filtering. / The first-order hysteresis autocorrelation is calculated using the first-order autocorrelation spectral density (ASD). These four autocorrelation parameters can be directly adapted to the traditional PPP (pulse pair processing) process to achieve radar variable estimation without modifying the original signal processing logic.

[0038] The technical concept of this invention is as follows: The weather signal is sampled by staggered pulses or single pulses to obtain the complex voltage sequence V(m) of the received echo signal, m=0,1,2…,M-1; M is the number of pulses.

[0039] Depending on whether the radar operates in a single-pulse repetition interval or a staggered-pulse repetition interval mode, a complex voltage sequence with uniform time intervals is reconstructed from the original sequence. It is the sequence length. It is a uniform sampling period, as detailed below: For radars operating within a single pulse repetition interval, the pulse repetition period is... ,Right now The complex voltage sequence of the original echo is... , It is the sequence length. Reconstruction. A uniformly sampled subsequence of length M, where M represents the number of pulses: , ;in .

[0040] For radars operating at staggered pulse repetition intervals, These represent the odd and even pulse repetition time intervals under staggered repetition intervals, respectively. Based on the original sequence, a uniformly sampled subsequence of length M is reconstructed. , , , ;in ;like Figure 2 As shown; With m=0, that is , , , For example, the process of reconstructing these four uniformly sampled subsequences is as follows: First subsequence Starting from the first sample of the original sequence, the sampling interval is... ; Second subsequence Starting from the second sample of the original sequence, the sampling interval is... ; Third subsequence Starting from the third sample of the original sequence, the sampling interval is... ; The fourth subsequence Starting from the fourth sample of the original sequence, the sampling interval is... ; Wherein, sequence V is a complex sequence of echo signals received by the radar, M is the number of pulses, m=0,1,2,…M-1; , , The stagger ratio of the staggered pulse sequence. ;and, Coprime.

[0041] For sequences with uneven pulse repetition intervals, These represent the time intervals of odd and even pulses under staggered repetition intervals, respectively. , ,Right now ,in , .

[0042] That is, uniform sampling subsequence , , , ; Based on the clutter ratio of the uniformly sampled subsequence, a data window with sidelobe attenuation greater than the signal-to-noise ratio and the largest sidelobe level is selected; For the selected data window, power preservation processing is performed, that is... , Selected data window; Calculate time shift Discrete Fourier transform of a uniformly sampled subsequence; Calculate the lag of the uniformly sampled subsequence ASD; Calculate the PSD of the uniformly sampled subsequence; Based on the above calculation results, phase spectrum clutter identification and amplitude spectrum clutter identification are performed, and corresponding clutter filtering is carried out to obtain the frequency domain signal with ground clutter filtered out. The frequency domain signal, after filtering out ground clutter, is converted into a time domain signal so that radar variable estimation can be performed using the traditional PPP method.

[0043] In some embodiments, time shift is calculated The discrete Fourier transform of the uniformly sampled subsequence is given by the following formula: (1); when =0、 , , That is, calculation , , , .

[0044] In some embodiments, the lag- of the uniformly sampled subsequence is calculated. ASD: For time shift and The sequence calculation lag is Autocorrelation spectral density ASD: ; For radars operating at uniform pulse repetition intervals, the hysteresis of their corresponding uniform sampling subsequences of ASD is calculated as follows ( ): ; in, It is the complex conjugate of the discrete Fourier transform of the reference subsequence (0th lag) after power-preserving window processing. It is a delay after power preservation window processing The discrete Fourier transform results of the subsequences, It is a frequency index. The lag order (a positive integer); Specifically, for radars operating at uniform pulse repetition intervals, the ASD with a lag of 1 is calculated, i.e. ASD: ; That is, the hysteresis-1ASD can be obtained by multiplying the DFT conjugate of the signal with its linearly phase-shifted DFT.

[0045] For radars operating at staggered pulse repetition intervals, for time shifts of... and The sequence calculation lag is Autocorrelation spectral density ASD: ; In particular, for radars operating in staggered pulse repetition intervals, the fractional-order ASD is calculated as follows: Lag is , The autocorrelation spectral density ASD of 1, , , They are respectively: (2); (3); (4); For radars operating at staggered pulse repetition intervals, the autocorrelation spectral density (ASD) of the reconstructed uniform sampled subsequence is an aliasing result of the ASD of its underlying uniform pulse repetition interval sequence, where the repetition time of the underlying uniform pulse repetition interval sequence (basic uniform PRT sequence) is... That is, the fractional lag ASD of the staggered pulse sequence. ASD corresponds to the integer-order lag ASD of the underlying uniform PRT sequence. ASD). For example, in a time series with a staggered pulse repetition interval of 2:3, its lag-2 / 5 ASD can be obtained by aliasing the lag-2 ASD of the basic uniform PRT sequence, and its lag-3 / 5 ASD can be obtained by aliasing the lag-3 ASD of the basic uniform PRT sequence. The lag-1 ASD corresponds to the aliasing of the lag-5 ASD of the basic uniform PRT sequence (e.g., ASD). Figure 4 (As shown).

[0046] In some embodiments, the PSD of the uniformly sampled subsequence is calculated: (5); (6).

[0047] In some embodiments, radar variable estimation using the traditional PPP method can be performed by inversely transforming the filtered spectral density function to obtain four autocorrelation coefficients according to the following formula: (7); (8); (9); (10).

[0048] In some embodiments, during phase spectrum clutter identification, the phase characteristics of lag-1ASD are used to identify the spectral points corresponding to clutter in the phase spectrum. Clutter signals have narrow spectral widths and large power gradients, and their lag-1ASD phases are relatively flat near the spectral peaks, exhibiting a significant deviation from the ideal linear phase. Based on this characteristic, the spectral leakage of the lag-1ASD phase is used to identify the clutter spectrum near zero velocity. In this process, a clutter model is created to predict the phase deviation range of lag-1ASD caused by its spectral leakage. Spectral components in the lag-1ASD phase obtained after applying a certain data window to the sequence, whose absolute value is less than the maximum phase of the clutter model with the same window, may belong to clutter components.

[0049] The clutter model is a model of an ideal clutter signal with a certain spectral width and Doppler velocity when different data windows are applied. The lag-1ASD (i.e., the clutter model) is obtained by convolving a finely sampled ideal clutter lag-1ASD with the power spectrum of a data window. The lag-1ASD of an ideal clutter signal exhibits a Gaussian distribution in its amplitude spectrum (mean Doppler velocity of 0 m / s, spectral width of 0.4 m / s) and a linear phase spectrum with a phase slope of 1 in the interval [-π, π]. Fine sampling requires that the number of sampling points for the ideal clutter signal be greater than the number of sampling points for the data (e.g., if the number of data points M=64, then 512 sampling points are chosen for fine sampling). Furthermore, the data window length must be consistent with the number of points used when calculating the ASD of the sequence, and it needs to be padded with zeros to the number of points for fine sampling.

[0050] In some embodiments, when using only phase as a basis to identify clutter in amplitude spectrum clutter identification, noisy or low-power weather signals will also be identified as clutter components, and ASD amplitude is introduced as another criterion. Calculate the spectral noise floor power of lag-1ASD Then, spectral components smaller than the noise threshold are removed from the clutter spectral components identified by the above phase to obtain the final clutter spectral components.

[0051] In some embodiments, amplitude clutter filtering involves removing the spectral points corresponding to ground clutter from the amplitude spectrum and then linearly interpolating the removed amplitude spectral points back into the amplitude spectrum. The amplitude interpolation is performed in logarithmic units.

[0052] In some embodiments, phase clutter filtering involves removing the spectral points corresponding to ground clutter from the phase spectrum and inserting the removed phase spectral points into the ideal phase line corresponding to those points.

[0053] Clutter filtering: The above operations have identified the spectral points corresponding to the ground clutter. This step is to filter out the ground clutter signal while retaining the weather radar signal at the location. This step requires two PSD operations (calculated by equations (5) and (6)) and two ASD operations (calculated by equations (2) and (3)).

[0054] In some embodiments, for radars operating in staggered repetition interval mode, when identifying phase spectrum clutter, determining the presence of clutter in the phase spectrum requires not only utilizing the phase characteristics of the autocorrelation spectral density (lag-1ASD) of lag-1 as described in the above embodiments, but also further combining the phase characteristics of the fractional lag ASD. The type of clutter contained in the signal is related to the fractional lag at and near zero velocity, i.e., lag- and lag- The correspondences of ASD phases are as follows: If clutter is present only or the clutter intensity is stronger than the weather (clutter is dominant): the clutter aliases to the ideal phase line corresponding to zero velocity (e.g., in the case of a stagger ratio of 2:3, the Nyquist interval of the basic uniform sequence is divided into 5 sub-intervals (IV), and the ideal phase line corresponding to zero velocity clutter is the ideal phase line of the III sub-interval). Weather only or weather intensity stronger than clutter (weather dominates): aliased onto the ideal phase line corresponding to the true velocity of the weather signal; Weather intensity overlaps with clutter but the intensity difference is small: it does not alias onto any ideal phase line; lag- and lag- If the phase of the ASD is on the ideal phase line corresponding to zero velocity, then clutter is considered to exist in the weather signal; if clutter is present, further clutter identification and filtering are required; otherwise, no processing is performed. A rapid method for identification is to determine whether the zero-frequency coefficient based on the fractional-order hysteresis autocorrelation spectral density is greater than... / If the value is greater than this, it is determined that there is no noise and no further filtering operation is performed.

[0055] In the technical solution of this invention, theoretical derivation proves that the lag of a uniformly sampled sequence... l The mean and variance of ASD are consistent with those of PSD, indicating that ASD can replace PSD (Power Spectral Density) as a new tool for signal spectrum analysis. Furthermore, ASD inherently carries phase information, which PSD lacks. The core of this application's solution lies in utilizing this phase information absent in PSD.

[0056] This method determines the presence of clutter based on the ASD phase of the current sequence. Sequences without clutter are left unprocessed; for sequences with clutter, the method automatically determines the clutter range by combining the ASD amplitude and dynamically adjusts the notch width of the filter, thus achieving relatively accurate clutter filtering. After filtering, linear interpolation is performed on the filtered position to compensate for any potential weather signals that may have been filtered out. The PSD and ASD spectral density functions after clutter filtering are then inversely transformed back to the original input form, adding a "modular" ground clutter processing step. This allows for normal PPP processing without altering the original signal processing details.

[0057] The technical solution of the present invention is applicable to weather radars operating in both staggered pulse mode (non-uniform sampling) and single pulse (uniform sampling) modes.

[0058] In the technical solution of this invention, if other data windows are selected in the step "Calculate ASD" -> "Select a suitable window function", this solution is equally effective.

[0059] The technical solution of this invention is described using a single-polarization method. However, this technical solution is equally effective if a multi-polarization method is used.

[0060] The embodiments described above are for illustrative purposes only and are not intended to limit the invention. Therefore, any changes in numerical values ​​or substitutions of equivalent elements should still fall within the scope of this invention.

[0061] The above detailed description will enable those skilled in the art to understand that the present invention can indeed achieve the aforementioned objectives and has complied with the provisions of the Patent Law.

[0062] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention. The above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

[0063] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0064] The basic concepts have been described above. Obviously, for those skilled in the art who have read this application, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this application.

[0065] Furthermore, this application uses specific terms to describe its embodiments. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different positions in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application can be appropriately combined.

[0066] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Therefore, aspects of this application can be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. All of the above hardware or software can be referred to as a “unit,” “module,” or “system.” Furthermore, aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0067] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, and Python; general programming languages ​​such as C; Visual Basic, Fortran2103, Perl, COBOL2102, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0068] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention have been discussed in the foregoing disclosure by way of various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a purely software solution, such as an installation on an existing server or mobile device.

[0069] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this approach of the present application should not be construed as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject of the invention should possess fewer features than in any single embodiment described above.

Claims

1. A ground clutter frequency domain suppression method based on correlation spectrum, characterized in that, include: (1) Signal sampling: Perform staggered pulse sampling or single pulse sampling on the weather signal to obtain the original complex voltage sequence of the received echo signal. The number of pulses in the original complex voltage sequence is denoted as M, and the value of the sequence index m ranges from 0 to M-1. (2) Subsequence reconstruction: According to the pulse repetition interval mode of the radar operation, when the radar operates in the mode of staggered pulse time interval, based on the original complex voltage sequence obtained in step (1), at least three uniform sampling subsequences of length M are reconstructed, and the sampling period of the uniform sampling subsequence is denoted as the uniform sampling period. (3) Data window processing: Based on the clutter ratio of the uniform sampled subsequence obtained in step (2), select the data window with sidelobe attenuation greater than the clutter ratio and the largest sidelobe level; perform power preservation processing on the selected data window so that the total signal power after weighting the data window is consistent with the original signal power; (4) Spectral density calculation: Based on the data window processed with power preservation in step (3), calculate the discrete Fourier transform of the uniformly sampled subsequence with a time shift of a specific interval; based on the discrete Fourier transform results, calculate the autocorrelation spectral density of the uniformly sampled subsequence with a lag of a specific order, and the power spectral density of the uniformly sampled subsequence. (5) Clutter identification: Based on the autocorrelation spectral density and power spectral density obtained in step (4), phase spectrum clutter identification and amplitude spectrum clutter identification are performed respectively; among them, phase spectrum clutter identification utilizes the phase characteristics of autocorrelation spectral density with a specific hysteresis order, and amplitude spectrum clutter identification combines the amplitude of autocorrelation spectral density with the power of spectral floor noise. (6) Clutter filtering: For the clutter spectrum components identified in step (5), clutter filtering is performed on the amplitude spectrum and the phase spectrum respectively; interpolation compensation is performed on the missing positions after removing clutter spectrum points in the amplitude spectrum, and the corresponding ideal phase line is inserted into the missing positions after removing clutter spectrum points in the phase spectrum. (7) Inverse transformation and variable estimation: Convert the frequency domain signal after filtering out clutter in step (6) into a time domain signal, calculate specific autocorrelation parameters based on the time domain signal, and use pulse pair processing method to estimate radar variables.

2. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 1, characterized in that, In step (2), when the radar is operating in the single pulse repetition interval mode, the sampling period of the uniform sampling subsequence in this mode is equal to the pulse repetition period; the original complex voltage sequence is directly used as the base sequence, i.e. the uniform sampling subsequence, and the autocorrelation spectral density is calculated through the uniform sampling subsequence with hysteresis order of 0 and 1. Each uniform sampling subsequence is obtained from samples with different starting indices in the original complex voltage sequence, with the pulse repetition period as the sampling interval, and the sample index m of each subsequence is in the range of 0 to M-1.

3. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 1, characterized in that, In step (2), when the radar operates in the staggered pulse repetition interval mode, this mode includes a first repetition interval corresponding to odd-numbered pulses and a second repetition interval corresponding to even-numbered pulses. The time ratio of the first repetition interval to the second repetition interval is equal to the staggered ratio ( : ),in , are positive integers and < , and Coprime, and The sum of the two is as follows Based on the original complex voltage sequence, four uniform sampling subsequences of length M are reconstructed. The four subsequences start from the first, second, third and fourth samples of the original complex voltage sequence, respectively, and are all extracted with the sum of the first repetition interval and the second repetition interval as the sampling interval. The sample index m of each subsequence ranges from 0 to M-1.

4. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 1, characterized in that, In step (3), the selected data window is one of the rectangular window, Hanning window, Blackman window, and Nuttor window; the specific method of power preservation processing is to make the sum of the squares of the weighted values ​​of the data window at all sample points equal to the number of pulses M of the uniform sampling subsequence, so as to ensure that no additional power attenuation or gain is introduced after the data window is weighted.

5. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 3, characterized in that, In step (4), the time shift of the uniformly sampled subsequence at a specific interval includes 0, / 1, 1+ / The corresponding time interval; the autocorrelation spectral density of the specific order includes hysteresis. / Autocorrelation spectral density and hysteresis of order / The autocorrelation spectral density of order 1, lag 1, and all The calculation of autocorrelation spectral density is based on the corresponding time shift. Complex conjugate and time shift of the discrete Fourier transform result of the sequence The product of the discrete Fourier transform results of the sequence is normalized by combining the sampling period and the number of pulses.

6. The correlation-spectrum-based geomyuic frequency-domain suppression method according to claim 1, characterized in that, In step (5), the specific method for phase spectrum clutter identification is as follows: using the phase characteristics of the hysteresis first-order autocorrelation spectral density obtained in step (4), the ground clutter signal has a narrow spectral width and a large power gradient. The phase of its hysteresis first-order autocorrelation spectral density shows a relatively flat feature near the spectral peak, which deviates significantly from the ideal linear phase. Based on this deviation characteristic, the clutter spectral components near zero velocity are identified. Among them, when the phase of the hysteresis first-order autocorrelation spectral density is less than the maximum absolute value of the phase of the hysteresis first-order autocorrelation spectral density obtained from the clutter model established based on the ideal Gaussian spectrum, it is considered that there is a clutter signal.

7. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 1, characterized in that, In step (5), the specific method for amplitude spectrum clutter identification is as follows: first, calculate the spectral noise power of the hysteresis first-order autocorrelation spectral density obtained in step (4); from the clutter spectral components identified by the phase spectrum in step (5), remove the spectral components with amplitudes smaller than the spectral noise power to obtain the final clutter spectral components, so as to avoid misjudging the weather signal with low noise or power as clutter.

8. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 1, characterized in that, In step (6), the interpolation compensation method for the amplitude spectrum is: linear interpolation is performed on the missing positions after removing clutter spectral points in logarithmic units to ensure the continuity of the amplitude spectrum; the ideal phase line insertion method for the phase spectrum is: the ideal phase line corresponding to each clutter spectral point is determined according to the true velocity characteristics of the weather signal, and the missing phase spectral points are inserted into the corresponding ideal phase line to ensure the rationality of the phase spectrum.

9. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 3, characterized in that, For weather radars operating at staggered pulse repetition intervals, in step (5) when identifying phase spectrum clutter, in addition to utilizing the phase characteristics of the hysteresis first-order autocorrelation spectral density, the hysteresis obtained in step (4) is also combined. / First-order autocorrelation spectral density and hysteresis / The phase characteristics of the fractional-order autocorrelation spectral density are analyzed. If the phases of both fractional-order autocorrelation spectral densities are on the ideal phase line corresponding to zero velocity, ground clutter is determined to exist. If they are not on the ideal phase line, no clutter is determined and no further filtering is performed. The analysis also considers whether the zero-frequency coefficient based on the fractional-order hysteresis autocorrelation spectral density is greater than a certain value. / If the value is greater than this, it is determined that there is no noise and no further filtering operation is performed.

10. The ground clutter frequency domain suppression method based on correlation spectrum according to claim 9, characterized in that, In step (7), the specific autocorrelation parameters include four: zero-order hysteresis autocorrelation calculated based on the power spectral density after clutter filtering, ... and hysteresis autocorrelation calculated based on the power spectral density after clutter filtering. / Zero-order hysteresis autocorrelation in power spectral density calculation, and hysteresis after clutter filtering / First-order hysteresis autocorrelation calculated from first-order autocorrelation spectral density, and hysteresis autocorrelation after clutter filtering / The first-order hysteresis autocorrelation is calculated using the first-order autocorrelation spectral density; these four autocorrelation parameters are used to directly adapt the pulse pair processing flow to achieve radar variable estimation.

Images