CN116679278B - Target radar detection method under strong ground clutter interference - Google Patents
Target radar detection method under strong ground clutter interference Download PDFInfo
- Publication number
- CN116679278B CN116679278B CN202310971009.6A CN202310971009A CN116679278B CN 116679278 B CN116679278 B CN 116679278B CN 202310971009 A CN202310971009 A CN 202310971009A CN 116679278 B CN116679278 B CN 116679278B
- Authority
- CN
- China
- Prior art keywords
- clutter
- signal
- target
- radar
- spectrum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 29
- 238000000034 method Methods 0.000 claims abstract description 59
- 239000013598 vector Substances 0.000 claims abstract description 38
- 238000001914 filtration Methods 0.000 claims abstract description 22
- 230000001629 suppression Effects 0.000 claims abstract description 19
- 230000003044 adaptive effect Effects 0.000 claims abstract description 6
- 230000003068 static effect Effects 0.000 claims abstract description 6
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 238000001228 spectrum Methods 0.000 claims description 34
- 239000011159 matrix material Substances 0.000 claims description 22
- 230000003595 spectral effect Effects 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 11
- 238000013527 convolutional neural network Methods 0.000 claims description 10
- 238000005311 autocorrelation function Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 230000006872 improvement Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000005764 inhibitory process Effects 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 abstract description 3
- 238000010276 construction Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000002592 echocardiography Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/414—Discriminating targets with respect to background clutter
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/36—Means for anti-jamming, e.g. ECCM, i.e. electronic counter-counter measures
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The application discloses a target radar detection method under strong ground clutter interference, which relates to the field of radar detection and mainly comprises the following steps: filtering clutter by adopting a cascade self-adaptive method of multi-feature vectors and detecting the target of the processed signal by adopting a method for constructing a network model; based on the blind speed phenomenon generated by an MTI filter, an optimization method of a variable pulse repetition interval working mode is provided, and clutter suppression methods are respectively carried out on moving clutter and clutter combined by movement and static clutter; the original signal clutter suppression method is improved by designing an adaptive filter of a multi-feature vector, and a filter coefficient is adaptively generated for filtering; detecting targets in radar signals by adopting a neural network method, and optimizing the structure construction and learning strategy of the neural network according to the characteristics of radar data; the combined application of the two can realize the positioning detection of the target in the strong clutter environment.
Description
Technical Field
The application relates to the technical field of radar detection, in particular to a target radar detection method under strong ground clutter interference.
Background
Radar refers to equipment for detecting and sensing by using radio, and in the process of continuous development of radar technology, physical quantities measured by radar sensing are gradually expanded from distance, azimuth and pitch angle of a target to the speed of measuring the movement of the target by using Doppler effect and the detection of the size and shape of the target by using synthetic aperture radar. The application range of the radar is gradually expanded from the military field to the civil field along with the development of corresponding radar signal processing technology. At present, the radar has important roles in the fields of military countermeasure, aerospace, missile manufacture, transportation, communication, measurement and control, weather and the like.
Along with the progress of scientific technology, the radar has more and more complex application scenes, and radar echoes contain clutter information of various surrounding environments besides scattering echoes of targets. The clutter can cause the adaptability of the radar system in the environment to be reduced, so that the perception capability of the radar system on the target in the dynamic environment is reduced, and the actual application requirement is difficult to meet.
Therefore, aiming at the application requirements of radar signal high-efficiency processing in a complex dynamic environment, how to reasonably and effectively mine and process the data characteristics received by the radar, and the method has important significance for improving the perception performance of the radar system on targets and clutter signals.
At present, an MTI filter is commonly used for suppressing clutter, but the MTI filter with constant conventional coefficient can only have effect on static clutter, the problem of suppressing slow moving clutter can be solved by adding a filter notch, but the MTI filter has almost no suppression effect on moving clutter with Doppler frequency far away from zero frequency; the target detection mainly adopts a CFAR detection algorithm which has the functions of self-adaption and false alarm control, but the target detection performance in a strong clutter environment is general, and the false alarm rate is high.
Therefore, the application provides a target radar detection method under the interference of strong ground clutter.
Disclosure of Invention
(one) solving the technical problems
Aiming at the defects of the prior art, the application provides a target radar detection method under strong ground clutter interference, which is characterized in that firstly, an existing MTI filter is optimized, and a characteristic vector method of a plurality of vectors is realized to carry out self-adaptive filtering; then, in order to enable target detection to be more intelligent, the method adopts a neural network method to detect signals and improve the signals. And further optimizes the problems in the background art.
(II) technical scheme
In order to achieve the above purpose, the application is realized by the following technical scheme: a target radar detection method under strong clutter interference comprises the following steps:
step 1, filtering clutter by adopting a cascade self-adaptive method of multi-feature vectors, obtaining clutter spectrum properties at different distance gates by using a clutter characteristic analysis method, generating corresponding multiple groups of coefficients, and respectively performing inhibition operation on the clutter;
step 2, detecting the target of the processed signal by adopting a method for constructing a network model; step 1 firstly, an optimization method of a variable pulse repetition interval working mode is provided based on a blind speed phenomenon generated by an MTI filter, and clutter suppression methods are respectively designed for moving clutter and clutter combined with motion and stillness.
Further, the variable pulse repetition interval operation mode refers to that the pulse repetition frequency is increasedThe method of the method improves the first blind speed to avoid the blind speed phenomenon, uses the pulses with different repetition frequencies to alternately work, the time difference between the adjacent two pulses emitted by the radar is unequal under the working mode,
wherein ,represents the average repetition frequency of N pulses, +.>Indicating Doppler frequency point corresponding to the first blind velocity after adopting the staggered pulse frequency,/and>representing a multiple of the first blind speed increase.
Further, the clutter suppression in motion is divided into the following three steps:
step S1, clutter spectrum parameter estimation: the center frequency of the signal is calculated by analyzing the power spectrum obtained by carrying out Fourier transform on the signal and using the following formulaAnd spectral width->：
Where f represents the frequency of the signal,representing the signal power spectrum after normalization processing;
step S2, designing an adaptive MTI filter through a feature vector method: modeling clutter by adopting a Gaussian model, and normalizing a power spectrum density function by a Gaussian spectrumThe method comprises the following steps:
and deriving an autocorrelation function of the clutter autocorrelation matrix from the power spectrum of the clutter by using the wiener-Xin Qin theorem, wherein the autocorrelation function is as follows:
wherein ,representing the correlation time of adjacent pulse signals. The clutter autocorrelation matrix is:
similarly, the autocorrelation function of the target signal is:
and then obtain:
if the input of the MTI filter is composed of clutterSum signal->Two sequences are composed, the weighting coefficient of the MTI filter is +.>After filtering, the power of clutter is +.>Power of and signalThe method comprises the following steps of:
wherein ,represents the clutter autocorrelation matrix, ">Representing the autocorrelation matrix of the signal. /> and />The input powers of clutter and signals at the input of MTI filter are respectively represented, and an improvement factor +.>The calculation formula is as follows:
and because ofIs an identity matrix, then there is
And thenIs>N/> wherein />Is characteristic value +.>The corresponding feature vector. When->In the case of an N square matrix, +.>Having N eigenvalues, arranged in order of magnitude as:
the energy of the hybrid wave is mainly concentrated by d large eigenvaluesCorresponding dFeature vector->In the tensed signal subspace; the remaining N-d small eigenvalues +.>Corresponding feature vector +.>The stretched noise subspace is orthogonal to the signal subspace; the feature vector of the noise subspace is selected to suppress clutter;
step S3, selecting a multi-feature vector: the one or two feature values with the largest removal are selected. And then filtering the feature vectors corresponding to other feature values as the weight coefficients of the MTI filter.
Further, the clutter suppression method for combining motion and static is as follows:
applying a delay canceller to the data after pulse compression to inhibit clutter with Doppler frequency being zero; for the data after the first stage filter, calculating the center frequency of each distance unitAnd spectral width->，/>Represents the center frequency of the nth range bin spectrum; />Representing the spectral width of the nth distance cell; solving for spectral center +.>Is 0, the spectrum broadening is +.>Covariance matrix under the condition->；
Calculating covariance matrixIs defined as the feature value and feature vector; removing the feature vector matched with the maximum feature value, and carrying out spectrum shifting on the rest feature vector to +.>Then, storing the data into a coefficient library; and filtering the coefficients in the coefficient library as weighting coefficients, and smoothing a plurality of filtering results.
Specifically, step 2, a convolutional neural network model is established to further judge a target radar, positive and negative sample distribution is performed to control the proportion of positive and negative samples in a grid division mode, and when a target falls into the grid, the target is divided into positive samples; otherwise, the grid is divided into negative samples; and then, the measured radar signal data after clutter suppression is processed according to the following steps of 7:3, randomly dividing a training set and a testing set according to the proportion, and training a CNN model; based on the accuracy of the matching probability and the mismatch probability analysis model,
wherein ,the method is characterized in that the target is detected on the premise that the echo signal has the target, namely the matching probability; />The method is characterized in that the method indicates that the unmatched probability is obtained on the premise that the echo signal does not have the target; />Indicating that radar echo signal contains target definition, < +.>Indicating absence of radar echo signalsDefining the target time; and finally, selecting 10% of data from the training set as a verification set, training 200 epochs on the target detection model, and selecting the model with the best performance in the verification set for testing.
(III) beneficial effects
The application provides a target radar detection method under strong ground clutter interference, which has the following beneficial effects:
1. in the aspect of clutter suppression, through optimizing an MTI filter, clutter suppression is respectively designed aiming at the moving clutter and the moving and static combined clutter, the phenomenon that a target is lost due to a multi-feature vector method is improved by detecting the number of targets by a multi-feature vector method, the number of false alarms is greatly reduced, and in the aspect of improving factors, the signal-to-noise ratio is increased while the false alarms are reduced.
2. In the aspect of target detection, firstly, positive and negative samples are distributed in a grid division mode, so that the problem of serious unbalance of the positive and negative samples is avoided, and secondly, a designed convolutional neural network target detection model is less influenced by clutter environment, so that a large number of false alarm phenomena are avoided; finally, the clutter suppression method is combined with the convolutional neural network target detection model, so that the target can be well positioned and detected in a strong clutter environment.
Drawings
Fig. 1 is a schematic flow chart of a target radar detection method according to the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Referring to fig. 1, a target radar detection method under strong clutter interference specifically includes the following steps:
step 1, filtering clutter by adopting a cascade self-adaption method of multiple eigenvectors, obtaining clutter spectrum properties at different distance gates by using a clutter characteristic analysis method, generating corresponding multiple groups of coefficients, and respectively performing suppression operation on the clutter.
The MTI filter is commonly used in industry for denoising, but the filter can only eliminate clutter signals with fixed frequency, cannot filter the clutter signals in a targeted way according to the characteristics of the clutter signals, and can generate a blind speed phenomenon, so that the method suppresses the clutter of motion, the clutter of rest and the clutter combined with the motion by improving the MTI method.
The MTI filter, when filtering, will also cancel echo information if its doppler frequency is exactly at an integer multiple of the pulse repetition frequency for a moving object, for which the radial velocity of the doppler frequency is called the blind velocity. The radial velocity where the doppler frequency is equal to the pulse repetition frequency is called the first blind velocity. Can be used to increase the pulse repetition frequencyThe method of (1) increases the first blind speed of the detection to avoid the occurrence of the blind speed phenomenon. In order to raise the first blind speed without reducing the fuzzy distance, the pulses with different repetition frequencies can be used for alternately working, the time difference between two adjacent pulses emitted by the radar is unequal under the working mode,
wherein ,represents the average repetition frequency of N pulses, +.>Indicating Doppler frequency point corresponding to the first blind velocity after adopting the staggered pulse frequency,/and>representing a multiple of the first blind speed increase. Through experiments, the working mode is repeated at intervals of equal pulseCompared with the prior art, the variable pulse repetition interval working mode can greatly improve the blind speed and avoid the frequency mixing phenomenon of the target speed.
Step 101, generating adaptive weight coefficients through counting the change of clutter characteristics to pertinently perform clutter suppression of motion and reduce attenuation of target information.
Step 1011, clutter spectrum parameter estimation
The frequency spectrum characteristics of the signal are obtained by a statistical signal method, specifically: the center frequency of the signal is calculated by analyzing the power spectrum obtained by carrying out Fourier transform on the signal and using the following formulaAnd spectral width->：
Where f represents the frequency of the signal,representing the normalized signal power spectrum.
Step 1012, designing an adaptive MTI filter by eigenvector method
After the clutter is subjected to the characteristic extraction of the frequency spectrum information, the notch center and the notch width can be adjusted in real time aiming at the spectrum center of the clutter, and the notch position is moved to the corresponding frequency. For the suppression of motion clutter such as cloud rain and noise waves, an adaptive MTI filter is generally selected for filtering.
Modeling clutter is generally carried out by adopting a Gaussian model, and a Gaussian spectrum normalized power spectrum density functionThe method comprises the following steps:
because the clutter is stationary and random, the wiener-Xin Qin theorem can be used to derive the autocorrelation function from the power spectrum of the clutter as:
wherein ,representing the correlation time of adjacent pulse signals. The clutter autocorrelation matrix is:
doppler information of the target signal is generally unknown, and is within the bandwidthAny position within the range is possible, and the power spectrum of the target signal can be expressed as:
similar to clutter signals, the echo information of the target is still considered as a smooth random sequence, so the autocorrelation function of the target signal is:
and then obtain:
if MTI filterIs composed of clutterSum signal->Two sequences are composed, the weighting coefficient of the MTI filter is +.>After filtering, the power of clutter is +.>Power of and signalThe method comprises the following steps of:
wherein ,represents the clutter autocorrelation matrix, ">Representing the autocorrelation matrix of the signal. /> and />The input powers of clutter and signals at the input of MTI filter are respectively represented, and an improvement factor +.>The calculation formula is as follows:
and because ofIs an identity matrix, then there is
And thenIs>N, wherein->Is characteristic value +.>The corresponding feature vector. When->In the case of an N square matrix, +.>Having N eigenvalues, arranged in order of magnitude as:
for echo components of the same range bin, generally the energy of the clutter is>Target energy>Noise energy, so that the noise energy is mainly concentrated by d large eigenvaluesD corresponding feature vectors +.>In the tensed signal subspace. The remaining N-d small eigenvalues +.>Corresponding feature vector +.>The stretched noise subspace is orthogonal to the signal subspace. The feature vector of the noise subspace is selected to suppress clutter. When->Is->Minimum feature value +.>When the corresponding feature vector is the feature vector, the value of the improvement factor I reaches the maximum.
Step 1013, selecting multiple feature vectors: in theory, the feature vector corresponding to the minimum feature value is selected to improve the improvement factor to the greatest extent, but if only the minimum feature is selected, the energy of the target is placed in the signal subspace, and in this case, the target signal may be classified into clutter and suppressed, so that the target information may be lost on the premise that only the minimum feature is selected, and therefore, the increase of the number of the selected feature vectors is considered.
The analysis of feature vector number selection is carried out by taking the pulse self-adaptive filtering with equal repetition intervals as an example, and experiments prove that the proportion of the total energy occupied by the maximum feature value and the minimum feature value is reduced along with the increase of the pulse number, which indicates that the phenomenon that only noise information is finally left when the minimum feature value is selected.
The conclusion is drawn through analysis of experimental results: one or both of the feature values that are the largest can be chosen for removal. And then filtering the feature vectors corresponding to other feature values as the weight coefficients of the MTI filter.
Step 102, if there are two different types of clutter, i.e. stationary clutter and moving clutter, the notch corresponding to the center position of the spectrum needs to be designed for filtering.
Two clutter information is filtered out by using a cascade of two filters. The first stage is used for suppressing fixed clutter, the second stage is used for suppressing moving clutter, and the cascade filter formed by the two stages uses the structure of an FIR filter. The method comprises the following specific steps:
(1) Applying a delay canceller to the data after pulse compression to inhibit clutter with Doppler frequency being zero;
(2) For the data after the first stage filter, calculating the center frequency of each distance unitAnd spectral width->，/>Represents the center frequency of the nth range bin spectrum; />Representing the spectral width of the nth distance cell;
(3) Solving for spectral centersIs 0, the spectrum broadening is +.>Covariance matrix under the condition->；
(4) Calculating covariance matrixIs defined as the feature value and feature vector;
(5) Removing the eigenvectors matched with the maximum eigenvalues, and carrying out spectrum shifting on the rest eigenvectors toThen, storing the data into a coefficient library;
(6) Filtering the coefficients in the coefficient library as weighting coefficients, and smoothing a plurality of filtering results;
and 2, detecting the target of the processed signal by adopting a method for constructing a network model.
The method takes a convolutional neural network as a technical means, combines the format characteristics of radar measured data, and constructs a target intelligent detection model which mainly aims at the target detection of the measured radar signals.
The target intelligent detection model comprises two parts: a CNN-based feature extractor and a target detection head. Wherein the feature extractor part is mainly composed of 10 convolution modules and 5 pooling layers, and each convolution module is composed of a 5×5 convolution module, a batch normalization module and a ReLu activation function.
Input radar signal dimension isN, in order to extract high-level semantic information while ensuring computational efficiency, the entire feature extraction process performs 32-fold downsampling operations along the distance unit dimension (N). Specifically, in order to avoid losing the correlation between pulses during feature extraction, the step size of the max pooling layer (MaxPool) is set to (1, 2), i.e., the pulse dimension (fr) is not downsampled, but only a 2-fold pooling operation along the distance bin dimension (N).
Because the radar signal distance is generally higher in the signal dimension (N), the number of targets is generally only 1-2, the remaining distance units are all environmental clutter or noise backgrounds, and the huge difference in the number can lead to the model to deviate from the environmental clutter background of a negative sample when the target detector is directly trained and optimized later when the gradient descent method is used for optimization, so that the model can easily judge the sample to be identified as the environmental clutter background.
The application controls the proportion of positive and negative samples by carrying out positive and negative sample distribution in a grid division mode, specifically, carries out grid division on a distance radar signal, and each grid needs to predict whether a target exists in the area as a characteristic. Then in the model training optimization process, when the target falls in the grid, dividing into positive samples; otherwise, the grid is divided into negative samples. Through the strategy of grid division, the difference between the number of positive and negative samples can be greatly reduced. The method uses 25 distance units as a grid length, and performs non-overlapping grid division on all distance unit signals in a distance direction.
Based on clutter suppression in step 1, 1000 groups of measured radar signal data after clutter suppression are used as a data set, and the following steps are carried out according to 7:3, randomly dividing the training set and the testing set according to the proportion, and training the CNN model. If the radar echo signal contains the target defined as the accuracy of the analysis model according to the matching probability and the unmatched probabilityDefined as +.>The calculation formulas of the matching probability and the non-matching probability are as follows:
wherein ,the method is characterized in that the target is detected on the premise that the echo signal has the target, namely the matching probability; />The method indicates that the target is detected on the premise that the echo signal does not have the target, namely the unmatched probability.
Model training is carried out by using the actual measurement radar signal data of 700 groups of wave positions, 10% proportion data is selected from the training set randomly to serve as a verification set, and the rest 300 groups of data serve as a test set. 200 epochs were trained on the target detection model altogether, and the model with the best performance in the validation set was selected for testing.
After the test is completed, the trained model is deployed into a target recognition system for practical application.
The above embodiments may be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, the above-described embodiments may be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application.
Claims (3)
1. A target radar detection method under strong clutter interference is characterized in that: the method comprises the following steps:
filtering clutter by adopting a cascade self-adaptive method of multiple eigenvectors, obtaining clutter spectrum properties at different range gates by using a clutter characteristic analysis method, generating corresponding multiple groups of coefficients, and respectively performing inhibition operation on the clutter;
detecting the target of the processed signal by adopting a method for constructing a convolutional neural network model;
based on the blind speed phenomenon generated by an MTI filter, an optimization method of a variable pulse repetition interval working mode is provided, and clutter suppression methods are respectively designed for moving clutter and clutter combined by movement and static clutter;
the variable pulse repetition interval mode of operation includes: by increasing pulse repetition frequencyThe method of the method improves the first blind speed to avoid the blind speed phenomenon, and uses the pulses with different repetition frequencies to alternately work, and under the working mode, the time difference between the adjacent two pulses emitted by the radar is unequal:
wherein ,represents the average repetition frequency of N pulses, +.>Indicating Doppler frequency point corresponding to the first blind velocity after adopting the staggered pulse frequency,/and>representing a multiple of the first blind speed increase;
the clutter suppression in motion is divided into the following steps:
clutter spectrum parameter estimation: the center frequency of the signal is calculated by analyzing the power spectrum obtained by carrying out Fourier transform on the signal and using the following formulaAnd spectral width->：
Where f represents the frequency of the signal,representing the signal power spectrum after normalization processing;
the adaptive MTI filter is designed through a feature vector method: modeling clutter by adopting a Gaussian model, and normalizing a power spectrum density function by a Gaussian spectrumThe method comprises the following steps:
and deriving an autocorrelation function of the clutter autocorrelation matrix from the power spectrum of the clutter by using the wiener-Xin Qin theorem, wherein the autocorrelation function is as follows:
wherein ,representing the correlation time of adjacent pulse signals;
the clutter autocorrelation matrix is:
the autocorrelation function of the target signal is:
and then obtain:
if the input of the MTI filter is composed of clutterSum signal->Two sequences are composed, the weighting coefficient of the MTI filter is +.>After filtering, the power of clutter is +.>Power of signal->The method comprises the following steps of:
wherein ,represents the clutter autocorrelation matrix, ">Representing an autocorrelation matrix of the signal;
and />The input powers of clutter and signals at the input of MTI filter are respectively represented, and an improvement factor +.>The calculation formula is as follows:
and because ofIs an identity matrix, then there is
And thenIs>N, wherein->Is characteristic value +.>The corresponding feature vector;
when (when)In the case of an N square matrix, +.>Having N eigenvalues, arranged in order of magnitude as:
the energy of the hybrid wave is mainly concentrated by d large eigenvaluesD corresponding feature vectors +.>In the tensed signal subspace; the remaining N-d small eigenvalues +.>Corresponding feature vector +.>The stretched noise subspace is orthogonal to the signal subspace; selecting feature vectors of the noise subspace to suppress clutter;
selecting a multi-feature vector: and selecting one or two characteristic values with the maximum removal, and then filtering the characteristic vector corresponding to the other characteristic values as an MTI filter weight coefficient.
2. The method for detecting the target radar under the strong clutter interference according to claim 1, wherein the method comprises the following steps: the clutter suppression method combining motion and static is as follows:
applying a delay canceller to the data after pulse compression to inhibit clutter with Doppler frequency being zero; for the data after the first stage filter, calculating the center frequency of each distance unitAnd spectral width->，/>Represents the center frequency of the nth range bin spectrum; />Representing the spectral width of the nth distance cell;
solving for spectral centersIs 0, the spectrum broadening is +.>Covariance matrix under the condition->The method comprises the steps of carrying out a first treatment on the surface of the Calculating covariance matrix->Is defined as the feature value and feature vector; removing the feature vector matched with the maximum feature value, and carrying out spectrum shifting on the rest feature vector to +.>Then, storing the data into a coefficient library; and filtering the coefficients in the coefficient library as weighting coefficients, and smoothing a plurality of filtering results.
3. The method for detecting the target radar under the strong clutter interference according to claim 2, wherein the method comprises the following steps:
establishing a convolutional neural network model to further determine a target radar, comprising:
positive and negative sample distribution is carried out in a grid division mode to control the proportion of positive and negative samples, and when a target falls in the grid, the positive and negative samples are divided into positive samples; otherwise, the grid is divided into negative samples; and then, the measured radar signal data after clutter suppression is processed according to the following steps of 7:3, randomly dividing the training set and the test set according to the proportion, and training the CNN model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310971009.6A CN116679278B (en) | 2023-08-03 | 2023-08-03 | Target radar detection method under strong ground clutter interference |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310971009.6A CN116679278B (en) | 2023-08-03 | 2023-08-03 | Target radar detection method under strong ground clutter interference |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116679278A CN116679278A (en) | 2023-09-01 |
| CN116679278B true CN116679278B (en) | 2023-10-20 |
Family
ID=87782301
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310971009.6A Active CN116679278B (en) | 2023-08-03 | 2023-08-03 | Target radar detection method under strong ground clutter interference |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116679278B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3775768A (en) * | 1971-12-06 | 1973-11-27 | Westinghouse Electric Corp | Adaptive mti clutter filter |
| CN109613505A (en) * | 2018-12-13 | 2019-04-12 | 航天南湖电子信息技术股份有限公司 | A kind of device and its suppressing method inhibiting dual clutter |
| CN109856609A (en) * | 2019-01-02 | 2019-06-07 | 成都纳雷科技有限公司 | A kind of Waveform generation method removing radar blind speed |
| CN111624574A (en) * | 2020-07-29 | 2020-09-04 | 绵阳市游仙区创新科技产业技术研究院 | Target detection method, system, storage medium and device for weak target detection |
| CN115032601A (en) * | 2022-03-29 | 2022-09-09 | 哈尔滨工程大学 | Marine radar target detection algorithm for inhibiting sea clutter in image sequence based on space-time combined filtering technology |
| CN116256746A (en) * | 2023-05-12 | 2023-06-13 | 江苏亿连通信技术有限公司 | Radar-based system and method for monitoring intrusion of foreign matters into perimeter airspace of preventive area |
-
2023
- 2023-08-03 CN CN202310971009.6A patent/CN116679278B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3775768A (en) * | 1971-12-06 | 1973-11-27 | Westinghouse Electric Corp | Adaptive mti clutter filter |
| CN109613505A (en) * | 2018-12-13 | 2019-04-12 | 航天南湖电子信息技术股份有限公司 | A kind of device and its suppressing method inhibiting dual clutter |
| CN109856609A (en) * | 2019-01-02 | 2019-06-07 | 成都纳雷科技有限公司 | A kind of Waveform generation method removing radar blind speed |
| CN111624574A (en) * | 2020-07-29 | 2020-09-04 | 绵阳市游仙区创新科技产业技术研究院 | Target detection method, system, storage medium and device for weak target detection |
| CN115032601A (en) * | 2022-03-29 | 2022-09-09 | 哈尔滨工程大学 | Marine radar target detection algorithm for inhibiting sea clutter in image sequence based on space-time combined filtering technology |
| CN116256746A (en) * | 2023-05-12 | 2023-06-13 | 江苏亿连通信技术有限公司 | Radar-based system and method for monitoring intrusion of foreign matters into perimeter airspace of preventive area |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116679278A (en) | 2023-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106469313B (en) | A kind of detection method of small target of caliber adaptive space-time domain filtering | |
| CN104569948B (en) | Sub-band adaptive GLRT LTD detection methods under sea clutter background | |
| CN111965632B (en) | Radar target detection method based on Riemann manifold dimensionality reduction | |
| CN112859014A (en) | Radar interference suppression method, device and medium based on radar signal sorting | |
| CN109669182B (en) | Passive bistatic SAR moving/static target joint sparse imaging method | |
| CN109324315A (en) | Space-time adaptive based on double level block sparsity handles radar clutter suppression method | |
| CN110161494B (en) | RD plane weak target detection tracking method and device based on velocity square filtering | |
| CN108957403B (en) | Gaussian fitting envelope time delay estimation method and system based on generalized cross correlation | |
| CN102879767A (en) | Detection method of interference target for space-time adaptive processing | |
| CN105158749A (en) | High-frequency radar sea-clutter amplitude statistical distribution test method | |
| CN107229040B (en) | high-frequency radar target detection method based on sparse recovery space-time spectrum estimation | |
| CN111562569A (en) | Weighted group sparse constraint-based multi-target constant false alarm detection method under Weibull background | |
| CN107479037A (en) | A kind of PD radar clutters area judging method | |
| CN117233706B (en) | Radar active interference identification method based on multilayer channel attention mechanism | |
| CN114117912A (en) | Sea clutter modeling and inhibiting method under data model dual drive | |
| CN113887398A (en) | GPR signal denoising method based on variational modal decomposition and singular spectrum analysis | |
| CN116679278B (en) | Target radar detection method under strong ground clutter interference | |
| Wen et al. | Sea clutter suppression based on selective reconstruction of features | |
| CN111368653A (en) | Low-altitude small target detection method based on R-D (R-D) graph and deep neural network | |
| CN114690143B (en) | Method and device for suppressing radar clutter, radar and medium | |
| CN116047504A (en) | Method for improving deconvolution to inhibit ground penetrating radar multiple | |
| CN113156392B (en) | Clutter suppression method based on pitching domain self-adaptive processing | |
| CN114152918A (en) | Anti-intermittent main lobe interference method based on compressed sensing | |
| CN113447916A (en) | Self-adaptive target detection method under non-stationary interference | |
| CN113869358A (en) | Bearing fault diagnosis method based on cyclic correlation entropy and one-dimensional shallow convolution neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |