A weather clutter suppression method based on distance-doppler two-dimensional features

Abstract

Embodiments of the present application relate to the technical field of data processing, and disclose a weather clutter suppression method based on distance-Doppler two-dimensional features, which comprises the following steps: obtaining the coordinates of training samples on a distance-Doppler two-dimensional plane, and taking the coordinates of the training samples as the center to cut out a two-dimensional matrix according to a preset size; extracting image features corresponding to the training samples from the two-dimensional matrix corresponding to the training samples; inputting the image features corresponding to the training samples into a support vector machine classifier, and iteratively training the support vector machine classifier until convergence based on the classification results of the training samples output by the support vector machine classifier and the labels of the training samples; inputting image features corresponding to a detection target into the support vector machine classifier, and obtaining the classification result of the detection target output by the support vector machine classifier; and if the classification result is weather clutter, the detection target is eliminated. The method can effectively eliminate low-altitude slow-moving weather clutter and eliminate false alarms.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of data processing, in particular to a weather clutter suppression method based on distance-Doppler two-dimensional features. BACKGROUND

[0002] Radar has the advantages of all-weather and all-day when detecting the unmanned aerial vehicle target flying at low altitude, and can obtain multi-dimensional information of the target. However, the unmanned aerial vehicle is a typical low, slow and small target, which flies at low altitude and at a slow speed, and has a small radar scattering cross section. These characteristics make it difficult to distinguish the echo of the unmanned aerial vehicle from the weather clutter at low altitude. When detecting, if the detection threshold is set too low, a large number of clutter false alarms will be formed, and if the detection threshold is set too high, the target will be missed. Therefore, how to effectively suppress the weather clutter is a key problem for radar to accurately detect the unmanned aerial vehicle target.

[0003] Common clutter suppression methods include an AMTI (Adaptive Moving Target Indication) based method and a clutter map based method. The AMTI based method mainly targets narrow spectrum moving clutter. First, the spectrum center of the clutter is estimated to calculate the frequency to be compensated, then the spectrum center of the clutter is moved to zero frequency through motion compensation, and finally the clutter is suppressed through an MTI filter. The clutter map based method divides the space scanned by the antenna into several clutter units, and then stores the processed clutter information according to the divided units. This method can record and update the information of the fixed ground clutter in the detection scene in real time, and can eliminate the fixed ground clutter by comparing the intensity of the echo in the current clutter unit with the historical clutter intensity recorded in the memory. However, the spectrum width of the weather clutter is usually wide, and the time variation of the spectrum is strong, so the effect of the above-mentioned clutter suppression methods is not ideal. SUMMARY

[0004] Therefore, the embodiments of the present application propose a weather clutter suppression method based on distance-Doppler two-dimensional features, which can effectively eliminate the low-altitude slow-moving weather clutter, and can ensure effective detection of the unmanned aerial vehicle target while eliminating the false alarm caused by the weather clutter.

[0005] To achieve the above object, embodiments of the present application provide a weather clutter suppression method based on distance-Doppler two-dimensional features, the method comprising the following steps: obtaining coordinates of training samples on a distance-Doppler two-dimensional plane, and taking the coordinates of the training samples as the center, a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size; extracting image features corresponding to the training samples from the two-dimensional matrix corresponding to the training samples; wherein the image features comprise distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features and energy proportion features; inputting the image features corresponding to the training samples into a support vector machine classifier, and iteratively training the support vector machine classifier to convergence based on a classification result of the training samples output by the support vector machine classifier and a label of the training samples; inputting image features corresponding to a detection target into the support vector machine classifier, and obtaining a classification result of the detection target output by the support vector machine classifier; and if the classification result of the detection target output by the support vector machine classifier is weather clutter, the detection target is eliminated.

[0006] To achieve the above object, embodiments of the present application also provide a weather clutter suppression system based on distance-Doppler two-dimensional features, comprising: a positioning module configured to obtain coordinates of training samples on a distance-Doppler two-dimensional plane, and take the coordinates of the training samples as the center, a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size; a feature extraction module configured to extract image features corresponding to the training samples from the two-dimensional matrix corresponding to the training samples, wherein the image features comprise distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features and energy proportion features; a training module configured to input the image features corresponding to the training samples into a support vector machine classifier, and iteratively train the support vector machine classifier to convergence based on a classification result of the training samples output by the support vector machine classifier and a label of the training samples; a use module configured to input image features corresponding to a detection target into the support vector machine classifier, and obtain a classification result of the detection target output by the support vector machine classifier, and in the case that the classification result of the detection target output by the support vector machine classifier is weather clutter, the detection target is eliminated.

[0007] To achieve the above object, embodiments of the present application also provide an electronic device, comprising: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory has instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform a weather clutter suppression method based on distance-Doppler two-dimensional features as described above.

[0008] To achieve the above object, the embodiment of the present application further provides a computer readable storage medium storing a computer program, which can realize the above-mentioned weather clutter suppression method based on distance-Doppler two-dimensional features when executed by a processor.

[0009] The embodiment of the present application provides a weather clutter suppression method based on distance-Doppler two-dimensional features, a support vector machine classifier is constructed and trained to perform a classification task of a detection target, and image features such as distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features and energy proportion features are taken as the basis for classification. The weather clutter has randomness and expansibility, the distance dimension waveform entropy and the Doppler dimension waveform entropy of the weather clutter are both large, the energy distribution does not present symmetry, and the energy is not as concentrated as the real target, so the image features such as the distance dimension waveform entropy features, the Doppler dimension waveform entropy features, the image symmetry features and the energy proportion features can well represent whether the detection target is the weather clutter. In the case that the support vector machine classifier determines that the classification result of the detection target is the weather clutter, the detection target is immediately eliminated, so that the false alarm phenomenon caused by the weather clutter is eliminated while the unmanned aerial vehicle target is effectively detected.

[0010] In some optional embodiments, the coordinates of the training sample on the distance-Doppler two-dimensional plane are (x i ,y i ), and i represents that the training sample is the ith training sample in the training sample set; and the two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane with the coordinates of the training sample as the center and according to a preset size, including: a two-dimensional matrix S is cut out on the distance-Doppler two-dimensional plane with (x i ,y i ) as the center and according to the sizes of the rows and columns being 2M+1 and 2N+1 respectively; wherein the starting point of the X axis is x i -N, the end point of the X axis is x i +N, the starting point of the Y axis is y i -M, and the end point of the Y axis is y i +M, and M and N are both integers greater than 0.

[0011] In some optional embodiments, the distance dimension waveform entropy features and the Doppler dimension waveform entropy features are extracted by the following steps:

[0012] The amplitude maximum value of each column of S is obtained to obtain a one-dimensional vector D, D=[d0, d1…, d 2M ];

[0013] The amplitude maximum value of each row of S is obtained to obtain a one-dimensional vector R, R=[r0, r1…, r 2M ];

[0014] The elements in the one-dimensional vector D are norm-normalized, and based on the waveform entropy calculation formula and the values of the norm-normalized elements, a distance dimension waveform entropy feature E is calculated D ;

[0015] The elements in the one-dimensional vector R are norm-normalized, and based on the waveform entropy calculation formula and the values of the norm-normalized elements, a Doppler dimension waveform entropy feature E is calculated R ;

[0016] The expression of the waveform entropy calculation formula is:

[0017]

[0018] Among them, represents the value of the kth element in the norm-normalized one-dimensional vector D, represents the value of the kth element in the norm-normalized one-dimensional vector R.

[0019] In some optional embodiments, the image symmetry feature is extracted by the following steps:

[0020] Flip S horizontally to obtain a two-dimensional matrix S hc ;

[0021] Flip S vertically to obtain a two-dimensional matrix S vc ;

[0022] Based on the values of each point in S, S hc and S vc , an image symmetry feature is calculated;

[0023] The expression of the formula for calculating the image symmetry feature is:

[0024]

[0025] Among them, S(m,n) represents the value of point (m,n) in S, S hc (m,n) represents the value of point (m,n) in S hc , S vc (m,n) represents the value of point (m,n) in s vc , and T diff represents the calculated image symmetry feature.

[0026] In some optional embodiments, the energy ratio feature is extracted by the following steps:

[0027] The sum value of the values of each point in S is calculated, and based on the value of the center point of S and the sum value, an energy ratio feature is calculated;

[0028] The formula of the energy proportion feature is expressed as:

[0029]

[0030] wherein S(M+1, N+1) represents the value of the center point of S, and CAR represents the calculated energy proportion feature.

[0031] In some optional embodiments, the support vector machine classifier outputs the classification result of the detected target through the following formula:

[0032]

[0033] wherein sgn(·) represents a sign function, σ 2 represents a Gaussian kernel function, g, a and b are model parameters of the support vector machine classifier, x represents the image feature corresponding to the detected target, f(x) = 1 represents that the classification result of the detected target output by the support vector machine classifier is a real target, f(x) = -1 and f(x) = 0 represent that the classification result of the detected target output by the support vector machine classifier is a weather clutter.

[0034] In some optional embodiments, the detected target is rejected, specifically, in the final target report, the detected target with the classification result of weather clutter is rejected. BRIEF DESCRIPTION OF DRAWINGS

[0035] One or more embodiments are exemplarily illustrated by pictures in the drawings corresponding thereto, which do not constitute a limitation on the embodiments.

[0036] Figure 1 is a flowchart of a weather clutter suppression method based on a distance-Doppler two-dimensional feature provided in an embodiment of the present application;

[0037] Figure 2 is a flowchart of classification and discrimination of a detected target provided in an embodiment of the present application;

[0038] Figure 3 is a distance-Doppler plane diagram of a weather clutter and a UAV target provided in an embodiment of the present application;

[0039] Figure 4 is a schematic diagram of multi-frame detection results before and after clutter suppression provided in an embodiment of the present application;

[0040] Figure 5 is a structural schematic diagram of a weather clutter suppression system based on a distance-Doppler two-dimensional feature provided in another embodiment of the present application;

[0041] Figure 6Fig. 1 is a structural schematic diagram of an electronic device according to another embodiment of the present application. DETAILED DESCRIPTION

[0042] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below with reference to the drawings. However, those skilled in the art can understand that, in the embodiments of the present application, many technical details are presented in order to make the readers better understand the present application. However, the technical solutions claimed by the present application can be implemented even without these technical details and based on various changes and modifications of the following embodiments. The division of the following embodiments is for the convenience of description, and should not constitute any limitation on the specific implementation of the present application, and the embodiments can be combined and referenced with each other without contradiction.

[0043] One embodiment of the present application provides a weather clutter suppression method based on distance-Doppler two-dimensional features, applied to an electronic device, wherein the electronic device can be a terminal or a server, and the electronic device is taken as an example of a server in the embodiment and the following embodiments. The implementation details of the weather clutter suppression method based on distance-Doppler two-dimensional features provided in the embodiment will be described in detail below. The following content is only provided for the convenience of understanding the implementation details, and is not necessary for implementing the present solution.

[0044] The specific process of the weather clutter suppression method based on distance-Doppler two-dimensional features provided in the embodiment can be as shown in Figure 1 , including:

[0045] In step 101, the coordinates of the training samples on the distance-Doppler two-dimensional plane are obtained, and a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size with the coordinates of the training samples as the center.

[0046] In a specific implementation, the server first obtains a training sample set, a test sample set and a verification sample set of a support vector machine classifier. For each training sample in the training set, the server needs to obtain the coordinates of the training sample on the distance-Doppler two-dimensional plane, and cut out a two-dimensional matrix on the distance-Doppler two-dimensional plane according to a preset size with the coordinates of the training sample as the center.

[0047] In one example, the server records the coordinates of the training sample on the distance-Doppler two-dimensional plane as (x i ,y i ), and i represents the i-th training sample in the training sample set. The server cuts out a two-dimensional matrix on the distance-Doppler two-dimensional plane according to a preset size with (x i ,y i) is centered, a two-dimensional matrix S is intercepted on the distance-Doppler two-dimensional plane according to the size of the row and column being 2M+1 and 2N+1 respectively. The starting point of the X axis is x i -N, the ending point of the X axis is x i +N, the starting point of the Y axis is y i -M, the ending point of the Y axis is y i +M, and M and N are both integers greater than 0.

[0048] In step 102, the image features corresponding to the training sample are extracted from the two-dimensional matrix corresponding to the training sample, wherein the image features include distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features, and energy proportion features.

[0049] In a specific implementation, after intercepting the two-dimensional matrix corresponding to the training sample, the server can extract the image features corresponding to the training sample from the two-dimensional matrix corresponding to the training sample, wherein the image features include distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features, and energy proportion features.

[0050] In an example, when extracting the distance dimension waveform entropy features and the Doppler dimension waveform entropy features, the server needs to find the amplitude maximum value of each column of the two-dimensional matrix S to obtain a one-dimensional vector D, D=[d0, d1…, d 2N ], and find the amplitude maximum value of each row of the two-dimensional matrix S to obtain a one-dimensional vector R, R=[r0, r1…, r 2M ]. The one-dimensional vector D is a distance dimension vector, and the one-dimensional vector R is a Doppler dimension vector.

[0051] For the one-dimensional vector D, the server performs norm normalization on each element in the one-dimensional vector D, and based on the waveform entropy calculation formula and the value of each element after norm normalization, the distance dimension waveform entropy features E D are calculated.

[0052] For the one-dimensional vector R, the server performs norm normalization on each element in the one-dimensional vector R, and based on the waveform entropy calculation formula and the value of each element after norm normalization, the Doppler dimension waveform entropy features E R are calculated.

[0053] The expression of the waveform entropy calculation formula is:

[0054]

[0055] wherein, represents the value of the kth element in the one-dimensional vector D after norm normalization, represents the value of the kth element in the one-dimensional vector R after norm normalization.

[0056] In one example, the echo of the real target presents a center symmetry in the range-Doppler plane, specifically, the energy distribution of the upper half matrix is similar to that of the lower half matrix, and the energy distribution of the left half matrix is similar to that of the right half matrix. The weather clutter has randomness and expansibility, and its energy distribution does not present such characteristics, so this feature can distinguish the weather clutter from the unmanned aerial vehicle target. When the server extracts the image symmetry feature, the two-dimensional matrix S is horizontally flipped to obtain a two-dimensional matrix S hc , and the two-dimensional matrix S is vertically flipped to obtain a two-dimensional matrix S vc . Then, the image symmetry feature is calculated based on the values (energy) of each point in S, S hc and S vc . The smaller the image symmetry feature is, the stronger the energy distribution center symmetry of the two-dimensional matrix S is.

[0057] The expression of the formula for calculating the image symmetry feature is:

[0058]

[0059] wherein R(m, n) represents the value of the point (m, n) in S, S hc (m, n) represents the value of the point (m, n) in S hc (m, n) represents the value of the point (m, n) in S vc (m, n) represents the value of the point (m, n) in S vc (m, n) represents the value of the point (m, n) in S diff , and T represents the calculated image symmetry feature.

[0060] In one example, the center point position of the two-dimensional matrix S is the strongest value of the target. When the server extracts the energy proportion feature, the sum value of the values of each point in the two-dimensional matrix S is calculated, and the energy proportion feature is calculated based on the value of the center point of the two-dimensional matrix S and the sum value. The larger the value of the energy proportion feature is, the more concentrated the energy is.

[0061] The expression of the formula for calculating the energy proportion feature is:

[0062]

[0063] wherein S(M+1, N+1) represents the value of the center point of S, and CAR represents the calculated energy proportion feature.

[0064] Step 103: input the image features corresponding to the training samples into the support vector machine classifier, and perform iterative training on the support vector machine classifier until convergence based on the classification results of the training samples output by the support vector machine classifier and the labels of the training samples.

[0065] In a specific implementation, after the server extracts the image features corresponding to each training sample in the training sample set, the server can input the image features corresponding to the training sample into the support vector machine classifier, and based on the classification result of the training sample output by the support vector machine classifier and the label of the training sample, iteratively train the support vector machine classifier to convergence to obtain the model parameters of the support vector machine classifier.

[0066] In one example, the server can also use the test sample set to test the model performance of the trained support vector machine classifier, and use the verification sample to verify the related indicators of the trained support vector machine classifier.

[0067] Step 104, input the image features corresponding to the detection target into the support vector machine classifier to obtain the classification result of the detection target output by the support vector machine classifier.

[0068] In a specific implementation, after the construction, training, testing and verification of the support vector machine classifier are completed, the model usage stage, that is, the classification and discrimination of the detection target, can be entered. The specific process of the classification and discrimination of the detection target can be as shown in Figure 2 The server needs to obtain the coordinates of the detection target on the distance-Doppler two-dimensional plane, and centering on the coordinates of the detection target, a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size, and then the image features corresponding to the detection target are extracted from the two-dimensional matrix corresponding to the detection target, including the distance dimension waveform entropy feature, the Doppler dimension waveform entropy feature, the image symmetry feature and the energy proportion feature. After the extraction is completed, the server inputs the image features corresponding to the detection target into the support vector machine classifier to obtain the classification result of the detection target output by the support vector machine classifier.

[0069] In one example, the distance-Doppler plane of the weather clutter and the unmanned aerial vehicle target can be as shown in Figure 3 From Figure 3 It can be clearly seen that the randomness and expansibility of the weather clutter.

[0070] In one example, the support vector machine classifier outputs the classification result of the detection target through the following formula:

[0071]

[0072] Wherein, sgn(·) represents the sign function, σ 2represents a Gaussian kernel function, g, a and b are model parameters of the support vector machine classifier, x represents image features corresponding to the detection target, f(x) = 1 represents that the classification result of the detection target output by the support vector machine classifier is a real target, f(x) = -1, f(x) = 0 represent that the classification result of the detection target output by the support vector machine classifier is a weather clutter.

[0073] In step 105, if the classification result of the detection target output by the support vector machine classifier is a weather clutter, the detection target is eliminated.

[0074] In a specific implementation, if the classification result of the detection target output by the support vector machine classifier is a weather clutter, the server immediately eliminates the detection target, that is, in the final target report, the detection target with the classification result of a weather clutter is eliminated. If the classification result of the detection target output by the support vector machine classifier is a real target, the server saves the detection target and presents it in the final target report.

[0075] In one example, Figure 4 The detection results of multiple frames before and after clutter suppression are shown. It can be seen that the clutter suppression effectively eliminates the false alarm phenomenon caused by weather clutters.

[0076] In this embodiment, the support vector machine classifier is constructed and trained to classify the detection target, and the image features such as the distance dimension waveform entropy feature, the Doppler waveform entropy feature, the image symmetry feature and the energy proportion feature are taken as the basis for classification. The weather clutter has randomness and expansibility, the distance dimension waveform entropy and the Doppler waveform entropy of the weather clutter are both large, the energy distribution does not present symmetry, and the energy is not as concentrated as the real target, so the image features such as the distance dimension waveform entropy feature, the Doppler waveform entropy feature, the image symmetry feature and the energy proportion feature can well represent whether the detection target is a weather clutter or a real target. In the case where the support vector machine classifier determines that the classification result of the detection target is a weather clutter, the detection target is eliminated, so as to ensure effective detection of the unmanned aerial vehicle target while eliminating the false alarm phenomenon caused by weather clutters.

[0077] The step division of the above various methods is only for the purpose of clear description, and in implementation, one step can be combined or some steps can be split and decomposed into multiple steps, as long as the same logical relationship is included, all are within the protection scope of the present application; adding irrelevant modifications or introducing irrelevant designs in the algorithm or flow, but not changing the core design of the algorithm and flow are within the protection scope of the present application.

[0078] Another embodiment of the present application provides a weather clutter suppression system based on distance-Doppler two-dimensional features. The following describes the implementation details of the weather clutter suppression system based on distance-Doppler two-dimensional features. The following implementation details are provided for the convenience of understanding and are not essential to implement the present application. The structure of the weather clutter suppression system based on distance-Doppler two-dimensional features is shown in FIG. 1, which includes: Figure 5

[0079] The positioning module 201 is configured to obtain the coordinates of the training sample on the distance-Doppler two-dimensional plane, and to obtain a two-dimensional matrix by centering on the coordinates of the training sample and cutting out a two-dimensional matrix of a preset size on the distance-Doppler two-dimensional plane.

[0080] The feature extraction module 202 is configured to extract image features corresponding to the training sample from the two-dimensional matrix corresponding to the training sample, wherein the image features include distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features, and energy proportion features.

[0081] The training module 203 is configured to input the image features corresponding to the training sample into a support vector machine classifier, and to perform iterative training on the support vector machine classifier until convergence based on the classification result of the training sample output by the support vector machine classifier and the label of the training sample.

[0082] The using module 204 is configured to input the image features corresponding to the detection target into the support vector machine classifier, to obtain the classification result of the detection target output by the support vector machine classifier, and to eliminate the detection target in the case where the classification result of the detection target output by the support vector machine classifier is weather clutter.

[0083] It is worth mentioning that each module involved in the present embodiment is a logical module. In actual application, one logical unit can be one physical unit, or a part of one physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of the present application, some units that are not closely related to solving the technical problems proposed in the present application are not introduced in the present embodiment, but this does not mean that there are no other units in the present embodiment.

[0084] It can be found that the present embodiment is a system embodiment corresponding to the above-mentioned method embodiment. The present embodiment can be implemented in cooperation with the above-mentioned method embodiment. The related technical details and technical effects mentioned in the above-mentioned embodiments are still valid in the present embodiment. In order to reduce repetition, they will not be described here. Accordingly, the related technical details mentioned in the present embodiment can also be applied in the above-mentioned embodiments.

[0085] Another embodiment of the present application provides an electronic device, which can have a specific structure as shown in FIG. 2.​Figure 6 As shown, the device includes at least one processor 301, and a memory 302 connected with the at least one processor 301; wherein the memory 302 stores instructions executable by the at least one processor 301, and the instructions are executed by the at least one processor 301 to enable the at least one processor 301 to perform the method of any one of the above method embodiments.

[0086] The bus connection between the memory and the processor can include any number of interconnecting buses and bridges, the bus connecting various circuits of the one or more processors and the memory together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, all of which are well known in the art, and therefore, not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as a plurality of receivers and transmitters, providing a means of communicating with various other apparatus over a transmission medium. Data processed by the processor is transmitted over the wireless medium via the antenna, and further, the antenna receives data and transmits the data to the processor.

[0087] The processor is responsible for managing the bus and general processing, and can also provide various functions including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory can be used to store data used by the processor during execution of operations.

[0088] Another embodiment of the present application relates to a computer readable storage medium, storing a computer program, wherein the computer program is executed by a processor to implement the method of any one of the above method embodiments.

[0089] That is, those skilled in the art can understand that all or part of the steps of the above-mentioned embodiment methods can be completed by programs instructing relevant hardware, the programs are stored in a storage medium, and include a plurality of instructions for enabling a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various storage medium capable of storing program codes.

[0090] Those skilled in the art can understand that the above-mentioned embodiments are specific embodiments for implementing the present application, and various changes can be made in form and details in actual application without departing from the spirit and scope of the present application.

Claims

1. A method of weather clutter suppression based on range-Doppler two-dimensional features, characterized in that, The method comprises the following steps: obtaining the coordinates of the training sample on the distance-Doppler two-dimensional plane, and taking the coordinates of the training sample as the center, a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size; extracting the image features corresponding to the training sample from the two-dimensional matrix corresponding to the training sample; wherein the image features comprise distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features and energy proportion features; inputting the image features corresponding to the training sample into a support vector machine classifier, and iteratively training the support vector machine classifier to convergence based on the classification result of the training sample output by the support vector machine classifier and the label of the training sample; inputting the image features corresponding to the detection target into the support vector machine classifier, and obtaining the classification result of the detection target output by the support vector machine classifier; if the classification result of the detection target output by the support vector machine classifier is weather clutter, the detection target is removed.

2. The method of claim 1, wherein the method is characterized by, The coordinates of the training sample on the distance-Doppler two-dimensional plane are (x i ,y i ), i represents that the training sample is the ith training sample in the training sample set; The two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to the preset size, taking the coordinates of the training sample as the center, comprising: with (x i ,y i ) as the center, a two-dimensional matrix S is intercepted on a distance-Doppler two-dimensional plane according to the size of rows and columns being 2M+1 and 2N+1 respectively; wherein, the starting point of X axis is x i -N, the ending point of X axis is x i +N, the starting point of Y axis is y i -M, the ending point of Y axis is y i +M, M and N are both integers greater than 0.

3. The method of claim 2, wherein the method is characterized by, The distance dimension waveform entropy features and the Doppler dimension waveform entropy features are extracted by the following steps: Find the maximum amplitude for each column of S to obtain a one-dimensional vector D, D = [d0, d1, ..., d2]. 2N ]; For each row of S, find the amplitude maximum, resulting in a one-dimensional vector R, R = [r0, r1…, r 2M ] The elements in the one-dimensional vector D are norm-normalized, and based on the waveform entropy calculation formula and the values of the norm-normalized elements, a distance dimension waveform entropy feature E is calculated D ; The elements in the one-dimensional vector R are norm-normalized, and based on the waveform entropy calculation formula and the values of the norm-normalized elements, Doppler wave form entropy features E are calculated R ; The expression of the waveform entropy calculation formula is: wherein, represents the value of the kth element in the norm-normalized one-dimensional vector D, represents the value of the kth element in the norm-normalized one-dimensional vector R.

4. The method of claim 2, wherein the method is characterized by, The image symmetry features are extracted by the following steps: S is horizontally flipped to obtain a two-dimensional matrix S hc ; S is vertically flipped to obtain a two-dimensional matrix S vc ; Based on S, S hc and S vc The image symmetry feature is calculated based on the values of the points. The expression of the formula for calculating the image symmetry features is: where S(m, n) represents the value of the S midpoint (m, n), S hc (m, n) represents the value of the S hc midpoint (m, n), S vc (m, n) represents the value of the S vc midpoint (m, n), T diff represents the calculated image symmetry feature.

5. The method of claim 2, wherein the method is characterized by, The energy proportion features are extracted by the following steps: The sum value of the values of each point in S is calculated, and the energy proportion features are calculated based on the value of the center point of S and the sum value; The expression of the formula for calculating the energy proportion features is: Wherein, S(M+1, N+1) represents the value of the center point of S, and CAR represents the calculated energy proportion features.

6. The method of claim 1 to 5, wherein, The support vector machine classifier outputs the classification result of the detection target by the following formula: where sgn(·) denotes the sign function, σ 2 denotes a Gaussian kernel function, g, a and b are model parameters of the support vector machine classifier, x denotes image features corresponding to the detection target, f(x) = 1 indicates that the classification result of the detection target output by the support vector machine classifier is a real target, f(x) = -1 and f(x) = 0 indicate that the classification result of the detection target output by the support vector machine classifier is a weather clutter.

7. The method of claim 1 to 5, wherein, The detection target is removed, specifically: in the final target report, the detection target with the class result of weather clutter is removed.

8. A range-Doppler two-dimensional feature based weather clutter mitigation system, characterized by, The method comprises the following steps: a positioning module is configured to obtain the coordinates of the training sample on the distance-Doppler two-dimensional plane, and take the coordinates of the training sample as the center, a two-dimensional matrix is cut out on the distance-Doppler two-dimensional plane according to a preset size; a feature extraction module is configured to extract the image features corresponding to the training sample from the two-dimensional matrix corresponding to the training sample; wherein the image features comprise distance dimension waveform entropy features, Doppler dimension waveform entropy features, image symmetry features and energy proportion features; a training module is configured to input the image features corresponding to the training sample into a support vector machine classifier, and iteratively train the support vector machine classifier to convergence based on the classification result of the training sample output by the support vector machine classifier and the label of the training sample; a use module is configured to input the image features corresponding to the detection target into the support vector machine classifier, and obtain the classification result of the detection target output by the support vector machine classifier, and remove the detection target if the classification result of the detection target output by the support vector machine classifier is weather clutter.

9. An electronic device, comprising: The method comprises the following steps: at least one processor; and a memory connected in communication with the at least one processor; wherein The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the weather clutter suppression method based on range-Doppler two-dimensional features according to any one of claims 1 to 7.

10. A computer readable storage medium storing a computer program, characterized in that, The computer program, when executed by a processor, implements the weather clutter suppression method based on range-Doppler two-dimensional features according to any one of claims 1 to 7.

Images