SAR working mode recognition method, device, equipment and medium

Abstract

The present invention provides a SAR working mode recognition method, device, equipment and medium. The method includes: obtaining a signal pulse string group of a spaceborne SAR, the signal pulse string group includes at least one signal pulse string; performing data preprocessing on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group; extracting the intensity variation characteristics along the azimuth time axis from the azimuth time-frequency diagram sequence, the intensity variation characteristics characterizing the receiving power intensity variation characteristics of the SAR working mode; extracting the Doppler time-frequency curve motion characteristics from the azimuth time-frequency diagram sequence, the Doppler time-frequency curve motion characteristics characterizing the Doppler course motion characteristics of the SAR working mode; determining the working mode of the spaceborne SAR according to at least one of the intensity variation characteristics and the Doppler time-frequency curve motion characteristics.

Description

Technical Field

[0001] The invention belongs to the technical field of radar electronic countermeasures, and in particular relates to a SAR working mode recognition method, device, equipment and medium. Background Art

[0002] Synthetic Aperture Radar (SAR) can break through the limitations of visible light and infrared, achieve all-day, long-distance and wide-band imaging, and play an important role in obtaining information such as terrain conditions and resource distribution. Therefore, obtaining parameter information of non-cooperative SAR systems through reconnaissance means and taking targeted interference measures to reduce the application efficiency of their SAR systems has become a research hotspot in the field of radar electronic countermeasures. The reconnaissance parameter information includes the range parameters, azimuth parameters and imaging working mode of the SAR signal. Among them, the key areas of SAR ground observation under different imaging working modes are different, and the imaging algorithms and interference measures used are also different. Therefore, using the SAR signals intercepted by reconnaissance aircraft to identify the imaging working mode is of great significance for the targeted selection of interference strategies to protect specific targets. Summary of the invention

[0003] In view of this, the present invention provides a SAR working mode recognition method, apparatus, device and medium.

[0004] The first aspect of the present invention provides a SAR working mode recognition method, comprising: obtaining a signal pulse string group of a space-borne SAR, the signal pulse string group including at least one signal pulse string; performing data preprocessing on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group; extracting an intensity variation feature along the azimuth time axis from the azimuth time-frequency diagram sequence, the intensity variation feature characterizing the received power intensity variation feature of the SAR working mode; extracting a Doppler time-frequency curve motion feature from the azimuth time-frequency diagram sequence, the Doppler time-frequency curve motion feature characterizing the Doppler course motion feature of the SAR working mode; and determining the working mode of the space-borne SAR according to at least one of the intensity variation feature and the Doppler time-frequency curve motion feature.

[0005] According to an embodiment of the present invention, obtaining a signal pulse string group of a space-borne SAR includes: obtaining a signal pulse string received by at least one receiver placed at a different azimuth position and the same distance position relative to the space-borne SAR; arranging each signal pulse string according to the start time sequence of each receiver receiving the signal pulse string to obtain a signal pulse string group.

[0006] According to an embodiment of the present invention, data preprocessing is performed on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group, including: rearranging each signal pulse string using a pulse repetition time to obtain a two-dimensional receiving data matrix; performing range pulse compression on the two-dimensional receiving data matrix using a frequency modulation rate, extracting the signal at each range signal peak position in the two-dimensional receiving data matrix to obtain an azimuth signal; performing a short-time Fourier transform on the azimuth signal and extracting the amplitude to obtain an azimuth time-frequency diagram corresponding to each signal pulse string, wherein each azimuth time-frequency diagram constitutes an azimuth time-frequency diagram sequence.

[0007] According to an embodiment of the present invention, extracting intensity change features along the azimuth time axis from the azimuth time-frequency graph sequence includes: performing highly oriented maximum pooling, feature reshaping, convolution, batch normalization, and activation processing on the first feature of the azimuth time-frequency graph sequence in sequence to generate an axial attention weight; performing weighted processing on the first feature through the axial attention weight, and then obtaining the intensity change feature after feature reshaping.

[0008] According to an embodiment of the present invention, extracting the Doppler time-frequency curve motion feature from the azimuth time-frequency diagram sequence includes: performing feature reshaping and slicing processing on the second feature of the azimuth time-frequency diagram sequence to obtain feature slices of different dimensions; performing feature connection, convolution, batch normalization and activation processing on the difference results between the feature slices, and then obtaining the Doppler time-frequency curve motion feature after feature reshaping.

[0009] According to an embodiment of the present invention, determining the operating mode of the spaceborne SAR based on at least one of the intensity change characteristics and the Doppler time-frequency curve motion characteristics includes: utilizing at least one of the intensity change characteristics and the Doppler time-frequency curve motion characteristics to enhance the feature extraction capability of the network, and determining the operating mode of the spaceborne SAR.

[0010] According to an embodiment of the present invention, the operating modes of the spaceborne SAR include a strip mode, a scanning mode, a spotlight mode, a sliding spotlight mode, a sliding scanning mode and a staggered mode.

[0011] The second aspect of the present invention provides a SAR working mode recognition device, including: an acquisition module, used to acquire a signal pulse string group of a space-borne SAR, the signal pulse string group including at least one signal pulse string; a preprocessing module, used to perform data preprocessing on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group; a first extraction module, used to extract intensity change characteristics along the azimuth time axis from the azimuth time-frequency diagram sequence, the intensity change characteristics characterizing the received power intensity change characteristics of the SAR working mode; a second extraction module, used to extract Doppler time-frequency curve motion characteristics from the azimuth time-frequency diagram sequence, the Doppler time-frequency curve motion characteristics characterizing the Doppler course motion characteristics of the SAR working mode; an identification module, used to determine the working mode of the space-borne SAR according to at least one of the intensity change characteristics and the Doppler time-frequency curve motion characteristics.

[0012] A third aspect of the present invention provides an electronic device, comprising: one or more processors; a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the above method.

[0013] The fourth aspect of the present invention further provides a computer-readable storage medium having a computer program or instructions stored thereon, which implements the steps of the above method when the computer program or instructions are executed by a processor.

[0014] According to the SAR working mode recognition method, device, equipment and medium of the embodiments of the present invention, by fully exploring the differences between different SAR working modes, the intensity change characteristics and the Doppler time-frequency curve motion characteristics are extracted from the azimuth time-frequency diagram sequence having both the receiving power intensity change characteristics and the Doppler process motion characteristics, so as to realize the recognition of multiple working modes, improve the recognition accuracy, and expand the application scope of the working mode recognition algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0016] Figure 1 A flowchart of a SAR working mode recognition method according to an embodiment of the present invention is schematically shown;

[0017] Figure 2 Schematically shows a layout diagram of multiple receivers according to an embodiment of the present invention;

[0018] Figure 3 Schematically shows a data preprocessing flow chart according to an embodiment of the present invention;

[0019] Figure 4A SAR working mode recognition model according to an embodiment of the present invention is schematically shown;

[0020] Figure 5 Schematically shows a flow chart of extracting intensity variation features according to an embodiment of the present invention;

[0021] Figure 6 Schematically shows a flow chart of extracting motion features of a Doppler time-frequency curve according to an embodiment of the present invention;

[0022] Figure 7 Schematically shows a flow chart of generating a SAR working mode recognition model according to an embodiment of the present invention;

[0023] Figure 8 A structural block diagram of a SAR working mode recognition device according to an embodiment of the present invention is schematically shown;

[0024] Fig. 9 A block diagram of an electronic device suitable for implementing a SAR working mode recognition method according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0025] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0026] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0027] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0028] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0029] Current SAR working mode recognition methods can be divided into two categories: traditional methods and neural network-based methods. Traditional SAR working mode recognition methods usually manually select the features to be extracted for working mode recognition. The main extracted features include received power morphology, beam ground scanning speed, and Doppler history parameters. The neural network-based method mainly uses the amplitude peak sequence of the received signal as input, and automatically extracts features through the neural network for SAR working mode recognition.

[0030] Although some progress has been made in SAR working mode recognition methods, there are still the following problems:

[0031] (1) Traditional methods can achieve working mode recognition, but the feature extraction results are easily affected by noise, the amount of calculation is large, the real-time performance is poor, and the application scope in actual scenarios is limited;

[0032] (2) Although the neural network-based method improves the speed and accuracy of SAR working mode recognition, it can only recognize traditional SAR working modes and has limited applicability to new SAR working modes;

[0033] (3) Existing neural network-based algorithms basically only use pulse amplitude information for SAR working mode recognition, and the mining of working mode information contained in the received signal is limited;

[0034] (4) The neural network structure used in existing methods is basically directly borrowed from the field of artificial intelligence and is not designed in combination with the characteristics of SAR signals.

[0035] Based on this, the present invention provides a SAR working mode recognition method, which fully explores the differences between different SAR working modes, extracts the intensity change characteristics and the Doppler time-frequency curve motion characteristics from the azimuth time-frequency diagram sequence that has both the received power intensity change characteristics and the Doppler history motion characteristics, and uses the classification neural network model to realize rapid and accurate recognition of multiple working modes. The SAR working mode recognition method will be described below in conjunction with the accompanying drawings.

[0036] Figure 1 The flowchart of the SAR working mode recognition method according to an embodiment of the present invention is schematically shown.

[0037] like Figure 1 As shown, the SAR working mode recognition method may include operations S110 to S150:

[0038] In operation S110, a signal pulse string group of a space-borne SAR is acquired, where the signal pulse string group includes at least one signal pulse string.

[0039] In operation S120, data preprocessing is performed on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group.

[0040] In operation S130, an intensity variation feature along the azimuth time axis is extracted from the azimuth time-frequency diagram sequence, where the intensity variation feature represents a received power intensity variation characteristic of the SAR working mode.

[0041] In operation S140, a Doppler time-frequency curve motion feature is extracted from the azimuth time-frequency graph sequence, where the Doppler time-frequency curve motion feature represents the Doppler history motion characteristic of the SAR working mode.

[0042] In operation S150, the operating mode of the spaceborne SAR is determined according to at least one of the intensity variation characteristic and the Doppler time-frequency curve motion characteristic.

[0043] It should be emphasized that operation S140 can be performed after operation S130, or before operation S130. That is, the intensity variation feature can be extracted first, and then the Doppler time-frequency curve motion feature can be extracted; or the Doppler time-frequency curve motion feature can be extracted first, and then the intensity variation feature can be extracted. Both can realize the recognition of the working mode of the spaceborne SAR.

[0044] The working mode of the spaceborne SAR may be, for example, a strip mode, a scanning mode, a beam mode, a sliding beam mode, a sliding scanning (Terrain Observation by Progressive Scans, TOPS) mode or a staggered mode. Different working modes have different ways of beam rotation and scanning, which will cause different changes in the intensity of the receiver's received power. The intensity change of the azimuth time-frequency diagram along the azimuth time axis will reflect the change in the intensity of the received power. This embodiment can better extract the characteristics of the change in the received power intensity by performing intensity enhancement processing on the features, so as to enhance the recognition effect of the SAR working mode. In addition to the intensity change characteristics, the relative changes in the Doppler frequency history of the received signal under different working modes are also different, and this relative change can be reflected by the movement of the Doppler time-frequency curve in the azimuth time-frequency diagram sequence. This embodiment can improve the accuracy of SAR working mode recognition by extracting the movement of the Doppler history.

[0045] This embodiment can realize the automatic extraction of intensity change characteristics and Doppler time-frequency curve motion characteristics by inserting different extraction modules into the neural network model used for classification, and after training, testing and verifying the neural network model, the neural network model can realize the rapid and accurate identification of the SAR working mode according to the extracted intensity change characteristics and Doppler time-frequency curve motion characteristics. The construction, training, testing and verification of the neural network model will be further described later. It should be emphasized that in this embodiment, the working mode of the spaceborne SAR can be identified according to the intensity change characteristics of the extracted azimuth time-frequency diagram sequence or one of the Doppler time-frequency curve motion characteristics, or the working mode of the spaceborne SAR can be identified by combining the intensity change characteristics and the Doppler time-frequency curve motion characteristics at the same time. According to a feature or a combination of features, the calculation amount and recognition accuracy of the neural network model will be affected accordingly, and different extraction modules can be reasonably selected according to actual needs.

[0046] Figure 2 The diagram schematically shows a layout of multiple receivers according to an embodiment of the present invention.

[0047] Based on the above embodiment, the above operation S110 may include: obtaining a signal pulse string received by at least one receiver placed at a different azimuth position and the same distance position relative to the satellite-borne SAR; arranging each signal pulse string according to the start time sequence of the signal pulse string received by each receiver to obtain a signal pulse string group.

[0048] By using the known flight direction of the SAR payload in advance and combining the different characteristics of the Doppler history changes of receivers at different azimuth positions and the same range positions under different working modes, this embodiment proposes a solution for placing multiple receivers along the azimuth direction, which can be as follows: Figure 2 In this embodiment, the number of receivers placed is T. The number T can be 1, 2 or any other positive integer, and the present invention does not specifically limit the number T.

[0049] like Figure 2 As shown, when the satellite-borne SAR passes by, the beam scans each receiver in turn, and the moment when the receiver first receives the SAR signal is the start moment of the signal reception, and the moment when all receivers cannot receive the SAR signal is the end moment of the signal reception. All receivers can receive signals independently without affecting each other. After the signal reception is completed, the signal pulse strings received by T receivers can be arranged in the time sequence of the first reception of the SAR signal to obtain a received signal pulse string group with a length of T. Then, operation S120 can be performed.

[0050] Figure 3 A data preprocessing flow chart according to an embodiment of the present invention is schematically shown.

[0051] Based on the above embodiment, the above operation S120 may include: rearranging each signal pulse string using the pulse repetition time to obtain a two-dimensional receiving data matrix; performing range pulse compression on the two-dimensional receiving data matrix using the frequency modulation rate, extracting the signal at each range signal peak position in the two-dimensional receiving data matrix to obtain an azimuth signal; performing short-time Fourier transform on the azimuth signal and extracting the amplitude to obtain an azimuth time-frequency diagram corresponding to each signal pulse string, and each azimuth time-frequency diagram constitutes an azimuth time-frequency diagram sequence.

[0052] By performing data preprocessing on each pulse string in the received signal pulse string group, a azimuth time-frequency diagram sequence can be obtained. The processing flow of a single pulse string can be as follows: Figure 3 As shown. First, the signal pulse train received by each receiver is rearranged into a two-dimensional receiving data matrix using the estimated pulse repetition time, and the two-dimensional receiving data matrix is ​​subjected to range pulse compression using the estimated frequency modulation, and then the signal at each range signal peak position is extracted to obtain the azimuth signal, and finally the azimuth signal is subjected to short-time Fourier transform and the amplitude is extracted to obtain the azimuth time-frequency diagram. T azimuth time-frequency diagrams constitute an azimuth time-frequency diagram sequence with a length of T. Then, operations S130 to S150 can be performed.

[0053] In operation S130 to operation S150, the intensity variation characteristics and the Doppler time-frequency curve motion characteristics are extracted from the azimuth time-frequency diagram sequence, and the SAR working mode is identified based on these characteristics, for example, based on a neural network model for classification. The neural network model can be, for example, a model using 2D ResNet50 as the backbone network, or other neural networks for classification such as RseNet18, ResNet34, VGG, etc. The present invention does not specifically limit the model used, as long as it has the ability to classify according to the extracted image features. In the neural network model used, different modules can be inserted to realize the extraction of intensity variation characteristics and Doppler time-frequency curve motion characteristics. The position where the module is inserted can be reasonably set as needed, and the present invention does not specifically limit it.

[0054] As an example, the embodiment of the present invention uses 2D ResNet50 as the backbone network for illustration, and designs an axial attention mechanism module and a motion feature extraction module according to the characteristics of the azimuth time-frequency diagram sequence and inserts them into 2D RseNet50 to form a SAR working mode recognition model. The axial attention mechanism module can be used to extract intensity change features, and the motion feature extraction module can be used to extract Doppler time-frequency curve motion features. Figure 4 The SAR working mode recognition model according to an embodiment of the present invention is schematically shown.

[0055] like Figure 4As shown in the figure, the SAR working mode recognition model extracts and enhances features of the azimuth time-frequency diagram sequence with a length of T read through multiple convolution operations, axial attention modules, motion feature extraction modules and fully connected layers. The final output feature results are averaged along the dimension of the number of receivers T to obtain the final working mode recognition result output.

[0056] In the SAR working mode recognition model, the convolution kernel size of Conv2D used in the axial attention mechanism and motion feature extraction module can be adjusted as long as the width and height of the feature map before and after the convolution remain unchanged. The activation functions used in the axial attention mechanism and motion feature extraction module can be tanh, ReLU and other activation functions commonly used in the field of deep learning.

[0057] The insertion positions of the axial attention mechanism module and the motion feature extraction module include but are not limited to Figure 4 Inserting the module into other locations of the network can also form a network model for SAR working mode recognition. Figure 4 The axial attention module and / or motion feature extraction module are inserted into the modules 1, 2 and 3 shown. One or more axial attention modules and / or motion feature extraction modules can also be inserted into any position of conv2_x, conv3_x, conv4_x and conv5_x in different stages of the neural network. The axial attention module can be processed before the motion feature extraction module or after the motion feature extraction module. It can be understood that the more axial attention modules and motion feature extraction modules there are, the higher the accuracy of SAR working mode recognition will be, but correspondingly, the recognition efficiency may be affected. Therefore, the number and position of the axial attention modules and motion feature extraction modules can be reasonably set according to actual needs.

[0058] Figure 5 The flowchart for extracting intensity variation features according to an embodiment of the present invention is schematically shown.

[0059] Based on the above embodiment, the above operation S130 may include: performing high-level maximum pooling, feature reshaping, convolution, batch normalization, and activation processing on the first feature of the azimuth time-frequency graph sequence in turn to generate an axial attention weight; performing weighted processing on the first feature through the axial attention weight, and then obtaining a feature with a stronger ability to express the received power change after feature reshaping. In some embodiments, it can be achieved by Figure 4 The axial attention module in implements the specific steps of operation S130.

[0060] The computational flow of the axial attention module can be as follows Figure 5As shown in Figure 2, the axial attention module can first compress the height information through a large pooling operation in the height direction, and highlight the feature information along the width direction, that is, the azimuth time axis. Suppose the input feature of the axial attention module is , where B represents the batch size, T represents the number of receivers placed, C represents the number of channels, and H and W represent the width and height of the feature map, respectively. The calculation formula for the height-wise maximum pooling operation is:

[0061]

[0062] Then, axial attention weights are generated through feature reshaping, convolution, batch normalization, activation, and sigmoid operations. Finally, the weights are used to weight the original features to obtain features that emphasize axial intensity changes.

[0063] Figure 6 The flowchart for extracting motion features of a Doppler time-frequency curve according to an embodiment of the present invention is schematically shown.

[0064] Based on the above embodiment, the above operation S140 may include: reshaping and slicing the second feature of the azimuth time-frequency diagram sequence to obtain feature slices of different dimensions; performing feature connection, convolution, batch normalization and activation processing on the difference results between the feature slices, and then obtaining the Doppler time-frequency curve motion feature after feature reshaping. In some embodiments, it can be achieved by Figure 4 The motion feature extraction module in implements the specific steps of operation S140.

[0065] The calculation process of the motion feature extraction module can be as follows: Figure 6 As shown. Assume that the input feature of the motion feature extraction module is , where B represents the batch size, T represents the number of receivers placed, C represents the number of channels, and H and W represent the width and height of the feature map, respectively. First, Y can be reshaped into a dimension of Then slice Y' along the second dimension T to get T dimensions. The feature slices of , subtract the first feature slice from T feature slices in turn and stack the difference results to get a dimension of The motion feature M' is reshaped into a dimension of The tensor is then transformed into a tensor and the motion feature expression capability is enhanced through convolution, batch normalization and activation operations. Finally, the motion feature is obtained through feature reshaping. .

[0066] According to the intensity variation characteristics and Doppler time-frequency curve motion characteristics extracted in the above embodiment, Figure 4In some embodiments, determining the working mode of the spaceborne SAR according to at least one of the intensity variation characteristics of the azimuth time-frequency diagram sequence and the motion characteristics of the Doppler time-frequency curve may include: enhancing the feature extraction capability of the network by using at least one of the intensity variation characteristics and the motion characteristics of the Doppler time-frequency curve to determine the working mode of the spaceborne SAR.

[0067] It can be understood that by pre-training the neural network model, the neural network model can be equipped with the ability to extract intensity change characteristics and Doppler time-frequency curve motion characteristics.

[0068] Figure 7 The flowchart of generating a SAR working mode recognition model according to an embodiment of the present invention is schematically shown.

[0069] like Figure 7 As shown, the process of generating a SAR working mode recognition model may include operations S710 to S740.

[0070] In operation S710, a SAR working mode recognition data set is generated.

[0071] The characteristics of different SAR working modes can be analyzed, the placement of multiple receivers can be set, the SAR pulse group data received by multiple receivers can be obtained, and the data preprocessing program can be input to obtain the azimuth time-frequency diagram sequence. The azimuth time-frequency diagram sequence can be obtained by executing operations S110 to S120 above, which will not be repeated here. Next, the characteristic parameters under different working modes are set. The list of characteristic parameters selected in this example is shown in Table 1. The simulation signal and preprocessing can obtain the azimuth time-frequency diagram sequence data sets under six working modes, namely strip mode, scanning mode, beam mode, sliding beam mode, TOPS mode and Staggered mode. The data sets are randomly shuffled and divided into training set, verification set and test set according to the ratio of 70%, 15% and 15%.

[0072] Table 1

[0073]

[0074] In operation S720, a SAR working mode recognition model is constructed.

[0075] The constructed SAR working mode recognition model can be, for example, Figure 4 As shown, no further description is given here.

[0076] In operation S730, the SAR working mode recognition model is trained and verified using the training set and the verification set.

[0077] As an example, parameters such as batch size, number of training rounds, optimizer type, learning rate, and initialization type can be set during the training process. The training set data generated in step S710 is read and input into the model constructed in step S720 for training. The loss function used in the training process can be a cross entropy loss function:

[0078]

[0079] Where N represents the batch size, C represents the total number of working mode categories, represents the true label, Represents the predicted label.

[0080] The model is then trained iteratively and the validation set is used to verify the performance of the current model until the model converges. After the model training is completed, the model weights with the highest recognition accuracy on the validation set are saved.

[0081] In operation S740, a test set is used to perform a performance test on the SAR working mode recognition model.

[0082] The saved model weights are loaded, and the test set data generated in step S710 can be input into the model. The SAR working mode recognition accuracy is calculated according to the predicted category and the actual category output by the model, thereby completing the test of the model performance.

[0083] According to the above-mentioned multiple embodiments, the present invention suppresses noise by compressing the distance of the received signal, and at the same time uses a neural network model with strong generalization ability to extract features, which can solve the problem that the feature extraction result is greatly affected by noise. At the same time, once the model is pre-trained, when the received data is input into the pre-trained network in actual applications, the weights loaded and saved by the model can quickly identify the current reconnaissance SAR system working mode, which has strong real-time performance and a wide range of applications in actual scenarios; the difference between different SAR working modes mainly lies in the difference in the change of received power intensity and the Doppler process motion. The present invention fully explores the difference between different working modes, uses the azimuth time-frequency diagram sequence with both the received power intensity change characteristics and the Doppler process motion characteristics to identify the SAR working mode, and can realize the recognition of multiple working modes, thereby expanding the scope of application of the working mode recognition algorithm; while drawing on the 2DResNet network in the field of artificial intelligence, the present invention designs an axial attention module and a motion feature extraction module according to the characteristics of the azimuth time-frequency diagram sequence obtained from the received SAR signal, so that the network is more suitable for the SAR working mode recognition task.

[0084] Based on the above SAR working mode recognition method, the present invention also provides a SAR working mode recognition device. Figure 8 The device is described in detail.

[0085] Figure 8 The structure block diagram of the SAR working mode recognition device according to an embodiment of the present invention is schematically shown.

[0086] like Figure 8 As shown, the SAR working mode recognition device 800 of this embodiment includes an acquisition module 810 , a preprocessing module 820 , a first extraction module 830 , a second extraction module 840 and a recognition module 850 .

[0087] The acquisition module 810 is used to acquire a signal pulse string group of the spaceborne SAR, where the signal pulse string group includes at least one signal pulse string. In one embodiment, the acquisition module 810 can be used to perform the operation S110 described above, which will not be described in detail here.

[0088] The preprocessing module 820 is used to perform data preprocessing on each signal pulse string to obtain an azimuth time-frequency diagram sequence of the signal pulse string group. In one embodiment, the preprocessing module 820 can be used to perform the operation S120 described above, which will not be described in detail here.

[0089] The first extraction module 830 is used to extract the intensity variation feature along the azimuth time axis from the azimuth time-frequency diagram sequence, and the intensity variation feature represents the received power intensity variation characteristic of the SAR working mode. In one embodiment, the first extraction module 830 can be used to perform the operation S130 described above, which will not be repeated here.

[0090] The second extraction module 840 is used to extract the Doppler time-frequency curve motion features from the azimuth time-frequency diagram sequence, and the Doppler time-frequency curve motion features represent the Doppler course motion characteristics of the SAR working mode. In one embodiment, the second extraction module 840 can be used to perform the operation S140 described above, which will not be repeated here.

[0091] The identification module 850 is used to determine the working mode of the spaceborne SAR according to at least one of the intensity variation characteristics of the azimuth time-frequency diagram sequence and the motion characteristics of the Doppler time-frequency curve. In one embodiment, the identification module 850 can be used to perform the operation S150 described above, which will not be repeated here.

[0092] According to an embodiment of the present invention, any multiple modules among the acquisition module 810, the preprocessing module 820, the first extraction module 830, the second extraction module 840 and the identification module 850 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present invention, at least one of the acquisition module 810, the preprocessing module 820, the first extraction module 830, the second extraction module 840 and the identification module 850 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware and firmware or by a suitable combination of any of them. Alternatively, at least one of the acquisition module 810, the preprocessing module 820, the first extraction module 830, the second extraction module 840 and the identification module 850 may be at least partially implemented as a computer program module, which may perform a corresponding function when executed.

[0093] Fig. 9 A block diagram of an electronic device suitable for implementing a SAR working mode recognition method according to an embodiment of the present invention is schematically shown.

[0094] like Fig. 9 As shown, the electronic device 900 according to an embodiment of the present invention includes a processor 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage part 908 to a random access memory (RAM) 903. The processor 901 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 901 may also include an onboard memory for caching purposes. The processor 901 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0095] In RAM 903, various programs and data required for the operation of electronic device 900 are stored. Processor 901, ROM 902 and RAM 903 are connected to each other via bus 904. Processor 901 performs various operations of the method flow according to the embodiment of the present invention by executing the program in ROM 902 and / or RAM 903. It should be noted that the program can also be stored in one or more memories other than ROM 902 and RAM 903. Processor 901 can also perform various operations of the method flow according to the embodiment of the present invention by executing the program stored in one or more memories.

[0096] According to an embodiment of the present invention, the electronic device 900 may further include an input / output (I / O) interface 905, which is also connected to the bus 904. The electronic device 900 may further include one or more of the following components connected to the input / output (I / O) interface 905: an input portion 906 including a keyboard, a mouse, etc.; an output portion 907 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 908 including a hard disk, etc.; and a communication portion 909 including a network interface card such as a LAN card, a modem, etc. The communication portion 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the input / output (I / O) interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as needed, so that a computer program read therefrom is installed into the storage portion 908 as needed.

[0097] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0098] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 902 and / or RAM 903 described above and / or one or more memories other than ROM 902 and RAM 903.

[0099] The embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the SAR working mode recognition method provided by the embodiment of the present invention.

[0100] The computer program executes the above functions defined in the system / device of the embodiment of the present invention when it is executed by the processor 901. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0101] In one embodiment, the computer program may be based on a tangible storage medium such as an optical storage device, a magnetic storage device, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 909, and / or installed from a removable medium 911. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0102] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 909, and / or installed from the removable medium 911. When the computer program is executed by the processor 901, the above functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0103] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0104] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0105] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present invention may be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention may be combined and / or combined in various ways. All of these combinations and / or combinations fall within the scope of the present invention.

[0106] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A SAR working mode recognition method, characterized in that: include: Acquire a signal pulse train group of a space-borne SAR, wherein the signal pulse train group includes at least one signal pulse train; Performing data preprocessing on each of the signal pulse trains to obtain an azimuth time-frequency diagram sequence of the signal pulse train group; performing data preprocessing on each of the signal pulse trains to obtain an azimuth time-frequency diagram sequence of the signal pulse train group includes: rearranging each of the signal pulse trains using pulse repetition time to obtain a two-dimensional receiving data matrix; performing range pulse compression on the two-dimensional receiving data matrix using frequency modulation, extracting the signal at each range signal peak position in the two-dimensional receiving data matrix to obtain an azimuth signal; performing short-time Fourier transform on the azimuth signal and extracting the amplitude to obtain an azimuth time-frequency diagram corresponding to each of the signal pulse trains, wherein each of the azimuth time-frequency diagrams constitutes the azimuth time-frequency diagram sequence; Extracting the intensity variation feature along the azimuth time axis from the azimuth time-frequency graph sequence, wherein the intensity variation feature characterizes the received power intensity variation characteristic of the SAR working mode; extracting the intensity variation feature along the azimuth time axis from the azimuth time-frequency graph sequence, comprising: performing height-wise maximum pooling, feature reshaping, convolution, batch normalization, and activation processing on the first feature of the azimuth time-frequency graph sequence in sequence to generate an axial attention weight; performing weighted processing on the first feature according to the axial attention weight, and obtaining the intensity variation feature after feature reshaping; Extracting Doppler time-frequency curve motion features from the azimuth time-frequency graph sequence, wherein the Doppler time-frequency curve motion features characterize the Doppler process motion characteristics of the SAR working mode; extracting Doppler time-frequency curve motion features from the azimuth time-frequency graph sequence comprises: performing feature reshaping and slicing processing on the second feature of the azimuth time-frequency graph sequence to obtain feature slices of different dimensions; performing feature connection, convolution, batch normalization and activation processing on the difference results between the feature slices, and then obtaining the Doppler time-frequency curve motion features after feature reshaping; The operating mode of the spaceborne SAR is determined according to at least one of the intensity variation characteristic and the Doppler time-frequency curve motion characteristic.

2. The identification method according to claim 1, characterized in that: The method of acquiring a signal pulse train group of a space-borne SAR comprises: Acquire a signal pulse train received by at least one receiver placed at a different azimuth position and the same distance position relative to the spaceborne SAR; Each of the signal pulse strings is arranged according to the start time sequence of each of the receivers receiving the signal pulse string to obtain the signal pulse string group.

3. The identification method according to claim 1, characterized in that: Determining the operating mode of the spaceborne SAR according to at least one of the intensity variation characteristic and the Doppler time-frequency curve motion characteristic includes: At least one of the intensity variation feature and the Doppler time-frequency curve motion feature is used to enhance the feature extraction capability of the network and determine the working mode of the spaceborne SAR.

4. The identification method according to claim 3, characterized in that: The working modes of the spaceborne SAR include strip mode, scanning mode, spotlight mode, sliding spotlight mode, sliding scanning mode and staggered mode.

5. A SAR working mode recognition device, characterized in that: include: An acquisition module, used for acquiring a signal pulse string group of a space-borne SAR, wherein the signal pulse string group includes at least one signal pulse string; A preprocessing module is used to perform data preprocessing on each of the signal pulse trains to obtain an azimuth time-frequency diagram sequence of the signal pulse train group; the preprocessing module is also used to rearrange each of the signal pulse trains using a pulse repetition time to obtain a two-dimensional receiving data matrix; after performing range pulse compression on the two-dimensional receiving data matrix using a frequency modulation rate, a signal at each range signal peak position in the two-dimensional receiving data matrix is ​​extracted to obtain an azimuth signal; the azimuth signal is short-time Fourier transformed and the amplitude is extracted to obtain an azimuth time-frequency diagram corresponding to each of the signal pulse trains, wherein each of the azimuth time-frequency diagrams constitutes the azimuth time-frequency diagram sequence; A first extraction module is used to extract the intensity change feature along the azimuth time axis from the azimuth time-frequency diagram sequence, wherein the intensity change feature characterizes the received power intensity change characteristic of the SAR working mode; the first extraction module is also used to perform height maximum pooling, feature reshaping, convolution, batch normalization, and activation processing on the first feature of the azimuth time-frequency diagram sequence in sequence to generate an axial attention weight; the first feature is weighted by the axial attention weight, and the intensity change feature is obtained after feature reshaping; A second extraction module is used to extract Doppler time-frequency curve motion features from the azimuth time-frequency diagram sequence, wherein the Doppler time-frequency curve motion features characterize the Doppler process motion characteristics of the SAR working mode; the second extraction module is also used to perform feature reshaping and slicing processing on the second features of the azimuth time-frequency diagram sequence to obtain feature slices of different dimensions; perform feature connection, convolution, batch normalization and activation processing on the difference results between the feature slices, and then obtain the Doppler time-frequency curve motion features after feature reshaping; The identification module is used to determine the working mode of the spaceborne SAR according to at least one of the intensity change characteristics and the Doppler time-frequency curve motion characteristics.

6. An electronic device comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images