Radar radiation source individual identification method and related device

Abstract

The invention discloses a radar radiation source individual identification method and a related device, and relates to the technical field of radar radiation source individual identification, and the method comprises the steps: obtaining a radar pulse signal; performing preliminary feature extraction on the radar pulse signal to obtain preliminary features; step-by-step feature extraction is carried out on the preliminary features to obtain target features, and when feature extraction is carried out on any level, feature extraction is carried out on the feature extraction object of the level, attention calculation is carried out on the extracted features, and the attention features obtained through attention calculation are fused with the feature extraction object of the level to obtain a feature extraction result of the level. The attention calculation is used for enhancing features related to a radar radiation source individual identification task and inhibiting features irrelevant to the radar radiation source individual identification task; and determining the type of the radar radiation source individual transmitting the radar pulse signal according to the target characteristics. According to the radar radiation source individual identification method disclosed by the invention, the type of the radar radiation source individual transmitting the radar pulse signal can be accurately identified.

Description

Technical Field

[0001] The present application relates to the technical field of radar radiation source individual identification, and in particular to a radar radiation source individual identification method and related devices. Background Art

[0002] Radar emitter individual identification refers to the process of uniquely identifying the category of the radar emitter that emits the radar pulse signal by analyzing the "fingerprint" features contained in the radar pulse signal.

[0003] In recent years, radar technology has advanced by leaps and bounds, and various new radars have emerged one after another, making the electromagnetic environment more complex. As the electromagnetic environment becomes increasingly complex, with dense signals and variable parameters, the challenges and difficulties of identifying individual radar emitters are becoming increasingly greater. Finding an accurate way to identify individual radar emitters has become an urgent problem to be solved in the field of radar emitters. Summary of the invention

[0004] In view of this, the present application provides a radar radiation source individual identification method and related device, which are used to accurately identify the category of the radar radiation source individual that emits the radar pulse signal according to the radar pulse signal. The technical solution is as follows:

[0005] The first aspect of the present application provides a method for identifying an individual radar radiation source, comprising:

[0006] Acquire target radar pulse signal;

[0007] Performing preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal;

[0008] Performing feature extraction on the preliminary features step by step to obtain the target features of the target radar pulse signal; wherein, the feature extraction object of the first level is the preliminary features, the feature extraction objects of each other level are the feature extraction results of the previous level, and the feature extraction result of the last level is the target features. When performing feature extraction at any level, extract features from the feature extraction objects of that level, and perform attention calculation on the extracted features, and fuse the attention features obtained by the attention calculation with the feature extraction objects of that level to obtain the feature extraction results of that level, wherein the attention calculation is used to strengthen the features related to the radar radiation source individual identification task and suppress the features unrelated to the radar radiation source individual identification task;

[0009] According to the target characteristics, the category of the radar radiation source individual that emits the target radar pulse signal is determined.

[0010] In a possible implementation, the performing preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal; performing step-by-step feature extraction on the preliminary features to obtain target features of the target radar pulse signal; and determining, according to the target features, the category of the radar radiation source individual that emits the target radar pulse signal, includes:

[0011] Based on the radar radiation source individual recognition model obtained by pre-training, preliminary feature extraction is performed on the target radar pulse signal to obtain preliminary features of the target radar pulse signal;

[0012] Based on the radar radiation source individual recognition model, the preliminary features are extracted step by step to obtain the target features of the target radar pulse signal;

[0013] Based on the radar emitter individual recognition model and the target features, predicting the category of the radar emitter individual that emits the target radar pulse signal;

[0014] The radar emitter individual recognition model is obtained by training with training radar pulse signals marked with radar emitter individual categories.

[0015] In a possible implementation, the step of extracting the preliminary features step by step based on the radar emitter individual recognition model to obtain the target features of the target radar pulse signal includes:

[0016] Based on a plurality of cascaded feature extraction modules included in the radar radiation source individual recognition model, feature extraction is performed on the preliminary features to obtain target features of the target radar pulse signal;

[0017] Among them, the input of the first feature extraction module in the multiple cascaded feature extraction modules is the preliminary feature, the input of each other feature extraction module is the feature output by the previous feature extraction module, and the feature output by the last feature extraction module is the target feature. Any feature extraction module further extracts features from the input features, and performs attention calculation on the further extracted features, and merges the attention features obtained by the attention calculation with the input features.

[0018] In a possible implementation, the process of extracting features at any level includes:

[0019] Extract features from the feature extraction object at this level, where the features extracted from the feature extraction object at this level are multi-channel features;

[0020] Determine, according to the multi-channel feature, an attention weight of each channel of the multi-channel feature;

[0021] Using the attention weight of each channel of the multi-channel feature, weighting the corresponding channel of the multi-channel feature to obtain a first attention feature;

[0022] Determining, according to the first attention feature, an attention weight for each spatial position of the first attention feature;

[0023] Using the attention weight of each spatial position of the first attention feature, weighting the feature of each spatial position of the first attention feature to obtain a second attention feature;

[0024] Fusing the second attention feature with the feature extraction object of this level to obtain a fused feature;

[0025] The fused features are post-processed.

[0026] In a possible implementation, determining the attention weight of each channel of the multi-channel feature according to the multi-channel feature includes:

[0027] Obtain the maximum feature and the average feature of each channel of the multi-channel feature;

[0028] Determine a first attention weight for each channel of the multi-channel feature according to a maximum feature of each channel of the multi-channel feature, and determine a second attention weight for each channel of the multi-channel feature according to an average feature of each channel of the multi-channel feature;

[0029] Fusing the first attention weight and the second attention weight of the same channel of the multi-channel feature to obtain a fused weight of each channel of the multi-channel feature;

[0030] The fused weight of each channel of the multi-channel feature is normalized to obtain the final attention weight of each channel of the multi-channel feature.

[0031] In a possible implementation, determining, according to the first attention feature, an attention weight of each spatial position of the first attention feature includes:

[0032] Obtaining the maximum feature among the features at the same spatial position of each channel of the first attention feature, and obtaining the average value of the features at the same spatial position of each channel of the first attention feature, to obtain the maximum feature and the average feature of each spatial position of the first attention feature;

[0033] Fusing the maximum feature and the average feature at the same spatial position of the first attention feature to obtain a fused feature at each spatial position of the first attention feature;

[0034] generating an initial attention weight for each spatial position of the first attention feature according to the fused features of each spatial position of the first attention feature;

[0035] The initial attention weight of each spatial position of the first attention feature is normalized to obtain the final attention weight of each spatial position of the first attention feature.

[0036] In a possible implementation manner, determining the category of the radar radiation source individual that transmits the target radar pulse signal according to the target feature includes:

[0037] According to the target characteristics, predict the probability that the category of the radar radiation source individual emitting the target radar pulse signal is the set category of each radar radiation source individual;

[0038] According to the predicted probability, the category of the individual radar emitter that emits the target radar pulse signal is determined from the set individual categories of the radar emitters.

[0039] A second aspect of the present application provides a radar radiation source individual identification device, comprising: a radar pulse signal acquisition module, a preliminary feature acquisition module, an advanced feature acquisition module and a radar radiation source individual category determination module;

[0040] The radar pulse signal acquisition module is used to acquire the target radar pulse signal;

[0041] The preliminary feature acquisition module is used to perform preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal;

[0042] The high-level feature acquisition module is used to perform step-by-step feature extraction on the preliminary features to obtain the target features of the target radar pulse signal; wherein the feature extraction object of the first level is the preliminary features, the feature extraction objects of each other level are the feature extraction results of the previous level, and the feature extraction result of the last level is the target feature. When performing feature extraction at any level, features are extracted from the feature extraction objects of that level, and attention calculation is performed on the extracted features, and the attention features obtained by the attention calculation are merged with the feature extraction objects of that level to obtain the feature extraction results of that level. The attention calculation is used to strengthen the features related to the radar radiation source individual identification task and suppress the features unrelated to the radar radiation source individual identification task;

[0043] The individual radiation source category determination module is used to determine the category of the individual radar radiation source that emits the target radar pulse signal according to the target characteristics.

[0044] A third aspect of the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0045] The memory is used to store computer programs;

[0046] The processor is used to execute the computer program so that the electronic device can implement the steps of any one of the above-mentioned radar radiation source individual identification methods.

[0047] In four aspects, the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement the steps of any one of the above-mentioned radar radiation source individual identification methods.

[0048] By means of the above technical scheme, the radar emitter individual identification method provided by the present application, after obtaining the target radar pulse signal, firstly performs preliminary feature extraction on the target radar pulse signal to obtain the preliminary features of the target radar pulse signal, and then further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal to obtain the target features of the target radar pulse signal, and finally determines the category of the radar emitter individual that emits the target radar pulse signal according to the target features of the target radar pulse signal. The radar emitter individual identification method provided by the present application, after obtaining the preliminary features of the target radar pulse signal, further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal, that is, further performs deep feature mining on the preliminary features, and when performing deep feature mining, in order to mine more critical features, an attention mechanism is introduced, and at the same time, in order to reduce information loss, when performing feature extraction at any level, the attention features obtained based on the attention mechanism are fused with the feature mining object of the level, and the above strategy can finally obtain high-level features that can more effectively represent the target radar pulse signal, that is, target features, and then, based on the target features, the category of the radar emitter individual that emits the target radar pulse signal can be more accurately determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0050] Figure 1 A schematic diagram of a system architecture involved in this application;

[0051] Figure 2A schematic diagram of a hardware structure of a terminal provided in an embodiment of the present application;

[0052] Figure 3 A schematic diagram of a hardware structure of a server provided in an embodiment of the present application;

[0053] Figure 4 A schematic diagram of a process for identifying a radar radiation source individual provided in an embodiment of the present application;

[0054] Figure 5 A schematic diagram of a process for extracting features at any level provided in an embodiment of the present application;

[0055] Figure 6 A schematic diagram of the structure of a feature extraction module provided in an embodiment of the present application;

[0056] Figure 7 A schematic diagram of the structure of a channel attention module provided in an embodiment of the present application;

[0057] Figure 8 A schematic diagram of the structure of a spatial attention module provided in an embodiment of the present application;

[0058] Fig. 9 A schematic diagram of the structure of a radar radiation source individual identification device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0060] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0061] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0062] Before introducing the radar radiation source individual identification method provided by the present application, the system architecture involved in the present application is first described.

[0063] In one possible implementation, Figure 1 As shown, the system architecture involved in the present application may include a terminal 101 and a server 102, and the terminal 101 may interact with the server 102 via a network (wired network or wireless network). The server 102 may include one or more servers ( Figure 1 The example includes a server. The terminal 101 can obtain the target radar pulse signal, transmit the target radar pulse signal to the server 102 through the network, and the server 102 uses the radar radiation source individual identification method provided in the present application to identify the category of the radar radiation source individual that emits the target radar pulse signal, and transmits the identification result to the terminal 101 through the network, and the terminal 101 can output the identification result.

[0064] In another possible implementation, the system architecture involved in the present application may include a terminal. The terminal has strong data processing capabilities. The terminal may use the radar emitter individual identification method provided in the present application to identify the category of the radar emitter individual that emits the target radar pulse signal, and then output the identification result.

[0065] Next, the product form of the above-mentioned terminal is described.

[0066] The above-mentioned terminal can be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc., and the embodiments of the present application do not impose any restrictions on this.

[0067] Figure 2 An optional hardware structure diagram of a terminal is shown.

[0068] refer to Figure 2 As shown, the terminal may include a radio frequency unit 210, a memory 220, an input unit 230, a display unit 240, a camera 250 (optional), an audio circuit 260 (optional), a speaker 261 (optional), a microphone 262 (optional), an earphone jack 263 (optional), a processor 270, an external interface 280, a power supply 290 and other components. Those skilled in the art will appreciate that Figure 2This is merely an example of a terminal and does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown in the figure, or a combination of certain components, or different components.

[0069] The input unit 230 can be used to receive input digital or character information, and generate key signal input related to the user settings and function control of the terminal. Specifically, the input unit 230 may include a touch screen 231 (optional) and / or other input devices 232. The touch screen 231 can collect the user's touch operations on or near it (such as the user's operation on or near the touch screen using any suitable object such as fingers, joints, stylus, etc.), and drive the corresponding connection device according to a pre-set program. The touch screen can detect the user's touch action on the touch screen, convert the touch action into a touch signal and send it to the processor 270, and can receive and execute the command sent by the processor 270; the touch signal at least includes the contact point coordinate information. The touch screen 231 can provide an input interface and an output interface between the terminal and the user. In addition, the touch screen can be implemented using multiple types such as resistive, capacitive, infrared, and surface acoustic waves. In addition to the touch screen 231, the input unit 230 can also include other input devices. Specifically, other input devices 232 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, and the like.

[0070] The display unit 240 may be used to display information input by a user or information provided to a user, various menus of the terminal, an interactive interface, file display, and / or playback of any multimedia file.

[0071] The memory 220 can be used to store instructions and data. The memory 220 can mainly include an instruction storage area and a data storage area. The data storage area can store various data, such as multimedia files, texts, etc. The instruction storage area can store software units such as operating systems, applications, and instructions required for at least one function, or their subsets and extensions. It can also include a non-volatile random access memory; it provides the processor 270 with hardware, software, and data resources including management of computing and processing equipment, and supports control software and applications. It is also used for the storage of multimedia files, and the storage of running programs and applications.

[0072] The processor 270 is the control center of the terminal. It uses various interfaces and lines to connect various parts of the entire terminal. By running or executing instructions stored in the memory 220 and calling data stored in the memory 220, it executes various functions of the terminal and processes data, thereby controlling the terminal as a whole. Optionally, the processor 270 may include one or more processing units; preferably, the processor 270 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application program, and the modem processor mainly processes wireless communication. It is understandable that the above-mentioned modem processor may not be integrated into the processor 270. In some embodiments, the processor and the memory may be implemented on a single chip, and in some embodiments, they may also be implemented separately on separate chips. The processor 270 may also be used to generate corresponding operation control signals, send them to corresponding components of the computing and processing device, read and process data in the software, especially read and process data and programs in the memory 220, so that each functional module therein performs corresponding functions, thereby controlling the corresponding components to act according to the requirements of the instructions.

[0073] Among them, the memory 220 can be used to store software codes related to the radar radiation source individual identification method, the processor 270 can execute the software codes in the memory 220, and can also schedule other units (such as the above-mentioned input unit 230 and display unit 240) to implement corresponding functions.

[0074] The radio frequency unit 210 (optional) can be used for receiving and sending information or receiving and sending signals during a call, for example, after receiving the downlink information of the base station, it is sent to the processor 270 for processing; in addition, the designed uplink data is sent to the base station. Generally, the radio frequency unit 210 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (LowNoiseAmplifier, LNA), a duplexer, etc. In addition, the radio frequency unit 210 can also communicate with network devices and other devices through wireless communication. The wireless communication can use any communication standard or protocol, including but not limited to Global System of Mobile communication (Global System of Mobile communication, GSM), General Packet Radio Service (General PacketRadio Service, GPRS), Code Division Multiple Access (Code Division Multiple Access, CDMA), Wideband Code Division Multiple Access (Wideband Code Division Multiple Access, WCDMA), Long Term Evolution (Long Term Evolution, LTE), email, Short Messaging Service (SMS), etc.

[0075] In the embodiment of the present application, the radio frequency unit 210 can send data to other devices, and can also receive data sent by other devices. It should be understood that the radio frequency unit 210 is optional, and can be replaced by other communication interfaces, such as a network port.

[0076] The terminal also includes a power supply 290 (such as a battery) for supplying power to various components. Preferably, the power supply can be logically connected to the processor 270 through a power management system, so that the power management system can manage charging, discharging, power consumption and other functions.

[0077] The terminal also includes an external interface 280, which may be a standard Micro USB interface or a multi-pin connector, and may be used to connect the terminal to communicate with other devices, or to connect a charger to charge the terminal.

[0078] Although not shown, the terminal may further include a flashlight, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which will not be described in detail here.

[0079] Next, the product form of the above server is described.

[0080] Figure 3 A structural diagram of the above server is provided, such as Figure 3As shown, the server may include a bus 301, a processor 302, a communication interface 303, and a memory 304. The processor 302, the memory 304, and the communication interface 303 communicate with each other via the bus 301.

[0081] The bus 301 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0082] The processor 302 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0083] The memory 304 may include a volatile memory, such as a random access memory (RAM). The memory 304 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0084] The memory 304 may be used to store software codes related to the radar emitter individual identification method. The processor 302 may call the software codes stored in the memory 304 or schedule other units to implement corresponding functions.

[0085] The processors in the above-mentioned terminals and servers (for example, processor 270 and processor 302) may be hardware circuits (such as application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), general-purpose processors, digital signal processors (DSP), microprocessors or microcontrollers, etc.), or combinations of these hardware circuits. For example, the processor may be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0086] In the process of implementing this case, the inventors of this case found that most of the current radar emitter individual identification methods are identification methods based on radar feature matching. Specifically, the target radar pulse signal is first preprocessed (such as noise reduction, multipath suppression, sorting, etc.), and then the artificially designed radar "fingerprint" features are extracted from the preprocessed signal, such as carrier frequency, envelope characteristics, high-order moments, etc., and then the extracted radar "fingerprint" features are matched with the features of each category of radar emitter individuals in the radar feature database, and the radar emitter individual category to which the features in the radar feature database that are most similar to the extracted radar "fingerprint" features belong is determined as the category of the radar emitter individual that emits the target radar pulse signal.

[0087] The inventors of this case studied the existing radar emitter individual identification methods and found that the existing radar emitter individual identification methods have many defects: in terms of feature extraction, due to the limitations of dimension and feature expression, the existing radar emitter individual identification methods only focus on some basic characteristics of radar pulse signals, such as amplitude, frequency, phase, etc., but ignore the higher-level characteristics of radar pulse signals. For complex radar emitter individuals, these basic characteristics cannot well represent the radar pulse signal, which leads to inaccurate matching results; in terms of identification results, relying on predefined templates, the existing radar emitter individual identification methods need to build a template library of radar emitter individuals in advance, which requires prior understanding of various possible radiation sources and preparation of corresponding templates. However, in actual applications, unknown radiation sources may be encountered, or changes and evolutions of radiation sources may occur. At this time, the template library needs to be updated and maintained, which increases the workload and complexity.

[0088] In view of the defects of the existing methods for individually identifying radar radiation sources, the inventors of this case conducted research and, through continuous research, finally proposed a method for individually identifying radar radiation sources with better effects. This method for individually identifying radar radiation sources overcomes the defects of the existing methods for individually identifying radar radiation sources.

[0089] Next, the radar radiation source individual identification method provided by the present application is introduced through the following embodiments.

[0090] See also Figure 4 , shows a schematic flow chart of a method for identifying individual radar emitters provided in an embodiment of the present application, which method may include:

[0091] Step S401: Acquire target radar pulse signal.

[0092] The target radar pulse signal is a radar pulse signal used to identify an individual radar radiation source, and the radar pulse signal may be a one-dimensional time series signal of a radar pulse.

[0093] Step S402: Perform preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal.

[0094] In a possible implementation, preliminary feature extraction may be performed on the target radar pulse signal based on a convolution module to obtain preliminary features of the target radar pulse signal.

[0095] Step S403: extracting the preliminary features of the target radar pulse signal step by step to obtain the target features of the target radar pulse signal.

[0096] When the preliminary features of the target radar pulse signal are extracted step by step, the feature extraction object of the first level is the preliminary features of the target radar pulse signal, the feature extraction object of each other level is the feature extraction result of the previous level, and the feature extraction result of the last level is the target feature of the target radar pulse signal.

[0097] When performing feature extraction at any level, extract features from the feature extraction object at that level, perform attention calculation on the extracted features, and fuse the attention features obtained by attention calculation with the feature extraction object at that level to obtain the feature extraction result at that level. Among them, attention calculation is used to strengthen the features related to the radar emitter individual identification task and suppress the features unrelated to the radar emitter individual identification task.

[0098] The target features obtained by extracting the preliminary features of the target radar pulse signal step by step are high-level features that can effectively represent the target radar pulse signal.

[0099] Step S404: Determine the category of the radar radiation source that transmits the target radar pulse signal according to the target feature of the target radar pulse signal.

[0100] Specifically, according to the target feature of the target radar pulse signal, the process of determining the category of the radar radiation source individual that emits the target radar pulse signal may include:

[0101] Step S4041: predicting the probability that the category of the radar radiation source individual emitting the target radar pulse signal is the set category of each radar radiation source individual according to the target feature of the target radar pulse signal.

[0102] Step S4042: According to the predicted probability, determine the category of the radar radiation source individual that transmits the target radar pulse signal from the set radar radiation source individual categories.

[0103] The radar emitter individual category corresponding to the maximum probability in the predicted probabilities can be determined as the category of the radar emitter individual that emits the target radar pulse signal.

[0104] The radar emitter individual identification method provided in the embodiment of the present application, after obtaining the target radar pulse signal, firstly performs preliminary feature extraction on the target radar pulse signal to obtain the preliminary features of the target radar pulse signal, and then further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal to obtain the target features of the target radar pulse signal, and finally determines the category of the radar emitter individual that emits the target radar pulse signal according to the target features of the target radar pulse signal. The radar emitter individual identification method provided in the embodiment of the present application, after obtaining the preliminary features of the target radar pulse signal, further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal, that is, performs deep feature mining on the preliminary features, and when performing deep feature mining, in order to mine more critical features, an attention mechanism is introduced, and in order to reduce information loss, when performing feature extraction at any level, the attention features obtained based on the attention mechanism are fused with the feature mining object of the level, and the above strategy can finally obtain high-level features that can more effectively represent the target radar pulse signal, that is, target features, and then, based on the target features, the category of the radar emitter individual that emits the target radar pulse signal can be more accurately determined.

[0105] In another embodiment of the present application, the above-mentioned “step S403: extracting the preliminary features of the target radar pulse signal step by step to obtain the target features of the target radar pulse signal” is introduced.

[0106] The above embodiment mentioned that when the preliminary features of the target radar pulse signal are extracted step by step, the feature extraction object of the first level is the preliminary features of the target radar pulse signal, the feature extraction object of each other level is the feature extraction result of the previous level, and the feature extraction result of the last level is the target feature of the target radar pulse signal.

[0107] It should be noted that when performing feature extraction at each level, the feature extraction objects at each level are different, but the feature extraction processes at each level are similar. The process of feature extraction at any level is introduced below.

[0108] See also Figure 5 , showing a schematic diagram of the process of performing feature extraction at any level, which may include:

[0109] Step S501: extracting features from the feature extraction object at this level, wherein the extracted features are multi-channel features.

[0110] Features may be extracted from the feature extraction object at this level based on one or more convolutional layers.

[0111] Step S502: Determine the attention weight of each channel of the multi-channel feature according to the multi-channel feature.

[0112] In a possible implementation, according to the multi-channel feature, the process of determining the attention weight of each channel of the multi-channel feature may include:

[0113] Step a1: Obtain the maximum feature and average feature of each channel of the multi-channel feature.

[0114] A global maximum pooling operation and a global average pooling operation may be performed on each channel of the multi-channel feature to obtain a maximum feature and an average feature of each channel of the multi-channel feature.

[0115] Step a2: determine the first attention weight of each channel of the multi-channel feature according to the maximum feature of each channel of the multi-channel feature, and determine the second attention weight of each channel of the multi-channel feature according to the average feature of each channel of the multi-channel feature.

[0116] For any channel of the multi-channel feature, the attention weight of the channel is determined according to the maximum feature of the channel as the first attention weight of the channel, and the attention weight of the channel is determined according to the average feature of the channel as the second attention weight of the channel. In this way, the first attention weight and the second attention weight of each channel of the multi-channel feature can be obtained.

[0117] Step a3: Fuse the first attention weight and the second attention weight of the same channel of the multi-channel feature to obtain the fused weight of each channel of the multi-channel feature.

[0118] Specifically, for any channel of the multi-channel feature, the first attention weight of the channel is fused (for example, added) with the second attention weight of the channel to obtain the fused weight of the channel. In this way, the fused weight of each channel of the multi-channel feature can be obtained.

[0119] Step a4: normalize the fused weight of each channel of the multi-channel feature to obtain the final attention weight of each channel of the multi-channel feature.

[0120] The fused weight of each channel of the multi-channel feature is normalized so that the fused weight of each channel of the multi-channel feature is between 0 and 1.

[0121] Step S503: Use the final attention weight of each channel of the multi-channel feature to weight the corresponding channels of the multi-channel feature to obtain the first attention feature.

[0122] The final attention weight of each channel of the multi-channel feature reflects the importance of the corresponding channel. Using the attention weight of each channel of the multi-channel feature to weight the corresponding channel of the multi-channel feature can strengthen the channels containing important information and weaken the channels containing unimportant information, that is, highlight the channels related to the radar emitter individual identification task and suppress the channels irrelevant to the radar emitter individual identification task, thereby improving the accuracy of radar emitter individual identification.

[0123] Step S504: Determine the attention weight of each spatial position of the first attention feature according to the first attention feature.

[0124] In a possible implementation, according to the first attention feature, the process of determining the attention weight of each spatial position of the first attention feature may include:

[0125] Step b1, obtain the maximum feature among the features of the same spatial position of each channel of the first attention feature, and obtain the average value of the features of the same spatial position of each channel of the first attention feature, to obtain the maximum feature and average feature of each spatial position of the first attention feature.

[0126] In this step, the maximum feature among the features at the same spatial position and the average value of the features at the same spatial position are obtained for the first attention feature along the channel dimension to obtain the maximum feature and the average feature at each spatial position of the first attention feature.

[0127] Step b2: fuse the maximum feature and the average feature at the same spatial position of the first attention feature to obtain the fused feature at each spatial position of the first attention feature.

[0128] The maximum feature and the average feature at the same spatial position of the first attention feature can be concatenated to obtain the fused feature at each spatial position of the first attention feature.

[0129] Step b3: Generate an initial attention weight for each spatial position of the first attention feature based on the fused features of each spatial position of the first attention feature.

[0130] Step b4: normalize the initial attention weight of each spatial position of the first attention feature to obtain the final attention weight of each spatial position of the first attention feature.

[0131] The initial attention weight of each spatial position of the first attention feature is normalized so that the fused weight of each spatial position of the first attention feature is between 0 and 1.

[0132] Step S505: Using the final attention weight of each spatial position of the first attention feature, weight the features of each spatial position of the first attention feature to obtain the second attention feature.

[0133] The final attention weight of each spatial position of the first attention feature can reflect the importance of the features at each spatial position of the first attention feature. By using the final attention weight of each spatial position of the first attention feature and weighting the features at each spatial position of the first attention feature, important feature areas can be highlighted and the influence of unimportant feature areas can be reduced, thereby improving the accuracy of individual identification of radar emitters.

[0134] Step S506: Fuse the second attention feature with the feature extraction object of this level to obtain a fused feature, and post-process the fused feature.

[0135] The second attention feature can be fused with the feature extraction object of this level by adding them together. Fusion of the second attention feature with the feature extraction object of this level can reduce information loss and improve the individual recognition effect of radar emitters.

[0136] It should be noted that this embodiment does not limit the use of channel attention mechanism and spatial attention mechanism to determine the attention features. For example, only the channel attention mechanism can be used to determine the attention features (that is, the first attention feature obtained through step S503 is used as the final attention feature, and then the attention feature is fused with the feature extraction object of this level, and the fused feature is post-processed), or only the spatial attention mechanism can be used to determine the attention features (that is, according to the feature obtained in step S501, the attention weight of each spatial position of the feature is determined, and the features of each spatial position of the feature are weighted by the attention weight of each spatial position of the feature to obtain the final attention feature, and then the attention feature is fused with the feature extraction object of this level, and the fused feature is post-processed).

[0137] In a possible implementation, the radar emitter individual identification method provided in the present application may be implemented based on a model. Specifically, the process of implementing the radar emitter individual identification based on the model may include:

[0138] Step c1: Based on the pre-trained radar radiation source individual recognition model, preliminary feature extraction is performed on the target radar pulse signal to obtain preliminary features of the target radar pulse signal.

[0139] Among them, the radar emitter individual recognition model is trained using training radar pulse signals labeled with radar emitter individual categories.

[0140] The training process of the radar emitter individual recognition model may include: inputting the training radar pulse signal into the radar emitter individual recognition model, obtaining the radar emitter individual category prediction result corresponding to the training radar pulse signal output by the radar emitter individual recognition model, determining the prediction loss according to the radar emitter individual category prediction result and the radar emitter individual category marked by the training radar pulse signal, and updating the parameters of the radar emitter individual recognition model according to the prediction loss. The radar emitter individual recognition model is iteratively trained in the above manner until the preset end training condition is met (such as model convergence, reaching the preset number of training iterations, etc.).

[0141] In a possible implementation, the radar emitter individual identification model includes a convolution module, and preliminary feature extraction can be performed on the target radar pulse signal based on the convolution module of the radar emitter individual identification model to obtain preliminary features of the target radar pulse signal.

[0142] Step c2: Based on the radar radiation source individual recognition model, the preliminary features of the target radar pulse signal are extracted step by step to obtain the target features of the target radar pulse signal.

[0143] In a possible implementation, the radar emitter individual recognition model includes a plurality of cascaded feature extraction modules, and preliminary features of the target radar pulse signal may be extracted based on the plurality of cascaded feature extraction modules included in the radar emitter individual recognition model.

[0144] Among them, the input of the first feature extraction module in multiple cascaded feature extraction modules is the preliminary features of the target radar pulse signal, the input of each other feature extraction module is the features output by the previous feature extraction module, and the features output by the last feature extraction module are the target features of the target radar pulse signal. Any feature extraction module further extracts features from the input features, and performs attention calculation on the further extracted features, and merges the attention features obtained by the attention calculation with the input features.

[0145] The structures of each feature extraction module are the same, see Figure 6 , shows an example of a feature extraction module, such as Figure 6 As shown, the feature extraction module may include a first convolution layer 601 (e.g., 3×3 convolution), a first batch of normalization layers (BN) 602, a first activation layer (ReLu) 603, a second convolution layer 604 (e.g., 3×3 convolution), and a second batch of normalization layers (BN) 605. The input features are sequentially processed by the first convolution layer 601, the first batch of normalization layers 602, the first activation layer 603, the second convolution layer 604, and the second batch of normalization layers 605 to obtain features that are further extracted from the input features.

[0146] like Figure 6 As shown, the feature extraction module may further include an attention module, and the attention module may include a channel attention module 606 and a spatial attention module 607. The channel attention module 606 obtains a first attention feature according to the process of step S502 to step S503, and the spatial attention module 607 obtains a second attention feature according to the process of step S504 to step S505.

[0147] See also Figure 7, showing an example of a channel attention module 606, which may include: a maximum pooling layer 701a, an average pooling layer 701b, a shared fully connected layer 702, an adder 703, a weight normalization module 704 and a feature weighting module 705. Among them, the multi-channel features output by the second batch normalization layer 605 are respectively input into the maximum pooling layer 701a and the average pooling layer 701b. The maximum pooling layer 701a performs a global maximum pooling operation on each channel of the multi-channel features and outputs the maximum feature of each channel of the multi-channel features. The average pooling layer 701b performs a global average pooling operation on each channel of the multi-channel features and outputs the average feature of each channel of the multi-channel features. The maximum feature of each channel of the multi-channel features is input into the shared fully connected layer 702. The shared fully connected layer 702 determines a first attention weight according to the maximum feature of each channel of the multi-channel features and outputs the first attention weight of each channel of the multi-channel features. The average feature of each channel of the multi-channel features is input into the shared fully connected layer 702. The shared fully connected layer 702 determines a second attention weight according to the average feature of each channel of the multi-channel features and outputs the second attention weight of each channel of the multi-channel features. The first attention weight and the second attention weight of each channel of the feature are input into the adder 703. The adder 703 adds the first attention weight and the second attention weight of the same channel, and outputs the fused weight of each channel of the multi-channel feature. The fused weight of each channel of the multi-channel feature is input into the weight normalization module 704. The weight normalization module 704 uses the activation function sigmoid to normalize the fused weight of each channel of the multi-channel feature (processes the fused weight into a weight between 0 and 1), and outputs the normalized weight of each channel of the multi-channel feature, that is, the final attention weight of each channel of the multi-channel feature. The multi-channel feature and the final attention weight of each channel of the multi-channel feature are input into the feature weighting module 705. The feature weighting module 705 uses the final attention weight of each channel of the multi-channel feature to weight the corresponding channel of the multi-channel feature, and outputs the weighted feature, that is, the first attention feature.

[0148] See also Figure 8, a schematic diagram of the structure of the spatial attention module 607 is shown, which may include: a maximum pooling layer 801a, an average pooling layer 801b, a feature splicing module 802, a convolution layer 803, a weight normalization module 804 and a feature weighting module 805. The first attention feature is input into the maximum pooling layer 801a and the average pooling layer 801b respectively, the maximum pooling layer 801a performs a maximum pooling operation on the first attention feature along the channel dimension to obtain the maximum feature of each spatial position of the first attention feature, the average pooling layer 801b performs an average pooling operation on the first attention feature along the channel dimension to obtain the average feature of each spatial position of the first attention feature, the maximum feature and the average feature of each spatial position of the first attention feature are input into the feature fusion module 802, the feature fusion module 802 splices the maximum feature and the average feature of the same spatial position of the first attention feature, and outputs the spliced ​​feature of each spatial position of the first attention feature, that is, the fused feature of each spatial position of the first attention feature, the fused feature of each spatial position of the first attention feature is input into the convolution layer 803, the convolution Layer 803 generates an initial attention weight for each spatial position of the first attention feature according to the fused features of each spatial position of the first attention feature. The initial attention weight for each spatial position of the first attention feature is input into the weight normalization module 804. The weight normalization module 804 normalizes the initial attention weight for each spatial position of the first attention feature and outputs the normalized weight for each spatial position of the first attention feature, i.e., the final attention weight for each spatial position of the first attention feature. The first attention feature and the final attention weight for each spatial position of the first attention feature are input into the feature weighting module 805. The feature weighting module 805 uses the final attention weight for each spatial position of the first attention feature to weight the features of each spatial position of the first attention feature and outputs the weighted features, i.e., the second attention feature.

[0149] like Figure 6 As shown, the feature extraction module may further include a feature fusion module 608 and a second activation layer (ReLu) 609. The feature fusion module 608 fuses the second attention feature with the input feature (the input feature of the feature extraction module), and the feature output by the feature fusion module 608 is input to the second activation layer 609 for processing and then output. Fusion of the second attention feature with the input feature can reduce information loss and improve the individual recognition effect of radar emitters.

[0150] It should be noted that the feature extraction module in this embodiment includes a main path and a skip path. The main path includes a first convolutional layer 601, a first batch of normalization layers 602, a first activation layer 603, a second convolutional layer 604 (for example, 3×3 convolution), a second batch of normalization layers 605 and an attention module (a channel attention module 606 and a spatial attention module 607). The skip connection directly skips the input over the main path and passes it to the subsequent layer in the form of an identity mapping. The feature extraction module provided in this embodiment realizes a "jump" from input to output, reducing information loss.

[0151] The skip path of the feature extraction module can effectively solve the gradient vanishing and degradation problems that occur as the network depth increases. This structure not only simplifies the learning process of the model and enhances gradient propagation, but also allows the model to maintain performance improvements while increasing depth, thereby improving the generalization ability of the model.

[0152] Step c3: based on the radar emitter individual recognition model and taking the target features of the target radar pulse signal as a basis, predict the category of the radar emitter individual that emits the target radar pulse signal.

[0153] The radar radiation source individual identification model in this embodiment includes a category prediction module. After obtaining the target features of the target radar pulse signal, the target features of the target radar pulse signal are input into the category prediction module of the radar radiation source individual identification model. The category prediction module predicts the probability that the category of the radar radiation source individual that transmits the target radar pulse signal is one of the set radar radiation source individual categories based on the target features of the target radar pulse signal, and then determines the category of the radar radiation source individual that transmits the target radar pulse signal from the set radar radiation source individual categories based on the predicted probability.

[0154] The radar emitter individual identification method provided in the embodiment of the present application realizes radar emitter individual identification based on a radar emitter individual identification model. An attention module is introduced into the feature extraction module of the radar emitter individual identification model. The introduction of the attention module enables the model to focus on important features, thereby improving the accuracy of radar emitter individual identification. In addition, a jump path is introduced into the feature extraction module of the radar emitter individual identification model. The introduction of the jump path can improve the generalization ability of the model, so that the radar emitter individual can be effectively identified even when a situation that has not been seen during training occurs. Furthermore, the radar emitter individual identification model is an end-to-end model, that is, the radar pulse signal is directly input into the radar emitter individual identification model, and the radar emitter individual identification model can output the radar emitter individual identification result. The end-to-end model is used to realize radar emitter individual identification, which reduces the need for manual intervention and adjustment and improves the automation and efficiency of radar emitter individual identification.

[0155] Compared with the existing radar emitter individual identification method, the present application can extract high-level features that effectively represent radar pulse signals. Based on the high-level features, the category of the radar emitter individual that emits radar pulse signals can be identified more accurately. The present application implements radar emitter individual identification based on an end-to-end model rather than a template matching method. Therefore, there is no need to construct a template library of radar emitter individual, and accordingly, there is no need to maintain the template library of radar emitter individual.

[0156] The above describes the radar radiation source individual identification method provided in the embodiment of the present application. The following describes the device corresponding to the above radar radiation source individual identification method.

[0157] See also Fig. 9 , Fig. 9 A schematic diagram of the structure of a radar radiation source individual identification device provided in an embodiment of the present application, the radar radiation source individual identification device may include: a radar pulse signal acquisition module 901, a preliminary feature acquisition module 902, an advanced feature acquisition module 903 and a radar radiation source individual category determination module 904.

[0158] The radar pulse signal acquisition module 901 is used to acquire the target radar pulse signal.

[0159] The preliminary feature acquisition module 902 is used to perform preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal.

[0160] The high-level feature acquisition module 903 is used to extract the preliminary features of the target radar pulse signal step by step to obtain the target features of the target radar pulse signal.

[0161] Among them, the feature extraction object of the first level is the preliminary feature of the target radar pulse signal, the feature extraction object of each other level is the feature extraction result of the previous level, and the feature extraction result of the last level is the target feature of the target radar pulse signal. When performing feature extraction at any level, features are extracted from the feature extraction object of that level, and attention calculation is performed on the extracted features. The attention features obtained by the attention calculation are fused with the feature extraction objects of that level to obtain the feature extraction results of that level. The attention calculation is used to strengthen the features related to the individual identification task of the radar radiation source and suppress the features that are not related to the individual identification task of the radar radiation source.

[0162] The radar emitter individual category determination module 904 is used to determine the category of the radar emitter individual that emits the target radar pulse signal according to the target feature of the target radar pulse signal.

[0163] In one possible implementation, when the preliminary feature acquisition module 902 performs preliminary feature extraction on the target radar pulse signal to obtain the preliminary features of the target radar pulse signal, it is specifically used to perform preliminary feature extraction on the target radar pulse signal based on a pre-trained radar radiation source individual recognition model to obtain the preliminary features of the target radar pulse signal.

[0164] The advanced feature acquisition module 903 performs step-by-step feature extraction on the preliminary features of the target radar pulse signal to obtain the target features of the target radar pulse signal. Specifically, it is used to perform step-by-step feature extraction on the preliminary features of the target radar pulse signal based on the radar radiation source individual recognition model to obtain the target features of the target radar pulse signal.

[0165] When determining the category of the individual radar emitter that emits the target radar pulse signal according to the target characteristics of the target radar pulse signal, the radar emitter individual category determination module 904 is specifically used to predict the category of the individual radar emitter that emits the target radar pulse signal based on the radar emitter individual identification model and the target characteristics of the target radar pulse signal.

[0166] Among them, the radar emitter individual recognition model is trained using training radar pulse signals labeled with radar emitter individual categories.

[0167] In one possible implementation, the advanced feature acquisition module 903 performs step-by-step feature extraction on the preliminary features of the target radar pulse signal based on the radar radiation source individual recognition model to obtain the target features of the target radar pulse signal. Specifically, it is used to extract the preliminary features of the target radar pulse signal based on multiple cascaded feature extraction modules included in the radar radiation source individual recognition model to obtain the target features of the target radar pulse signal.

[0168] Among them, the input of the first feature extraction module in multiple cascaded feature extraction modules is the preliminary features of the target radar pulse signal, the input of each other feature extraction module is the features output by the previous feature extraction module, and the features output by the last feature extraction module are the target features of the target radar pulse signal. Any feature extraction module further extracts features from the input features, and performs attention calculation on the further extracted features, and merges the attention features obtained by the attention calculation with the input features.

[0169] In a possible implementation, when performing feature extraction at any level, the high-level feature acquisition module 903 is specifically used to:

[0170] Extract features from the feature extraction object at this level, where the features extracted from the feature extraction object at this level are multi-channel features;

[0171] According to the multi-channel features, determine the attention weight of each channel of the multi-channel features;

[0172] Use the attention weight of each channel of the multi-channel feature to weight the corresponding channel of the multi-channel feature to obtain the first attention feature;

[0173] Determine, according to the first attention feature, an attention weight for each spatial position of the first attention feature;

[0174] Using the attention weight of each spatial position of the first attention feature, weight the features of each spatial position of the first attention feature to obtain a second attention feature;

[0175] The second attention feature is fused with the feature extraction object of this level to obtain the fused feature, and the fused feature is post-processed.

[0176] In a possible implementation, when determining the attention weight of each channel of the multi-channel feature according to the multi-channel feature, the high-level feature acquisition module 903 is specifically used to:

[0177] Get the maximum feature and average feature of each channel of the multi-channel feature;

[0178] Determine a first attention weight of each channel of the multi-channel feature according to the maximum feature of each channel of the multi-channel feature, and determine a second attention weight of each channel of the multi-channel feature according to the average feature of each channel of the multi-channel feature;

[0179] The first attention weight and the second attention weight of the same channel of the multi-channel feature are fused to obtain the fused weight of each channel of the multi-channel feature;

[0180] The fused weights of each channel of the multi-channel features are normalized to obtain the final attention weight of each channel of the multi-channel features.

[0181] In a possible implementation, when the high-level feature acquisition module 903 determines the attention weight of each spatial position of the first attention feature according to the first attention feature, it is specifically used to:

[0182] Obtain the maximum feature among the features of the same spatial position of each channel of the first attention feature, and obtain the average value of the features of the same spatial position of each channel of the first attention feature, to obtain the maximum feature and average feature of each spatial position of the first attention feature;

[0183] The maximum feature and the average feature at the same spatial position of the first attention feature are fused to obtain the fused feature at each spatial position of the first attention feature;

[0184] Generate an initial attention weight for each spatial position of the first attention feature according to the fused features of each spatial position of the first attention feature;

[0185] The initial attention weight of each spatial position of the first attention feature is normalized to obtain the final attention weight of each spatial position of the first attention feature.

[0186] In a possible implementation, when the radar emitter individual category determination module 904 determines the category of the radar emitter individual that emits the target radar pulse signal according to the target feature of the target radar pulse signal, it is specifically configured to:

[0187] According to the target characteristics of the target radar pulse signal, the probability of predicting that the category of the radar radiation source individual emitting the target radar pulse signal is the set category of each radar radiation source individual;

[0188] According to the predicted probability, the category of the radar emitter individual that emits the target radar pulse signal is determined from the set radar emitter individual categories.

[0189] The radar radiation source individual identification device provided in the embodiment of the present application, after obtaining the target radar pulse signal, first performs preliminary feature extraction on the target radar pulse signal to obtain the preliminary features of the target radar pulse signal, and then further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal to obtain the target features of the target radar pulse signal, and finally determines the category of the radar radiation source individual that emits the target radar pulse signal according to the target features of the target radar pulse signal. After obtaining the preliminary features of the target radar pulse signal, the radar radiation source individual identification device provided in the embodiment of the present application further performs step-by-step feature extraction on the preliminary features of the target radar pulse signal, that is, performs deep feature mining on the preliminary features. When performing deep feature mining, in order to mine more critical features, an attention mechanism is introduced. In order to reduce information loss, when performing feature extraction at any level, the attention features obtained based on the attention mechanism are fused with the feature mining object of the level. Through the above strategy, a high-level feature that can more effectively represent the target radar pulse signal, that is, the target feature, can be finally obtained. Then, based on the target feature, the category of the radar radiation source individual that emits the target radar pulse signal can be more accurately determined.

[0190] An embodiment of the present application also provides an electronic device, which may include: at least one processor, at least one communication interface, at least one memory and at least one communication bus.

[0191] In the embodiment of the present application, the number of the processor, the communication interface, the memory, and the communication bus is at least one, and the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0192] The processor may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application;

[0193] The memory may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;

[0194] The memory stores a program, and the processor can call the program stored in the memory, wherein the program is used to implement the steps of the radar radiation source individual identification method provided in the above embodiment.

[0195] An embodiment of the present application also provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement the steps of the radar radiation source individual identification method provided in the above embodiment.

[0196] An embodiment of the present application also provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements the steps of the radar radiation source individual identification method provided in the above embodiment.

[0197] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0198] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0199] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0200] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. A method for identifying individual radar radiation sources, characterized in that: include: Acquire target radar pulse signal; Performing preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal; Performing feature extraction on the preliminary features step by step to obtain the target features of the target radar pulse signal; wherein, the feature extraction object of the first level is the preliminary features, the feature extraction objects of each other level are the feature extraction results of the previous level, and the feature extraction result of the last level is the target features. When performing feature extraction at any level, extract features from the feature extraction objects of that level, and perform attention calculation on the extracted features, and fuse the attention features obtained by the attention calculation with the feature extraction objects of that level to obtain the feature extraction results of that level, wherein the attention calculation is used to strengthen the features related to the radar radiation source individual identification task and suppress the features unrelated to the radar radiation source individual identification task; According to the target characteristics, the category of the radar radiation source individual that emits the target radar pulse signal is determined.

2. The radar radiation source individual identification method according to claim 1, characterized in that: The target radar pulse signal is subjected to preliminary feature extraction to obtain preliminary features of the target radar pulse signal; the preliminary features are subjected to step-by-step feature extraction to obtain target features of the target radar pulse signal; Determining the category of the radar radiation source individual that emits the target radar pulse signal according to the target feature includes: Based on the radar radiation source individual recognition model obtained by pre-training, preliminary feature extraction is performed on the target radar pulse signal to obtain preliminary features of the target radar pulse signal; Based on the radar radiation source individual recognition model, the preliminary features are extracted step by step to obtain the target features of the target radar pulse signal; Based on the radar emitter individual recognition model and the target features, predicting the category of the radar emitter individual that emits the target radar pulse signal; The radar emitter individual recognition model is obtained by training with training radar pulse signals marked with radar emitter individual categories.

3. The radar radiation source individual identification method according to claim 2, characterized in that: The step of extracting the preliminary features step by step based on the radar radiation source individual recognition model to obtain the target features of the target radar pulse signal includes: Based on a plurality of cascaded feature extraction modules included in the radar radiation source individual recognition model, feature extraction is performed on the preliminary features to obtain target features of the target radar pulse signal; Among them, the input of the first feature extraction module in the multiple cascaded feature extraction modules is the preliminary feature, the input of each other feature extraction module is the feature output by the previous feature extraction module, and the feature output by the last feature extraction module is the target feature. Any feature extraction module further extracts features from the input features, and performs attention calculation on the further extracted features, and merges the attention features obtained by the attention calculation with the input features.

4. The radar radiation source individual identification method according to claim 1, characterized in that: The process of extracting features at any level includes: Extract features from the feature extraction object at this level, where the features extracted from the feature extraction object at this level are multi-channel features; Determine, according to the multi-channel feature, an attention weight of each channel of the multi-channel feature; Using the attention weight of each channel of the multi-channel feature, weighting the corresponding channel of the multi-channel feature to obtain a first attention feature; Determining, according to the first attention feature, an attention weight for each spatial position of the first attention feature; Using the attention weight of each spatial position of the first attention feature, weighting the feature of each spatial position of the first attention feature to obtain a second attention feature; Fusing the second attention feature with the feature extraction object of this level to obtain a fused feature; The fused features are post-processed.

5. The method for identifying individual radar radiation sources according to claim 4, characterized in that: The step of determining the attention weight of each channel of the multi-channel feature according to the multi-channel feature comprises: Obtain the maximum feature and the average feature of each channel of the multi-channel feature; Determine a first attention weight for each channel of the multi-channel feature according to a maximum feature of each channel of the multi-channel feature, and determine a second attention weight for each channel of the multi-channel feature according to an average feature of each channel of the multi-channel feature; Fusing the first attention weight and the second attention weight of the same channel of the multi-channel feature to obtain a fused weight of each channel of the multi-channel feature; The fused weight of each channel of the multi-channel feature is normalized to obtain the final attention weight of each channel of the multi-channel feature.

6. The method for identifying individual radar radiation sources according to claim 4, characterized in that: The step of determining the attention weight of each spatial position of the first attention feature according to the first attention feature includes: Obtaining the maximum feature among the features at the same spatial position of each channel of the first attention feature, and obtaining the average value of the features at the same spatial position of each channel of the first attention feature, to obtain the maximum feature and the average feature of each spatial position of the first attention feature; Fusing the maximum feature and the average feature at the same spatial position of the first attention feature to obtain a fused feature at each spatial position of the first attention feature; generating an initial attention weight for each spatial position of the first attention feature according to the fused features of each spatial position of the first attention feature; The initial attention weight of each spatial position of the first attention feature is normalized to obtain the final attention weight of each spatial position of the first attention feature.

7. The method for identifying individual radar radiation sources according to claim 1, characterized in that: The step of determining the category of the radar radiation source individual that emits the target radar pulse signal according to the target feature includes: According to the target characteristics, predict the probability that the category of the radar radiation source individual emitting the target radar pulse signal is the set category of each radar radiation source individual; According to the predicted probability, the category of the individual radar emitter that emits the target radar pulse signal is determined from the set individual categories of the radar emitters.

8. A radar radiation source individual identification device, characterized in that: include: Radar pulse signal acquisition module, preliminary feature acquisition module, advanced feature acquisition module and radar radiation source individual category determination module; The radar pulse signal acquisition module is used to acquire the target radar pulse signal; The preliminary feature acquisition module is used to perform preliminary feature extraction on the target radar pulse signal to obtain preliminary features of the target radar pulse signal; The high-level feature acquisition module is used to perform step-by-step feature extraction on the preliminary features to obtain the target features of the target radar pulse signal; wherein the feature extraction object of the first level is the preliminary features, the feature extraction objects of each other level are the feature extraction results of the previous level, and the feature extraction result of the last level is the target feature. When performing feature extraction at any level, features are extracted from the feature extraction objects of that level, and attention calculation is performed on the extracted features, and the attention features obtained by the attention calculation are merged with the feature extraction objects of that level to obtain the feature extraction results of that level. The attention calculation is used to strengthen the features related to the radar radiation source individual identification task and suppress the features unrelated to the radar radiation source individual identification task; The individual radiation source category determination module is used to determine the category of the individual radar radiation source that emits the target radar pulse signal according to the target characteristics.

9. An electronic device, characterized in that: The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement the steps of the radar radiation source individual identification method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: The storage medium carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement the steps of the radar radiation source individual identification method as described in any one of claims 1 to 7.

Images