Ship radar image feature regulation and control method and system based on generative artificial intelligence

Abstract

The invention provides a ship radar image feature regulation and control method and system based on generative artificial intelligence, and the method comprises the steps: employing a regulation and control strategy system, employing a generative adversarial network architecture, employing a generator and a discriminator to carry out the discrimination and training iteration of a feature sample generated through regulation and control, introducing a generated variable through random noise, and carrying out the recognition and training iteration of the generated feature sample; the generated ship radar image features have the capabilities of simulating random real environment parameters and randomly observing apertures on the basis of analog simulation characteristics, and the finally output ship radar image features are highly close to real data, so that a radar detection and identification discriminator is difficult to distinguish, and an ideal feature regulation and control effect is achieved. The software and hardware regulation and control system based on the generative artificial intelligence ship radar image feature regulation and control strategy can realize an intelligent ship radar camouflage function, is used for modern ship design, improves the concealment, survivability and safety of ships, and has excellent engineering application value.

Description

Technical Field

[0001] The present invention relates to the technical field of ship artificial intelligence, and particularly to a method and system for regulating ship radar image features based on generative artificial intelligence. Background Art

[0002] In the modern ship confrontation radar sensor scenario, the observable features of ship radar images are a very important technical means. Modern radar sensors integrate long-distance detection, imaging, and recognition technologies. Radars equipped on vehicle platforms such as early warning aircraft, satellites, and ships can detect and identify target ships at long distances and use artificial intelligence algorithms to enhance the recognition effect. The radar detection and the affiliated precision strike system pose a great threat to the survival of ships, forcing more and more consideration of radar stealth characteristics design and real-time adjustable observable radar feature design in ship design.

[0003] Research status shows that in the current design of real-time adjustable observable radar features, researchers have paid more attention to real-time adjustable materials for radar wave reflection characteristics and have made progress in the design of various material characteristics such as radar wave reflection amplitude, reflection phase, and reflection polarization. Based on the idea of automatic regulation of radar observable features, the frequency adaptive radar wave absorption and scattering regulation functions have been realized, but there is still a considerable gap from intelligence and interaction with the recognition scenario. The modern electromagnetic feature confrontation conditions tend to be complex. The current adjustable radar feature design architecture is difficult to reach the level of adapting to the actual confrontation scenario and has not considered the interaction process with the detection and recognition algorithm. There is an urgent need to develop a regulation strategy system with the characteristics of generative artificial intelligence.

[0004] Artificial intelligence is a new technology science that studies, develops theories, methods, and technologies for simulating, extending, and expanding human intelligence, and has begun to be applied in multiple fields such as multi-domain battlefields, image recognition, and computer vision. Generative artificial intelligence is an important branch of the field of artificial intelligence, which generates new content with logic and learns to improve the generation strategy. Typical generative artificial intelligence models include probabilistic graphical models, generative adversarial networks, autoregressive generative models, diffusion models, etc. Currently, generative artificial intelligence has not been used in the design of real-time adjustable observable radar features. Summary of the Invention

[0005] The present invention provides a method and system for regulating ship radar image features based on generative artificial intelligence, aiming to solve the defect in the prior art that generative artificial intelligence is not applied to observable ship radar features, and to generate random ship radar image features according to input conditions, and significantly improve the authenticity level of the generated features through adversarial training with a discriminator, so as to ensure that the regulated features can achieve the effect that is difficult to distinguish by radar detection and recognition discriminators.

[0006] In a first aspect, the present invention provides a method for regulating the characteristics of ship radar images based on generative artificial intelligence, including: Obtain an electromagnetic feature training set of ship radar images, collect random noise signals and environmental parameters related to the electromagnetic feature training set, and generate real-time electromagnetic features; Construct an adversarial network for regulating the characteristics of ship radar images, where the adversarial network includes a feature generator and a feature discriminator; Input the real-time electromagnetic features into the feature generator and then output a generation result, and have the feature discriminator judge the generation result to obtain the characteristics of the ship radar image.

[0007] According to the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention, obtaining an electromagnetic feature training set of ship radar images includes: Determine that the electromagnetic feature training set includes different types of ship target templates, and each type of ship target template contains corresponding classification features and description parameters; Set the electromagnetic feature training set as the first number of parameters, and the electromagnetic feature training set also includes expandability and updateability.

[0008] According to the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention, collecting random noise signals and environmental parameters related to the electromagnetic feature training set includes: Collect random noise of different factors, determine the type of noise function according to the mechanism of each factor, perform noise superposition, and generate a total random noise signal, where the total random noise signal is the second number of parameters; Collect detection radar parameters and maritime environment parameters. The detection radar parameters are obtained by a radar wave environment perception module perceiving the radar detection scene, the detection radar parameters are the third number of parameters, and the maritime environment parameters determine the type of ship to which the generation template of the electromagnetic features belongs, and the maritime environment parameters are the fourth number of parameters.

[0009] According to the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention, generating real-time electromagnetic features includes: Generate real-time electromagnetic features that are based on the training set template and have randomness from the feature control input variables composed of the electromagnetic feature training set, the random noise signal, and the environmental parameters.

[0010] According to the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention, constructing an adversarial network for regulating the characteristics of ship radar images, where the adversarial network includes a feature generator and a feature discriminator, includes: Determine the types of regulatory features, determine the generation of the feature generator based on the types of regulatory features, and determine the discrimination process of the feature discriminator according to the main feature analysis method; The main framework of the feature generator includes a given parameter generation function, the given parameter generation function has parameter extensibility, and the generation function parameter of the given parameter generation function is the fifth number of parameters; The feature discriminator discriminates the electromagnetic features generated by the feature generator.

[0011] According to a method for regulating ship radar image features based on generative artificial intelligence provided by the present invention, after inputting the real-time electromagnetic features into the feature generator, the generated result is output, and the feature discriminator judges the generated result to obtain ship radar image features, including: If the judgment result output by the feature discriminator is false, then the feature generator regenerates electromagnetic features again according to the logical given, and the feature discriminator judges again until the judgment result is true, and the ship radar image features are output.

[0012] According to a method for regulating ship radar image features based on generative artificial intelligence provided by the present invention, it further includes: Use the preset number of training iterations to train and debug the adversarial network for regulating ship radar image features, so that the passing probability of the feature discriminator exceeds the preset threshold.

[0013] In a second aspect, the present invention further provides a system for regulating ship radar image features based on generative artificial intelligence, including: A generation module, configured to obtain an electromagnetic feature training set of a ship radar image, collect random noise signals and environmental parameters related to the electromagnetic feature training set, and generate real-time electromagnetic features; A construction module, configured to construct an adversarial network for regulating ship radar image features, and the adversarial network includes a feature generator and a feature discriminator; A regulation module, configured to input the real-time electromagnetic features into the feature generator and then output the generated result, and the feature discriminator judges the generated result to obtain ship radar image features.

[0014] In a third aspect, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the method for regulating ship radar image features based on generative artificial intelligence as described in any one of the above.

[0015] Fourthly, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the method for regulating the characteristics of ship radar images based on generative artificial intelligence as described in any one of the above.

[0016] The method and system for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention utilize a generative adversarial network composed of a feature generator and a feature discriminator, and train the feature regulation generation system with an electromagnetic feature training set, significantly improving the authenticity of the generated electromagnetic features. Then, the regulatory strategy system architecture has a complete set of parameter inputs, ensuring the richness of the generated characteristics and the regulatory change space. Finally, through each simulation training and actual use, the feature regulation strategy system continuously adjusts, self-learns, and self-optimizes the feature generation strategy, featuring self-learning and self-evolution, achieving a leapfrog development from adaptability to artificial intelligence functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 is a schematic flowchart of the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention; Figure 2 is a schematic architecture diagram of the generative artificial intelligence ship radar image feature regulation strategy system provided by the present invention; Figure 3 is a schematic module structure diagram of the electromagnetic feature training set provided by the present invention; Figure 4 is a schematic module structure diagram of the environmental parameter input provided by the present invention; Figure 5 is a schematic structural diagram of the system for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the present invention; Figure 6 is a schematic structural diagram of the electronic device provided by the present invention.

[0019] REFERENCE SIGNS: A1: Electromagnetic feature training set; A2: Random noise signal; A3: Environmental parameter input; A4: Feature generator; A5: Feature discriminator; A6: Detection radar parameters; A7: Maritime environmental parameters. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0021] Aiming at the deficiencies in the existing technology, the present invention proposes a method for regulating the characteristics of ship radar images based on generative artificial intelligence. Based on the system architecture of the generative artificial intelligence ship radar image feature regulation strategy, it provides a typical principle architecture and design implementation idea with application value for the artificial intelligence radar stealth and camouflage system, thereby supporting the design of high-performance camouflage skins for modern equipment and the design of radar targets for training.

[0022] Figure 1 It is a schematic flowchart of the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the embodiments of the present invention. As Figure 1 shown, it includes: Step 100: Obtain the electromagnetic feature training set of the ship radar image, collect the random noise signals and environmental parameters related to the electromagnetic feature training set, and generate real-time electromagnetic features; Step 200: Construct an adversarial network for regulating the characteristics of ship radar images, and the adversarial network includes a feature generator and a feature discriminator; Step 300: Input the real-time electromagnetic features into the feature generator and then output the generated result, and the feature discriminator judges the generated result to obtain the characteristics of the ship radar image.

[0023] Specifically, the system architecture of the generative artificial intelligence ship radar image feature regulation strategy proposed by the embodiments of the present invention consists of 4 main modules, including the ship electromagnetic feature training set, the parameter input module, the electromagnetic feature generator, and the feature discriminator. The 4 modules generate random ship radar image features according to the given logical relationship. Specifically, it includes the ship electromagnetic feature training set (A1), the random noise signal (A2), the environmental parameter input (A3), the electromagnetic feature generator (A4), the electromagnetic feature discriminator (A5), and the logical relationship followed by the module operation.

[0024] The electromagnetic feature training set (A1) contains the data of the electromagnetic feature training library of ship-like targets that have been pre-verified. The data of the electromagnetic feature training library is determined by multiple parameters, including radar wave frequency, electromagnetic wave polarization direction, observation angle, etc. The electromagnetic feature training set contains the electromagnetic features of multiple different types of ships for backup.

[0025] The random noise signal (A2) contains batch differences and working state differences of ships of the same type, the actual attitude swing error of the ship caused by sea waves, the non-ideality of the observation scene, the randomness of the radar aperture, etc., ensuring that the features generated each time have a certain degree of difference, which is also the source of feature authenticity. If the features adjusted each time are highly consistent, it is easy to cause abnormalities in the feature discriminator, and the determination that the features are not naturally generated or the traces of artificial control are too obvious, resulting in the failure of feature simulation.

[0026] The environmental parameter input (A3) contains the detection radar parameters (A6) and the maritime environmental parameters (A7). In the actual working system, the detection radar parameters (A6) are the working parameters of the detection radar given by the radar wave environment perception module, including frequency, incoming wave angle, etc. The maritime environmental parameters (A7) can obtain the types and quantities of navigation ships in the nearby sea area from the global ship trajectory query website, and give the types of camouflage target ships with strong generality, large quantity, and relatively reasonable running trajectories.

[0027] The electromagnetic feature generator (A4) and the electromagnetic feature training set (A1) together generate real-time electromagnetic features based on the training set template with a certain degree of randomness according to the environmental parameter input given by the random noise signal (A2) and the environmental parameter input (A3).

[0028] The electromagnetic feature discriminator (A5) discriminates the electromagnetic features generated by the feature generator, and the discrimination basis can include the statistical characteristics, randomness, etc. of the electromagnetic features. Commonly used radar image statistical characteristics include similarity, correlation coefficient, feature point amplitude, feature point spacing, randomness index, etc. If the judgment result given by the feature discriminator (A5) is false, the logic gives the electromagnetic features to be generated again by the feature generator, and then judged by the feature discriminator until the judgment result is true and the output is completed.

[0029] The present invention is based on the system architecture of the generative artificial intelligence ship radar image feature regulation strategy, generates ship radar image features with randomness according to the input conditions, and significantly improves the authenticity level of the generated features through adversarial training with the discriminator, so as to ensure that the regulated features achieve the effect that is difficult to distinguish by radar detection and recognition discriminators.

[0030] In one embodiment, obtaining the electromagnetic feature training set of the ship radar image includes: Determining that the electromagnetic feature training set includes different types of ship target templates, and each type of ship target template contains corresponding classification features and description parameters; Setting the electromagnetic feature training set as the first parameter quantity, and the electromagnetic feature training set also includes expandability and updateability.

[0031] Specifically, after the preliminary designs of the electromagnetic feature generator (A4) and the electromagnetic feature discriminator (A5) are completed in the embodiments of the present invention, the main data structure of the electromagnetic feature training set (A1) and the logical relationship between the control variable parameters are then determined. The electromagnetic feature training set (A1) needs to load the data of the electromagnetic feature training library. The electromagnetic features include scattering features, radar image features, etc. The unit of the scattering feature can be dBsm (dB square meter), and the typical reference range of the scattering feature is 0 dBsm to 70 dBsm. The data of the electromagnetic feature training library needs to be pre-verified by means such as simulation and actual measurement before being loaded into the training library. The training library data needs to contain various types of ship target templates, such as Figure 3 as shown, for example, common military ships, commercial transport ships, commercial yachts, etc. Each type of ship target template is described by its corresponding classification features and parameters. The number of control parameters of the electromagnetic feature training library data can vary from 2 to 50, including ship type, model, observed radar wave frequency, electromagnetic wave polarization direction, observation angle, etc. The electromagnetic feature training set stores the electromagnetic features of multiple types of ships and has good scalability to update the database to accommodate more ship type templates.

[0032] In one embodiment, random noise signals and environmental parameters related to the electromagnetic feature training set are collected, including: Collect random noise of different factors, determine the type of noise function according to the mechanism of each factor, and perform noise superposition to generate a total random noise signal, where the total random noise signal is the number of second parameters; Collect detection radar parameters and maritime environment parameters. The detection radar parameters are obtained by the radar wave environment perception module perceiving the radar detection scene. The detection radar parameters are the number of third parameters, and the maritime environment parameters determine the ship type to which the electromagnetic feature generation template belongs. The maritime environment parameters are the number of fourth parameters.

[0033] Specifically, in the design of the generation process of the random noise signal (A2) in the embodiments of the present invention, it is necessary to include random noises caused by various factors, including batch differences and working state differences of ships of the same model, the actual attitude swing error of the ship caused by sea waves, the non-ideality of the observation scene, the randomness of the radar aperture, etc. The type of noise function is determined according to the mechanism of each factor, and then noise superposition is performed to generate the total random noise signal, which is output to the electromagnetic feature generator (A4). The random noise signal (A2) can ensure that the features generated each time have a certain degree of difference, and effective feature authenticity is generated under repeated observation conditions. The amplitude and superposition method of the random noise signal (A2) can be optimized during multiple trainings to adapt to the discrimination process of the electromagnetic feature discriminator (A5). The ideal effect is that the determined feature of the electromagnetic feature discriminator (A5) is naturally generated, that is, the result is true. The typical value of the number of random noise signal parameters is from 2 to 10.

[0034] In the design of the generation process of the environmental parameter input (A3), it is necessary to include the detection radar parameters (A6) and the maritime environmental parameters (A7). As Figure 4 shown, in the actual radar detection scene where the feature regulation module works, the detection radar parameters (A6) are sensed by the radar wave environment perception module for the radar detection scene, and the working parameters of the detection radar are given, including frequency, incoming wave angle, etc. The electromagnetic features corresponding to different detection radar parameters may have very large differences, and this difference is determined by the electromagnetic feature training set. Therefore, it is necessary to determine the generation template of the electromagnetic features according to the detection radar parameters. The reference range of the number of detection radar parameters is from 3 to 30.

[0035] The maritime environmental parameters (A7) are mainly used to determine the type of ship to which the generation template of the electromagnetic features belongs. According to the seasonal and regional laws of ship operation, the type of ship corresponding to the generation template of the electromagnetic features should be selected as the type with a lower degree of attention arousal. For example, ship templates with a large number and a strong association with regional operating ships are less likely to attract the attention of observers, so as to achieve the determination feature of the electromagnetic feature discriminator (A5) being naturally generated. The maritime environmental parameters can have a real-time update input interface, access the types and quantities of navigation ships in the nearby sea area obtained from sources such as the global ship trajectory query website, and give the types of camouflage target ships with strong generality, large quantity, and relatively reasonable operation trajectories. The reference range of the number of maritime environmental parameters is from 1 to 15.

[0036] In one embodiment, during the feature generation process of the electromagnetic feature generator (A4), the feature control input variables given by the electromagnetic feature training set (A1), the random noise signal (A2), and the environmental parameter input (A3) are used to generate real-time electromagnetic features based on the training set template with a certain degree of randomness.

[0037] In one embodiment, an adversarial network for regulating the characteristics of ship radar images is constructed. The adversarial network includes a feature generator and a feature discriminator, and includes: Determine the types of regulated features, determine the generation of the feature generator based on the types of regulated features, and determine the discrimination process of the feature discriminator according to the main feature analysis method; The main framework of the feature generator includes a given-parameter generation function, which has parameter extensibility, and the generation function parameter of the given-parameter generation function is the fifth parameter quantity; The feature discriminator discriminates the electromagnetic features generated by the feature generator.

[0038] Specifically, in the embodiment of the present invention, the central module of the generative artificial intelligence ship radar image feature regulation strategy system architecture is an adversarial network formed by an electromagnetic feature generator (A4) and an electromagnetic feature discriminator (A5). First, the types of regulated features should be determined. Typical features include electromagnetic scattering features, infrared features, electric field features, magnetic field features, wake electromagnetic features, etc. According to the regulated features, determine the generation method of the feature generator (A4), and determine the discrimination method and process of the electromagnetic feature discriminator (A5) according to the common main feature analysis method.

[0039] Furthermore, the main framework of the electromagnetic feature generator (A4) is a generation function with given parameters, and the function has parameter extensibility. The generation function parameters are jointly determined by the electromagnetic feature training set (A1), the random noise signal (A2), and the environmental parameter input (A3), and the typical parameter quantity range is from 4 to 60.

[0040] In one embodiment, after inputting the real-time electromagnetic features into the feature generator, the generated result is output, and the feature discriminator judges the generated result to obtain the ship radar image features, including: If the judgment result output by the feature discriminator is false, then according to the logical given, the feature generator regenerates the electromagnetic features again, and the feature discriminator judges again until the judgment result is true, and the ship radar image features are output.

[0041] Specifically, in the embodiment of the present invention, the electromagnetic feature discriminator (A5) is used to discriminate the electromagnetic features generated by the feature generator. The discrimination basis may include the statistical characteristics, randomness, etc. of the electromagnetic features. Common radar image statistical characteristics include similarity, correlation coefficient, feature point amplitude, feature point spacing, randomness index, etc. If the judgment result given by the feature discriminator (A5) is false, then the logical given causes the feature generator to generate electromagnetic features again, and the feature discriminator judges again until the judgment result is true and the output is completed.

[0042] In one embodiment, it further includes: Train and debug the adversarial network for regulating the features of ship radar images using a preset number of training iterations, so that the passing probability of the feature discriminator exceeds a preset threshold.

[0043] Specifically, after the implementation software of the generative artificial intelligence ship radar image feature regulation strategy system architecture is completed, the regulation strategy software needs to be trained and debugged. Through training and debugging, if the passing probability of the electromagnetic feature discriminator (A5) exceeds a relatively high threshold, the architecture software meets the practical conditions. In the embodiments of the present invention, the reference range of the number of training iterations is usually from 10 to 100,000 times, and the reference threshold range is usually from 80% to 100%.

[0044] It can be understood that the electromagnetic feature regulation system generally operates in a fully automatic mode and generally does not adopt the manual discrimination mode, because the manual discrimination mode consumes a long time, which will cause a significant delay in the regulation strategy and is relatively easy to expose its true features. In the training stage, the manual discrimination mode can be adopted to improve the authenticity and effectiveness of the radar image feature regulation strategy system.

[0045] The generative artificial intelligence ship radar image feature regulation strategy system architecture in the present invention provides a basic architecture for the radar image feature regulation software. The regulation computer software can be written, debugged, packaged according to this architecture and then put into actual use. The regulation computer software, together with the sensing hardware module, the regulation hardware module, and the feature regulation materials, serves as the main modules to form the radar image feature regulation system. The radar image feature regulation system can be integrated into the ship platform for radar feature regulation during navigation, improving the stealth, survivability, and safety of the ship.

[0046] Therefore, for the ship radar image feature regulation method based on generative artificial intelligence provided by the present invention, the regulation strategy system uses the generative adversarial network architecture, uses the generator and the discriminator to discriminate and train iteratively the feature samples generated by the regulation. By introducing random noise into the generated variables, the generated ship radar image features have the ability to simulate random real environment parameters and random observation apertures on the basis of the simulation characteristics. Finally, the output ship radar image features are highly close to the real data, making it difficult for the radar detection and recognition discriminator to distinguish, achieving an ideal feature regulation effect. The software and hardware regulation system based on the generative artificial intelligence ship radar image feature regulation strategy realizes the intelligent ship radar camouflage function, is used in modern ship design, improves the stealth, survivability, and safety of the ship, and has excellent engineering application value.

[0047] The ship radar image feature regulation system based on generative artificial intelligence provided by the present invention will be described below. The ship radar image feature regulation system based on generative artificial intelligence described below can be correspondingly referred to the ship radar image feature regulation method based on generative artificial intelligence described above.

[0048] Figure 5 It is a schematic structural diagram of the ship radar image feature regulation system based on generative artificial intelligence provided by an embodiment of the present invention. As Figure 5 shown, it includes: a generation module 51, a construction module 52, and a regulation module 53, wherein: The generation module 51 is used to obtain an electromagnetic feature training set of ship radar images, collect random noise signals and environmental parameters related to the electromagnetic feature training set, and generate real-time electromagnetic features; the construction module 52 is used to construct an adversarial network for ship radar image feature regulation, and the adversarial network includes a feature generator and a feature discriminator; the regulation module 53 is used to input the real-time electromagnetic features into the feature generator and then output a generation result, and the feature discriminator judges the generation result to obtain ship radar image features.

[0049] Figure 6 It exemplifies a schematic structural diagram of an electronic device. As Figure 6 shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640. Among them, the processor 610, the communication interface 620, and the memory 630 complete mutual communication through the communication bus 640. The processor 610 can call the logical instructions in the memory 630 to execute the ship radar image feature regulation method based on generative artificial intelligence, and the method includes: obtaining an electromagnetic feature training set of ship radar images, collecting random noise signals and environmental parameters related to the electromagnetic feature training set, and generating real-time electromagnetic features; constructing an adversarial network for ship radar image feature regulation, and the adversarial network includes a feature generator and a feature discriminator; inputting the real-time electromagnetic features into the feature generator and then outputting a generation result, and the feature discriminator judges the generation result to obtain ship radar image features.

[0050] In addition, when the logical instructions in the above-mentioned memory 630 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0051] On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is used to implement the method for regulating the characteristics of ship radar images based on generative artificial intelligence provided by the above-mentioned various methods. The method includes: obtaining an electromagnetic feature training set of ship radar images, collecting random noise signals and environmental parameters related to the electromagnetic feature training set, and generating real-time electromagnetic features; constructing an adversarial network for regulating the characteristics of ship radar images, where the adversarial network includes a feature generator and a feature discriminator; inputting the real-time electromagnetic features into the feature generator and then outputting a generation result, and having the feature discriminator judge the generation result to obtain the characteristics of the ship radar image.

[0052] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. One can select some or all of the modules according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0053] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for regulating the characteristics of ship radar images based on generative artificial intelligence, characterized in that, Including: Obtain an electromagnetic feature training set of ship radar images, collect random noise signals and environmental parameters related to the electromagnetic feature training set, and generate real-time electromagnetic features; Construct an adversarial network for regulating ship radar image features, where the adversarial network includes a feature generator and a feature discriminator; Input the real-time electromagnetic features into the feature generator and then output a generation result, and have the feature discriminator judge the generation result to obtain ship radar image features.

2. The method for regulating the characteristics of ship radar images based on generative artificial intelligence according to claim 1, wherein, Obtain an electromagnetic feature training set of ship radar images, including: Determine that the electromagnetic feature training set includes different types of ship target templates, and each type of ship target template contains corresponding classification features and description parameters; Set the electromagnetic feature training set to a first parameter quantity, and the electromagnetic feature training set also includes expandability and updateability.

3. The method for regulating the characteristics of ship radar images based on generative artificial intelligence according to claim 2, characterized in that, Collect random noise signals and environmental parameters related to the electromagnetic feature training set, including: Collect random noise of different factors, determine the type of noise function according to the mechanism of each factor, perform noise superposition, and generate a total random noise signal, where the total random noise signal is a second parameter quantity; Collect detection radar parameters and maritime environment parameters, where the detection radar parameters are obtained by a radar wave environment perception module perceiving the radar detection scene, the detection radar parameters are a third parameter quantity, and the maritime environment parameters determine the ship type to which the generation template of the electromagnetic features belongs, and the maritime environment parameters are a fourth parameter quantity.

4. The method for regulating ship radar image features based on generative artificial intelligence according to claim 3, wherein, Generate real-time electromagnetic features, including: Use the feature control input variables composed of the electromagnetic feature training set, the random noise signal, and the environmental parameters to generate real-time electromagnetic features based on the training set template and with randomness.

5. The method for regulating ship radar image features based on generative artificial intelligence according to claim 1, wherein Construct an adversarial network for regulating ship radar image features, where the adversarial network includes a feature generator and a feature discriminator, including: Determine the types of regulated features, determine the generation of the feature generator based on the types of regulated features, and determine the discrimination process of the feature discriminator according to the main feature analysis method; The main framework of the feature generator includes a given parameter generation function, the given parameter generation function has parameter expandability, and the generation function parameters of the given parameter generation function are a fifth parameter quantity; The feature discriminator discriminates the electromagnetic features generated by the feature generator.

6. The method for regulating the characteristics of ship radar images based on generative artificial intelligence according to claim 1, wherein Input the real-time electromagnetic features into the feature generator and then output a generation result, and have the feature discriminator judge the generation result to obtain ship radar image features, including: If the judgment result output by the feature discriminator is false, then regenerate the electromagnetic features by the feature generator again according to the logical given, and have the feature discriminator judge again until the judgment result is true, and output the ship radar image features.

7. The method for regulating the characteristics of ship radar images based on generative artificial intelligence according to claim 1, characterized in that, Also include: Use a preset number of training iterations to train and debug the adversarial network for regulating ship radar image features, so that the passing probability of the feature discriminator exceeds a preset threshold.

8. A ship radar image feature regulation system based on generative artificial intelligence, characterized in that, Including: A generation module for obtaining an electromagnetic feature training set of ship radar images, collecting random noise signals and environmental parameters related to the electromagnetic feature training set, and generating real-time electromagnetic features; A building module for building an adversarial network for regulating the characteristics of ship radar images, the adversarial network including a feature generator and a feature discriminator; A regulation module for inputting the real-time electromagnetic features into the feature generator and then outputting a generation result, and the feature discriminator judges the generation result to obtain the characteristics of the ship radar image.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein When the processor executes the program, it implements the method for regulating the characteristics of ship radar images based on generative artificial intelligence according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method for regulating the characteristics of ship radar images based on generative artificial intelligence according to any one of claims 1 to 7.

Images