Model training method and shape design method for designing shape of stealth material

Abstract

The invention provides a model training method and a shape design method for designing the shape of an invisible material, which can be applied to the technical field of computers. The method comprises the following steps: inputting a plurality of two-dimensional coordinate points of a to-be-hidden sample projected on a horizontal plane into an initial shape design model to obtain respective normal vectors of the plurality of two-dimensional coordinate points; for each two-dimensional coordinate point, determining a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point; based on a preset radar direction, the three-dimensional coordinate points corresponding to the plurality of two-dimensional coordinate points and the normal vectors of the plurality of two-dimensional coordinate points, predicting an initial imaging result of the millimeter wave radar on the to-be-stealthy sample covered with the stealth material; obtaining a target loss value based on the initial imaging result, the background representation data and a preset height threshold value for the stealth material; and training the initial shape design model based on the target loss value to obtain a target shape design model.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a model training method, device, equipment, medium and program product for designing the shape of a stealth material, and a stealth material shape design method, device, equipment, medium and program product. Background Art

[0002] In recent years, the shape design technology of stealth materials has developed rapidly and has shown important application prospects in many fields. For example, in the field of aerospace, advanced stealth materials can significantly reduce the radar cross section of aircraft and improve their survivability. In the field of communications, specially designed stealth materials can reduce electromagnetic interference and improve signal transmission quality. In the civilian field, stealth materials can also be used for electromagnetic shielding of buildings to protect privacy and sensitive information.

[0003] However, with the popularization of millimeter-wave radar imaging systems, stealth material shape design technology faces new challenges. Since millimeter waves have very short wavelengths and can provide higher resolution and finer target details, traditional stealth material shape design methods are often powerless when facing such high-frequency electromagnetic waves. In addition, stealth material design methods based on optimization algorithms in related technologies can usually only handle simple geometric shapes and have poor stealth effects on complex three-dimensional targets. Summary of the invention

[0004] In view of the above problems, the present invention provides a model training method, device, equipment, medium and program product for designing the shape of stealth materials, and a stealth material shape design method, device, equipment, medium and program product.

[0005] According to one aspect of the present invention, a model training method for designing the shape of a stealth material is provided, comprising: inputting a plurality of two-dimensional coordinate points of a sample to be hidden projected on a horizontal plane into an initial shape design model to obtain normal vectors of each of the plurality of two-dimensional coordinate points, wherein the normal vector defines the surface geometry of the stealth material; for each two-dimensional coordinate point, determining a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points; predicting initial imaging results of a millimeter-wave radar for a sample to be hidden covered with a stealth material based on a preset radar direction, the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points, and the normal vectors of the plurality of two-dimensional coordinate points; obtaining a target loss value based on the initial imaging result, background characterization data, and a preset height threshold for the stealth material; and training the initial shape design model based on the target loss value to obtain a target shape design model.

[0006] Another aspect of the present invention provides a method for designing the shape of a stealth material, comprising: inputting multiple two-dimensional test coordinate points of the object to be invisible projected on a horizontal plane into a target shape design model to obtain normal vectors of each of the multiple two-dimensional test coordinate points; determining three-dimensional test coordinate points corresponding to each of the multiple two-dimensional test coordinate points based on the normal vectors of each of the multiple two-dimensional test coordinate points and the multiple two-dimensional test coordinate points; determining the shape of the stealth material based on the three-dimensional test coordinate points corresponding to each of the multiple two-dimensional test coordinate points.

[0007] Another aspect of the present invention provides a model training device for designing the shape of a stealth material, comprising: a first input module, used to input multiple two-dimensional coordinate points of the sample to be hidden projected on a horizontal plane into an initial shape design model to obtain normal vectors of each of the multiple two-dimensional coordinate points, wherein the normal vector defines the surface geometry of the stealth material; a first determination module, used to determine, for each two-dimensional coordinate point, a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points; a result prediction module, used to predict the initial imaging result of the millimeter wave radar on the sample to be hidden covered with the stealth material based on a preset radar direction, the three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points, and the normal vectors of each of the multiple two-dimensional coordinate points; a loss value determination module, used to obtain a target loss value based on the initial imaging result, background characterization data, and a preset height threshold for the stealth material; a training module, used to train the initial shape design model based on the target loss value to obtain a target shape design model.

[0008] Another aspect of the present invention provides a device for designing the shape of a stealth material, comprising: a second input module, used to input multiple two-dimensional test coordinate points of the object to be hidden projected on a horizontal plane into a target shape design model to obtain the normal vectors of each of the multiple two-dimensional test coordinate points; a second determination module, used to determine the three-dimensional test coordinate points corresponding to each of the multiple two-dimensional test coordinate points based on the normal vectors of each of the multiple two-dimensional test coordinate points and the multiple two-dimensional test coordinate points; and a shape determination module, used to determine the shape of the stealth material based on the three-dimensional test coordinate points corresponding to each of the multiple two-dimensional test coordinate points.

[0009] Another aspect of the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the above method.

[0010] Another aspect of the present invention further provides a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions implement the steps of the above method when executed by a processor.

[0011] Another aspect of the present invention further provides a computer program product, including a computer program or instructions, which implement the steps of the above method when executed by a processor.

[0012] According to the model training method for designing the shape of stealth materials of the present invention, a plurality of two-dimensional coordinate points representing the projection of the sample to be hidden on the horizontal plane are input into the initial shape design model to obtain the normal vector of each two-dimensional coordinate point, that is, the definition of the shape of the stealth material for the shape of the sample to be hidden is realized through the initial shape design model, and the two-dimensional coordinate point is converted into a three-dimensional coordinate point through the normal vector to determine the shape of the stealth material, so that the initial imaging result of the sample to be hidden covered with the stealth material by the millimeter wave radar can be predicted based on the plurality of three-dimensional coordinate points, the normal vectors of the plurality of two-dimensional coordinate points and the preset radar direction, and the target loss value is calculated through the initial imaging result, the target loss value is calculated, and the initial shape design model is trained based on the target loss value, so that the imaging result of the constraint model supervised by the initial imaging result is realized, so that the initial shape design model is optimized to obtain the target shape design model, so that a more effective stealth material shape can be obtained through the target shape design model, at least partially solving the technical problem of poor stealth effect for three-dimensional targets existing in the related art, and realizing the customization of the stealth material shape and effective stealth of samples to be hidden of any shape. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above contents and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0014] Figure 1 A diagram showing an application scenario of a model training method, apparatus, device, medium and program product for designing the shape of a stealth material and a stealth material shape design method, apparatus, device, medium and program product according to an embodiment of the present invention.

[0015] Figure 2 A flow chart of a model training method for designing the shape of stealth materials according to an embodiment of the present invention is shown.

[0016] Figure 3 A schematic diagram of determining an initial imaging result according to an embodiment of the present invention is shown.

[0017] Figure 4 A flow chart of a method for designing a shape of a stealth material according to an embodiment of the present invention is shown.

[0018] Figure 5 A first schematic diagram of a stealth material shape design method according to an embodiment of the present invention is shown.

[0019] Figure 6 A second schematic diagram of the stealth material shape design method according to an embodiment of the present invention is shown.

[0020] Figure 7 A first stealth effect schematic diagram of a stealth material shape design method according to an embodiment of the present invention is shown.

[0021] Figure 8 A second stealth effect schematic diagram of the stealth material shape design method according to an embodiment of the present invention is shown.

[0022] Fig. 9 A third stealth effect schematic diagram of the stealth material shape design method according to an embodiment of the present invention is shown.

[0023] Fig.10 A structural block diagram of a model training device for designing the shape of stealth materials according to an embodiment of the present invention is shown.

[0024] Fig.11 A structural block diagram of a stealth material shape design device according to an embodiment of the present invention is shown.

[0025] Fig.12 A block diagram of an electronic device suitable for implementing a method for designing a stealth material shape and a stealth material shape designing method according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0026] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0027] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0028] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0029] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0030] It should be noted that the model training method for designing the shape of stealth materials and the stealth material shape design method of the present invention can be used in the field of computer technology, and can also be used in any field other than the field of computer technology, such as the field of stealth material design technology, etc. The present invention does not limit the application field of the model training method for designing the shape of stealth materials and the stealth material shape design method.

[0031] An embodiment of the present invention provides a model training method for designing the shape of a stealth material, comprising: inputting a plurality of two-dimensional coordinate points of a sample to be hidden projected on a horizontal plane into an initial shape design model to obtain normal vectors of each of the plurality of two-dimensional coordinate points, wherein the normal vector defines the surface geometry of the stealth material; for each two-dimensional coordinate point, determining a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points; predicting the initial imaging result of a millimeter-wave radar on a sample to be hidden covered with a stealth material based on a preset radar direction, the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points, and the normal vectors of the plurality of two-dimensional coordinate points; obtaining a target loss value based on the initial imaging result, background characterization data, and a preset height threshold for the stealth material; and training the initial shape design model based on the target loss value to obtain a target shape design model.

[0032] Figure 1 A diagram schematically shows an application scenario of a model training method, apparatus, device, medium and program product for designing the shape of a stealth material and a stealth material shape design method, apparatus, device, medium and program product according to an embodiment of the present invention.

[0033] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0034] The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0035] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0036] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0037] It should be noted that the model training method for designing the shape of stealth materials and the stealth material shape design method provided in the embodiment of the present invention can generally be executed by the server 105. Accordingly, the model training device for designing the shape of stealth materials and the stealth material shape design device provided in the embodiment of the present invention can generally be set in the server 105. The model training method for designing the shape of stealth materials and the stealth material shape design method provided in the embodiment of the present invention can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Correspondingly, the model training device for designing the shape of stealth materials and the stealth material shape design device provided in the embodiment of the present invention can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0038] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0039] The following will be based on Figure 1 The scene described by Figure 2~Figure 9The model training method for designing the shape of a stealth material and the stealth material shape design method according to an embodiment of the invention are described in detail.

[0040] Figure 2 The flowchart of the model training method for designing the shape of stealth materials according to an embodiment of the present invention is schematically shown.

[0041] like Figure 2 As shown, the method includes operations S210 to S250.

[0042] In operation S210, a plurality of two-dimensional coordinate points of the sample to be cloaked projected on a horizontal plane are input into an initial shape design model to obtain normal vectors of the plurality of two-dimensional coordinate points, wherein the normal vectors define the surface geometry of the cloaking material.

[0043] In operation S220, for each two-dimensional coordinate point, a three-dimensional coordinate point corresponding to the two-dimensional coordinate point is determined based on a normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points.

[0044] In operation S230, an initial imaging result of a millimeter wave radar on a sample to be cloaked covered with a cloaking material is predicted based on a preset radar direction, three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points, and normal vectors of each of the plurality of two-dimensional coordinate points.

[0045] In operation S240, a target loss value is obtained based on the initial imaging result, the background characterization data, and a preset height threshold for the stealth material.

[0046] In operation S250, the initial shape design model is trained based on the target loss value to obtain a target shape design model.

[0047] According to the embodiment of the present invention, there is no limitation on the sample to be invisible, and it can be an object of any shape.

[0048] According to an embodiment of the present invention, the horizontal plane may be an xy plane.

[0049] According to an embodiment of the present invention, a projection surface can be obtained after the sample to be invisible is projected on the xy plane, and a plurality of two-dimensional coordinate points can be obtained based on the coordinate points included in the projection surface.

[0050] According to an embodiment of the present invention, there is no limitation on the implementation method of the initial shape design model and the target shape design model, and they can be implemented using a neural network, such as a multilayer perceptron (MLP), a feedforward neural network (FNN), etc.

[0051] According to an embodiment of the present invention, normal vectors corresponding to each of the plurality of two-dimensional coordinate points define the surface geometry of the stealth material shape, which is crucial for operating radar reflection.

[0052] According to an embodiment of the present invention, by inputting multiple two-dimensional coordinate points representing the projection of the sample to be invisible on the horizontal plane into the initial shape design model to obtain the normal vectors of the multiple two-dimensional coordinate points, the shape of the stealth material of each sample to be invisible can be customized to obtain the stealth material shape for the sample to be invisible.

[0053] According to an embodiment of the present invention, a three-dimensional coordinate point of the two-dimensional coordinate point may be determined by using a normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point itself.

[0054] According to an embodiment of the present invention, the abscissa and ordinate included in the three-dimensional coordinate point corresponding to the two-dimensional coordinate point may be the same as the abscissa and ordinate of the two-dimensional coordinate point.

[0055] According to an embodiment of the present invention, the shape of the stealth material can be obtained by fitting three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points, that is, multiple three-dimensional coordinate points.

[0056] According to an embodiment of the present invention, the reflection situation of the millimeter-wave radar on the sample to be invisible and covered with stealth material can be predicted by presetting the radar direction and the normal vectors of each of the multiple two-dimensional coordinate points, and the initial imaging result can be obtained by the reflection situation and the predicted imaging situation of each three-dimensional coordinate point by the millimeter-wave radar, wherein the reflection situation can be characterized by the reflectivity, and the imaging situation of each three-dimensional coordinate point by the millimeter-wave radar can be characterized by the preset imaging result of the target corresponding to each three-dimensional coordinate point.

[0057] According to an embodiment of the present invention, the background characterization data is not limited and may be the same as the imaging result when the millimeter wave radar does not detect any object to be detected. The background characterization data may be a preset value, and different millimeter wave radars may correspond to different preset background characterization data.

[0058] According to an embodiment of the present invention, the target loss function for the initial shape design model can be determined through the initial imaging results, background characterization data and a preset height threshold, and when it is determined that the target loss function has not converged, the parameters of the initial shape design model are iteratively adjusted until it is determined that the target loss function has converged, and the initial shape design model corresponding to the current target loss function is determined as the target shape design model.

[0059] According to an embodiment of the present invention, training and optimizing the initial shape design model to obtain the target shape design model may be a self-supervised learning method, thereby being able to optimize the shape design of the stealth material even without labeled data.

[0060] According to the model training method for designing the shape of stealth materials of the present invention, a plurality of two-dimensional coordinate points representing the projection of the sample to be hidden on the horizontal plane are input into the initial shape design model to obtain the normal vector of each two-dimensional coordinate point, that is, the definition of the shape of the stealth material for the shape of the sample to be hidden is realized through the initial shape design model, and the two-dimensional coordinate point is converted into a three-dimensional coordinate point through the normal vector to determine the shape of the stealth material, so that the initial imaging result of the sample to be hidden covered with the stealth material by the millimeter wave radar can be predicted based on the plurality of three-dimensional coordinate points, the normal vectors of the plurality of two-dimensional coordinate points and the preset radar direction, and the target loss value is calculated through the initial imaging result, the target loss value is calculated, and the initial shape design model is trained based on the target loss value, so that the imaging result of the constraint model supervised by the initial imaging result is realized, so that the initial shape design model is optimized to obtain the target shape design model, so that a more effective stealth material shape can be obtained through the target shape design model, at least partially solving the technical problem of poor stealth effect for three-dimensional targets existing in the related art, and realizing the customization of the stealth material shape and effective stealth of samples to be hidden of any shape.

[0061] According to an embodiment of the present invention, based on a preset radar direction, three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points, and the normal vectors of each of the multiple two-dimensional coordinate points, predicting the initial imaging result of the millimeter wave radar on a sample to be hidden and covered with a hidden material may include the following operations.

[0062] For each three-dimensional coordinate point, a preset imaging result of a target corresponding to the three-dimensional coordinate point is determined from a plurality of preset imaging results, wherein the preset imaging result is determined based on the response of the millimeter-wave radar to the point to be measured in space, and the radar emission direction of the millimeter-wave radar is the preset radar direction; based on the normal vector of the two-dimensional coordinate point corresponding to the three-dimensional coordinate point, the preset radar direction and the preset imaging result of the target, an initial imaging result for the three-dimensional coordinate point is determined; and the initial imaging results of each of the plurality of three-dimensional coordinate points are superimposed to obtain an initial imaging result for a sample to be invisible and covered with a stealth material.

[0063] According to an embodiment of the present invention, the preset imaging result is determined based on the response of the millimeter-wave radar to the point to be measured in the space, and each preset imaging result may correspond to a point to be measured in the space. Specifically, when the imaging target, i.e., the object to be detected, is located and only located at the point to be measured or there is only one single point, i.e., the point to be detected, in the space, the imaging result of the millimeter-wave radar is the preset imaging result corresponding to the point to be measured.

[0064] According to the embodiment of the present invention, there is no limitation on the preset radar direction, and it can be set according to demand.

[0065] According to an embodiment of the present invention, a target point to be measured having coordinates identical to the three-dimensional coordinate point can be determined from preset imaging results corresponding to each of the plurality of points to be measured, and the preset imaging result corresponding to the target point to be measured can be determined as the target preset imaging result.

[0066] According to an embodiment of the present invention, by superimposing the initial imaging results corresponding to each three-dimensional coordinate point, the initial imaging result for the sample to be invisible covered with the invisible material can be obtained more accurately and quickly.

[0067] According to an embodiment of the present invention, determining the initial imaging result for a three-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point corresponding to the three-dimensional coordinate point, a preset radar direction and a preset imaging result of the target may include the following steps.

[0068] The radar reflectivity is determined based on the normal vector of the two-dimensional coordinate point corresponding to the three-dimensional coordinate point and the preset radar direction; and the initial imaging result for the three-dimensional coordinate point is determined based on the radar reflectivity and the preset imaging result of the target.

[0069] According to an embodiment of the present invention, the radar reflectivity is related to the angle between the normal vector and the preset radar direction. The radar reflectivity at the three-dimensional coordinate point can be obtained by determining the angle between the normal vector of the two-dimensional coordinate point corresponding to the three-dimensional coordinate point and the preset radar direction, and determining the cosine value of the angle.

[0070] According to an embodiment of the present invention, the initial imaging result for the three-dimensional coordinate point may be obtained by multiplying the radar reflectivity and the preset imaging result of the target.

[0071] According to an embodiment of the present invention, based on the radar reflectivity and the preset imaging result of the target, the initial imaging result for the three-dimensional coordinate point is determined by the following formula (1).

[0072] ; (1)

[0073] Among them, n i is a two-dimensional coordinate point (x i ,yi ), i=1, 2, …N, (x i ,y i , z i ) is the two-dimensional coordinate point (x i ,y i ) corresponds to the three-dimensional coordinate point, PSF (x i ,y i , z i ) is the three-dimensional coordinate point (x i ,y i , z i ) corresponding to the preset imaging result of the target, n r To preset radar direction. Represents a three-dimensional coordinate point (x i ,y i , z i ), whose value depends on the normal vector n of the two-dimensional coordinate point corresponding to this point i With radar direction n r Angle.

[0074] According to an embodiment of the present invention, by analyzing the above formula (1), it can be known that in some embodiments, in order to determine a more accurate final imaging result, the present invention establishes a differentiable relationship between the shape of the stealth material and the final imaging result, that is, formula (1) can be considered as an imaging model, and the imaging model is differentiable.

[0075] According to an embodiment of the present invention, through the linear system characteristics of the millimeter-wave radar imaging system, the final imaging result, that is, the initial imaging result of the sample to be hidden covered with stealth material, is expressed as the superposition of the response of the millimeter-wave radar for each point, that is, the superposition of the initial imaging results of multiple three-dimensional coordinate points, thereby greatly accelerating the process of determining the final imaging result, and then accelerating the optimization process of the initial shape design model, and achieving a highly concealed and precisely controlled stealth effect by finely controlling the reflection and scattering of the millimeter-wave radar wave at the physical level. At the same time, since the final imaging result obtained by the above method has the advantage of high precision, the target shape design model finally trained also has a more accurate characteristic, that is, the precision required for designing the shape of the stealth material is maintained.

[0076] According to an embodiment of the present invention, the process of determining the target preset imaging result for each three-dimensional coordinate point can be called a PSF retrieval operation, which can be processed by a straight-through estimator (Straight-Through Estimato, STE). By using the straight-through estimator to process the PSF retrieval operation, the gradient can flow through the PSF retrieval process, thereby optimizing the shape design of the stealth material through back propagation.

[0077] According to the embodiment of the present invention, the whole training process of the initial shape design model can be made differentiable through the through-estimator and the imaging model.

[0078] According to an embodiment of the present invention, the shape design of stealth materials for samples to be invisible is implemented based on the initial shape design model and the millimeter-wave radar imaging system. This can combine neural rendering technology with electromagnetic theory to achieve efficient and accurate stealth material design suitable for the millimeter-wave radar band.

[0079] According to an embodiment of the present invention, the preset imaging result is determined in the following manner.

[0080] Based on the transmission signal of the millimeter-wave radar and the distance between the millimeter-wave radar and the point to be measured, the receiving signal for the transmission signal is determined; the transmission signal and the receiving signal are mixed to obtain an intermediate frequency signal; based on the intermediate frequency signal and a preset radar imaging model, a preset imaging result for the point to be measured is obtained.

[0081] According to an embodiment of the present invention, the preset imaging result may be determined according to a point spread function (PSF), where the point spread function is a response of a millimeter wave radar imaging system to a point to be measured.

[0082] According to an embodiment of the present invention, in some embodiments, the millimeter wave radar in the millimeter wave radar imaging system may be a frequency modulated continuous wave (FMCW) radar, and the transmission signal of the FMCW radar may be determined by the following formula (2).

[0083] ; (2)

[0084] in, is the amplitude, is the carrier frequency, is the frequency sweep rate, s tx (t) is the transmitted signal at time t.

[0085] According to an embodiment of the present invention, the distance from the transmitting antenna of the millimeter wave radar to the point to be measured can be determined, and the distance from the point to be measured to the receiving antenna can be determined, so as to obtain the round-trip distance of the signal, and the round-trip delay of the signal can be determined by the round-trip distance. The received signal can be obtained based on the round-trip delay and the transmitted signal, which can be specifically shown in the following formula (3).

[0086] ; (3)

[0087] Among them, s rx (t) is the received signal at time t, is the received amplitude, Indicates that the signal is transmitted from the mth transmitting antenna to the test point (x i ,y i , z i ) and the round trip delay from the point to be tested to the nth receiving antenna, The representation indicates that the transmitted signal is from the mth transmitting antenna to the test point (x i ,y i , z i ) and then from the point to be measured (x i ,y i , z i ) is the distance to the nth receiving antenna, and c is the speed of light.

[0088] According to an embodiment of the present invention, the process of mixing the transmit signal and the receive signal to obtain the intermediate frequency signal can be shown as the following formula (4).

[0089] ; (4)

[0090] in, is the amplitude after mixing, x IF (t) is the intermediate frequency signal at time t.

[0091] According to the embodiment of the present invention, the preset radar imaging model is not limited and can be obtained by any range migration algorithm, for example, RMA (Range Migration Algorithm) or a back projection algorithm can be used.

[0092] According to an embodiment of the present invention, in some embodiments, the point spread function (PSF) determined by a preset radar imaging model determined by RMA may be as shown in the following formula (5).

[0093] ; (5)

[0094] Among them, RMA() represents the preset radar imaging model, PSF(x i ,y i , z i ) is characterized as the point to be tested (x i ,y i , z i ) preset imaging results.

[0095] According to an embodiment of the present invention, the preset imaging results of all points to be tested in the space can be determined in advance before training the initial shape design model or before predicting the initial imaging results of the millimeter-wave radar on the sample to be hidden covered with stealth material, so that the preset imaging results of the target corresponding to the three-dimensional coordinate point can be directly determined in the subsequent determination of the initial imaging results, thereby accelerating the efficiency of the online training of the model.

[0096] Figure 3 A schematic diagram of determining an initial imaging result according to an embodiment of the present invention is schematically shown.

[0097] like Figure 3 As shown in the formula (1), in the process of determining the initial imaging result of the millimeter wave radar on the sample to be hidden covered with the hidden material, the three-dimensional coordinate point (x i ,y i , z i ), i=1, 2, ..., N corresponding to the initial imaging result PSF (x i ,y i , z i ), and through the normal vector n corresponding to the three-dimensional coordinate point i , preset radar direction n r Determine the radar reflectivity cos for the three-dimensional coordinate point <n i , n r >, and by presetting the imaging result PSF (x i ,y i , z i ) and radar reflectivity cos <n i , n r >It is possible to determine the coordinates of each 3D point (x i ,y i , z i ) initial imaging results.

[0098] According to an embodiment of the present invention, by superimposing the initial imaging results of all three-dimensional coordinate points, the final initial imaging result Y of the millimeter-wave radar on the sample to be invisible covered with the invisible material can be obtained.

[0099] According to an embodiment of the present invention, determining a three-dimensional coordinate point corresponding to a two-dimensional coordinate point based on a normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point may include the following steps.

[0100] Determine the height coordinate of the two-dimensional coordinate point based on the first horizontal coordinate of the two-dimensional coordinate point, the second horizontal coordinate of the neighbor coordinate point of the two-dimensional coordinate point, the first vertical coordinate of the two-dimensional coordinate point, the second vertical coordinate of the neighbor coordinate point, and the normal vector of the two-dimensional coordinate point; determine the three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the first horizontal coordinate, the first vertical coordinate and the height coordinate.

[0101] According to an embodiment of the present invention, a three-dimensional coordinate point corresponding to the two-dimensional coordinate point may be determined by using the two-dimensional coordinate point and the coordinates of its neighbors.

[0102] According to an embodiment of the present invention, the horizontal and vertical coordinates of the two-dimensional coordinate point and the three-dimensional coordinate point corresponding to the two-dimensional coordinate point are the same.

[0103] According to an embodiment of the present invention, the neighbor coordinate point of a two-dimensional coordinate point is the coordinate point closest to the two-dimensional coordinate point. Specifically, the neighbor coordinate point of each two-dimensional coordinate point can be determined by placing multiple two-dimensional coordinate points in a two-dimensional coordinate system, or the neighbor coordinate point of each two-dimensional coordinate point can be determined by calculating the distance between each two-dimensional coordinate point.

[0104] According to an embodiment of the present invention, determining the height coordinate of a two-dimensional coordinate point based on the first horizontal coordinate of the two-dimensional coordinate point, the second horizontal coordinate of a neighbor coordinate point of the two-dimensional coordinate point, the first vertical coordinate of the two-dimensional coordinate point, the second vertical coordinate of the neighbor coordinate point, and the normal vector of the two-dimensional coordinate point may include the following steps.

[0105] Based on the first horizontal coordinate, the second horizontal coordinate and the horizontal coordinate component of the target normal vector, the first sub-height coordinate is obtained; based on the first vertical coordinate, the second vertical coordinate and the vertical coordinate component of the target normal vector, the second sub-height coordinate is obtained; based on the first sub-height coordinate and the second sub-height coordinate, the height coordinate is obtained.

[0106] According to an embodiment of the present invention, the second horizontal coordinate may be subtracted from the first horizontal coordinate to obtain a first calculation result, and the first calculation result may be multiplied by the horizontal coordinate component of the target normal vector to obtain a first sub-height coordinate.

[0107] According to an embodiment of the present invention, the second ordinate may be subtracted from the first ordinate to obtain a second calculation result, and the second calculation result may be multiplied by the ordinate component of the target normal vector to obtain a second sub-height coordinate.

[0108] According to an embodiment of the present invention, the first sub-height coordinate and the second sub-height coordinate may be added together to obtain a height coordinate, which may be specifically as shown in formula (6).

[0109] ; (6)

[0110] Among them, x j+1 Represented as the second horizontal axis, x j Represented as the first horizontal axis, y j+1 Represented as the second horizontal axis, y j Characterized by the first horizontal axis n j (x) represents the target normal vector nj The horizontal axis component, n j (y) represents the target normal vector n j The vertical axis component of .

[0111] According to an embodiment of the present invention, in some embodiments, in the process of determining the height coordinates of each of a plurality of two-dimensional coordinate points, z1=0 may be set as a starting point, and then the z coordinate of each two-dimensional coordinate point may be calculated.

[0112] According to an embodiment of the present invention, the complete shape of the stealth material can be obtained through multiple three-dimensional coordinate points. Specifically, a continuous surface of the stealth material can be obtained by fitting multiple three-dimensional coordinate points, which can then be converted into a standard file format through a related renderer and 3D printed using a related 3D printer to generate a stealth material that can be attached to the sample to be invisible.

[0113] According to an embodiment of the present invention, by determining the height coordinate corresponding to the two-dimensional coordinate point through the normal vector corresponding to the two-dimensional coordinate point, and then obtaining the three-dimensional coordinate point, it is possible to customize the shape of the stealth material based on the shape of the sample to be invisible, and define the surface geometry of the stealth material by determining the normal vector, thereby achieving the imaging result of the stealth material in the direction of the millimeter-wave radar prediction radar by controlling the normal vector, and then achieving the invisibility of the sample to be invisible through the stealth material.

[0114] According to an embodiment of the present invention, obtaining a target loss value based on an initial imaging result, background characterization data, and a preset height threshold for a material to be cloaked may include the following steps.

[0115] Based on the initial imaging results, background characterization data and a preset signal-to-noise ratio, a first loss value is obtained; based on the height coordinates of each of the multiple three-dimensional coordinate points and a preset height threshold, a second loss value is obtained; based on the first loss value and the second loss value, a target loss value is determined.

[0116] According to an embodiment of the present invention, a first loss value is obtained based on an initial imaging result, background characterization data, and a preset signal-to-noise ratio, which can be shown in the following formula (7).

[0117] ; (7)

[0118] Among them, L img is the first loss value, Y represents the initial imaging result of the millimeter-wave radar on the sample to be stealthed covered with stealth material, B represents the background characterization data, and γ represents the signal-to-noise ratio.

[0119] According to an embodiment of the present invention, based on the height coordinates of each of the multiple three-dimensional coordinate points and a preset height threshold, a second loss value is obtained, which can be shown in the following formula (8).

[0120] ; (8)

[0121] Among them, L z is the second loss value, z i Represents the height coordinate of the i-th three-dimensional coordinate point, z max is the preset height threshold.

[0122] According to an embodiment of the present invention, the process of determining the target loss value based on the first loss value and the second loss value may be as shown in formula (9).

[0123] ; (9)

[0124] Among them, L is the target loss value, is the weight parameter.

[0125] According to an embodiment of the present invention, ,γ,z max is a hyperparameter, and the value of the hyperparameter is not limited. The appropriate value can be determined by heuristic algorithms, exhaustive methods, etc. according to the actual situation. For example, in some embodiments, Set to 0.1, γ to 20dB, z max Set to 10cm.

[0126] According to an embodiment of the present invention, in some embodiments, the Adam optimizer (Adaptive Moment Estimation) is used in the training process, the initial learning rate is set to 1e-4, and the cosine annealing strategy is used to adjust the learning rate, and the training continues for 100 cycles.

[0127] According to an embodiment of the present invention, the first loss value is intended to reduce the target intensity in the initial imaging result to ensure the effectiveness and concealment of the shape of the stealth material. The second loss value constrains the maximum height of the shape of the stealth material to ensure its practicality and physical feasibility. By jointly constraining the initial shape design model with the first loss value and the second loss value, the effect of improving the normal vector prediction accuracy of the target shape design model can be achieved, thereby obtaining the shape of the stealth material with better stealth effect.

[0128] According to an embodiment of the present invention, a plurality of two-dimensional coordinate points are determined in the following manner.

[0129] The sample to be invisible is mapped into a three-dimensional coordinate system; the projection surface of the sample to be invisible on the horizontal plane of the three-dimensional coordinate system is determined; the projection surface is divided into multiple grids according to a preset step size; based on the vertex coordinates of the multiple grids, multiple two-dimensional coordinate points representing the projection of the sample to be invisible on the horizontal plane are determined.

[0130] According to an embodiment of the present invention, considering that the millimeter wave radar system usually operates the antenna array on the xy plane, the present invention adopts a method of projecting the 3D target object, i.e., the sample to be invisible, onto the xy plane to obtain the projection of the sample to be invisible on the horizontal plane to obtain the shape of the sample to be invisible.

[0131] According to an embodiment of the present invention, the sample to be invisible can be fixed in the three-dimensional coordinate system by mapping the sample to be invisible to the three-dimensional coordinate system. In some embodiments, the sample to be invisible can be mapped in a space where z is greater than or equal to 0 to facilitate determining its projection surface on the horizontal plane.

[0132] According to an embodiment of the present invention, there is no limitation on the method of determining the projection surface of the sample to be invisible in the horizontal plane of the three-dimensional coordinate system. For example, a top view of the sample to be invisible can be determined, the top view can be determined as the projection surface, and the boundary points of the projection surface can be determined, and the boundary points can be connected to obtain the projection surface.

[0133] According to an embodiment of the present invention, the preset step size is not limited and may be 1 millimeter (mm), or 2 millimeters, etc. For example, by dividing the projection surface into multiple grids according to a preset step size of 1 mm, a projection surface containing multiple grids with an area or resolution of 1 mm×1 mm can be obtained.

[0134] According to an embodiment of the present invention, by determining the vertex coordinates of each mesh in a three-dimensional coordinate system, multiple coordinates (x i ,y i , 0), and then multiple two-dimensional coordinate points (x i ,y i ), i=1, 2, ....N, that is, the discrete representation of the sample to be hidden is obtained: .

[0135] According to an embodiment of the present invention, a plurality of two-dimensional coordinate points representing the projection of the sample to be invisible on the horizontal plane are determined, and used to represent the sample to be invisible. It is possible to maintain computational efficiency while capturing the details of the geometric shape of the object, provide an effective input for the initial shape design model, and adapt to the characteristics of the millimeter-wave radar when detecting objects.

[0136] According to an embodiment of the present invention, the initial shape design model includes a coding layer and a mapping layer; inputting multiple two-dimensional coordinate points representing the projection of the sample to be hidden on the horizontal plane into the initial shape design model to obtain the normal vectors of the multiple two-dimensional coordinate points may include the following steps.

[0137] Input multiple two-dimensional coordinate points into the encoding layer for feature encoding to obtain encoding vectors corresponding to each of the multiple two-dimensional coordinate points; input the encoding vectors corresponding to each of the multiple two-dimensional coordinate points into the mapping layer to obtain normal vectors of each of the multiple two-dimensional coordinate points.

[0138] According to an embodiment of the present invention, in some embodiments, the initial shape design model G is composed of two associated components including: a coding layer G h and mapping layer G n .

[0139] According to an embodiment of the present invention, the encoding layer is a multi-layer perceptron, which represents the space of the sample to be hidden, that is, multiple two-dimensional coordinate points Mapping to the latent code, i.e., the encoding vector This transformation from the spatial domain to the latent field enables the network to efficiently capture and process the complex geometric features of the samples to be cloaked.

[0140] According to the embodiment of the present invention, feature dimension reduction and feature encoding of multiple two-dimensional coordinate points can be achieved through the encoding layer.

[0141] According to an embodiment of the present invention, the mapping layer may also be a multi-layer perceptron, which receives the potential code h sent by the encoding layer as input and generates a normal vector field of the invisible material. The output is a set of normal vectors , where each Corresponding to a discrete point on the xy plane These normal vectors define the surface geometry of the stealth material's shape and are critical for manipulating radar reflections.

[0142] According to an embodiment of the present invention, since the initial shape design model can be implemented by a neural network, that is, the initial shape design model can also be called an initial neural rendering network, that is, the target shape design model is a target neural rendering network, and the initial neural rendering network G is composed of G h and G n The final output is n∈R （N×3） .

[0143] According to an embodiment of the present invention, the model structure of the initial shape design model is not limited, for example: Can contain 2 hidden layers, It can contain 4 hidden layers, with 256 neurons in each layer.

[0144] According to an embodiment of the present invention, the initial shape design model output encodes the surface normal vector of the stealth material and the shape of the stealth material is specially customized for a specific target object.

[0145] According to an embodiment of the present invention, based on the shape of the sample to be invisible, a neural rendering network G is introduced to accurately control the shape of the invisible material covering it, and a differentiable imaging model is used to self-supervisorily constrain the imaging result under the shape, thereby iteratively optimizing the parameters of G to obtain the final shape of the invisible material, that is, the invisible material can be finely designed on the physical level to achieve excellent invisible effect.

[0146] Figure 4 The flowchart of the method for designing the shape of a stealth material according to an embodiment of the present invention is schematically shown.

[0147] like Figure 4 As shown, the method includes operations S410 to S430.

[0148] In operation S410, a plurality of two-dimensional test coordinate points of the object to be hidden projected on a horizontal plane are input into a target shape design model to obtain normal vectors of the plurality of two-dimensional test coordinate points.

[0149] In operation S420 , three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points are determined based on the normal vectors of each of the plurality of two-dimensional test coordinate points and the plurality of two-dimensional test coordinate points.

[0150] In operation S430, a shape of the stealth material is determined based on the three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points.

[0151] According to the embodiment of the present invention, there is no limitation on the object to be hidden, and it can be an object of any shape.

[0152] According to an embodiment of the present invention, the target shape design model is obtained by training the model training method for designing the shape of stealth materials according to an embodiment of the present invention.

[0153] According to an embodiment of the present invention, the horizontal plane may be an xy plane in a three-dimensional coordinate system.

[0154] According to an embodiment of the present invention, a projection surface can be obtained after the sample to be invisible is projected on the xy plane, and a plurality of two-dimensional test coordinate points can be obtained according to the coordinate points included in the projection surface.

[0155] According to an embodiment of the present invention, there is no limitation on the implementation method of the target shape design model, and a neural network may be used for implementation, for example, a multilayer perceptron (MLP), a feedforward neural network (FNN), and the like.

[0156] According to an embodiment of the present invention, normal vectors corresponding to each of the plurality of two-dimensional test coordinate points define the surface geometry of the stealth material shape, which is essential for operating radar reflection.

[0157] According to an embodiment of the present invention, by inputting multiple two-dimensional coordinate points representing the projection of the sample to be invisible on the horizontal plane into the initial shape design model to obtain the normal vectors of the multiple two-dimensional coordinate points, the shape of the stealth material of each sample to be invisible can be customized to obtain the stealth material shape for the sample to be invisible.

[0158] According to an embodiment of the present invention, a three-dimensional test coordinate point of the two-dimensional test coordinate point may be determined by using a normal vector of the two-dimensional test coordinate point and the two-dimensional test coordinate point itself.

[0159] According to an embodiment of the present invention, the abscissa and ordinate included in the three-dimensional coordinate point corresponding to the two-dimensional coordinate point may be the same as the abscissa and ordinate of the two-dimensional coordinate point.

[0160] According to an embodiment of the present invention, the shape of the stealth material can be obtained by fitting the three-dimensional test coordinate points corresponding to the multiple two-dimensional test coordinate points, that is, fitting the multiple three-dimensional test coordinate points.

[0161] According to an embodiment of the present invention, since the target shape design model can be a neural network, the method can also be called a stealth material shape design method based on neural network rendering.

[0162] According to an embodiment of the present invention, by inputting a plurality of two-dimensional test coordinate points containing shape information of an object to be concealed into a target shape design model, a plurality of normal vectors defining the surface geometric shape of a stealth material are obtained, and a plurality of three-dimensional coordinate points are determined by the normal vectors, thereby determining the shape of the stealth material. This enables the shape of the stealth material for any object to be concealed to be determined, and better stealth of the object to be concealed to be achieved through the stealth material.

[0163] Figure 5 A first schematic diagram of a stealth material shape design method according to an embodiment of the present invention is schematically shown.

[0164] like Figure 5As shown, the initial shape design model G is optimized by self-supervised learning, and in the process of optimizing G, the initial imaging result F of the millimeter-wave radar on the sample to be invisible covered with stealth material is predicted by multiple normal vectors output by G, so that the shape of the stealth material determined by the multiple normal vectors generated by the target shape design model can make the object to be invisible T successfully invisible under the detection of the millimeter-wave radar imaging system.

[0165] Figure 6 A second schematic diagram of the stealth material shape design method according to an embodiment of the present invention is schematically shown.

[0166] like Figure 6 As shown, the object to be invisible, represented by the car model, can be discretized and projected on a horizontal plane to obtain multiple two-dimensional test coordinate points T' pos ∈R （N×2） , T' pos ∈R （N×2） Input to the encoding layer G of the target shape design model G' h 'Perform feature encoding, and you can get the encoding vectors corresponding to multiple two-dimensional test coordinate points, that is, the potential encoding h'∈R （N×1） , then input h' into the mapping layer G of the target shape design model G' n ', get the normal vectors corresponding to multiple two-dimensional test coordinate points , , thus through T' pos Included (x i ,y i ) and n i Determine the three-dimensional test coordinate points (x i ,y i , z i ), and then obtain the shape of the stealth material for the object to be hidden through multiple three-dimensional test coordinate points.

[0167] Figure 7 A schematic diagram of a first stealth effect of a stealth material shape design method according to an embodiment of the present invention is schematically shown.

[0168] like Figure 7 As shown in 710, it can be known that the object to be invisible is a car model. The millimeter wave radar imaging result of the object to be invisible without using stealth material can be shown in 720. From 720, it can be seen that the outline of the car is clearly visible, showing a clear elliptical shape, and the color gradually transitions from red to yellow, green, and finally to blue from the center to the outside. This clear outline and strong color contrast indicate that the unprocessed car model has high visibility and recognizability under the radar system.

[0169] According to an embodiment of the present invention, the millimeter wave radar imaging result of the object to be cloaked using the stealth material can be shown as 730, wherein the shape of the stealth material is determined by the present invention for the object to be cloaked. As can be seen from 730, compared with 720, the image here is almost completely blue, with no obvious target contour or intensity change. This shows that the stealth material designed by the present invention successfully reduces the visibility of the target, i.e., the object to be cloaked, in the radar image, and achieves an effective stealth effect.

[0170] According to an embodiment of the present invention, the millimeter wave radar imaging result when the millimeter wave radar does not detect any object may be shown as 740. 740 shows the imaging result of the radar system when there is no target object, showing a uniform blue background. This image provides a reference benchmark for the present invention to evaluate the quality of the stealth effect.

[0171] According to the embodiment of the present invention, the millimeter wave radar imaging system used by 720, 730, and 740 is the same.

[0172] According to the embodiment of the present invention, the stealth effect of the shape of the stealth material of the present invention can be clearly seen by comparing the three images in 720, 730, and 740. The untreated car model is very eye-catching in the radar image as shown in 720, but after using the stealth material designed by the present invention, the car model almost completely disappears in the radar image as shown in 730, and its imaging result is very similar to the background when there is no target object as shown in 740, which verifies the effectiveness of the present invention.

[0173] Figure 8 A second stealth effect schematic diagram of the stealth material shape design method according to an embodiment of the present invention is schematically shown.

[0174] like Figure 8 As shown, it can be known from 810 that the object to be invisible is a black pet model. Through 810, the standing posture of the pet can be seen, and the limbs and tail are clearly discernible.

[0175] According to an embodiment of the present invention, the millimeter wave radar imaging result of the object to be cloaked without using cloaking material may be as shown in 820. Figure 8 As can be seen in the 820 in the image, the pet's outline is clearly visible, showing a distinct shape, and the color gradually transitions from red to yellow, green, and finally to blue from the center outward. In particular, the outlines of the head, body, and limbs can be clearly identified. This clear outline and strong color contrast show that the unprocessed pet model has high visibility and identifiability under the radar system.

[0176] According to an embodiment of the present invention, the millimeter wave radar imaging result of the object to be cloaked using the cloaking material may be shown as 830, wherein the shape of the cloaking material is determined by the present invention for the object to be cloaked. Compared with 820, the image here presents a uniform blue state without any recognizable target contour or intensity change. This shows that the cloaking material designed by the present invention successfully reduces the visibility of the target in the radar image and achieves a very effective cloaking effect.

[0177] According to an embodiment of the present invention, the millimeter wave radar imaging result when the millimeter wave radar does not detect any object may be shown as 840. It presents a uniform blue background. This image provides a reference benchmark for the present invention to evaluate the quality of the stealth effect.

[0178] According to an embodiment of the present invention, by comparing Figure 8 The three images 820, 830, and 840 in the figure clearly show the stealth effect of the present invention. The untreated pet model is very eye-catching in the radar image, as shown in 820, but after using the stealth material designed by the present invention, the pet model completely disappears in the radar image, as shown in 830, and the imaging result is almost the same as the background without the target object, as shown in 840.

[0179] According to an embodiment of the present invention, by Figure 8 The embodiments of the present invention further demonstrate the effectiveness and versatility of the present invention in designing millimeter-wave band stealth materials. Through the neural network rendering technology, the present invention can successfully design efficient stealth material shapes for biological targets with complex shapes and small sizes, such as pets, making them completely invisible to millimeter-wave radar systems, thereby achieving excellent stealth effects.

[0180] Fig. 9 A third stealth effect schematic diagram of the stealth material shape design method according to an embodiment of the present invention is schematically shown.

[0181] like Fig. 9 As shown, it can be known from 910 that the object to be invisible is an aircraft model. In 910, the main structures of the aircraft such as the fuselage, wings, and tail can be clearly seen, and the aircraft is in a take-off or landing posture.

[0182] According to an embodiment of the present invention, the millimeter wave radar imaging result of the object to be cloaked without using a stealth material may be shown as 920. As can be seen from 920, the outline of the aircraft is clearly visible, showing a distinct elongated shape, and the color gradually transitions from red to yellow, green, and finally to blue from the center to the outside. In particular, the outlines of the fuselage and wings can be clearly identified, forming two obvious highlight areas. This clear outline and strong color contrast indicate that the unprocessed aircraft model has high visibility and recognizability under the radar system.

[0183] According to an embodiment of the present invention, the millimeter wave radar imaging result of the object to be cloaked using the stealth material can be shown as 930, wherein the shape of the stealth material is determined by the present invention for the object to be cloaked. Compared with 920, the image here presents a uniform blue state without any recognizable target contour or intensity change. This shows that the stealth material designed by the present invention successfully reduces the visibility of the aircraft target in the radar image, achieving a very effective stealth effect.

[0184] According to an embodiment of the present invention, the millimeter wave radar imaging result when the millimeter wave radar does not detect any object may be shown as 940. The image in 940 presents a uniform blue background.

[0185] According to an embodiment of the present invention, by comparing Fig. 9 The three images 920, 930, and 940 in the figure can clearly show the stealth effect of the present invention. The untreated aircraft model is very eye-catching in the radar image, such as 920, but after using the stealth material designed by the present invention, the aircraft model completely disappears in the radar image, such as 930, and its imaging result is almost the same as the background without the target object, such as 940, which verifies the effectiveness of the present invention.

[0186] According to an embodiment of the present invention, by Figure 7~Figure 9 It can be seen from the embodiments that the present invention can effectively handle different types of targets, greatly reduce the visibility of targets in millimeter-wave radar images, and excel in the design of finely controlled stealth materials. For different stealth objects, the stealth materials designed by the present invention can always successfully reduce the visibility of targets in radar images. Compared with the related art, the present invention has significant improvements in stealth effect and concealment. In particular, for synthetic aperture radar (SAR) imaging systems, the stealth materials designed by the present invention can effectively reduce the radar scattering cross section of the target object, making it difficult to be effectively detected in the radar image, verifying the wide applicability of the stealth material design of the present invention in the millimeter wave band.

[0187] Based on the above-mentioned model training method for designing the shape of stealth materials and the method for designing the shape of stealth materials, the present invention also provides a model training device for designing the shape of stealth materials and a device for designing the shape of stealth materials. Figures 10-11 The above-mentioned device is described in detail.

[0188] Fig.10 The structural block diagram of a model training device for designing the shape of stealth materials according to an embodiment of the present invention is schematically shown.

[0189] like Fig.10As shown, the model training device 1000 for designing the shape of stealth materials in this embodiment includes a first input module 1010, a first determination module 1020, a result prediction module 1030, a loss value determination module 1040 and a training module 1050.

[0190] The first input module 1010 is used to input multiple two-dimensional coordinate points of the sample to be invisible projected on the horizontal plane into the initial shape design model to obtain normal vectors of the multiple two-dimensional coordinate points, wherein the normal vectors define the surface geometry of the invisible material.

[0191] The first determination module 1020 is used to determine, for each two-dimensional coordinate point, a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points.

[0192] The result prediction module 1030 is used to predict the initial imaging result of the millimeter wave radar on the sample to be hidden covered with the hidden material based on the preset radar direction, the three-dimensional coordinate points corresponding to each of the multiple two-dimensional coordinate points, and the normal vectors of each of the multiple two-dimensional coordinate points.

[0193] The loss value determination module 1040 is used to obtain a target loss value based on the initial imaging result, the background characterization data and the preset height threshold for the stealth material.

[0194] The training module 1050 is used to train the initial shape design model based on the target loss value to obtain a target shape design model.

[0195] According to an embodiment of the present invention, any multiple modules of the first input module 1010, the first determination module 1020, the result prediction module 1030, the loss value determination module 1040 and the training module 1050 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present invention, at least one of the first input module 1010, the first determination module 1020, the result prediction module 1030, the loss value determination module 1040 and the training module 1050 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the first input module 1010, the first determination module 1020, the result prediction module 1030, the loss value determination module 1040 and the training module 1050 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is executed.

[0196] Fig.11 The structural block diagram of the stealth material shape design device according to an embodiment of the present invention is schematically shown.

[0197] like Fig.11 As shown, the stealth material shape design device 1100 of this embodiment includes a second input module 1110 , a second determination module 1120 and a shape determination module 1130 .

[0198] The second input module 1110 is used to input a plurality of two-dimensional test coordinate points of the object to be invisible projected on a horizontal plane into the target shape design model to obtain normal vectors of the plurality of two-dimensional test coordinate points.

[0199] The second determining module 1120 is configured to determine the three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points based on the normal vectors of each of the plurality of two-dimensional test coordinate points and the plurality of two-dimensional test coordinate points.

[0200] The shape determination module 1130 is used to determine the shape of the stealth material based on the three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points.

[0201] According to an embodiment of the present invention, any multiple modules of the second input module 1110, the second determination module 1120 and the shape determination module 1130 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present invention, at least one of the second input module 1110, the second determination module 1120 and the shape determination module 1130 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in any appropriate combination of any of them. Alternatively, at least one of the second input module 1110, the second determination module 1120 and the shape determination module 1130 can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding function can be executed.

[0202] Fig.12 A block diagram of an electronic device suitable for implementing a method for designing a stealth material shape and a method for designing a stealth material shape according to an embodiment of the present invention is schematically shown.

[0203] like Fig.12 As shown, the electronic device 1200 according to an embodiment of the present invention includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1202 or a program loaded from a storage part 1208 to a random access memory (RAM) 1203. The processor 1201 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1201 may also include an onboard memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0204] In RAM 1203, various programs and data required for the operation of electronic device 1200 are stored. Processor 1201, ROM 1202 and RAM 1203 are connected to each other via bus 1204. Processor 1201 performs various operations of the method flow according to the embodiment of the present invention by executing the programs in ROM 1202 and / or RAM 1203. It should be noted that the program can also be stored in one or more memories other than ROM 1202 and RAM 1203. Processor 1201 can also perform various operations of the method flow according to the embodiment of the present invention by executing the programs stored in one or more memories.

[0205] According to an embodiment of the present invention, the electronic device 1200 may further include an input / output (I / O) interface 1205, which is also connected to the bus 1204. The electronic device 1200 may further include one or more of the following components connected to the input / output (I / O) interface 1205: an input portion 1206 including a keyboard, a mouse, etc.; an output portion 1207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1208 including a hard disk, etc.; and a communication portion 1209 including a network interface card such as a LAN card, a modem, etc. The communication portion 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output (I / O) interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1210 as needed, so that a computer program read therefrom is installed into the storage portion 1208 as needed.

[0206] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0207] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 1202 and / or RAM 1203 described above and / or one or more memories other than ROM 1202 and RAM 1203.

[0208] The embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the method for designing the shape of a stealth material and the method for designing the shape of a stealth material provided in the embodiment of the present invention.

[0209] The computer program executes the above functions defined in the system / device of the embodiment of the present invention when it is executed by the processor 1201. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0210] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 1209, and / or installed from the removable medium 1211. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0211] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1209, and / or installed from the removable medium 1211. When the computer program is executed by the processor 1201, the above functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0212] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0213] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0214] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present invention may be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention may be combined and / or combined in various ways. All of these combinations and / or combinations fall within the scope of the present invention.

[0215] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A model training method for designing the shape of stealth materials, characterized in that: The model training method comprises: Inputting a plurality of two-dimensional coordinate points of the sample to be cloaked projected on a horizontal plane into the initial shape design model to obtain normal vectors of the plurality of two-dimensional coordinate points, wherein the normal vectors define the surface geometric shape of the cloaked material; For each of the two-dimensional coordinate points, based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point, determine a three-dimensional coordinate point corresponding to the two-dimensional coordinate point, wherein the shape of the stealth material is determined by the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points; Predicting an initial imaging result of the millimeter wave radar on the sample to be invisible covered with the invisible material based on a preset radar direction, three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points, and normal vectors of each of the plurality of two-dimensional coordinate points; Obtaining a target loss value based on the initial imaging result, the background characterization data, and a preset height threshold for the stealth material; The initial shape design model is trained based on the target loss value to obtain a target shape design model.

2. The model training method according to claim 1, characterized in that: The method predicts the initial imaging result of the millimeter wave radar on the sample to be invisible and covered with the invisible material based on the preset radar direction, the three-dimensional coordinate points corresponding to each of the plurality of two-dimensional coordinate points, and the normal vectors of each of the plurality of two-dimensional coordinate points, including: For each of the three-dimensional coordinate points, Determining a preset imaging result of a target corresponding to the three-dimensional coordinate point from a plurality of preset imaging results, wherein the preset imaging result is determined based on a response of the millimeter-wave radar to a point to be measured in space, and a radar emission direction of the millimeter-wave radar is the preset radar direction; Determining an initial imaging result for the three-dimensional coordinate point based on a normal vector of a two-dimensional coordinate point corresponding to the three-dimensional coordinate point, the preset radar direction, and the preset imaging result of the target; The initial imaging results of each of the plurality of three-dimensional coordinate points are superimposed to obtain an initial imaging result of the sample to be invisible covered with the invisible material.

3. The model training method according to claim 2, characterized in that: The determining of the initial imaging result for the three-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point corresponding to the three-dimensional coordinate point, the preset radar direction and the preset imaging result of the target includes: determining a radar reflectivity based on a normal vector of a two-dimensional coordinate point corresponding to the three-dimensional coordinate point and the preset radar direction; Based on the radar reflectivity and the preset imaging result of the target, an initial imaging result for the three-dimensional coordinate point is determined.

4. The model training method according to claim 2, characterized in that: The preset imaging result is determined in the following manner: Determining a received signal for the transmitted signal based on the transmitted signal of the millimeter wave radar and the distance between the millimeter wave radar and the point to be measured; Mixing the transmission signal and the reception signal to obtain an intermediate frequency signal; Based on the intermediate frequency signal and the preset radar imaging model, a preset imaging result for the point to be measured is obtained.

5. The model training method according to claim 1, characterized in that: The determining of a three-dimensional coordinate point corresponding to the two-dimensional coordinate point based on the normal vector of the two-dimensional coordinate point and the two-dimensional coordinate point comprises: Determine a height coordinate of the two-dimensional coordinate point based on a first horizontal coordinate of the two-dimensional coordinate point, a second horizontal coordinate of a neighboring coordinate point of the two-dimensional coordinate point, a first vertical coordinate of the two-dimensional coordinate point, a second vertical coordinate of the neighboring coordinate point, and a target normal vector of the two-dimensional coordinate point; Based on the first horizontal coordinate, the first vertical coordinate, and the height coordinate, a three-dimensional coordinate point corresponding to the two-dimensional coordinate point is determined.

6. The model training method according to claim 5, characterized in that: The step of determining the height coordinate of the two-dimensional coordinate point based on the first horizontal coordinate of the two-dimensional coordinate point, the second horizontal coordinate of the neighboring coordinate point of the two-dimensional coordinate point, the first vertical coordinate of the two-dimensional coordinate point, the second vertical coordinate of the neighboring coordinate point, and the target normal vector of the two-dimensional coordinate point comprises: Obtaining a first sub-height coordinate based on the first horizontal coordinate, the second horizontal coordinate, and the horizontal coordinate component of the target normal vector; Obtaining a second sub-height coordinate based on the first ordinate, the second ordinate, and the ordinate component of the target normal vector; The height coordinate is obtained based on the first sub-height coordinate and the second sub-height coordinate.

7. The model training method according to claim 1, characterized in that: The step of obtaining a target loss value based on the initial imaging result, the background characterization data and a preset height threshold for the stealth material comprises: Obtaining a first loss value based on the initial imaging result, the background characterization data, and a preset signal-to-noise ratio; Obtaining a second loss value based on the height coordinates of each of the plurality of three-dimensional coordinate points and the preset height threshold; The target loss value is determined based on the first loss value and the second loss value.

8. The model training method according to claim 1, characterized in that: The two-dimensional coordinate points are determined as follows: Mapping the sample to be invisible into a three-dimensional coordinate system; Determine the projection surface of the sample to be invisible on the horizontal plane of the three-dimensional coordinate system; Dividing the projection surface into a plurality of grids according to a preset step size; Based on the vertex coordinates of each of the plurality of grids, a plurality of two-dimensional coordinate points representing the projection of the sample to be hidden on a horizontal plane are determined.

9. The model training method according to claim 1, characterized in that: The initial shape design model includes a coding layer and a mapping layer; The step of inputting a plurality of two-dimensional coordinate points representing the projection of the sample to be invisible on the horizontal plane into the initial shape design model to obtain the normal vectors of the plurality of two-dimensional coordinate points includes: Inputting the plurality of two-dimensional coordinate points into the encoding layer for feature encoding to obtain encoding vectors corresponding to the plurality of two-dimensional coordinate points; The encoding vectors corresponding to the plurality of two-dimensional coordinate points are input into the mapping layer to obtain the normal vectors of the plurality of two-dimensional coordinate points.

10. A method for designing the shape of a stealth material, characterized in that: The method comprises: Input a plurality of two-dimensional test coordinate points of the object to be cloaked projected on a horizontal plane into a target shape design model to obtain normal vectors of each of the plurality of two-dimensional test coordinate points, wherein the target shape design model is obtained by training based on the model training method for designing the shape of a stealth material according to any one of claims 1 to 9; Determine three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points based on the normal vectors of each of the plurality of two-dimensional test coordinate points and the plurality of two-dimensional test coordinate points; The shape of the stealth material is determined based on the three-dimensional test coordinate points corresponding to each of the plurality of two-dimensional test coordinate points.

Images