Vector simulation method and device, electronic equipment and storage medium

Abstract

The invention discloses a vector simulation method and device, electronic equipment and a storage medium, and relates to the technical field of optical metasurs.The vector simulation method comprises the steps that mapping calculation is conducted according to a preset scalar phase to obtain structural parameters of a target metasurface optical element; the structure parameters are input into a preset vector simulation model, a vector near-field simulation result of the target metasurface optical element is obtained, far-field lighting effect information of the target metasurface optical element is obtained according to the vector near-field simulation result, and the vector simulation model is a trained neural network model. The vector simulation speed for the metasurface optical element can be increased, and meanwhile consumption of computing resources is reduced.

Description

Technical Field

[0001] The present application relates to the field of optical metasurface technology, and in particular to vector simulation methods, devices, electronic devices and storage media. Background Art

[0002] When dealing with large-scale MOE (Metasurface Optical Element) vector simulation tasks, due to the structural complexity and high accuracy requirements of metasurface optical elements, the traditional simulation methods are often time-consuming and consume a lot of computing resources, which affects the design and optimization process of metasurface optical elements.

[0003] In summary, how to improve the vector simulation speed of metasurface optical elements while reducing the consumption of computing resources has become a technical problem that needs to be urgently solved in this field. Summary of the invention

[0004] The main purpose of this application is to provide a vector simulation method, device, electronic device and storage medium, aiming to improve the vector simulation speed for metasurface optical elements while reducing the consumption of computing resources.

[0005] To achieve the above purpose, the present application proposes a vector simulation method, which includes:

[0006] The structural parameters of the target metasurface optical element are obtained by mapping and calculating according to the preset scalar phase;

[0007] The structural parameters are input into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element, and the far-field optical effect information of the target metasurface optical element is obtained according to the vector near-field simulation result, wherein the vector simulation model is a trained neural network model.

[0008] In one embodiment, before the step of inputting the structural parameters into a preset vector simulation model to obtain the vector near-field simulation results of the target metasurface optical element, the method further includes:

[0009] Establishing a model training data set, wherein the model training data set includes training structural parameters and training near-field light field data of each training metasurface optical element;

[0010] The neural network model is trained based on the model training data set to obtain a trained vector simulation model.

[0011] In one embodiment, before the step of training the neural network model based on the model training data set to obtain a trained vector simulation model, the step further includes:

[0012] For each of the training near-field light field data, decomposing the training near-field light field data according to electric field components and component forms to obtain a plurality of training near-field component data;

[0013] A neural network model is constructed, wherein the neural network model includes a plurality of fully connected neural networks corresponding one-to-one to each of the training near-field component data.

[0014] In one embodiment, the step of training the neural network model based on the model training data set to obtain a trained vector simulation model includes:

[0015] For the first training metasurface optical element in the model training data set, the first training metasurface optical element is divided into multiple groups of training units according to a preset unit division strategy, wherein the training unit includes a target training unit structure and multiple non-target training unit structures, and there is a structural coupling between the target training unit structure and each of the non-target training unit structures, and the first training metasurface optical element is obtained by combining the target training unit structures of each of the training units;

[0016] Acquire the unit training near-field component data of each training unit from each training near-field component data, and acquire the unit training structure parameters of each training unit from each training structure parameter;

[0017] Based on the unit training structure parameters of each training unit and the unit training near-field component data, the neural network model is trained to obtain a trained vector simulation model.

[0018] In one embodiment, the step of training the neural network model based on the unit training structure parameters of each training unit and the unit training near-field component data to obtain a trained vector simulation model includes:

[0019] For each of the training units, inputting the unit training structure parameters into a neural network model to obtain corresponding predicted training near-field component data;

[0020] A loss value is calculated based on the mean absolute error between the predicted training near-field component data and the unit training near-field component data;

[0021] The parameters of the neural network model are updated according to the loss value, and each training unit is iteratively trained to obtain a trained vector simulation model.

[0022] In one embodiment, the step of inputting the structural parameters into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element includes:

[0023] Inputting the structural parameters into a preset vector simulation model, dividing the target metasurface optical element into a plurality of groups of simulation units through the vector simulation model, wherein the simulation unit includes a target unit structure and a plurality of non-target unit structures, there is structural coupling between the target unit structure and each of the non-target unit structures, and the metasurface optical element is obtained by combining the target unit structures of each of the simulation units;

[0024] Generate predicted near-field component data of each target unit structure through the vector simulation model;

[0025] The predicted near-field component data are combined to obtain the predicted unit near-field light field data of each target unit structure, and the predicted unit near-field light field data are used as the vector near-field simulation result of the target metasurface optical element.

[0026] In one embodiment, the step of obtaining the far-field light effect information of the target metasurface optical element according to the vector near-field simulation result includes:

[0027] The near-field light field data of each prediction unit are spliced ​​according to the position arrangement of each target unit structure to obtain the target predicted near-field light field data of the target metasurface optical element;

[0028] The target predicted near-field light field data is converted into far-field light effect information of the target metasurface optical element by using Fourier transform.

[0029] In addition, to achieve the above purpose, the present application also proposes a vector simulation device, which includes:

[0030] A mapping module, used to perform mapping calculation according to a preset scalar phase to obtain the structural parameters of the target metasurface optical element;

[0031] A simulation module is used to input the structural parameters into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element, and obtain far-field optical effect information of the target metasurface optical element based on the vector near-field simulation result, wherein the vector simulation model is a trained neural network model.

[0032] In addition, to achieve the above objectives, the present application also proposes an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vector simulation method as described above.

[0033] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the vector simulation method as described above are implemented.

[0034] The present application proposes a vector simulation method, in which the structural parameters of a target metasurface optical element are calculated based on preset scalar phase information and a specific mapping algorithm; then, the calculated structural parameters are input into a pre-trained neural network model, which is a vector simulation model. The vector near-field simulation result of the target metasurface optical element is obtained through processing by this model, and the far-field optical effect information of the target metasurface optical element is obtained based on the vector near-field simulation result. The far-field optical effect information is crucial for evaluating and optimizing the performance of the target metasurface optical element.

[0035] In summary, the present application introduces a trained neural network model as a vector simulation model to fully utilize the advantages of neural networks in processing complex data and rapid prediction. The vector simulation model can quickly and accurately generate vector near-field simulation results based on the input structural parameters, thereby greatly shortening the simulation time and reducing the consumption of computing resources, thereby broadening the design and optimization space of metasurface optical elements. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 A schematic diagram of a process flow provided for the first embodiment of the vector simulation method of the present application;

[0039] Figure 2 A schematic diagram of the design process of a metasurface optical element provided in Embodiment 1 of the vector simulation method of the present application;

[0040] Figure 3 A schematic diagram of a neural network architecture provided for the second embodiment of the vector simulation method of the present application;

[0041] Figure 4 A schematic diagram of training near-field component data provided for the second embodiment of the vector simulation method of the present application;

[0042] Figure 5 A schematic diagram of the real component of Ex provided in Embodiment 2 of the vector simulation method of this application;

[0043] Figure 6 A schematic diagram of the imaginary component of Ex provided in Embodiment 2 of the vector simulation method of this application;

[0044] Figure 7 A schematic diagram of predicted near-field light field data provided for the second embodiment of the vector simulation method of the present application;

[0045] Figure 8 A schematic diagram of far-field light effect information provided in Embodiment 2 of the vector simulation method of this application;

[0046] Fig. 9 This is a schematic diagram of the module structure of the vector simulation device according to an embodiment of the present application;

[0047] Fig.10 Schematic diagram of the device structure of the hardware operating environment involved in the vector simulation method in the embodiment of the present application.

[0048] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0049] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0050] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0051] Currently, when dealing with large-scale MOE vector simulation tasks, due to the structural complexity and high accuracy requirements of metasurface optical elements, the traditional simulation methods are often time-consuming and consume a lot of computing resources, which affects the design and optimization process of metasurface optical elements.

[0052] An embodiment of the present application provides a solution, which performs mapping calculation according to a preset scalar phase to obtain the structural parameters of the target metasurface optical element; inputs the structural parameters into a preset vector simulation model to obtain the vector near-field simulation result of the target metasurface optical element, and obtains the far-field optical effect information of the target metasurface optical element based on the vector near-field simulation result, wherein the vector simulation model is a trained neural network model.

[0053] The embodiment of the present application introduces a trained neural network model as a vector simulation model to fully utilize the advantages of neural networks in processing complex data and rapid prediction. The vector simulation model can quickly and accurately generate vector near-field simulation results based on the input structural parameters, thereby greatly shortening the simulation time and reducing the consumption of computing resources, thereby broadening the design and optimization space of metasurface optical elements.

[0054] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a computer, a server, etc., or an electronic device capable of realizing the above functions, etc. The following takes a vector simulation terminal as an example to illustrate this embodiment and the following embodiments.

[0055] Based on this, the embodiment of the present application provides a vector simulation method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the vector simulation method of the present application.

[0056] In this embodiment, the vector simulation method includes steps S10 to S40:

[0057] Step S10, performing mapping calculation according to a preset scalar phase to obtain structural parameters of the target metasurface optical element;

[0058] It should be noted that scalar phase refers to the change in the phase of light waves during propagation, and is an important reference for designing metasurface optical elements.

[0059] According to the preset scalar phase, a mapping algorithm is used to convert the scalar phase information into the structural parameters of the target metasurface optical element. The structural parameters include the arrangement and size of the unit structure in the target metasurface optical element. In this way, the modeling of the target metasurface optical element structure can be completed through the generated structural parameters.

[0060] Taking the cylindrical unit structure in the metasurface optical element as an example, the phase information is calculated by the GS algorithm (Gerchberg-Saxton, phase recovery algorithm), and then the phase information is mapped to the cylinder radius through the established metasurface unit structure. After obtaining the cylinder radius information, the metasurface structure is modeled.

[0061] Step S20, inputting the structural parameters into a preset vector simulation model to obtain the vector near-field simulation results of the target metasurface optical element, and obtaining the far-field optical effect information of the target metasurface optical element according to the vector near-field simulation results, wherein the vector simulation model is a trained neural network model.

[0062] After obtaining the structural parameters of the target metasurface optical element, they are input into a pre-trained neural network model, namely the vector simulation model. After a large amount of data training and verification, the neural network model can accurately simulate the behavior of the metasurface optical element under the action of light waves. The vector simulation model not only considers the phase information of the light wave, but also covers vector characteristics such as amplitude and polarization, thereby being able to more comprehensively reflect the optical performance of the metasurface optical element.

[0063] The vector near-field simulation results of the target metasurface optical element are obtained through the operation of the model. The results are used to display the training near-field light field data of the target metasurface optical element. Furthermore, the far-field light effect information of the target metasurface optical element can be calculated according to the vector near-field simulation results. The far-field light effect information refers to the light effect generated by the target metasurface optical element at a position far away from its surface, such as the shape, size, brightness distribution of the light spot, etc. This information is crucial for evaluating the performance of the target metasurface optical element in practical applications.

[0064] In addition, in a feasible implementation scheme, taking the cylindrical unit structure in the metasurface optical element as an example, the design process of the metasurface optical element is as follows: Figure 2 As shown, the phase information is first calculated by the GS algorithm and other methods, and then the phase information is mapped to the cylinder radius through the established metasurface unit structure. After the cylinder radius information is obtained, the metasurface structure is modeled, the metasurface structure is vector-simulated, and finally the far-field light effect is reconstructed.

[0065] In this embodiment, a trained neural network model is introduced as a vector simulation model to fully utilize the advantages of neural networks in processing complex data and rapid prediction. The vector simulation model can quickly and accurately generate vector near-field simulation results based on the input structural parameters, thereby greatly shortening the simulation time and reducing the consumption of computing resources, thereby broadening the design and optimization space of metasurface optical elements.

[0066] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. On this basis, steps A10 to A20 may also be included before step S20:

[0067] Step A10, establishing a model training data set, wherein the model training data set includes training structural parameters and training near-field light field data of each training metasurface optical element;

[0068] A model training data set is constructed, which includes the training structural parameters and training near-field light field data of each training metasurface optical element, wherein the training metasurface optical element refers to the metasurface optical element used for model training, and its training structural parameters describe the geometric shape, size, material properties, unit structure and other characteristics of the element, which are the objects that the neural network model needs to learn and predict; its training near-field light field data refers to the light field distribution generated by light waves near the surface of the element, including vector information such as the intensity, phase, polarization and other vector information of the electric and magnetic fields, which are the targets that the neural network model needs to predict.

[0069] Step A20, training the neural network model based on the model training data set to obtain a trained vector simulation model.

[0070] After obtaining the model training data set, it is input into the neural network model for model training. The training process is an iterative optimization process. By continuously adjusting the weights and bias parameters of the neural network, the predicted output of the model gradually approaches the actual training near-field light field data.

[0071] During the training process, the loss function and optimization algorithm are selected according to the application scenario to guide the training of the model. The loss function is used to measure the difference between the model's predicted output and the actual output, and the optimization algorithm is used to minimize this difference. Through multiple iterations of training, a trained vector simulation model is obtained, which can accurately predict the near-field light field data of metasurface optical elements.

[0072] In a feasible embodiment, before step A20, steps A30 to A40 may also be included:

[0073] Step A30, for each of the training near-field light field data, decomposing the training near-field light field data according to electric field components and component forms to obtain a plurality of training near-field component data;

[0074] For each training near-field light field data in the data set, the training near-field light field data is decomposed to improve the training efficiency and prediction accuracy of the model. Specifically, the training near-field light field data is decomposed according to the electric field component and component form to obtain multiple training near-field component data, wherein the electric field component can be divided into various components of the electric field, such as electric field components Ex, Ey, and Ez, and each component is in complex form, and each electric field component is decomposed into real and imaginary parts based on the component form to obtain six training near-field component data, namely the real and imaginary parts of three electric field components.

[0075] Step A40, constructing a neural network model, wherein the neural network model includes a plurality of fully connected neural networks corresponding one-to-one to each of the training near-field component data.

[0076] A neural network model is constructed to process these training near-field component data. The model includes multiple fully-connected neural networks corresponding to each training near-field component data. Each fully-connected neural network is responsible for processing a specific training near-field component data and outputting a prediction result of the component. These fully-connected neural networks can adopt the same structure or different structures, depending on the complexity of the problem and the characteristics of the data set.

[0077] In this way, by decomposing the training near-field light field data, complex vector information can be converted into multiple simple scalar information, thereby reducing the complexity of the model. At the same time, since each component data contains specific information about the light field, they are input into different fully connected neural networks respectively, which can enable the model to better learn and predict the training near-field light field data.

[0078] In a feasible embodiment, step A20 may include steps A201 to A203:

[0079] Step A201, for the first training metasurface optical element in the model training data set, divide the first training metasurface optical element into multiple groups of training units according to a preset unit division strategy, wherein the training unit includes a target training unit structure and multiple non-target training unit structures, there is a structural coupling between the target training unit structure and each of the non-target training unit structures, and the first training metasurface optical element is obtained by combining the target training unit structures of each of the training units;

[0080] It should be noted that the unit division strategy is determined based on factors such as the physical size, structural characteristics or functional requirements of the metasurface optical element. In the present embodiment, considering the mutual coupling effect between the structures of the metasurface optical element, it is preferred to divide the element into multiple unit structures to form a 3×3 training unit. There is a structural coupling between the target training unit structure in each group of training units (i.e., the unit structure at the center of the 3×3 training unit) and the surrounding non-target training unit structures.

[0081] For any training metasurface optical element in the model training data set, that is, the first training metasurface optical element, the first training metasurface optical element is divided into multiple groups of training units according to the unit division strategy. The training unit includes a target training unit structure and multiple non-target training unit structures. For example, when the training unit is in the form of 3×3, the unit structure at the center position is the target training unit structure, and the remaining 8 unit structures around the center are non-target training unit structures. There is structural coupling between the target training unit structure and each non-target training unit structure, and the first training metasurface optical element is obtained by combining the target training unit structures of each of the training units.

[0082] It should also be noted that for the unit structure at the edge of the first training metasurface optical element, since this part of the edge unit structure accounts for a small proportion and there are insufficient coupled units, this part of the edge unit structure is usually randomly distributed, rather than being trained separately as a target training unit structure.

[0083] Step A202, acquiring the unit training near-field component data of each training unit from each training near-field component data, and acquiring the unit training structure parameters of each training unit from each training structure parameter;

[0084] From the training near-field component data, for each training unit, the corresponding unit training near-field component data is extracted. These data reflect the distribution and changes of the electromagnetic field in the near-field area. At the same time, from the training structure parameters, the unit training structure parameters of each training unit are obtained. These parameters describe the size, shape, material properties and other characteristics of the training unit.

[0085] Step A203: training the neural network model based on the unit training structure parameters and the unit training near-field component data of each training unit to obtain a trained vector simulation model.

[0086] The unit training structural parameters and unit training near-field component data of each training unit are used to train the neural network model to obtain a trained vector simulation model, which can quickly predict the near-field light field information of the metasurface optical element based on the given structural parameters. The neural network model can be a convolutional neural network, a recurrent neural network, or a deep neural network, etc., which is customized according to the complexity of the problem and the characteristics of the data, and is not specifically limited in this embodiment.

[0087] For example, the neural network model used in this embodiment can be as follows: Figure 3 The neural network architecture shown includes an input layer, an output layer and eight intermediate layers. Based on the actual application scenario, the number of layers and / or the weight matrix of each layer of the neural network can be adjusted, which is not specifically limited in this embodiment.

[0088] In a feasible embodiment, step A203 may include steps A2031 to A2033:

[0089] Step A2031, for each of the training units, input the unit training structure parameters into a neural network model to obtain corresponding predicted training near-field component data;

[0090] For each training unit, the unit training structure parameters of these training units are used as input information and input into the constructed neural network model. The model performs internal calculations and processing based on these input parameters, and finally outputs the corresponding predicted training near-field component data. These predicted training near-field component data reflect the electromagnetic field near-field characteristics of the training unit.

[0091] Step A2032, calculating a loss value according to the mean absolute error between the predicted training near-field component data and the unit training near-field component data;

[0092] The obtained predicted training near-field component data are compared with the known, actual unit training near-field component data, and the mean absolute error between them is calculated. This error value is used as a loss value to measure the prediction accuracy of the neural network model, reflecting the degree of difference between the model prediction results and the actual data.

[0093] Step A2033, updating the parameters of the neural network model according to the loss value, and iteratively training each training unit to obtain a trained vector simulation model.

[0094] Based on the calculated loss value, the internal parameters of the neural network model are adjusted and updated to reduce the prediction error of the model and improve the prediction accuracy of the model. Subsequently, the updated model parameters are used to conduct a new round of iterative training for all training units. Through multiple such iterative processes, the model is gradually optimized until the predetermined training effect is achieved, and finally a fully trained vector simulation model with excellent performance is obtained.

[0095] For example, the six training near-field component data obtained by neural network training are as follows: Figure 4 As shown in the figure, the real part of Ex (Re), the imaginary part of Ex (Im), the real part of Ey, the imaginary part of Ey, the real part of Ez and the imaginary part of Ez. The horizontal and vertical axes in the figure represent the position coordinates (x, y) of each unit structure. According to the figure, the near-field component of each training unit can be determined.

[0096] Then, taking the real and imaginary components of Ex as an example, the unit training near-field component data of the real part of the Ex component is compared with the predicted training near-field component data, such as Figure 5 As shown, Exrealtest represents the unit training near-field component data of the real part of the Ex component, Exrealpred represents the predicted training near-field component data of the real part of the Ex component, and Exrealmae represents the average absolute error between the predicted training near-field component data and the unit training near-field component data. The horizontal and vertical coordinates in the figure represent the position coordinates (x, y). The unit training near-field component data of the imaginary part of the Ex component is compared with the predicted training near-field component data, as shown in FIG. Figure 6As shown, Eximagtest represents the unit training near-field component data of the imaginary part in the Ex component, Eximagpred represents the predicted training near-field component data of the imaginary part in the Ex component, Eximagmae represents the average absolute error between the predicted training near-field component data and the unit training near-field component data, and the horizontal and vertical coordinates in the figure represent the position coordinates (x, y).

[0097] In a feasible embodiment, the step of "inputting the structural parameters into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element" in step S20 may include steps S201 to S203:

[0098] Step S201, inputting the structural parameters into a preset vector simulation model, dividing the target metasurface optical element into a plurality of groups of simulation units through the vector simulation model, wherein the simulation unit includes a target unit structure and a plurality of non-target unit structures, there is structural coupling between the target unit structure and each of the non-target unit structures, and the metasurface optical element is obtained by combining the target unit structures of each of the simulation units;

[0099] The structural parameters of the target metasurface optical element are input into the vector simulation model, and the target metasurface optical element is divided into multiple groups of simulation units according to the unit division strategy. The simulation unit includes a target unit structure and multiple non-target unit structures. For example, when the simulation unit is in the form of 3×3, the unit structure at the center is the target unit structure, and the remaining 8 unit structures around the center are non-target unit structures. There is structural coupling between the target unit structure and each non-target unit structure, and the target metasurface optical element is obtained by combining the target unit structures of each simulation unit.

[0100] Step S202, generating predicted near-field component data of each target unit structure through the vector simulation model;

[0101] Step S203, combining the predicted near-field component data to obtain the predicted unit near-field light field data of each target unit structure, and using the predicted unit near-field light field data as the vector near-field simulation result of the target metasurface optical element.

[0102] The predicted near-field component data of each target unit structure are generated through a vector simulation model. The predicted near-field component data include the real and imaginary parts of the three electric field components Ex, Ey, and Ez, that is, a total of six predicted near-field component data. The predicted unit near-field light field data of each target unit structure are obtained by combining them. The predicted unit near-field light field data include the electromagnetic field information and optical properties of each target unit structure. Finally, these predicted unit near-field light field data are output as the vector near-field simulation results of the target metasurface optical element.

[0103] In a feasible embodiment, the step of "obtaining the far-field light effect information of the target metasurface optical element according to the vector near-field simulation result" in step S20 may include steps S204 to S205:

[0104] Step S204, splicing the near-field light field data of each of the prediction units according to the position arrangement of each of the target unit structures to obtain target predicted near-field light field data of the target metasurface optical element;

[0105] The near-field light field data of each prediction unit are spliced ​​according to the positions of their corresponding target unit structures on the metasurface optical element, so as to integrate the local near-field light field data to form a complete and accurate predicted near-field light field data of the target metasurface optical element.

[0106] Step S205, using Fourier transform to convert the target predicted near-field light field data into far-field light effect information of the target metasurface optical element.

[0107] After completing the splicing of the predicted near-field light field data, the Fourier transform method is used to convert this near-field data into far-field light effect information for the subsequent design and optimization of the target metasurface optical elements.

[0108] For example, the near-field light field data of each prediction unit are spliced ​​according to the positions of their corresponding target unit structures on the metasurface optical element, and the predicted near-field light field data of the spliced ​​target metasurface optical element is obtained as follows: Figure 7 As shown, Figure 7 In the example, each simulation unit is 3×3, and then the target predicted near-field light field data is converted into the far-field light effect information of the target metasurface optical element through Fourier transform. Figure 8 As shown in the figure, the horizontal and vertical axes represent the position coordinates (x, y) of the far-field light effect information. According to the figure, the far-field light effect of the target metasurface optical element can be determined.

[0109] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the vector simulation method of the present application. More simple transformations based on this technical concept are all within the protection scope of the present application.

[0110] The present application also provides a vector simulation device, please refer to Fig. 9 , the vector simulation device comprises:

[0111] A mapping module 10 is used to perform mapping calculation according to a preset scalar phase to obtain the structural parameters of the target metasurface optical element;

[0112] The simulation module 20 is used to input the structural parameters into a preset vector simulation model to obtain the vector near-field simulation results of the target metasurface optical element, and obtain the far-field optical effect information of the target metasurface optical element based on the vector near-field simulation results, wherein the vector simulation model is a trained neural network model.

[0113] Optionally, the vector simulation device further includes a model training module (not shown), and the model training module is used to:

[0114] Establishing a model training data set, wherein the model training data set includes training structural parameters and training near-field light field data of each training metasurface optical element;

[0115] The neural network model is trained based on the model training data set to obtain a trained vector simulation model.

[0116] Optionally, the model training module is also used to:

[0117] For each of the training near-field light field data, decomposing the training near-field light field data according to electric field components and component forms to obtain a plurality of training near-field component data;

[0118] A neural network model is constructed, wherein the neural network model includes a plurality of fully connected neural networks corresponding one-to-one to each of the training near-field component data.

[0119] Optionally, the model training module is also used to:

[0120] For the first training metasurface optical element in the model training data set, the first training metasurface optical element is divided into multiple groups of training units according to a preset unit division strategy, wherein the training unit includes a target training unit structure and multiple non-target training unit structures, and there is a structural coupling between the target training unit structure and each of the non-target training unit structures, and the first training metasurface optical element is obtained by combining the target training unit structures of each of the training units;

[0121] Acquire the unit training near-field component data of each training unit from each training near-field component data, and acquire the unit training structure parameters of each training unit from each training structure parameter;

[0122] Based on the unit training structure parameters of each training unit and the unit training near-field component data, the neural network model is trained to obtain a trained vector simulation model.

[0123] Optionally, the model training module is also used to:

[0124] For each of the training units, inputting the unit training structure parameters into a neural network model to obtain corresponding predicted training near-field component data;

[0125] A loss value is calculated based on the mean absolute error between the predicted training near-field component data and the unit training near-field component data;

[0126] The parameters of the neural network model are updated according to the loss value, and each training unit is iteratively trained to obtain a trained vector simulation model.

[0127] Optionally, the simulation module 20 is further configured to:

[0128] Inputting the structural parameters into a preset vector simulation model, dividing the target metasurface optical element into a plurality of groups of simulation units through the vector simulation model, wherein the simulation unit includes a target unit structure and a plurality of non-target unit structures, there is structural coupling between the target unit structure and each of the non-target unit structures, and the metasurface optical element is obtained by combining the target unit structures of each of the simulation units;

[0129] Generate predicted near-field component data of each target unit structure through the vector simulation model;

[0130] The predicted near-field component data are combined to obtain the predicted unit near-field light field data of each target unit structure, and the predicted unit near-field light field data are used as the vector near-field simulation result of the target metasurface optical element.

[0131] Optionally, the simulation module 20 is further configured to:

[0132] The near-field light field data of each prediction unit are spliced ​​according to the position arrangement of each target unit structure to obtain the target predicted near-field light field data of the target metasurface optical element;

[0133] The target predicted near-field light field data is converted into far-field light effect information of the target metasurface optical element by using Fourier transform.

[0134] The vector simulation device provided in the embodiment of the present application adopts the vector simulation method in the above embodiment, which can improve the vector simulation speed for the metasurface optical element and reduce the consumption of computing resources. Compared with the prior art, the beneficial effects of the vector simulation device provided in the embodiment of the present application are the same as the beneficial effects of the vector simulation method provided in the above embodiment, and other technical features in the vector simulation device are the same as the features disclosed in the vector simulation method in the above embodiment, which will not be described in detail here.

[0135] An embodiment of the present application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the vector simulation method in the above-mentioned embodiment 1.

[0136] Reference below Fig.10 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present application. The electronic device in the embodiments of the present application may include but is not limited to mobile terminals such as laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.10 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0137] like Fig.10As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the electronic device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows an electronic device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0138] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0139] The electronic device provided in the embodiment of the present application adopts the vector simulation method in the above embodiment, which can improve the vector simulation speed for the metasurface optical element and reduce the consumption of computing resources. Compared with the prior art, the beneficial effects of the electronic device provided in the embodiment of the present application are the same as the beneficial effects of the vector simulation method provided in the above embodiment, and the other technical features in the electronic device are the same as the features disclosed in the vector simulation method in the previous embodiment, which will not be described in detail here.

[0140] It should be understood that the various parts disclosed in the embodiments of the present application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0141] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0142] An embodiment of the present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the vector simulation method in the above-mentioned embodiment.

[0143] The computer-readable storage medium provided in the embodiment of the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0144] The computer-readable storage medium may be included in the electronic device, or may exist independently without being installed in the electronic device.

[0145] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by an electronic device, the electronic device: performs mapping calculation according to a preset scalar phase to obtain the structural parameters of the target metasurface optical element; inputs the structural parameters into a preset vector simulation model to obtain the vector near-field simulation results of the target metasurface optical element, and obtains the far-field optical effect information of the target metasurface optical element according to the vector near-field simulation results, wherein the vector simulation model is a trained neural network model.

[0146] The computer program code for performing the operation of the embodiment of the present application can be written in one or more programming languages ​​or a combination thereof, and the above-mentioned programming languages ​​include object-oriented programming languages, such as Java, Smalltalk, C++, and also include conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer, partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN: Local Area Network) or a wide area network (WAN: Wide Area Network), or it can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0147] 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 application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the 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 square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square 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 square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / 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.

[0148] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0149] The readable storage medium provided in the embodiment of the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned vector simulation method, and can improve the vector simulation speed for metasurface optical elements while reducing the consumption of computing resources. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the embodiment of the present application are the same as the beneficial effects of the vector simulation method provided in the above-mentioned embodiment, and will not be described in detail here.

[0150] The above are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A vector simulation method, characterized in that: The vector simulation method comprises: The structural parameters of the target metasurface optical element are obtained by mapping and calculating according to the preset scalar phase; The structural parameters are input into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element, and the far-field optical effect information of the target metasurface optical element is obtained according to the vector near-field simulation result, wherein the vector simulation model is a trained neural network model.

2. The vector simulation method according to claim 1, characterized in that: Before the step of inputting the structural parameters into a preset vector simulation model to obtain the vector near-field simulation result of the target metasurface optical element, the method further includes: Establishing a model training data set, wherein the model training data set includes training structural parameters and training near-field light field data of each training metasurface optical element; The neural network model is trained based on the model training data set to obtain a trained vector simulation model.

3. The vector simulation method according to claim 2, characterized in that: Before the step of training the neural network model based on the model training data set to obtain a trained vector simulation model, the method further includes: For each of the training near-field light field data, decomposing the training near-field light field data according to electric field components and component forms to obtain a plurality of training near-field component data; A neural network model is constructed, wherein the neural network model includes a plurality of fully connected neural networks corresponding one-to-one to each of the training near-field component data.

4. The vector simulation method according to claim 3, characterized in that: The step of training the neural network model based on the model training data set to obtain a trained vector simulation model includes: For the first training metasurface optical element in the model training data set, the first training metasurface optical element is divided into multiple groups of training units according to a preset unit division strategy, wherein the training unit includes a target training unit structure and multiple non-target training unit structures, and there is a structural coupling between the target training unit structure and each of the non-target training unit structures, and the first training metasurface optical element is obtained by combining the target training unit structures of each of the training units; Acquire the unit training near-field component data of each training unit from each training near-field component data, and acquire the unit training structure parameters of each training unit from each training structure parameter; Based on the unit training structure parameters of each training unit and the unit training near-field component data, the neural network model is trained to obtain a trained vector simulation model.

5. The vector simulation method according to claim 4, characterized in that: The step of training the neural network model based on the unit training structure parameters of each training unit and the unit training near-field component data to obtain a trained vector simulation model comprises: For each of the training units, inputting the unit training structure parameters into a neural network model to obtain corresponding predicted training near-field component data; A loss value is calculated based on the mean absolute error between the predicted training near-field component data and the unit training near-field component data; The parameters of the neural network model are updated according to the loss value, and each training unit is iteratively trained to obtain a trained vector simulation model.

6. The vector simulation method according to claim 1, characterized in that: The step of inputting the structural parameters into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element comprises: Inputting the structural parameters into a preset vector simulation model, dividing the target metasurface optical element into a plurality of groups of simulation units through the vector simulation model, wherein the simulation unit includes a target unit structure and a plurality of non-target unit structures, there is structural coupling between the target unit structure and each of the non-target unit structures, and the metasurface optical element is obtained by combining the target unit structures of each of the simulation units; Generate predicted near-field component data of each target unit structure through the vector simulation model; The predicted near-field component data are combined to obtain the predicted unit near-field light field data of each target unit structure, and the predicted unit near-field light field data are used as the vector near-field simulation result of the target metasurface optical element.

7. The vector simulation method according to claim 6, characterized in that: The step of obtaining the far-field light effect information of the target metasurface optical element according to the vector near-field simulation result comprises: The near-field light field data of each prediction unit are spliced ​​according to the position arrangement of each target unit structure to obtain the target predicted near-field light field data of the target metasurface optical element; The target predicted near-field light field data is converted into far-field light effect information of the target metasurface optical element by using Fourier transform.

8. A vector simulation device, characterized in that: The device comprises: A mapping module, used to perform mapping calculation according to a preset scalar phase to obtain the structural parameters of the target metasurface optical element; A simulation module is used to input the structural parameters into a preset vector simulation model to obtain a vector near-field simulation result of the target metasurface optical element, and obtain far-field optical effect information of the target metasurface optical element based on the vector near-field simulation result, wherein the vector simulation model is a trained neural network model.

9. An electronic device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vector simulation method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the vector simulation method according to any one of claims 1 to 7 are implemented.

Images