Design method of reconfigurable metasurface unit based on current distribution and image segmentation

Abstract

The invention relates to the technical field of electromagnetic metasurface devices, and discloses a design method of a reconfigurable metasurface unit based on current distribution and image segmentation, comprising the following steps: S1, graying a current distribution image: the current distribution of a reconfigurable metasurface is a color image and has information of three channels, the information amount is large, and the graying time is short; and the passive metal surface of the reconfigurable metasurface only has two conditions of'existence 'and'absence' of a metal patch, and the information capable of being represented is less, so that the obtained current distribution diagram needs to be processed, and useful information needs to be extracted. According to the method, a new structure topology is generated for the metasurface unit. The method focuses on the topological design of the reconfigurable metasurface, the current distribution is incorporated into the topological design of the reconfigurable metasurface unit, and the current distribution is effectively mapped into the metasurface unit structure by using the image segmentation technology, so that the miniaturization of the reconfigurable metasurface unit is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of electromagnetic metasurface devices, and in particular to a design method for a reconfigurable metasurface unit based on current distribution and image segmentation. Background Art

[0002] Reconfigurable metasurface design has broad application prospects in many fields. In the field of wireless communications, it can improve the capacity and coverage of communication systems and realize smart antenna functions. In the field of radar and remote sensing, it can be used for target detection, imaging, and interference suppression. In the design of reconfigurable metasurfaces, it mainly covers two aspects: parameter optimization and topology optimization. Among them, parameter optimization requires a comprehensive exploration of the parameter space of the initial structure to determine whether it meets specific design goals.

[0003] At present, the reconfigurable metasurface topology optimization methods are mainly divided into two categories: the first category is topology optimization based on machine learning, which uses the data set obtained from electromagnetic simulation to train the machine learning model, and then uses the model to obtain the reconfigurable metasurface topology design that meets specific design standards. The core of this method is machine learning modeling, so a large amount of electromagnetic simulation data is required to establish the relationship between different topology designs and electromagnetic responses. The certainty of this method is that the computational cost is relatively high.

[0004] The second type is a topological design method based on a multi-port network. This method divides the reconfigurable metasurface into multiple rectangular metal patches, connects the metal patches with discrete multi-ports, extracts the impedance parameters / reflection parameters of the multi-ports, and analyzes the load parameters of the multi-ports to establish the relationship between the multi-port load and the reflection coefficient of the reconfigurable metasurface. This method can reduce the reliance on electromagnetic simulation for the topological design of the metasurface unit, thereby shortening the design time. However, the reconfigurable metasurface generated by this method is similar to the original design, making it difficult to miniaturize the unit structure.

[0005] Therefore, it is necessary to design a practical design method for reconfigurable metasurface units based on current distribution and image segmentation. Summary of the invention

[0006] The purpose of the present invention is to provide a design method for a reconfigurable metasurface unit based on current distribution and image segmentation to solve the problems raised in the above background technology. In the design of a reconfigurable metasurface, the area with higher current density has a greater impact on the performance of the reconfigurable metasurface. Therefore, eliminating the metal patch with lower current has less impact on the overall performance of the metasurface, and the miniaturization of the reconfigurable metasurface can be achieved.

[0007] In order to solve the above technical problems, the present invention provides the following technical solution: a design method of a reconfigurable metasurface unit based on current distribution and image segmentation, comprising the following steps:

[0008] S1: Grayscale current distribution image

[0009] The current distribution of the reconfigurable metasurface is a color image with three channels of information, which has a large amount of information. However, the passive metal surface of the reconfigurable metasurface has only two conditions: "with" and "without" metal patches, which can represent less information. Therefore, it is necessary to process the obtained current distribution map to extract useful information. By graying the current distribution, useful information in the current distribution map of the reconfigurable metasurface unit can be extracted.

[0010] S2: Current distribution image segmentation

[0011] Since the range of each pixel value of the grayscale processed image is 0-255, the amount of information is still large for the passive metal surface of the reconfigurable metasurface unit, so it is necessary to further simplify the image information. Through image segmentation technology, the pixel value of the image is simplified to two possibilities: "0" and "1";

[0012] S3: Transformation of current distribution images into reconfigurable metasurface structures

[0013] In order to convert a two-dimensional image into a reconfigurable metasurface structure, the image domain information needs to be converted into the physical structure of the metal patch. The segmented image pixel matrix consists of two elements "0" and "1". Through linear mapping, the pixel value "1" is mapped to the metal structure of the specified size.

[0014] S4: Reconfigurable metasurface unit design and system design in different working states

[0015] Once the topological structure of the passive metal patch on the top is determined, the structure of the reconfigurable metasurface unit is constructed by adding a substrate and a ground layer. In order to achieve dynamic regulation of electromagnetic waves, the quasi-Newton gradient optimization algorithm is used to obtain metasurface units in different working states. The generalized Snell's theorem is used to construct a reconfigurable metasurface system based on the regulation requirements of electromagnetic waves.

[0016] According to the above technical solution, the image segmentation technology adopts OTSU threshold segmentation technology.

[0017] According to the above technical solution, the goal of the OTSU threshold calculation is to maximize the ratio of the between-class variance to the within-class variance, as shown below:

[0018]

[0019] In the formula, is the between-class variance, is the intra-class variance, μ1 and μ2 are the means of the two classes, δ1 and δ2 are the variances of the two classes;

[0020] OTSU threshold segmentation calculates the intra-class variance and inter-class variance for each gray level in the image, and selects the gray level that maximizes the inter-class variance as the threshold; the OTSU threshold calculation formula is as follows:

[0021]

[0022] In the formula, is the optimal threshold.

[0023] Compared with the prior art, the beneficial effects achieved by the present invention are: the method of the present invention generates a new structural topology for the metasurface unit, transcends geometric constraints, and can identify and establish the optimal structural topology. The present invention focuses on the topological design of the reconfigurable metasurface, incorporates the current distribution into the topological design of the reconfigurable metasurface unit, and uses image segmentation technology to effectively map the current distribution to the metasurface unit structure, thereby realizing the miniaturization of the reconfigurable metasurface unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0025] Figure 1 is the original structure and component diagram of the reconfigurable metasurface unit of the present invention;

[0026] Figure 2 is a schematic diagram of current distribution of the reconfigurable metasurface of the present invention under different working states;

[0027] Figure 3 is a current distribution diagram of the present invention;

[0028] Figure 4 is a schematic diagram of the relationship between the pixel matrix of the present invention and the topology of the reconfigurable metasurface;

[0029] Figure 5 is a design flow chart of the reconfigurable metasurface of the present invention;

[0030] Figure 6 is an initial structural parameter diagram of the reconfigurable metasurface unit of the present invention;

[0031] Figure 7 is a current distribution diagram of the grayscale processing of the present invention;

[0032] Figure 8 is a current distribution diagram after segmentation of the present invention;

[0033] Fig. 9 is a new topological structure diagram of the reconfigurable metasurface unit of the present invention;

[0034] Fig.10 It is the amplitude curve diagram of eight states of the present invention;

[0035] Fig.11 is a phase curve diagram of eight states of the present invention;

[0036] Fig.12 is a diagram of a reconfigurable metasurface system of the present invention;

[0037] Fig.13 It is the far-field pattern of the “0000222244446666” code of the present invention;

[0038] Fig.14 It is the far-field pattern of the "4432210077655433" code of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] See also Figure 1-14 ,The present invention provides a technical solution: a design method of a reconfigurable metasurface unit based on current distribution and image segmentation, comprising the following steps;

[0041] S1: Grayscale current distribution image

[0042] Taking the reflective rectangular patch composed of a reconfigurable metasurface unit as an example, the schematic diagram of its structural model is as follows Figure 1 As shown, it includes adjustable devices, metal patch layers, dielectric board layers, metal grounding layers, etc.

[0043] First, the current distribution of the reconfigurable metasurface under different working conditions at different times changes dynamically. Figure 2 (a), (b) and (c) show the current distribution under different working conditions. Figure 2 (d), (e) and (f) are the current distributions at 0, T / 8, and T / 4. In order to comprehensively consider the working performance of the reconfigurable metasurface under various working states, the maximum value of the current distribution under different states and at different times is selected, namely

[0044]

[0045] Where t represents time, s represents working state, and T is the working period of the metasurface unit (the inverse of the working frequency).

[0046] (a), (b) and (c) represent the current distribution under different working conditions, (d), (e) and (f) represent the current distribution at time 0, T / 8 and T / 4.

[0047] Secondly, the current distribution map Ds of the reconfigurable metasurface unit is determined by three components: R(i,j), G(i,j) and B(i,j), where the value range of i is from 1 to I and the value range of j is from 1 to J. Here, I and J represent the number of rows and columns in the matrix, respectively, and the value range of each component is from 0 to 255. Since the amount of information in the current distribution map Ds is large, the grayscale technology is used to process the current distribution image, and the effective information of the topology design is extracted from the three channels of R(i,j), G(i,j) and B(i,j), such as Figure 3 As shown. Because the red information in the colored current distribution diagram represents the intensity of the current, the red channel information is selected as the grayscale processing object, which is shown as follows:

[0048] G=R (2)

[0049] Where G is the grayscale image matrix and R is the pixel matrix of the red channel.

[0050] S2: Current distribution image segmentation

[0051] According to the grayscale image G obtained in the first step, the value range of each pixel in the image matrix is ​​0-255. In the reconfigurable metasurface, the passive metal structure on the top only has two cases: "with" and "without" metal patches. The grayscale image G is segmented so that each pixel in the image has a value of "0" or "1".

[0052] The present invention uses OTSU threshold segmentation technology to achieve image segmentation. OTSU threshold segmentation is an automatic image segmentation technology that maximizes the inter-class variance between the target object and the background while minimizing the intra-class variance. The inter-class variance is used to evaluate the degree of difference in pixel grayscale values ​​between different classes, and the intra-class variance quantifies the degree of dispersion of pixel grayscale values ​​within a single class. The goal of OTSU threshold calculation is to maximize the ratio of inter-class variance to intra-class variance, as shown below:

[0053]

[0054] In the formula, is the between-class variance, is the intra-class variance. μ1 and μ2 are the means of the two classes, respectively. δ1 and δ2 are the variances of the two classes, respectively.

[0055] OTSU threshold segmentation calculates the intra-class variance and inter-class variance for each gray level in the image, and selects the gray level that maximizes the inter-class variance as the threshold. The OTSU threshold calculation formula is as follows:

[0056]

[0057] In the formula, is the optimal threshold.

[0058] The pixel G(i,j) in the grayscale image matrix has a value range of 0-255, which contains a total of 256 different pixel value information. The reconfigurable metasurface structure only needs 0 and 1 to indicate the presence or absence of metal patch areas. By analyzing the intensity characteristics of the pixel value, the object and background in the image are distinguished and the threshold processing is completed. G(i,j) is compared with For comparison, the image information is converted into binary information of the pixels in the i-th row and j-th column, which is defined as:

[0059]

[0060] S3: Transformation of current distribution images into reconfigurable metasurface unit structures

[0061] The two-dimensional image processed in step S2 is converted into the structure of a reconfigurable metasurface unit. The segmentation image matrix P(i, j) consists of "0" and "1". P(i, j) = 1 is mapped to the metal structure of the reconfigurable metasurface, and P(i, j) = 0 is mapped to the metal-free area.

[0062] First, due to the large amount of pixel data in P(i,j), the complexity of directly converting the topological structure is very high. Sampling sparse technology, for the P(i,j) matrix of size I×J, obtain the sparse matrix P t , the sparse matrix P t The size is where γ x , γ y are the sparse factors of the rows and columns of the matrix P, respectively.

[0063] Then, for P t The pixel value of 1 is converted into a metal patch in the reconfigurable metasurface structure, such as Figure 4 As shown. The length and width of the metal patch are set to l x , l y , as shown below:

[0064] l x =γ x ·L d / I (6)

[0065] l y =γ y ·W d / J (7)

[0066] Where, L d , Wd Respectively represent the maximum length and maximum width of the reconfigurable metasurface structure, I and J represent the number of rows and columns in the matrix P

[0067] S4: Reconfigurable metasurface unit design and system design in different working states

[0068] Based on the metal patch topology obtained in step S3, the structure of the reconfigurable intelligent metasurface unit is obtained by adding a dielectric substrate and a metal ground. In order to optimize the parameters of the adjustable devices of the unit for different working states of the metasurface unit, the i-th operating state of the reconfigurable metasurface is represented as S i (x), the optimization target of the adjustable device parameters in this state is:

[0069]

[0070] Where x is the design variable of the adjustable device, x * is the optimal parameter of the adjustable device, S 11 (x) is the reflection coefficient of the metasurface unit, ||·|| 2 is the bi-norm function of .

[0071] Since the equivalent parameter dimensions of the adjustable device are relatively small and the optimization problem is relatively simple, the quasi-Newton gradient optimization algorithm is used for optimization, and the parameter equal division values ​​of the adjustable device are used as the initial values, as follows:

[0072]

[0073] The quasi-Newton gradient optimization algorithm does not need to calculate the second-order derivative and its inverse. The iterative formula of the algorithm is as follows:

[0074]

[0075] In the formula, and They represent the optimal solutions of the t+1th and tth generations respectively, and H t+1 is the Hessen matrix of the t+1th generation, is the change in the design variable x.

[0076] Finally, using the generalized Snell's theorem, a reconfigurable metasurface system is constructed based on metasurface units in different working states according to the needs of electromagnetic wave control.

[0077] Experimental example:

[0078] The present invention has been verified by electromagnetic simulation data, and the reconstructible metasurface is used as an object, and its initial structure is as follows Figure 6As shown, the operating frequency is 2.6GHz, with 8 phase states: 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°. In these 8 phase states, the amplitude of the reflection coefficient is less than -3dB, as shown in Table 1. Considering the changes in the passive structure and the working state of the adjustable device, the current distribution of the reconfigurable metasurface will change accordingly. The metal patch at the top center of the 3-bit metasurface is connected by a varactor diode. The equivalent circuit of the varactor diode is an RLC series circuit, R = 0.3Ω, L = 0.7nH, and the equivalent capacitance ranges from 0.6pF to 2.6pF. Within the extreme working range of the varactor diode, the maximum current distribution diagram of the reconfigurable metasurface unit is obtained.

[0079] The length and width of the current distribution map are the size of the reconfigurable metasurface unit. The color current distribution image information consists of three channels: R, G, and B. The value of each channel is 0-255. The information of the red channel represents the strong current distribution area of ​​the reconfigurable metasurface. Image grayscale is used for image data processing. The grayscale processed current distribution image is shown in the figure. Figure 7 As shown in the figure, it is composed of white area and black area, and the pixel value of its image information is 0-255. The current distribution image segmentation uses the OTSU threshold image segmentation method to convert the image information into a binary image, as shown in Figure 8 shown.

[0080] First, take the 3-bit phase-modulated reconfigurable metasurface as an example. Its structure consists of a metal stratum, a dielectric plate layer, and a metal patch layer. The topological structure design is applied to the metal patch layer. Electromagnetic simulation is performed on the initial structure to obtain the current distribution map. Considering that the design of the reconfigurable metasurface requires the analysis of the current distribution in all possible states and at all times. Under the condition of the adjustable device working limit, the maximum current distribution map is selected as the initial topological map of the topological design.

[0081] Based on the OTSU threshold image segmentation method, the grayscale current distribution map is segmented and the pixel information value of 0-255 is transformed into a value of 0-1, such as Figure 7 As shown. Further, considering mapping the image to the physical structure of the reconfigurable metasurface, the image data volume is large, resulting in a high complexity of the topological structure directly converted to the reconfigurable metasurface. Sparse processing is used, and the sparse matrix is ​​converted into the structure of the reconfigurable metasurface unit using linear mapping, and then the dielectric plate and the floor are added to form the metasurface unit, as shown in Fig. 9 shown.

[0082] Table 1 Amplitude and phase of 8 phase states of reconfigurable metasurface unit

[0083]

[0084] exist Fig. 9 In the figure, the upper and lower metal patches are connected by an adjustable device, which is a varactor diode. Its effective circuit is a series R, L, C circuit, where R = 0.3Ω, L is 0.7nH, and the value range of C is 0.6pF-2.6pF. The C parameters of the adjustable device of the reconfigurable metasurface unit are optimized to achieve a 3-bit phase state. C is evenly divided into 8 values ​​as the initial optimization value of the optimization algorithm. The quasi-Newton gradient algorithm is used to obtain the capacitance values ​​for 8 phase states of 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°, as shown in Table 1. The amplitude curves and phase curves of the reflection coefficient of the eight working states of the reconfigurable metasurface unit are shown in Table 1. Fig.10 and 11 shown.

[0085] Finally, based on the designed reconfigurable metasurface unit, the design of the metasurface system is completed, such as Fig.12 As shown. According to the obtained 8 discrete phase states, given different array codes, the beam direction can be effectively controlled. Among them, the far-field pattern of the "0000222244446666" code is as follows Fig.13 As shown in Figure 1, a 15° beam deflection is achieved. The far-field pattern of the “4432210077655433” code is shown in Figure 1. Fig.14 As shown, a beam deflection of -20° is achieved.

[0086] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0087] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A design method for a reconfigurable metasurface unit based on current distribution and image segmentation, characterized in that: The following steps are involved: S1: Grayscale current distribution image The current distribution of the reconfigurable metasurface is a color image with three channels of information, which has a large amount of information. However, the passive metal surface of the reconfigurable metasurface has only two conditions: "with" and "without" metal patches, which can represent less information. Therefore, it is necessary to process the obtained current distribution map to extract useful information. By graying the current distribution, useful information in the current distribution map of the reconfigurable metasurface unit can be extracted. S2: Current distribution image segmentation Since the range of each pixel value of the grayscale processed image is 0-255, the amount of information is still large for the passive metal surface of the reconfigurable metasurface unit, so it is necessary to further simplify the image information. Through image segmentation technology, the pixel value of the image is simplified to two possibilities: "0" and "1". S3: Transformation of current distribution images into reconfigurable metasurface structures In order to convert a two-dimensional image into a reconfigurable metasurface structure, the image domain information needs to be converted into the physical structure of the metal patch. The segmented image pixel matrix consists of two elements "0" and "1". Through linear mapping, the pixel value "1" is mapped to the metal structure of the specified size. S4: Reconfigurable metasurface unit design and system design in different working states Once the topological structure of the passive metal patch on the top is determined, the structure of the reconfigurable metasurface unit is constructed by adding a substrate and a ground layer. In order to achieve dynamic regulation of electromagnetic waves, the quasi-Newton gradient optimization algorithm is used to obtain metasurface units in different working states. The generalized Snell's theorem is used to construct a reconfigurable metasurface system based on the regulation requirements of electromagnetic waves.

2. The design method of a reconfigurable metasurface unit based on current distribution and image segmentation according to claim 1, characterized in that: The image segmentation technology adopts OTSU threshold segmentation technology.

3. The design method of a reconfigurable metasurface unit based on current distribution and image segmentation according to claim 2, characterized in that: The goal of the OTSU threshold calculation is to maximize the ratio of the between-class variance to the within-class variance, as shown below: In the formula, is the between-class variance, is the intra-class variance, μ1 and μ2 are the means of the two classes, δ1 and δ2 are the variances of the two classes; OTSU threshold segmentation calculates the intra-class variance and inter-class variance for each gray level in the image, and selects the gray level that maximizes the inter-class variance as the threshold; the OTSU threshold calculation formula is as follows: In the formula, is the optimal threshold.

Images