A color filter array based on genetic algorithm and its implementation method

Abstract

The present invention discloses a genetic algorithm-based color filter array and its implementation method. The color filter array comprises a substrate and a nanoantenna disposed on the substrate, wherein the nanoantenna is made of a phase-change material. The genetic algorithm is used to optimize the dimensional parameters of the substrate and the nanoantenna to obtain a color filter array with optimal structural parameters. The present invention utilizes a genetic algorithm to search for optimal metasurface structural parameters within the visible light band, facilitating efficient and rapid structural color design. This approach also allows for a high degree of freedom in parameter design, enhancing the designability of the color filter array and enabling the generation of a high reflection peak at a specified wavelength.

Description

Technical Field

[0001] The present invention relates to the technical field of optoelectronic components, and in particular to a color filter array based on a genetic algorithm and an implementation method thereof. Background Art

[0002] Structural color, also known as physical color, is the result of light amplitude modulation caused by the scattering, reflection, and interference effects of the geometry, size, and arrangement of micro- and nanostructures with specific wavelengths. Compared to chemical colors, which are toxic and difficult to maintain, structural color offers advantages such as high durability, environmental friendliness, and high resolution. More importantly, it is compatible with complementary metal oxide semiconductor (CMOS) manufacturing processes. Its rapid development in applications across many vision-related fields has sparked significant interest and led to extensive research. Research on dielectric nanostructures based on Mie resonances has achieved significant breakthroughs over the past decade, providing an effective path to low-loss, low-cost, and lightweight optical systems. By varying the geometric parameters of their structural units to support electric dipoles, magnetic dipoles, and higher-order modes, dielectric nanostructures enable the free manipulation of color performance and functionality, thereby enabling different colors to be assigned to different frequencies of light in the visible light band.

[0003] However, research has shown that conventional forward-designed structural color methods can only optimize a single parameter. Due to the limited number of adjustable parameters, a systematic search of the parameter space is impossible, limiting the improvement of the display color gamut. Furthermore, the resulting devices can only display a single color, lacking dynamic control, which limits the application of structural color devices. The unavoidable high loss of metal materials such as gold (Au) and silver (Ag) results in low saturation, low brightness, and a narrow color gamut in the resulting structural colors, which cannot meet the requirements of current display applications. Furthermore, the high cost of preparing metal materials also makes it difficult for these devices to be put into practical use.

[0004] The application of phase-change materials can address the problem of color not being dynamically adjustable after structural design, enabling the application of structural color devices in sensors. Popular phase-change materials include vanadium dioxide (VO2) and germanium antimony telluride (Ge2Sb2Te5, GST). Although VO2 exhibits rapid phase transitions and a low transition temperature of 68°C, its metallic state is volatile at room temperature and requires continuous stimulation to maintain its state, limiting its application in technologies such as optical storage. GST exhibits nanosecond-scale phase transitions and a large refractive index difference between its crystalline and amorphous states. It also maintains its phase state without continuous stimulation, demonstrating strong non-volatility. While GST has been used to tune the near-infrared spectrum, it is not suitable for use in the visible light band due to its high absorption at visible frequencies and relatively small change in the real component of the refractive index. Minimizing the thickness of the GST layer to minimize excessive absorption of visible light is the only approach, but this further limits the change in optical path length caused by the phase transition. Therefore, further research is needed on phase-change materials suitable for visible light tuning. Summary of the Invention

[0005] Based on this, the present invention provides a color filter array based on a genetic algorithm and an implementation method thereof to solve the problem that the traditional forward design method of structural color can only optimize a single parameter and cannot be dynamically adjusted, and to overcome the problems of high absorption and small phase change frequency shift of existing structural color devices in the visible light band.

[0006] To achieve the above objectives, the present invention provides a color filter array based on a genetic algorithm, comprising: a substrate and a nanoantenna arranged on the substrate, wherein the material of the nanoantenna is a phase change material; the size parameters of the substrate and the nanoantenna are optimized by a genetic algorithm to obtain a color filter array with optimal structural parameters.

[0007] Preferably, the phase change material is antimony sulfide.

[0008] Preferably, the substrate adopts a multi-layer thin film structure, and the multi-layer thin film structure consists of a reflective layer and an insulating layer covering the reflective layer.

[0009] Preferably, the reflective layer is made of silver, and the insulating layer is made of silicon dioxide; the insulating layer and the reflective layer have the same width, and the insulating layer is twice as thick as the reflective layer.

[0010] Preferably, the nanoantenna adopts a circular ring structure or a cylindrical structure, and the size parameters of the nanoantenna include the inner diameter of the ring, the outer diameter of the ring and the height.

[0011] In addition, the present invention also provides a method for implementing a color filter array based on a genetic algorithm, comprising:

[0012] Determining a target wavelength, generating multiple sets of metasurface structure parameters, and constructing an initial population; the initial population includes multiple genetic individuals, each of which corresponds to a color filter array consisting of a substrate, a nanoantenna using a phase change material, and a set of the metasurface structure parameters;

[0013] Obtaining the reflectivity corresponding to the phase change material in the crystalline state and the reflection peak position corresponding to the phase change material in the amorphous state at the target wavelength for each genetic individual, inputting a preset fitness function, and obtaining the fitness value of each genetic individual;

[0014] Performing genetic operations on the initial population to generate a new generation population; the genetic operations include replication, selection, crossover and mutation operations;

[0015] By detecting whether the fitness value of the best individual in the new generation population tends to be stable after multiple optimizations, it is determined whether the genetic algorithm meets the stopping optimization condition; wherein the best individual is the genetic individual with the best fitness value;

[0016] If the conditions are met, the color filter array corresponding to the optimal individual is obtained.

[0017] Preferably, after obtaining the color filter array corresponding to the optimal individual, the method further includes:

[0018] Obtaining structural parameters of the color filter array corresponding to the optimal individual and inputting them into finite-difference time-domain software for simulation calculation to obtain reflection spectrum information;

[0019] The color performance evaluation and / or color difference performance characterization of the color filter array is performed according to the reflection spectrum information.

[0020] Preferably, generating multiple sets of metasurface structure parameters and constructing an initial population includes:

[0021] Determining variables to be optimized in a genetic algorithm according to size parameters of the substrate and the nanoantenna;

[0022] According to the value range of the variable, N groups of metasurface structure parameters that meet the conditions are generated through a random function;

[0023] Taking each group of surface structure parameters as the chromosome genes of a genetic individual, an initial population of genetic individuals with a population size of N is generated.

[0024] Preferably, obtaining the reflectivity corresponding to the phase change material in the crystalline state and the reflection peak position corresponding to the phase change material in the amorphous state at the target wavelength for each genetic individual, inputting a preset fitness function, and obtaining the fitness value of each genetic individual includes:

[0025] The fitness function is constructed with the color filter array satisfying the requirements of simultaneously reflecting highly saturated colors and producing large color differences during phase change as the optimization goal;

[0026] Calculating the reflectivity and the corresponding reflection peak position of the phase change material in the crystalline state at the target wavelength for each genetic individual by using a finite-difference time-domain method;

[0027] Inputting the reflectivity and the reflection peak position into a fitness function to obtain a fitness value of each genetic individual;

[0028] Wherein, the fitness function is:

[0029] Fit=Fit ad / 1000+Fit cf

[0030] Where, Fit is the fitness value of the genetic individual; Fit ad is the distance between the reflection peak position of the phase change material in the amorphous state and the target wavelength; cf An evaluation value of the saturation and sub-wave suppression degree of the color represented by the phase change material at the target wavelength when the phase change material is in a crystalline state;

[0031] Fit ad =|i-Peak(R a )|,

[0032] Where i is the target wavelength, Peak(R a ) is the reflection peak position when the phase change material is in the amorphous state;

[0033]

[0034] Where R(i) is the reflectivity of the phase change material at the target wavelength when it is in the crystalline state; is the sum of the reflectances of wavelengths other than the target wavelength; is the average of the sum of reflectances; ε is the imbalance weight.

[0035] Preferably, performing genetic operations on the initial population to generate a new generation population includes:

[0036] Back up the initial population to obtain a backup population;

[0037] After arranging the initial population in descending order according to the fitness value, selecting a first number of genetic individuals from top to bottom, and selecting a second number of genetic individuals from the backup population by a roulette wheel selection method to form a first candidate population;

[0038] Performing chromosome single-point crossover on genetic individuals in the first candidate population to obtain a second candidate population;

[0039] The genetic individuals in the second candidate population are mutated according to a preset mutation rate to obtain a new generation population after the mutation.

[0040] As can be seen from the above, the present invention provides a genetic algorithm-based color filter array and its implementation method. The color filter array is composed of a substrate and nanoantennas using phase change materials. The dimensional parameters of the substrate and nanoantennas are optimized using a genetic algorithm to obtain a color filter array with optimal structural parameters. Compared to traditional forward structural color design methods, the present invention uses a genetic algorithm to search for optimal metasurface structural parameters within the visible light band, facilitating efficient and rapid structural color design. It also provides a high degree of freedom in parameter design, making the color filter array more designable and capable of producing a high reflection peak at a specified wavelength. Furthermore, the genetic algorithm-based color filter array provided by the present invention uses antimony sulfide material, which has a higher refractive index, lower absorption, and greater phase change optical path length variation than germanium antimony telluride alloy in the visible light band. This enables the top nanoantenna to achieve high-brightness colors, large color difference, and a wide color gamut, and also provides the color filter array with dynamic color adjustability. At the same time, the top nanoantenna adopts a circular ring structure, which provides greater freedom for parameter optimization, can obtain a wider color gamut range, and makes the color filter array more manufacturable. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 is a schematic structural diagram of a color filter array in one embodiment of the present invention;

[0043] Figure 2 for Figure 1 Schematic diagram of the structure at A in the middle;

[0044] Figure 3 The process of implementing a color filter array based on a genetic algorithm in one embodiment of the present invention is as follows Figure 1 ;

[0045] Figure 4 The process of implementing a color filter array based on a genetic algorithm in one embodiment of the present invention is as follows Figure 2 ;

[0046] FIG5 is a schematic diagram of the maximum fitness convergence process of the metasurface during genetic algorithm optimization at different target wavelengths;

[0047] FIG6 shows the reflection spectra of the metasurface at different target wavelengths when the phase change material is in the crystalline and amorphous states;

[0048] FIG7 is a characterization of the color difference performance of different optimized structures in the CIE 1931 color space;

[0049] Figure 8 Representation of the color gamut achieved for a color filter array in the CIE 1931 color space. DETAILED DESCRIPTION

[0050] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] refer to Figure 1 and Figure 2 An embodiment of the present invention provides a color filter array based on a genetic algorithm, comprising: a substrate 1 and a nano-antenna 2 arranged on the substrate 1, wherein the material of the nano-antenna 2 is a phase change material; the size parameters of the substrate 1 and the nano-antenna 2 are optimized by a genetic algorithm to obtain a color filter array with optimal structural parameters.

[0052] In a preferred embodiment, the substrate 1 adopts a multi-layer thin film structure, and the multi-layer thin film structure is composed of a reflective layer 11 and an insulating layer 12 covering the reflective layer 11 .

[0053] In a preferred embodiment, the reflective layer 11 is made of silver (Ag), and the insulating layer 12 is made of silicon dioxide (SiO2). The insulating layer 12 and the reflective layer 11 have the same width P, and the thickness of the insulating layer 12 is twice the thickness of the reflective layer 11. The width P of the reflective layer 11 and the insulating layer 12 is equal to the period of the substrate 1.

[0054] In a preferred embodiment, the nanoantenna 2 has a ring structure or a cylindrical structure, and the size parameters of the nanoantenna 2 include the ring inner diameter r, the ring outer diameter R, and the height H. It is understandable that when the ring inner diameter r is 0, the nanoantenna 2 has a cylindrical structure.

[0055] In a preferred embodiment, the phase change material is antimony sulfide (Sb2S3).

[0056] This embodiment comprises a color filter array composed of a bottom-layer multilayer thin-film structure and a top-layer nanoantenna 2. The bottom-layer multilayer thin-film structure consists of a reflective layer and an insulating layer. The reflective layer is made of Ag with a width of P and a thickness of 100 nanometers to ensure that no radiation passes through the metasurface structure, thereby improving its reflective capability. The insulating layer is made of SiO2 with a width of P and a thickness of 200 nanometers to maintain the resonance of the metasurface structure. The dielectric constants of the Ag and SiO2 layers are derived from Palik data in the finite-difference time-domain software (i.e., FDTD Solutions). The top-layer nanoantenna 2 is a circular ring or cylindrical structure made of Sb2S3 with an inner diameter of r, an outer diameter of R, and a height of H.

[0057] As can be seen from the foregoing, the genetic algorithm-based color filter array provided in this embodiment comprises a substrate 1 comprising an Ag layer and a SiO2 layer, on which a phase change material is disposed. The phase change material adopts a circular or cylindrical structure. By optimizing the period P of the substrate 1, the inner diameter r, outer diameter R, and height H of the phase change material through a genetic algorithm, a color filter array with optimal structural parameters at a specified wavelength can be obtained. This embodiment utilizes a genetic algorithm to search for optimal metasurface structural parameters within the visible light band, facilitating efficient and rapid structural color design. It also provides a high degree of freedom in parameter design, making the color filter array more designable and capable of producing a higher reflection peak at a specified wavelength.

[0058] Furthermore, the genetic algorithm-based color filter array provided in this embodiment uses Sb2S3 material, which has a higher refractive index, lower absorption, and greater phase change optical path length variation than GST in the visible light band. This enables the top nanoantenna to achieve high-brightness colors, large color difference, and a wide color gamut, and also provides the color filter array with dynamic color adjustability. Furthermore, the top nanoantenna adopts a circular ring structure, which provides greater freedom in parameter optimization, enabling a wider color gamut and enhancing the manufacturability of the color filter array.

[0059] In addition, if Figure 3 As shown, an embodiment of the present invention further provides a method for implementing a color filter array based on a genetic algorithm, which specifically includes the following steps:

[0060] Step S10, determine the target wavelength, generate multiple sets of metasurface structure parameters, and construct an initial population; the initial population includes multiple genetic individuals, each of which corresponds to a color filter array composed of a substrate, a nanoantenna using a phase change material, and a set of the metasurface structure parameters.

[0061] In step S10, after selecting the target wavelength according to the required color, N groups of metasurface structure parameters are randomly generated. Each group of metasurface structure parameters may include the period P of the substrate, the inner diameter r of the nanoantenna ring, the outer diameter R of the ring, and the height H. Furthermore, based on the N groups of metasurface structure parameters, an initial population consisting of N genetic individuals is constructed.

[0062] The chromosomal genes of each genetic individual are determined by a set of metasurface structure parameters, and the metasurface structure corresponding to the genetic individual is the color filter array described in the aforementioned embodiment. Optionally, the color filter array comprises: a substrate composed of an Ag layer with a width P and a thickness of 100 nm, a SiO2 layer with a width P and a thickness of 200 nm covering the Ag layer, and a circular nanoantenna made of Sb2S3 material with an inner diameter r, an outer diameter R, and a height H.

[0063] Preferably, in step S10, multiple sets of metasurface structure parameters are generated to construct an initial population, which specifically includes the following steps:

[0064] Step S101, determining variables to be optimized in a genetic algorithm according to size parameters of the substrate and the nanoantenna;

[0065] Step S102, generating N groups of metasurface structure parameters that meet the conditions according to the value range of the variable and through a random function;

[0066] Step S103: using each set of surface structure parameters as the chromosome genes of a genetic individual, an initial population of genetic individuals with a population size of N is generated.

[0067] In this embodiment, the population size of the initial population depends on the number of groups of metasurface structure parameters. Optionally, the population size is 100.

[0068] Specifically, after determining the target wavelength, the substrate period P, the nanoantenna's circular inner radius r, outer diameter R, and height H are used as variables to be optimized in the genetic algorithm. The value ranges of the variables are shown in Table 1. A random function in Matlab is then used to generate 100 sets of metasurface structure parameters that meet the constraints. That is, metasurface structure parameters are generated in which each variable falls within the corresponding value range. Finally, based on each set of metasurface structure parameters, a corresponding genetic individual with a binary chromosome gene is generated, thus forming an initial population of 100. Optionally, the target wavelength is one of 450 nm, 540 nm, and 660 nm.

[0069] Table 1

[0070]

[0071] Step S20, obtaining the reflectivity corresponding to the phase change material in the crystalline state and the corresponding reflection peak position in the amorphous state of each genetic individual at the target wavelength, inputting a preset fitness function, and obtaining the fitness value of each genetic individual.

[0072] In this example, the finite-difference time-domain method (FDTD) is used to calculate the reflectivity of each genetic individual in the population when the Sb2S3 material is in the crystalline state and the corresponding reflection peak position when it is in the amorphous state at the target wavelength. These two data points are then input into a pre-defined fitness function to obtain the fitness value of the genetic individual.

[0073] Preferably, step S20 includes the following steps:

[0074] Step S201, constructing a fitness function with the color filter array satisfying the requirements of reflecting highly saturated colors and generating large color differences during phase change as optimization goals;

[0075] Step S202, calculating the reflectivity and the corresponding reflection peak position of the phase change material in the crystalline state at the target wavelength for each genetic individual by using the finite difference time domain method;

[0076] Step S203: input the reflectivity and the reflection peak position into a fitness function to obtain the fitness value of each genetic individual.

[0077] In this embodiment, the optimization target of the color filter array is used, and the optimization target can be to reflect highly saturated colors while phase changing to produce large color differences, to construct a fitness function; then, for each genetic individual, the chromosome genes of the genetic individual are converted from binary to decimal, and then imported into the FDTD Solutions software to establish a corresponding structural model for simulation calculation to obtain the reflectivity of the Sb2S3 material in the crystalline state and the reflection peak position in the amorphous state at wavelengths of 450nm, 540nm and 660nm respectively; then the two data of reflectivity and reflection peak position are imported into Matlab to calculate the fitness function. It can be understood that this embodiment combines FDTD Solutions and Matlab, and utilizes the superior algorithm function of MATLAB and the efficient calculation advantages of the finite difference time domain software to automatically generate the structural model corresponding to the genetic individual and calculate the fitness.

[0078] Preferably, the fitness function can be the reflectivity of the phase change material at the target wavelength when it is in the crystalline state and the distance between the reflection peak position and the target wavelength when it is in the amorphous state. The sum of these two data can be expressed as:

[0079] Fit=Fit ad / 1000+Fit cf ,

[0080] Where, Fit is the fitness value of the genetic individual; Fit ad is the distance between the reflection peak position of the phase change material in the amorphous state and the target wavelength; cf The saturation and sub-wave suppression degree of the color represented by the phase change material at the target wavelength when the phase change material is in the crystalline state are evaluated.

[0081] The distance between the reflection peak position of the phase change material in the amorphous state and the target wavelength is Fit ad It can be expressed as:

[0082] Fit ad =|i-Peak(R a )|,

[0083] Where i is the target wavelength, Peak(R a ) is the reflection peak position when the phase change material is in the amorphous state.

[0084] The evaluation value of the saturation and sub-wave suppression degree of the color represented by the phase change material at the target wavelength when it is in the crystalline state Fit cf , which can be expressed as:

[0085]

[0086] Where R(i) is the reflectivity of the phase change material at the target wavelength when it is in the crystalline state; is the sum of the reflectances of wavelengths other than the target wavelength; is the average of the sum of the reflectances; ε is the imbalance weight. It can be understood that by setting the imbalance weight ε, the reflection peak at the target wavelength can be increased and the resonance peaks in other bands that affect color purity can be suppressed.

[0087] Step S30, performing genetic operations on the initial population to generate a new generation population; the genetic operations include replication, selection, crossover and mutation operations.

[0088] In step S30, the initial population is subjected to replication, selection, crossover and mutation operations in sequence to generate the next generation population, so that the population continuously evolves forward, thereby updating the population.

[0089] Preferably, step S30 specifically includes the following steps:

[0090] Step S301, backing up the initial population to obtain a backup population;

[0091] Step S302, after arranging the initial population in descending order according to the fitness value, selecting a first number of genetic individuals from top to bottom, and selecting a second number of genetic individuals from the backup population by roulette wheel selection to form a first candidate population; wherein the sum of the first number and the second number is the same as the population size of the initial population;

[0092] Step S303, performing chromosome single-point crossover on the genetic individuals in the first candidate population to obtain a second candidate population;

[0093] Step S304: Mutate the genetic individuals in the second candidate population according to a preset mutation rate to obtain a new generation population after the mutation.

[0094] In this embodiment, the crossover rate and mutation rate are set according to requirements. Optionally, the crossover rate is 70% and the mutation rate is 10%. The first number and the second number are set according to a preset ratio. Optionally, the preset ratio is 1:4. For example, when the size of the initial population is 100, the first number is 20 and the second number is 80.

[0095] In the population update process, the initial population is first backed up to obtain a backup population to complete the replication operation; then the initial population is arranged from large to small according to the fitness value, and after sorting, the 20 genetic individuals with the best fitness are selected from large to small, and the roulette wheel selection method is used to select another 80 genetic individuals from the backup population to form the first candidate population to complete the selection operation; then, 70 genetic individuals are selected from the first candidate population with a probability of 70%, and chromosome single-point crossover is performed on these 70 genetic individuals in pairs to obtain the second candidate population after crossover to complete the crossover operation. The crossover operation effectively ensures the diversity of the population in the evolutionary process and enhances the algorithm's ability to search for the global optimal solution; then, 10 genetic individuals are selected from the second candidate population with a probability of 10%, and the binary chromosome genes of these 10 genetic individuals are mutated with a probability of 10% to complete the mutation operation, and the local optimization ability is enhanced through the mutation operation.

[0096] It should be noted that the roulette wheel selection method is a commonly used selection method.

[0097] Step S40, determining whether the genetic algorithm meets the optimization stop condition by detecting whether the fitness value of the best individual in the new generation population tends to be stable after multiple optimizations; wherein the best individual is the genetic individual with the best fitness value.

[0098] In step S40, if it is detected that the fitness value of the best individual in the population no longer increases after multiple optimizations, that is, it tends to be stable, the optimization is stopped, and it is judged that the genetic algorithm meets the conditions for stopping optimization; otherwise, continue to execute step S20 to calculate the fitness value of each genetic individual in the current population until it is judged that the genetic algorithm meets the conditions for stopping optimization.

[0099] Step S50: If the conditions are met, obtain the color filter array corresponding to the optimal individual.

[0100] In this embodiment, after stopping the optimization, the optimal solution of the genetic algorithm, that is, the optimal individual, is obtained. The optimal metasurface structure parameters can be determined based on the chromosome genes of the optimal individual, thereby determining a color filter array with the optimal structure parameters.

[0101] In an alternative embodiment, if Figure 4 As shown, after step S50, the following steps are further included:

[0102] Step S602, obtaining structural parameters of the color filter array corresponding to the optimal individual and inputting them into finite-difference time-domain software for simulation calculation to obtain reflection spectrum information;

[0103] Step S602 : performing color performance evaluation and color difference performance characterization on the color filter array according to the reflectance spectrum information.

[0104] In this example, the binary chromosome genes of the optimal individual are converted to decimal to obtain the four parameter values ​​for optimizing the color filter array. These four parameter values ​​are then input into FDTD Solutions for simulation calculations to obtain the required reflectance spectrum information. Based on this reflectance spectrum information, the color performance of the color filter array with the optimal structural parameters can be evaluated. Alternatively, the reflectance spectrum information can be converted into color information for color difference performance characterization in the CIE 1931 color space.

[0105] More specifically, the reflectance spectrum obtained from the FDTD simulation is converted into a color representation to evaluate the color performance of the color filter array, which can be expressed as:

[0106] X=∫R(n)×CIEX(n)×d(n),

[0107] Y=∫R(n)×CIEY(n)×d(n),

[0108] Z=∫R(n)×CIEZ(n)×d(n),

[0109] Where R(η) is the reflectance spectrum obtained by the finite-difference time-domain method, CIEX(η), CIEY(η), and CIEZ(η) are the CIE color matching functions (CMFs), and d(η) is the incident light intensity.

[0110] Then, by normalizing the XYZ tristimulus values ​​to a scale of 0-1, we get the chromaticity coordinates x, y, and z, which can be expressed as:

[0111] x=X / (X+Y+X),

[0112] y=Y / (X+Y+Z),

[0113] z=1-xy,

[0114] Wherein, the (x, y) coordinates can be represented on the CIE 1931 chromaticity diagram.

[0115] Finally, the color difference is quantified by CIEDE2000, that is, the lightness and chromaticity are first calculated according to the color performance of the color filter array, which can be expressed as:

[0116]

[0117]

[0118]

[0119] Where, L * is brightness, a * 、b * The chromaticity of .

[0120]

[0121]

[0122] Next, according to L * 、a * 、b * Determine the color difference ΔE.

[0123] In this embodiment, for the three target wavelengths of 450 nm, 540 nm, and 660 nm, the genetic algorithm is terminated after 50 iterations. The color filter array models corresponding to the three optimal individuals obtained at the three target wavelengths are as follows: Figure 1 The structural parameters of the optimized color filter array model are shown in Table 2.

[0124] Table 2

[0125]

[0126] As can be seen from Figure 6, the reflectivity is 37%, 25% and 71% at 450nm, 540nm and 660nm respectively. As can be seen from Figure 7, the three optimized color filter arrays achieve large color differences of 64.7, 27.8 and 44.4 at 450nm, 540nm and 660nm respectively, indicating strong color contrast. Figure 8 As can be seen in the figure, a color gamut of 27.8% of sRGB is achieved, which basically meets the requirements of the visible light band. Figure 5 shows that the genetic algorithm has a very rapid initial optimization speed, which then gradually slows down and finally reaches convergence. At this point, the metasurface structural parameters have reached the optimal value.

[0127] In summary, the genetic algorithm-based color filter array implementation method of this embodiment utilizes a genetic algorithm to search for optimal structural parameters for the color filter array in the visible light band, facilitating efficient and rapid structural color design. This method also provides a high degree of freedom in parameter design, making the color filter array more designable and capable of producing a high reflection peak at a specified wavelength. Furthermore, this embodiment utilizes Sb2S3 material, which has a higher refractive index, lower absorption, and greater phase change optical path length variation than GST in the visible light band. This enables the nanoantenna to achieve high-brightness colors and large color differences, and also provides the color filter array with dynamic color adjustability. Furthermore, the nanoantenna's circular ring structure provides greater freedom in parameter optimization, enabling a wide color gamut and enhancing the manufacturability of the color filter array.

[0128] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Within the scope of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations in different aspects of the embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0129] The embodiments of the present invention are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of the present invention. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for implementing a color filter array based on a genetic algorithm, characterized in that: The color filter array based on the genetic algorithm includes: a substrate and a nano-antenna disposed on the substrate, wherein the material of the nano-antenna is a phase change material; the size parameters of the substrate and the nano-antenna are optimized by the genetic algorithm to obtain a color filter array with optimal structural parameters. The implementation method includes: Determining a target wavelength, generating multiple sets of metasurface structure parameters, and constructing an initial population; the initial population includes multiple genetic individuals, each of which corresponds to a color filter array consisting of a substrate, a nanoantenna using a phase change material, and a set of the metasurface structure parameters; Obtaining the reflectivity corresponding to the phase change material in the crystalline state and the reflection peak position corresponding to the phase change material in the amorphous state at the target wavelength for each genetic individual, inputting a preset fitness function, and obtaining the fitness value of each genetic individual; Performing genetic operations on the initial population to generate a new generation population; the genetic operations include replication, selection, crossover and mutation operations; By detecting whether the fitness value of the best individual in the new generation population tends to be stable after multiple optimizations, it is determined whether the genetic algorithm meets the stopping optimization condition; wherein the best individual is the genetic individual with the best fitness value; If satisfied, the color filter array corresponding to the optimal individual is obtained; The step of obtaining the reflectivity corresponding to the phase change material in the crystalline state and the reflection peak position corresponding to the phase change material in the amorphous state at the target wavelength for each genetic individual, inputting a preset fitness function, and obtaining the fitness value of each genetic individual includes: The fitness function is constructed with the color filter array satisfying the requirements of simultaneously reflecting highly saturated colors and producing large color differences during phase change as the optimization goal; Calculating the reflectivity and the corresponding reflection peak position of the phase change material in the crystalline state at the target wavelength for each genetic individual by using a finite-difference time-domain method; Inputting the reflectivity and the reflection peak position into a fitness function to obtain a fitness value of each genetic individual; Wherein, the fitness function is: Where, is the fitness value of the genetic individual; is the distance between the reflection peak position of the phase change material when it is in an amorphous state and the target wavelength; An evaluation value of the saturation and sub-wave suppression degree of the color represented by the phase change material at the target wavelength when the phase change material is in a crystalline state; ， Where, is the target wavelength, is the reflection peak position when the phase change material is in the amorphous state; ， Where, is the reflectivity of the phase change material at the target wavelength when it is in the crystalline state; 、 is the sum of the reflectances of wavelengths other than the target wavelength; 、 To average the sum of reflectances; Unbalanced weight.

2. The method for implementing a color filter array based on a genetic algorithm according to claim 1, wherein: The phase change material is antimony sulfide.

3. The method for implementing a color filter array based on a genetic algorithm according to claim 1, wherein: The substrate adopts a multi-layer thin film structure, and the multi-layer thin film structure consists of a reflective layer and an insulating layer covering the reflective layer.

4. The method for implementing a color filter array based on a genetic algorithm according to claim 3, wherein: The reflective layer is made of silver, and the insulating layer is made of silicon dioxide. The insulating layer and the reflective layer have the same width, and the insulating layer is twice as thick as the reflective layer.

5. The method for implementing a color filter array based on a genetic algorithm according to claim 2, wherein: The nanoantenna adopts a circular ring structure or a cylindrical structure, and the size parameters of the nanoantenna include the inner diameter of the ring, the outer diameter of the ring and the height.

6. The method for implementing a color filter array based on a genetic algorithm according to claim 1, wherein: After obtaining the color filter array corresponding to the optimal individual, the method further includes: Obtaining structural parameters of the color filter array corresponding to the optimal individual and inputting them into finite-difference time-domain software for simulation calculation to obtain reflection spectrum information; The color performance evaluation and / or color difference performance characterization of the color filter array is performed according to the reflection spectrum information.

7. The method for implementing a color filter array based on a genetic algorithm according to claim 1, wherein: The generating of multiple sets of metasurface structure parameters and constructing an initial population includes: Determining variables to be optimized in a genetic algorithm according to size parameters of the substrate and the nanoantenna; According to the value range of the variable, N groups of metasurface structure parameters that meet the conditions are generated through a random function; Taking each group of surface structure parameters as the chromosome genes of a genetic individual, an initial population of genetic individuals with a population size of N is generated.

8. The method for implementing a color filter array based on a genetic algorithm according to claim 1, wherein: The performing of genetic operations on the initial population to generate a new generation population includes: Back up the initial population to obtain a backup population; After arranging the initial population in descending order according to the fitness value, selecting a first number of genetic individuals from top to bottom, and selecting a second number of genetic individuals from the backup population by a roulette wheel selection method to form a first candidate population; Performing chromosome single-point crossover on genetic individuals in the first candidate population to obtain a second candidate population; The genetic individuals in the second candidate population are mutated according to a preset mutation rate to obtain a new generation population after the mutation.

Images