Metasurface implementation of mid-wave infrared defocus plane filter and reverse design method

Abstract

This invention discloses a metasurface-based mid-wave infrared focusing planar filter and its reverse design method. Based on the light field modulation characteristics of metasurfaces, a single-layer all-silicon dielectric metasurface can achieve spectral routing and focusing functions for a specified mid-wave infrared band, thus achieving a filtering effect similar to a Bayer filter, while significantly improving spectral energy utilization. The design utilizes an end-to-end reverse approach, employing genetic algorithms and binary search algorithms (DBS) to optimize structural parameters in stages, fully leveraging the advantages of each algorithm, lowering the design threshold, increasing optimization freedom, and thus improving optimization performance. Therefore, this invention can provide key device support for low-light mid-infrared night vision color imaging applications and expands new design ideas in the field of integration and miniaturization.

Description

Technical Field

[0001] This invention belongs to the fields of optical imaging and spectral signal processing, and particularly relates to a mid-infrared spectral router based on a Bayer filter-like structure in an all-silicon dielectric metasurface and its structural inverse design using a multi-objective genetic algorithm. Background Technology

[0002] Mid-wave infrared imaging, due to its ability to operate day and night and its passive imaging capabilities, is widely used in military reconnaissance, security surveillance, unmanned vehicles, and aerospace remote sensing. Mid-wave infrared filters are one of the key components of detectors in this band. Especially in low-light night vision environments, filter design directly affects the spectral and color imaging quality of the detector. Common design methods include pre-filter wheels or back-end image enhancement and pseudo-color algorithms. However, traditional hardware designs such as filter wheels are limited by low energy efficiency and difficulties in mechanical system design, while back-end algorithms suffer from long training times and overfitting issues. These methods are insufficient to meet the increasing demands for miniaturization and low latency in integrated systems, thus severely limiting the development of mid-wave night vision imaging.

[0003] Metasurfaces are novel two-dimensional planar metamaterials composed of a single-layer subwavelength metallic or dielectric nanoantenna array. With wavelength-level thickness, metasurfaces allow for arbitrary manipulation of the phase, polarization, intensity, and other parameters of light. Currently, the ultra-high resolution, multi-channel, and multi-dimensional characteristics of metasurface light field manipulation have enabled various functionalities, including aberration-correcting microlenses, high-quality holography, and beam deflection and generation. Metasurface filters are promising candidates for improving spectral energy utilization under low illumination and achieving spectral routing.

[0004] Traditional metasurface design is achieved by determining the structural parameters and arrangement of nanoantennas. However, such forward design requires extensive electromagnetic theory and engineering experience, and the design process necessitates a step-by-step, point-by-point scanning of parameters to determine the optimal structure. This significantly increases the design threshold and cost, and also makes it impossible to achieve complex functions.

[0005] In summary, it is necessary to propose a metasurface filter that realizes multiple channels in the mid-infrared band and a targeted reverse design method for the metasurface, so as to provide key device support for spectral routing and low-light night vision color imaging applications. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes for the first time a mid-wave infrared (3μm~5μm) metasurface filter and provides a targeted reverse design algorithm. By utilizing the metasurface's light field modulation characteristics, it enables the routing of four different spectral bands to the focal plane of the specified point. Unlike the unidirectional spectral scattering mechanism, this metasurface significantly improves energy utilization and segmented spectral collection efficiency, thereby demonstrating higher imaging intensity than common imaging sensors in low-light night vision imaging and mid-wave infrared color imaging.

[0007] The design method for mid-wave infrared metasurface filters based on reverse design includes the following steps:

[0008] Step 1: Under plane incident light with the same polarization conditions in the mid-infrared band, the geometric parameters of the elliptical superatoms used to construct the metasurface are calculated using the finite-difference time-domain method. The optical response in the corresponding dimension is obtained, thus revealing the relationship between the superatoms and the amplitude and phase modulation of the light field under different geometric parameters. The geometric parameters of the superatoms include the major axis length *a*, minor axis length *b*, columnar height *h*, and distribution period *p*. The ellipse is a standard ellipse without rotation.

[0009] Step 2: In the design phase, the metasurface is divided into four regions of equal area, each mapped to a target distribution region of four wavebands on the target plane. Then, the entire metasurface is "pixelated," with each pixel cell having a size of p. Each pixel cell (square micro / nano pillar) is assigned either a superatom or air of the size selected in Step 1, represented by "1" or "0" respectively. This maps the metasurface plane into a square matrix containing only "0" and "1".

[0010] Step 3: Using a genetic algorithm, randomly generate k initial populations containing the binary coding patterns of the metasurface, based on the matrix selected in Step 2 with row and column sizes. These coded numbers are defined as either material or free space. Define the fitness function, or fitness F, for the routing properties of the metasurface for each binary coding pattern.

[0011]

[0012] Where, λ m1 and λ m2 Used to indicate the minimum and maximum wavelengths, the subscript m represents the corresponding four bands, namely, the minimum and maximum wavelengths of a are 3μm and 3.5μm, b is 3.5μm and 4μm, c is 4μm and 4.5μm, and d is 4.5μm and 5μm. T a T b T c T dThese are the ratios of light intensity in each band region on the imaging plane to the light intensity incident on the metasurface region, respectively. A, B, C, and D are adaptive weights for the integral of each band. The purpose of this function is to maximize the light energy of characteristic bands within different band regions, thereby achieving the goal of transmitting light of different bands to the designated regions via specific routes and improving the overall spectral utilization.

[0013] Step 4: Using Lumerical FDTD solutions and Python co-simulation, set the initial weighting factor and calculate the fitness of k metasurface individuals under the initial population binary encoding mode.

[0014] Step 5: Set the number of iterations to g and the crossover probability to p. c mutation probability p a Based on the parent generation obtained in step 4, generate k probability values ​​for selection. According to the principle of roulette wheel, the higher the fitness function value, the greater the probability of being selected.

[0015] Step 6: Select two parent generations in a loop and perform crossover mutations with a certain probability to obtain a child generation. Replace a certain number of poorly fit individuals from the parent generations with the new individuals to obtain a new generation. Simulate and calculate the fitness.

[0016] Step 7: Repeat Step 6 until the iteration ends or the fitness function converges. The values ​​of the four weight factors are adaptively changed according to certain rules to ensure that the objective function is continuously optimized during the iteration process. Output the hypersurface binary encoding matrix with the highest fitness after optimization.

[0017] Step 8: To further optimize the fitness function and increase the degrees of freedom, the improved binary search algorithm DBS is used for further iterative optimization. The iteration number g1 is set, and an initial cell is selected as the starting point for optimization. The state in the cell is changed, including the flipping of "0" and "1". The elliptical superatoms are rotated around the center of the ellipse by three angles (45°, 90°, 135°). The fitness function of each state is calculated by co-simulation. The state with the maximum fitness is retained. Each cell is scanned step by step until the fitness function converges. The weight factor remains unchanged. Finally, the hypersurface encoding pattern with the maximum fitness is output.

[0018] It also includes step 9: Based on the optimal geometric structure parameters obtained in step 8, fabricate a metasurface for mid-wave infrared filters, apply it in the field of mid-wave infrared detection, and improve low-light imaging performance.

[0019] The beneficial effects of this invention are as follows:

[0020] (1) This invention proposes a novel metasurface filter routing design that achieves spectral routing and filtering effects using only a single-layer metasurface structure, rather than losing too much spectral energy information as with common filters. Therefore, this invention significantly improves the spectral energy utilization rate of mid-infrared imaging in terms of hardware, while greatly reducing the bulkiness of the filtering system. It also provides a new design approach for mid-wave night vision color imaging applications.

[0021] (2) This invention provides a phased hybrid reverse design algorithm that optimizes metasurface structure parameters in stages using a genetic algorithm and a binary brute-force search algorithm (DBS). Combining the characteristics of both greatly improves the freedom of reverse design. Compared with traditional forward design, which requires complex algorithms and steps, this design method is simpler and has higher design accuracy.

[0022] (3) Compared with common metasurface design methods, this method achieves end-to-end design, requiring only the determination of target requirements without requiring users to spend time on analysis during the design process. All processes are automatically optimized using FDTD software and Python, significantly reducing manual design costs and improving ease of use. Attached Figure Description

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will now be described in detail with reference to the accompanying drawings, wherein:

[0024] Figure 1 A schematic diagram of the routing scheme for the four mid-wave infrared bands provided by this invention;

[0025] Figure 2 A schematic diagram illustrating the construction of a superatomic simulation model using the electromagnetic simulation method of this invention;

[0026] Figure 3 Initial structure for the design of the metasurface for the binary encoding mode provided by this invention;

[0027] Figure 4 This is a schematic diagram of the design process in Embodiment 1 of the present invention. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0029] This embodiment uses a hybrid reverse-engineered all-silicon metasurface routing filter based on a genetic algorithm and a DBS algorithm. Both the substrate material and the metaatomic material are silicon.

[0030] (1) In this embodiment, the incident light source is set to a 3μm-5μm wavelength band, with a phase of 0, a normalized amplitude of 1, and an x-polarized plane light. An elliptical cylindrical silicon superatomic model is established using Lumerical FDTD solutions, with periodic PML boundary conditions set. Since the size of a single pixel in the mid-infrared detector is 15×15μm... 2 The period of the superatom was set to 1.5 μm. Simulation scans were used to obtain the amplitude and phase of the transmitted and scattered fields corresponding to different sizes of the superatom (major axis a, minor axis b, column height h, period p). Superatoms that could cover the 2π phase across the entire wavelength range and had high transmittance were selected.

[0031] (2) Divide the metasurface into 20×20 cells, each cell having two states, "0" and "1", representing the remaining air after etching and the superatoms selected in step (1), respectively. Using a genetic algorithm, set the maximum number of iterations and randomly generate k initial populations containing the binary coding pattern of the metasurface. For each random structure, the substrate is uniformly set to silicon in the simulation software; four monitors are placed to receive the transmission field data, corresponding to the four pixel surfaces of the detector; the obtained simulation data are defined as T... a T b T c T d Four adaptive weighting factors are set, which are functions of the four transmittance values ​​of the iteration number i and the previous generation i-1, i.e.:

[0032]

[0033] Furthermore, the fitness function F is defined as:

[0034]

[0035] Thus, the corresponding fitness values ​​are obtained through simulation calculations based on k random initial structures.

[0036] (3) Set the crossover probability p c mutation probability p a Based on the parent generation obtained in step (2), k selection probability values ​​are generated. According to the principle of roulette wheel, the higher the fitness function value, the greater the probability of being selected. Therefore, the probability of the i-th parent generation being selected is p. i , is defined as:

[0037]

[0038] (4) Select two parent generations in a loop according to the probability values ​​in step (3) and perform crossover and mutation according to the crossover and mutation probabilities to obtain a offspring. Replace a certain number of individuals with poor fitness in the parent generation with the new individuals. Finally, a new first generation is obtained.

[0039] (5) Set the maximum number of iterations g, and repeat step (3). During this period, sort the fitness of each generation of the population, and plot the iteration convergence curve based on the maximum fitness of each generation to monitor the optimization trend in real time. Stop repeating when the curve converges or the maximum number of iterations g is reached. Obtain the metasurface structure parameters with the maximum fitness at this time.

[0040] (6) Based on the structural parameters obtained in step (5), perform improved DBS processing, set the iteration number g1, select the first initial cell in the upper left corner as the optimization starting point, change its state, including flipping "0" and "1", and rotating the elliptical superatoms around the center of the ellipse at three angles (45°, 90°, 135°), and record the states corresponding to these angles as "2", "3", and "4". Use Lumerical FDTD solutions and Python co-simulation to calculate the fitness function of each state and retain the state with the maximum fitness. The weight factors are all set to 1. Then, scan and calculate each cell point by point from left to right and from top to bottom, and finally generate the next generation.

[0041] (7) Repeat step (6), during which time an iterative convergence plot is generated based on the number of iterations and the maximum fitness of each generation. Repeat step (6) until the curve converges.

[0042] The final output is the metasurface encoding pattern with the highest fitness, where 0 to 4 each have a corresponding superatomic state.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A design method for a mid-wave infrared metasurface filter based on reverse design, characterized in that: Includes the following steps: Step 1: Under plane incident light with the same polarization conditions, the geometric parameters of the elliptical superatoms used to construct the metasurface are scanned and calculated using the finite-difference time-domain method simulation. The superatom A that can achieve the best light field transmittance and cover the 2π phase is found. The geometric parameters of the superatoms include the major axis a, minor axis b, height g, and distribution period p of the ellipse. Step 2: "Pixelate" the entire metasurface, with each pixel cell having a size of p. Each pixel cell is assigned as either a superatom or air, represented by "1" or "0" respectively. Map the metasurface plane as a square matrix containing only "0" and "1". Step 3: Based on the genetic algorithm, define the fitness of the routing attribute of the metasurface for each binary encoding pattern. Based on the row and column size, randomly generate k initial populations containing the binary encoding patterns of the metasurface from Step 2. The fitness function represents the weighted sum of the light transmittance of the metasurface in four bands, as described below: Where, λ m1 and λ m2 Used to indicate the minimum and maximum wavelengths, the subscript m represents the corresponding four bands, namely, the minimum and maximum wavelengths of band a are 3μm and 3.5μm, b is 3.5μm and 4μm, c is 4μm and 4.5μm, d is 4.5μm and 5μm, and T... a T b T c T d These are the ratios of light intensity in each band region on the imaging plane to the light intensity incident on the metasurface region, where A, B, C, and D are the adaptive weights for the integral of each band. Step 4: Set the initial weighting factor and calculate the fitness of k hypersurface individuals under the initial population binary encoding pattern; Step 5: Based on the parent generation obtained in Step 4, generate k probability values ​​for selection using the roulette wheel principle; Step 6: Select two parent generations in a loop and perform crossover mutation based on the probability values ​​generated in Step 5 to obtain a child generation. Replace the parent generation individuals whose fitness is lower than the custom threshold with the new individuals to obtain a new generation, and simulate and calculate the fitness. Step 7: Repeat step 6 until the iteration ends or the fitness function converges. The values ​​of the four weight factors are changed according to the adaptive weight rule to ensure that the objective function is continuously optimized during the iteration process. Output the hypersurface binary encoding matrix with the highest fitness after optimization. Step 8: Further iteratively optimize using the improved binary search algorithm. Select an initial cell as the starting point for optimization and change its state, including flipping "0" and "1" and rotating the elliptical superatoms around the center of the ellipse by three angles, namely 45°, 90°, and 135°. Simulate and calculate the fitness function for each state, with the weight factor remaining unchanged and retaining the state with the highest fitness. Gradually scan each cell to generate a new generation until the fitness function converges, and finally output the hypersurface encoding pattern with the highest fitness.

2. The design method for a mid-wave infrared metasurface filter based on reverse design as described in claim 1, characterized in that, It also includes step 9, which involves fabricating a metasurface for a mid-wave infrared filter based on the optimal geometric parameters obtained in step 8, and applying it in the field of mid-wave infrared detection to improve low-light imaging performance.

3. The design method for a mid-wave infrared metasurface filter based on reverse design as described in claim 1, characterized in that, In step 5, the probability of an individual being selected according to the roulette wheel betting principle is defined as follows:

4. The design method for a mid-wave infrared metasurface filter based on reverse design as described in claim 1, characterized in that, In step 7, the adaptive weighting factor is a function of the four transmittance values ​​of the iteration number i and the previous generation i-1, i.e.:

Images