Autonomously tunable metasurface structures, imagers, and target recognition methods

Abstract

The present invention relates to the technical field of optical devices, and provides an autonomously tuned metasurface structure, an imager, and a target recognition method, wherein the structure includes a polarizer, an electrode layer having a plurality of sub-electrodes, a liquid crystal layer having a plurality of pixel units, and a metasurface layer having a plurality of periodically distributed nanocolumns; by controlling the applied voltage of each sub-electrode on the corresponding pixel unit, the polarizer is coordinated to perform amplitude adjustment, and then the phase adjustment of the corresponding nanocolumns in the metasurface layer is combined to achieve independent adjustment of the brightness of each light spot in the light spot matrix; a multi-channel B×B light spot matrix is ​​generated by spatial multiplexing to determine the emitted characteristic image through the light spot matrix; different light spot brightnesses in the light spot matrix are used to characterize different convolution kernel weights. In the present invention, the liquid crystal layer is light amplitude adjustable, and the metasurface layer is phase regulated, which improves the imaging quality and minimum resolution, and can realize the processing of various coded information, thereby improving the accuracy and robustness of object recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of optical devices, and in particular to an autonomously tuned metasurface structure, an imager and a target recognition method. Background Art

[0002] The rise of Convolutional Neural Networks (CNNs) and the emergence of large-scale training datasets have laid a solid foundation for many artificial intelligence applications. Although these advances have significantly improved performance, they usually come at the cost of increasing model complexity, which in turn requires more computing resources.

[0003] The growth in the use and complexity of neural networks has led to an increase in energy consumption, especially when computing power is limited, which poses challenges for real-time decision-making. For example, for machine vision, processors and imagers need to be compact, lightweight and energy-efficient to support onboard processing while ensuring low power consumption and high throughput parallel operation. Therefore, these conflicting requirements highlight the need to develop innovative software and hardware solutions in the context of the ever-expanding needs of artificial intelligence.

[0004] Optical neural networks can increase computational speed and improve energy efficiency in a variety of applications. Machine vision computation can be significantly accelerated by integrating computational tasks into front-end imaging optics. Free-space optical computation, especially Fourier optics, predates modern digital circuits and enables highly parallel convolution operations common in machine vision architectures.

[0005] Optical neural networks have promoted breakthrough advances in machine vision, smart healthcare, and autonomous driving systems. However, fixed optical network architectures have encountered great challenges in achieving large-scale parameter adjustments without modifying physical components, limiting their ability to handle complex application scenarios, and the accuracy and stability of optical neural networks are still lacking. Summary of the invention

[0006] The present invention provides an autonomously tuned metasurface structure, an imager and a target recognition method, which are used to solve the technical problems of low accuracy and stability in traditional optical neural networks in the prior art.

[0007] The present invention provides an autonomous tuning metasurface structure, comprising a stacked polarizer, an electrode layer, a liquid crystal layer and a metasurface layer; by applying a voltage to the electrode layer, the liquid crystal molecules in the liquid crystal layer are controlled to turn in cooperation with the polarizer to adjust the amplitude of the light in the incident image; the metasurface layer performs phase adjustment on the light after the amplitude is adjusted and responds, and generates a multi-channel B×B spot matrix through spatial multiplexing to determine the emitted characteristic image through the spot matrix; B represents the number of rows and columns of the spot matrix; different spot brightnesses in the spot matrix are used to characterize different convolution kernel weights; wherein the electrode layer includes a plurality of sub-electrodes, the liquid crystal layer includes a plurality of pixel units, and the metasurface layer includes a plurality of periodically distributed nanocolumns, and there is a corresponding relationship between the sub-electrodes, the pixel units and the nanocolumns; by controlling the applied voltage of each sub-electrode on the corresponding pixel unit, the amplitude is adjusted in cooperation with the polarizer, and then combined with the phase adjustment of the corresponding nanocolumns in the metasurface layer, the brightness of each spot in the spot matrix can be independently adjusted; the complex amplitude field of the autonomous tuning metasurface structure It is expressed as:

[0008] ;

[0009] Wherein, M and N represent the total number of rows and columns of elements in the point spread function, n is the nth row number, and m is the mth column number; Indicates the corresponding weight of each element; represents the working wavelength; x and y represent the position information of the abscissa and ordinate in the plane of the metasurface layer, respectively; represents the pre-designed light transmission angle along the x direction; represents the pre-designed light transmission angle along the y direction; i represents a constant.

[0010] According to an autonomous tuning metasurface structure provided by the present invention, the thickness of the liquid crystal layer ranges from 1 micron to 10 microns, the width of the sub-electrode ranges from 500 nanometers to 100 microns; the length of the sub-electrode ranges from 500 nanometers to 100 microns; the cross-sectional shape of the nanocolumn is square, cylindrical or cross-shaped; the length of the nanocolumn cross-sectional shape ranges from 100 nanometers to 1000 nanometers; the height of the nanocolumn ranges from 100 nanometers to 1000 nanometers; and the distribution period of the nanocolumn ranges from 200 nanometers to 1000 nanometers.

[0011] According to an autonomous tuning metasurface structure provided by the present invention, the properties of the liquid crystal layer are expressed as follows:

[0012] ;

[0013] in, Indicates the direction of rotation of liquid crystal molecules; represents the constant axis refractive index; represents the extraordinary refractive index; Represents the effective refractive index.

[0014] According to an autonomous tuning metasurface structure provided by the present invention, the electrode layer includes a first electrode layer and a second electrode layer; a polarizer is arranged on the metasurface layer; the first electrode layer is arranged on the polarizer; the liquid crystal layer is arranged on the first electrode layer; and the second electrode layer is arranged on the liquid crystal layer.

[0015] According to an autonomous tuning metasurface structure provided by the present invention, it also includes a first light-transmitting substrate layer and a second light-transmitting substrate layer; the metasurface layer is arranged on the first light-transmitting substrate layer; and the second light-transmitting substrate layer is arranged on the second electrode layer.

[0016] According to an autonomous tuning metasurface structure provided by the present invention, the metasurface layer is used to perform a dot product operation on an incident image and a numerical value represented in a light spot matrix to obtain a characteristic image.

[0017] According to an autonomous tuning metasurface structure provided by the present invention, materials of the first electrode layer and the second electrode layer are both indium tin oxide.

[0018] According to an autonomous tuning metasurface structure provided by the present invention, materials of the first light-transmitting substrate layer and the second light-transmitting substrate layer are both glass.

[0019] The present invention also provides an imager based on an optical neural network, comprising a photodetector and the above-mentioned autonomously tuned metasurface structure; wherein the autonomously tuned metasurface structure is used to tune the light of the received image to obtain a characteristic image, and the photodetector is used to receive the light of the characteristic image of the autonomously tuned metasurface structure and form an image.

[0020] The present invention also provides a target recognition method based on an optoelectronic hybrid convolutional neural network, using the above-mentioned autonomous tuning metasurface structure. The target recognition method based on an optoelectronic hybrid convolutional neural network includes: performing optical neural network convolution processing on the target to be identified through the autonomous tuning metasurface structure to obtain a feature image after feature enhancement; converting the feature image from an optical signal to an electrical signal through a photoelectric detector; inputting the feature image of the electrical signal into a digital neural network for target recognition, and obtaining a recognition result output by the digital neural network.

[0021] The present invention provides an autonomous tuning metasurface structure, an imager and a target recognition method, wherein the autonomous tuning metasurface structure comprises a stacked polarizer, an electrode layer, a liquid crystal layer and a metasurface layer; by applying a voltage to the electrode layer, the liquid crystal molecules in the liquid crystal layer are controlled to turn in cooperation with the polarizer to adjust the amplitude of the light in the incident image; the metasurface layer performs phase adjustment on the light after the amplitude is adjusted and responds, and generates a multi-channel B×B spot matrix through spatial multiplexing to determine the emitted characteristic image through the spot matrix; B represents the number of rows and columns of the spot matrix; different spot brightnesses in the spot matrix are used to characterize different convolution kernel weights; wherein the electrode layer comprises a plurality of sub-electrodes, the liquid crystal layer comprises a plurality of pixel units, and the metasurface layer comprises a plurality of periodically distributed nanocolumns, and there is a corresponding relationship between the sub-electrodes, the pixel units and the nanocolumns; by controlling the applied voltage of each sub-electrode on the corresponding pixel unit, the amplitude is adjusted in cooperation with the polarizer, and then combined with the phase adjustment of the corresponding nanocolumns in the metasurface layer, the brightness of each spot in the spot matrix can be independently adjusted. Through the above-mentioned method, the liquid crystal layer in the present invention is light amplitude-adjustable, and the metasurface layer is phase-controlled, which improves the imaging quality and minimum resolution, and can process various coded information, thereby improving the accuracy and robustness of object recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the present invention or 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 It is one of the structural schematic diagrams of the autonomous tuning metasurface structure provided in an embodiment of the present invention.

[0024] Figure 2 It is a schematic diagram of the structure of the super surface layer provided in an embodiment of the present invention.

[0025] Figure 3 This is the second structural schematic diagram of the autonomous tuning metasurface structure provided in an embodiment of the present invention.

[0026] Figure 4 It is a schematic diagram of the structure of an imager based on an optical neural network provided in an embodiment of the present invention.

[0027] Figure 5 It is a schematic diagram of the principle of an imager based on an optical neural network provided in an embodiment of the present invention.

[0028] Figure 6 It is a flow chart of a target recognition method based on an optoelectronic hybrid convolutional neural network provided in an embodiment of the present invention.

[0029] Figure 7 It is a structural schematic diagram of the optoelectronic hybrid convolutional neural network architecture provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0031] In the description of the embodiments of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limitations on the embodiments of the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0032] In the description of the embodiments of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific circumstances.

[0033] In the embodiments of the present invention, unless otherwise clearly specified and limited, the first feature being "above" or "below" the second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, the first feature being "above", "above" and "above" the second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. The first feature being "below", "below" and "below" the second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0034] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0035] The present invention provides an autonomously tunable metasurface structure, see Figure 1 , Figure 1 It is one of the structural schematic diagrams of the self-tuning metasurface structure provided by the embodiment of the present invention. The self-tuning metasurface structure may include an electrode layer 120, a liquid crystal layer 130, a polarizer 100 and a metasurface layer 110 which are stacked.

[0036] By applying voltage to the electrode layer 120, the turning direction of the liquid crystal molecules in the liquid crystal layer 130 is controlled to cooperate with the polarizer 100 to adjust the amplitude of the light in the incident image; the metasurface layer 110 performs phase adjustment on the light after the amplitude is adjusted and responds, and generates a multi-channel B×B spot matrix through spatial multiplexing to determine the emitted characteristic image through the spot matrix.

[0037] Wherein, B represents the number of rows and columns of the spot matrix; different spot brightnesses in the spot matrix are used to represent different convolution kernel weights.

[0038] Optionally, the electrode layer 120 may include a first electrode layer 121 and a second electrode layer 122; the polarizer 100 may be disposed on the supersurface layer 110, and the first electrode layer 121 may be disposed on the polarizer 100; the liquid crystal layer 130 may be disposed on the first electrode layer 121; and the second electrode layer 122 may be disposed on the liquid crystal layer 130.

[0039] The material of the first electrode layer 121 and the second electrode layer 122 may be a transparent conductive material, such as indium tin oxide (ITO), zinc oxide (ZnO) doped material, silver nanowires, and the like.

[0040] Preferably, the materials of the first electrode layer 121 and the second electrode layer 122 are both indium tin oxide.

[0041] Specifically, the electrode layer 120 includes a plurality of sub-electrodes, the liquid crystal layer 130 includes a plurality of pixel units, the super-surface layer 110 includes a plurality of periodically distributed nano-columns, and there is a corresponding relationship between the sub-electrodes, the pixel units and the nano-columns.

[0042] Specifically, the number of sub-electrodes in the first electrode layer 121 and the second electrode layer 122 may be the same and correspond one to one.

[0043] In this embodiment, the applied voltage of each sub-electrode is controlled on the corresponding pixel unit, the amplitude is adjusted in conjunction with the polarizer, and then the phase of the corresponding nanocolumns in the super surface layer is adjusted to achieve independent adjustment of the brightness of each light spot in the light spot matrix.

[0044] The liquid crystal layer is mainly composed of liquid crystal molecules, which have special arrangement and optical properties. They can change their arrangement under the action of electric field, thereby affecting the light waves passing through them. Pixel unit means that the display area of ​​the entire liquid crystal layer is divided into several small areas, each area corresponds to a pixel unit. By changing the voltage applied to the sub-electrode, the liquid crystal molecules in the corresponding pixel unit can be controlled, and then the light amplitude characteristics of the pixel unit can be controlled.

[0045] In this embodiment, the liquid crystal layer in the self-tuning metasurface structure is light amplitude-adjustable, and the metasurface layer is phase-regulated, thereby improving imaging quality and minimum resolution.

[0046] Specifically, autonomous tuning of the complex amplitude field of the metasurface structure It can be expressed as:

[0047] ;

[0048] Wherein, M and N represent the total number of rows and columns of elements in the point spread function, n is the nth row number, and m is the mth column number; Indicates the corresponding weight of each element; represents the working wavelength; x and y represent the position information of the abscissa and ordinate in the plane of the metasurface layer, respectively; represents the pre-designed light transmission angle along the x direction; represents the pre-designed light transmission angle along the y direction; i represents a constant.

[0049] In some embodiments, the thickness of the liquid crystal layer ranges from 1 micrometer to 10 micrometers, the width of the sub-electrode ranges from 500 nanometers to 100 micrometers; and the length of the sub-electrode ranges from 500 nanometers to 100 micrometers.

[0050] Liquid crystal molecules have birefringence, and have different refractive indices for light with different polarization directions. By changing the voltage applied to the liquid crystal molecules through the electrode layer, the director of the liquid crystal molecules and the polarization direction of the incident light form a certain angle, thereby changing the effective refractive index of the liquid crystal, and then changing the optical path of the light, achieving amplitude modulation.

[0051] Exemplarily, the properties of the liquid crystal layer can be expressed as:

[0052] ;

[0053] in, Indicates the direction of rotation of liquid crystal molecules; represents the constant axis refractive index; represents the extraordinary refractive index; Represents the effective refractive index.

[0054] This embodiment uses the electro-optical effect of liquid crystal molecules to change the mixed field effect of the liquid crystal layer through the action of an external electric field, thereby changing the polarization state of the output light. Then, with the appropriate polarizer, the amplitude of the output light is modulated. When the voltage applied to the liquid crystal layer is different, the liquid crystal molecules turn differently, thereby producing different birefringence effects and twisted nematic effects, and the polarization state and direction of the transmitted light are different. After matching with a polarizer in a specific direction, the intensity of the output light will also be different, thus achieving amplitude modulation of the light.

[0055] See also Figure 2 , Figure 2 It is a schematic diagram of the structure of the super surface layer provided in an embodiment of the present invention.

[0056] In this embodiment, the metasurface is a nanoscale optical element with polarization-independent and spatial multiplexing characteristics. In the design of the metasurface, nanopillars are the smallest functional units that constitute the entire surface, and they can produce a specific phase response to the incident electromagnetic wave.

[0057] Specifically, polarization independence means that the metasurface layer has the same optical properties for light with different polarization directions, such as consistent focusing, imaging or other optical functions. Spatial multiplexing refers to the simultaneous processing of multiple independent optical functions on the same metasurface layer. Some optical components can usually only allocate a single optical function (such as focusing or imaging) in space, while metalenses can achieve different phase responses at different spatial locations by designing complex nanostructures on their surface.

[0058] Optionally, the cross-sectional shape of the nanocolumn can be square, cylindrical or cross-shaped; the length of the nanocolumn cross-sectional shape ranges from 100 nanometers to 1000 nanometers; the height of the nanocolumn ranges from 100 nanometers to 1000 nanometers; the distribution period of the nanocolumn ranges from 200 nanometers to 1000 nanometers.

[0059] In this embodiment, the phase response of the nanopillars to the incident light can be adjusted by changing the height, width or shape of the nanopillars. For example, nanopillars of different heights can produce different phase delays, thereby achieving precise control of the phase of the light wave.

[0060] As a functional unit, nanopillars can be used as subunits to construct metasurfaces. Their geometry, size, and material properties directly affect their response to input light, thereby achieving phase control of light.

[0061] Optionally, the metasurface layer is used to perform a dot product operation on the incident image and the numerical value represented in the spot matrix to obtain a characteristic image.

[0062] In the above manner, the embodiment of the present invention combines the metasurface and liquid crystal to obtain an autonomous tuning structure, which can be adjusted by an external voltage. By electrically controlling the orientation of the liquid crystal molecules, the continuous intensity adjustment of the point spread function (PSF) of the two spiral channels is achieved, thereby realizing spatial multiplexing and parallel multi-channel convolution; further, by integrating the liquid crystal layer with adjustable light amplitude and the metasurface layer with phase regulation with the photodetector, an optical convolution kernel can be generated, which can realize the processing of various coded information, thereby improving the accuracy and robustness of object recognition.

[0063] The metasurface variation capability based on liquid crystal polarization control combined with polarization multiplexing is limited. Therefore, in this embodiment, the cascade control characteristics are utilized. The liquid crystal layer is light amplitude adjustable, and the metasurface is phase controlled, thereby realizing a complete optoelectronic hybrid convolutional neural network with autonomous tuning.

[0064] It should be emphasized that, in this embodiment, the electrode layer includes a plurality of sub-electrodes, the liquid crystal layer includes a plurality of pixel units, the super-surface layer includes a plurality of periodically distributed nano-columns, and there is a corresponding relationship between the sub-electrodes, the pixel units and the nano-columns.

[0065] Sub-electrodes refer to the electrode layer being divided into multiple independent small areas, each of which can be independently applied with a voltage. These sub-electrodes correspond to the pixel units in the liquid crystal layer and are used to control the display state of each pixel unit.

[0066] In this embodiment, each pixel unit is controlled by a separate sub-electrode, which can achieve precise control of each pixel unit. Combined with the design of nanocolumns in the metasurface layer, the brightness of each light spot in the light spot matrix can be independently adjusted, further improving the autonomous tuning accuracy of the autonomous tuning metasurface structure.

[0067] Based on any of the above embodiments, the autonomous tuning metasurface structure also includes a first light-transmitting substrate layer and a second light-transmitting substrate layer; the metasurface layer is arranged on the first light-transmitting substrate layer; and the second light-transmitting substrate layer is arranged on the second electrode layer.

[0068] Preferably, the first light-transmitting substrate layer and the second light-transmitting substrate layer are both made of glass.

[0069] See also Figure 3 , Figure 3 This is the second structural schematic diagram of the autonomous tuning metasurface structure provided in an embodiment of the present invention.

[0070] In this embodiment, the self-tuning metasurface structure includes a first light-transmitting substrate layer 341, a metasurface layer 310, a first electrode layer 321, a liquid crystal layer 330, a second electrode layer 322, a second light-transmitting substrate layer 342 and a polarizer (not shown) stacked in sequence.

[0071] Specifically, the supersurface layer 310 is arranged on the first light-transmitting substrate layer 341, the polarizer is arranged on the supersurface layer 310, the first electrode layer 321 is arranged on the polarizer, the liquid crystal layer 330 is arranged on the first electrode layer 321, the second electrode layer 322 is arranged on the liquid crystal layer 330, and the second light-transmitting substrate layer 342 is arranged on the second electrode layer 322.

[0072] This embodiment implements optical convolution through a metasurface, offloading the computationally expensive convolution operation to a high-speed, low-power optical device. The anisotropy of the liquid crystal at the optoelectronic front end can change the polarization of light.

[0073] The present invention also provides an imager based on an optical neural network, see Figure 4-Figure 5 , Figure 4 is a schematic diagram of the structure of an imager based on an optical neural network provided in an embodiment of the present invention, Figure 5 It is a schematic diagram of the principle of an imager based on an optical neural network provided in an embodiment of the present invention.

[0074] The imager based on optical neural network includes a photodetector 420 and the above-mentioned autonomous tuning metasurface structure 410; wherein the autonomous tuning metasurface structure 410 is used to tune the light of the received image to obtain a characteristic image, and the photodetector 420 is used to receive the light of the characteristic image of the autonomous tuning metasurface structure 410 and form an image.

[0075] The self-tuning metasurface structure 410 includes a liquid crystal module 411 and a metasurface module 412. The liquid crystal module 411 further includes the aforementioned polarizer, electrode layer and liquid crystal layer, and the metasurface module 412 includes the aforementioned metasurface layer. The liquid crystal module 411 is used to tune the optical amplitude of the input light, and the metasurface module 412 is used to tune the phase of the input light.

[0076] In deep learning, convolution usually involves multiplying the target image with a B×B matrix, where each pixel in the matrix has an independent weight. This matrix is ​​applied to the image through a dot product operation and is shifted step by step to produce a single feature map.

[0077] In this embodiment, under the condition of incoherent illumination, PSY (x, y) is set to produce B × B foci, each with a specific intensity corresponding to the desired digital core weight. By adjusting the rotation of the liquid crystal molecules, the polarization state of the incoherent light can be modified to achieve nonlinear brightness adjustment of the foci. These foci produce overlapping B × B images on a sensor such as a photodetector, which are spatially moved according to their respective positions. Therefore, these weighted images are rasterized and summed by superimposing them on a sensor such as a photodetector.

[0078] The self-tuning metasurface structure achieves multi-channel signal processing through voltage regulation, rather than the convolution operation commonly performed in digital neural networks. The self-tuning metasurface structure consists of a voltage-regulated encapsulated liquid crystal layer and a pre-designed nuclear encoding two-layer metasurface. It uses a non-correlated light source to illuminate the object, and in the liquid crystal module, it adjusts the direction of rotating liquid crystal by voltage changes to tune the light amplitude. In the metasurface module, phase tuning is performed, and the photodetector can capture the feature map related to the object.

[0079] Continue reading Figure 5 , gives the comparison process of optical convolution and digital convolution, the input object is the light of the target to be identified with different intensities (Intensiyt). If it is processed by optical convolution (Optical), the spot matrix is ​​obtained, and finally the optical convolution result (Optical output) is output; if it is processed by digital convolution (Digital), the digital matrix is ​​obtained, and finally the digital convolution result (Digital output) is output. In optical convolution, each point in the spread function (PSF) has a brightness distribution corresponding to the weight of the digital convolution kernel.

[0080] The present invention also provides a target recognition method based on an optoelectronic hybrid convolutional neural network, see Figure 6-7 , Figure 6 is a flow chart of a target recognition method based on an optoelectronic hybrid convolutional neural network provided by an embodiment of the present invention, Figure 7 It is a structural schematic diagram of the optoelectronic hybrid convolutional neural network architecture provided by an embodiment of the present invention.

[0081] In this embodiment, the target recognition method based on the optoelectronic hybrid convolutional neural network uses the above-mentioned autonomous tuning metasurface structure and may include steps S610 to S630, each of which is specifically as follows:

[0082] S610: Perform convolution processing of the optical neural network on the target to be identified by autonomously tuning the metasurface structure to obtain a feature image after feature enhancement.

[0083] S620: Convert the characteristic image from an optical signal to an electrical signal through a photoelectric detector.

[0084] S630: Inputting the characteristic image of the electrical signal into the digital neural network for target recognition, and obtaining a recognition result output by the digital neural network.

[0085] Combination Figure 5 and Figure 7 This embodiment includes a self-tuning metasurface structure of a liquid crystal module (LC) and a metasurface module (MS1) as an optoelectronic front end, which is a composite optical device (Compound meta-optic) including a meta-optical function, which can replace the convolution layer of a digital neural network. A photodetector (photodetector) including gratings (Gratings) as a digital back end can further identify feature images.

[0086] In this embodiment, the designed self-tuning metasurface structure is used as the optoelectronic front end, while the digital back end includes a digital maximum pooling module, a rectified linear unit activation function, and a fully connected layer. The optical convolution layer can use 12 designed convolution channels to generate feature maps, while the fixed convolution channel is nonlinearly changed by adjusting the voltage. Each kernel consists of N=3 pixels, and the self-tuning metasurface structure introduces an additional voltage dimension to enhance object recognition in different scenarios.

[0087] For classification tasks, the feature map generated by the autonomously tuned metasurface structure is input into the back-end of the digital neural network for recognition, and the maximum probability can be selected from the preliminary recognition results as the recognition result.

[0088] In summary, the embodiments of the present invention provide an autonomously tuned metasurface structure, an imager, and a target recognition method, wherein the autonomously tuned metasurface structure, as an optoelectronic front end, replaces the traditional optical imaging element, and can process various coded information, thereby improving the classification accuracy and robustness of various objects. The system performs light amplitude tuning through a liquid crystal layer and phase control through a metasurface layer, thereby achieving flexible tuning of the entire system, achieving independent adjustment of the brightness of each light spot in the light spot matrix, and facilitating nonlinear brightness adjustment of multiple fixed information channels to generate multiple optical convolution channels; by integrating the autonomously tuned metasurface structure with a digital backend based on a photodetector / chip, etc., the speed can be increased, and data reading and transmission can be achieved without an analog-to-digital converter, which will help achieve ultra-fast and low-latency effects. In addition, the optical convolution layer overcomes the limitations of the optical front end of the traditional imager in terms of depth or number of layers, paving the way for large-scale models in neural network applications. Therefore, this autonomously tuned metasurface structure provides flexible parameter adjustment while generating rich information channels in a limited metasurface area, and has great potential in accelerating matrix operations.

[0089] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0090] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An autonomously tunable metasurface structure, characterized in that: It includes a polarizer, an electrode layer, a liquid crystal layer and a super surface layer which are stacked; By applying voltage to the electrode layer, the liquid crystal molecules in the liquid crystal layer are controlled to turn in coordination with the polarizer to adjust the amplitude of the light in the incident image; the metasurface layer performs phase adjustment on the light after the amplitude is adjusted and responds, and generates a multi-channel B×B spot matrix through spatial multiplexing to determine the emitted characteristic image through the spot matrix; B represents the number of rows and columns of the spot matrix; different spot brightness in the spot matrix is ​​used to characterize different convolution kernel weights; The electrode layer includes a plurality of sub-electrodes, the liquid crystal layer includes a plurality of pixel units, the super-surface layer includes a plurality of periodically distributed nano-columns, and the sub-electrodes, the pixel units and the nano-columns have a corresponding relationship; by controlling the applied voltage of each sub-electrode on the corresponding pixel unit, the amplitude is adjusted in conjunction with the polarizer, and then combined with the phase adjustment of the corresponding nano-columns in the super-surface layer, the brightness of each light spot in the light spot matrix can be independently adjusted; The complex amplitude field of the autonomously tuned metasurface structure It is expressed as: ； Wherein, M and N represent the total number of rows and columns of elements in the point spread function, n is the nth row number, and m is the mth column number; Indicates the corresponding weight of each element; represents the operating wavelength; x and y represent the position information of the abscissa and ordinate in the plane of the metasurface layer respectively; represents the pre-designed light transmission angle along the x direction; represents the pre-designed light transmission angle along the y direction; i represents a constant.

2. The autonomously tunable metasurface structure according to claim 1, characterized in that: The thickness of the liquid crystal layer ranges from 1 micrometer to 10 micrometers, the width of the sub-electrode ranges from 500 nanometers to 100 micrometers; the length of the sub-electrode ranges from 500 nanometers to 100 micrometers; The cross-sectional shape of the nanocolumn is square, cylindrical or cross-shaped; the length of the cross-sectional shape of the nanocolumn ranges from 100 nanometers to 1000 nanometers; the height of the nanocolumn ranges from 100 nanometers to 1000 nanometers; and the distribution period of the nanocolumn ranges from 200 nanometers to 1000 nanometers.

3. The autonomously tunable metasurface structure according to claim 1, characterized in that: The properties of the liquid crystal layer are expressed as: ； in, Indicates the direction of rotation of liquid crystal molecules; represents the constant axis refractive index; represents the extraordinary refractive index; Represents the effective refractive index.

4. The autonomously tunable metasurface structure according to claim 1, characterized in that: The electrode layer comprises a first electrode layer and a second electrode layer; The polarizer is arranged on the super surface layer; the first electrode layer is arranged on the polarizer; the liquid crystal layer is arranged on the first electrode layer; and the second electrode layer is arranged on the liquid crystal layer.

5. The autonomously tunable metasurface structure according to claim 4, characterized in that: Also includes a first light-transmitting substrate layer and a second light-transmitting substrate layer; The super surface layer is arranged on the first light-transmitting substrate layer; and the second light-transmitting substrate layer is arranged on the second electrode layer.

6. The autonomously tunable metasurface structure according to claim 1, characterized in that: The metasurface layer is used to perform a dot product operation on the incident image and the numerical value represented in the light spot matrix to obtain the characteristic image.

7. The autonomously tunable metasurface structure according to claim 4, characterized in that: The materials of the first electrode layer and the second electrode layer are both indium tin oxide.

8. The autonomously tunable metasurface structure according to claim 5, characterized in that: The first light-transmitting substrate layer and the second light-transmitting substrate layer are both made of glass.

9. An imager based on an optical neural network, characterized in that: A method comprising: comprising a photodetector and an autonomously tunable metasurface structure according to any one of claims 1 to 8; The autonomously tuned metasurface structure is used to tune the light of the received image to obtain a characteristic image, and the photodetector is used to receive the light of the characteristic image of the autonomously tuned metasurface structure and form an image.

10. A target recognition method based on an optoelectronic hybrid convolutional neural network, characterized in that: Using the autonomous tuning metasurface structure according to any one of claims 1 to 8, the target recognition method based on the optoelectronic hybrid convolutional neural network comprises: The self-tuning metasurface structure is used to perform convolution processing of an optical neural network on the target to be identified, so as to obtain a feature image after feature enhancement; Converting the characteristic image from an optical signal to an electrical signal through a photodetector; The characteristic image of the electrical signal is input into a digital neural network for target recognition, and a recognition result output by the digital neural network is obtained.

Images