A metasurface-based encryption method

Abstract

The present invention provides an encryption method based on a metasurface, and belongs to the field of optical encryption technology. The present invention electronically encrypts a pattern to be encrypted and makes it into a mask. Under the action of a light field, the mask is used to encode the encrypted pattern to a metasurface. A metasurface with linear operation mapping capability is designed by a reverse design method, and the encrypted pattern is modulated by the metasurface to restore the original pattern. The present invention uses a metasurface to perform the function of a decryption matrix, making full use of the advantages of high integration, strong modulation capability, and high design freedom of the metasurface. It can be arranged in different spatial positions, and the encryption combination form is increased by coupling the wavelength of the light field and the spatial information of the metasurface, thereby improving the security of the system, and can also increase the information capacity in the encryption process without the need to store and transmit complex keys.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical encryption, and in particular relates to an encryption method based on a metasurface. Background Art

[0002] Optical encryption utilizes the high parallel processing capability, anti-electromagnetic interference and large-capacity data processing advantages of optical technology to achieve information encryption through Fourier transform, phase encryption, spatial light modulation and holographic technology. It is widely used in military communications, data storage, financial transactions and medical data protection.

[0003] In related technologies, optical encryption mostly uses a dual random phase encoding method, which introduces random phase masks in the spatial domain and frequency domain respectively to encrypt the image. The encrypted image is random noise, and the original image cannot be directly identified. Only when two random phase masks are simultaneously available can the original image be correctly decrypted and restored. This method improves security through a dual encryption mechanism, but it has the disadvantages of limited information capacity and complex key management.

[0004] Therefore, it is necessary to provide an encryption method based on metasurface to solve the above problems. Summary of the invention

[0005] The present invention provides an encryption method based on a metasurface, which utilizes the metasurface to perform the function of a decryption matrix, making full use of the advantages of the metasurface such as high integration, strong modulation capability, and high degree of design freedom, and can effectively improve the security of the system and also increase the information capacity during the encryption process without the need to store and transmit complex keys, thereby solving at least one technical problem involved in the background technology.

[0006] In order to solve the above-mentioned technical problems, the present invention is achieved as follows:

[0007] A metasurface-based encryption method comprises the following steps:

[0008] Constructing a proxy model: Based on the inverse design method, a proxy model is constructed to simulate the modulation process of the metasurface on the light field, and learn the mapping relationship between the structural parameters of the metasurface and the optical response under the action of light fields of different wavelengths;

[0009] Metasurface design: Set an encryption matrix, randomly generate a training matrix with an amplitude value range of 0-1, use the product of the encryption matrix and the training matrix as the input of the metasurface, and minimize the output of the metasurface and the loss of the training matrix as the optimization goal. Optimize the structural parameters of the metasurface under the action of light fields of different wavelengths to obtain the optimal structural parameters of the metasurface required under the action of light fields of each wavelength;

[0010] Metasurface fabrication: Select the wavelength of the light field and use the optimal structural parameters required under the action of the light field of this wavelength to fabricate the metasurface;

[0011] Electronic encryption: Provide the original pattern and use the encryption matrix to electronically encrypt the original pattern;

[0012] Mask design: Encode the encrypted pattern into a mask;

[0013] Optical decryption: Under the action of a light field of a selected wavelength, a mask is used to transfer the encrypted pattern to the metasurface, and the original pattern is restored after modulation by the metasurface.

[0014] As a preferred improvement, the metasurface operates in the visible light band and is made of titanium dioxide material. The structural parameters of the metasurface include the size data of each nanounit.

[0015] As a preferred improvement, the proxy model is a fully connected neural network model, and the structural parameters of the metasurface and the wavelength of the light field are used as the input of the fully connected neural network model, and the complex amplitude of the output light field of the metasurface is used as the output target of the fully connected neural network model. FDTD is used to simulate the modulation process of the metasurface, and the results of the FDTD simulation are used as the true value. The output target of the fully connected neural network model is used as the predicted value, and the mean absolute error between the predicted value and the true value is used as the loss function. Stochastic gradient descent is used as the optimization algorithm, and iterative training is performed with the goal of minimizing the loss function until the model is fitted or the maximum number of iterations is reached.

[0016] As a preferred improvement, the metasurface is composed of a plurality of rectangular nanounits, and the structural parameters include the length and width of each nanounit. The physical model of the fully connected neural network model is expressed as:

[0017] ;

[0018] ;

[0019] ;

[0020] In the formula, Indicates the i The complex transmission coefficient of each nanometer unit; Represents the first i The spatial coordinates of the nanounits; represents the wavelength of the light field; represents the amplitude of the light field; represents the imaginary part of the light field; Represents the phase of the light field; Respectively represent iThe length and width of the nanounit; Represents fully connected neural network model processing; Represents the diffraction coefficient of the light field propagating in free space; Represents the output of the fully connected neural network model; Represents the input of the fully connected neural network model; where:

[0021] ;

[0022] In the formula, represents the spatial coordinates of the light field after propagation in free space, .

[0023] As a preferred improvement, under the action of each light field of a specific wavelength, each encryption matrix has one and only one set of metasurfaces with optimal structural parameters corresponding to it.

[0024] As a preferred improvement, a plurality of modulation areas are arranged on the metastructure surface, each modulation area is set to act as a light field of a different wavelength, and a plurality of modulation channels are arranged on each modulation area, each modulation channel corresponds to an encryption matrix, and is used to perform an optical decryption process matching the encryption matrix under the action of a light field of a specific wavelength in the modulation area.

[0025] As a preferred improvement, the encryption and decryption process specifically includes the following steps:

[0026] Electronic encryption: Provide an original pattern, divide the original pattern into multiple sub-patterns, each sub-pattern corresponds to a modulation area, and use the encryption matrix corresponding to any modulation channel in the modulation area to electronically encrypt the sub-pattern;

[0027] Mask design: Encode the encrypted pattern of each sub-pattern into a mask;

[0028] Optical decryption: Under the action of the light field of the selected wavelength in each modulation area, the encrypted pattern is transmitted to the metasurface using a mask, and decrypted by the modulation channel in the corresponding modulation area. Finally, the decrypted pattern is spliced ​​to restore the original pattern.

[0029] As a preferred improvement, the metasurface and the mask adopt a separate architecture, and the metasurface and the mask are manufactured on different transparent substrates.

[0030] As a preferred improvement, the metasurface and the mask adopt an integrated architecture. The metasurface and the mask are integrated onto the same transparent substrate to form an optical modulator, and then the optical modulator is glued and fixed to the CMOS pattern processor to achieve on-chip integration. The CMOS pattern processor is used to collect the output light field of the metasurface and read its light intensity. When in use, a light field of a specific wavelength is incident on a specific modulation area to directly restore the original pattern.

[0031] As a preferred improvement, the preparation process of the optical modulator includes the following steps:

[0032] Step S1, using ITO glass as a substrate, coating a layer of aluminum film on the surface of the substrate by an ion beam sputtering process and peeling off the shape of the mask;

[0033] Step S2, spin coating an isolation layer on the mask surface;

[0034] Step S3, spin coating photoresist on the surface of the isolation layer, and producing the designed metasurface pattern on the isolation layer by exposure and development;

[0035] Step S4, using atomic layer deposition technology to alternately and cyclically deposit titanium tetrachloride and deionized water on the surface of the isolation layer, forming titanium oxide through a gas-solid chemical adsorption reaction, and ensuring that the exposure pattern is completely filled with titanium oxide;

[0036] Step S5, removing excess titanium oxide on the photoresist by ion beam etching technology to obtain a titanium oxide nanorod array wrapped by the photoresist;

[0037] Step S6, using reactive ion beam etching to slowly react the photoresist that wraps the nanorods with oxygen to release the wrapped structure, completing the manufacture of the optical modulation device.

[0038] The beneficial effects of the present invention are:

[0039] The present invention electronically encrypts the pattern to be encrypted and makes it into a mask. Under the action of the light field, the mask is used to encode the encrypted pattern information to the metasurface. The metasurface with linear operation mapping capability is designed by the reverse design method. After the encrypted pattern is modulated by the metasurface, the original pattern can be restored. The present invention uses the metasurface to perform the function of the decryption matrix, making full use of the advantages of the high integration, strong modulation capability and high design freedom of the metasurface. It can be arranged in different spatial positions, and the encryption combination form is increased by coupling the wavelength of the light field and the spatial information of the metasurface, thereby improving the security of the system, and can also increase the information capacity in the encryption process without the need to store and transmit complex keys. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0041] Figure 1 A diagram showing the architecture of the fully connected neural network model in Example 1;

[0042] Figure 2 A schematic diagram showing the distribution of modulation areas on the surface of the metastructure in Example 1;

[0043] Figure 3 A schematic diagram showing the segmentation of the original pattern in Example 1;

[0044] Figure 4 A design diagram of a mask in Example 1 is shown. DETAILED DESCRIPTION

[0045] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0046] Example 1

[0047] like Figure 1-Figure 4 As shown, this embodiment provides an encryption method based on a metasurface, comprising the following steps:

[0048] Constructing a proxy model: Based on the inverse design method, a proxy model is constructed to simulate the modulation process of the metasurface on the light field, and learn the mapping relationship from the structural parameters of the metasurface to the optical response under the action of light fields of different wavelengths.

[0049] The metasurface is an optical modulation element composed of multiple nanounits. The optical response data of each nanounit is discretely distributed. The overall optical response of the metasurface cannot be expressed by analytical expressions like traditional diffraction elements. Therefore, the present invention constructs a proxy model of the optical response of the metasurface under different wavelength light fields with respect to the structural parameters, uses the proxy model to characterize the mapping relationship from the structural parameters to the optical response of each nanounit under different wavelength light fields, and uses the proxy model to predict the optical response of the metasurface under different light fields.

[0050] The structural parameters of the metasurface include the size data of each nanounit. In this embodiment, the metasurface is composed of a plurality of nanounits of rectangular structure, and the structural parameters include the length and width of each nanounit. In other embodiments, the nanounit can also be selected as a circular structure, a triangular structure, an elliptical structure or a polygonal structure (with more than 4 sides). When it is a circular structure, the structural parameters include the diameter; when it is a triangular structure and a polygonal structure, the structural parameters include the length of each side; when it is an elliptical structure, the structural parameters include the length of the major axis and the length of the minor axis.

[0051] In this embodiment, the proxy model is selected as a fully connected neural network model, and its architecture is as follows: Figure 1 As shown. The structural parameters of the metasurface and the wavelength of the light field are used as the input of the fully connected neural network model, and the complex amplitude of the output light field of the metasurface is used as the output target of the fully connected neural network model. The modulation process of the metasurface is simulated using FDTD (Finite Difference TimeDomain). The results of the FDTD simulation are used as the true value, and the output target of the fully connected neural network model is used as the predicted value. The mean absolute error between the predicted value and the true value is used as the loss function. The stochastic gradient descent method is used as the optimization algorithm. Iterative training is performed with the goal of minimizing the loss function until the model is fitted or the maximum number of iterations is reached. After the model training is completed, the complex amplitude of the output light field can be predicted given the wavelength of the light field and the structural parameters of the metasurface.

[0052] The physical model of the fully connected neural network model is expressed as:

[0053] ;

[0054] ;

[0055] ;

[0056] In the formula, Indicates the i The complex transmission coefficient of each nanometer unit; Represents the first i The spatial coordinates of the nanounits; represents the wavelength of the light field; represents the amplitude of the light field; represents the imaginary part of the light field; Represents the phase of the light field; Respectively represent i The length and width of the nanounit; Represents fully connected neural network model processing; Represents the diffraction coefficient of the light field propagating in free space; Represents the output of the fully connected neural network model; Represents the input of the fully connected neural network model; where:

[0057] ;

[0058] In the formula, represents the spatial coordinates of the light field after propagation in free space, .

[0059] Metasurface design: Set an encryption matrix and randomly generate a training matrix with an amplitude value range of 0-1. Use the product of the encryption matrix and the training matrix as the input of the metasurface. Take minimizing the output of the metasurface and the loss of the training matrix as the optimization goal, optimize the structural parameters of the metasurface under the action of light fields of different wavelengths, and obtain the optimal structural parameters of the metasurface required under the action of light fields of each wavelength.

[0060] In the field of encryption, the encryption matrix and the decryption matrix are a pair of mutually inverse matrices. The encryption matrix is ​​used to encrypt the data plaintext to obtain the ciphertext, and the decryption matrix is ​​used to decrypt the ciphertext to restore the data plaintext. The encryption process of the encryption matrix and the decryption process of the decryption matrix are essentially processes of performing linear operations. Similarly, the modulation process of the metasurface is essentially a process of performing linear operations, so the metasurface can be used to replace the decryption matrix to perform the decryption process. In order to make the modulation process of the metasurface match the encryption matrix, the metasurface needs to be designed; in addition, under the action of light fields of different wavelengths, the modulation effect of the metasurface is also different. Therefore, the structural parameters of the metasurface under the action of light fields of different wavelengths are optimized so that it can match a specific encryption matrix and perform the optical decryption process corresponding to the encryption matrix.

[0061] The optimization process belongs to the conventional technology in this field. Since the metasurface performs the decryption process, its expected output should be close to the training matrix. Therefore, the structural parameters of the metasurface are optimized by calculating the loss between the output of the metasurface and the training matrix to obtain the optimal structural parameters of the metasurface required under the light field of each wavelength. In order to ensure the unique correspondence between the encryption matrix and the decryption matrix, under the light field of each specific wavelength, each encryption matrix has and only has one set of metasurfaces with optimal structural parameters corresponding to it. Under the light field of a specific wavelength, the metasurface with the optimal structural parameters can perform the decryption function of the encrypted data of the encryption matrix. In other words, the metasurfaces with other structural parameters cannot perform the decryption function; similarly, under the light field of other wavelengths, the metasurface with the optimal structural parameters cannot perform the decryption function.

[0062] Metasurface production: Select the wavelength of the light field and use the optimal structural parameters required under the action of the light field of this wavelength to produce the metasurface.

[0063] In this embodiment, the metasurface operates in the visible light band, so a titanium dioxide material with high refraction and low loss in the visible light band is used to prepare the metasurface, and the preparation process can adopt the exposure and development technology in the art.

[0064] Electronic encryption: Provide the original pattern and use the encryption matrix to electronically encrypt the original pattern.

[0065] The original image can be represented by a matrix with amplitude values ​​in the range of 0-1, which retains the same properties as the training matrix and can be modulated by the designed metasurface.

[0066] Mask design: Encode the encrypted pattern into a mask.

[0067] The plane wave light field is digitally encoded through a designed mask, with no light passing through encoded as 0 and light passing through encoded as 1, so that the encrypted pattern can be encoded into the metasurface. After encoding, the input of the metasurface not only carries the information of the original image, but also carries the encrypted matrix information of a specific wavelength at a specific position.

[0068] Optical decryption: Under the action of a light field of a selected wavelength, a mask is used to transfer the encrypted pattern to the metasurface, and the original pattern is restored after modulation by the metasurface.

[0069] Furthermore, in order to enrich the form of encryption, multiple modulation areas are set on the metasurface, each modulation area is set to have a light field effect of a different wavelength, and multiple modulation channels are set on each modulation area, each modulation channel corresponds to an encryption matrix, and is used to perform an optical decryption process matching the encryption matrix under the light field effect of a specific wavelength in the modulation area. It should be noted that the encryption matrix corresponding to different modulation channels is generated by randomly disrupting the order of rows or columns of the unit matrix to ensure the sparsity and orthogonality of the encryption matrix.

[0070] This setting method increases the encryption combination form by coupling the wavelength of the light field and the spatial information of the metasurface, such as Figure 2 As shown, in this embodiment, the metasurface is provided with four different modulation areas P1, P2, P3, and P4. The four decryption areas have the ability to modulate the light fields with wavelengths of 450nm, 530nm, 580nm, and 640nm respectively; each modulation area has four modulation channels, and each modulation area randomly selects a modulation channel to play a role. The four modulation areas each undergo a modulation channel selection once, and a total of There are different encryption matrix combinations, but only one set of encryption matrices is actually selected, and the other combinations act as interference items. With the increase of modulation areas and modulation channels, the encryption combinations that can be provided will be richer.

[0071] On this basis, the encryption and decryption process specifically includes the following steps:

[0072] Electronic encryption: Provide an original pattern and divide the original pattern into multiple sub-patterns, each sub-pattern corresponds to a modulation area, and use the encryption matrix corresponding to any modulation channel in the modulation area to electronically encrypt the sub-pattern.

[0073] The original pattern is divided into four sub-patterns I1, I2, I3, and I4, which are sent to modulation areas P1, P2, P3, and P4 for modulation one by one, and the sub-pattern is electronically encrypted using an encryption matrix corresponding to any modulation channel in the modulation area.

[0074] Mask design: Encode the encrypted pattern of each sub-pattern into a mask.

[0075] The generated mask is Figure 4 As shown, O1, O2, O3, and O4 are masks corresponding to sub-patterns I1, I2, I3, and I4, respectively.

[0076] Optical decryption: Under the action of the light field of the selected wavelength in each modulation area, the encrypted pattern is transmitted to the metasurface using a mask, and decrypted by the modulation channel in the corresponding modulation area. Finally, the decrypted pattern is spliced ​​to restore the original pattern.

[0077] The present invention makes full use of the advantages of high integration, strong modulation capability and high design freedom of metasurfaces. Metasurfaces in different modulation areas are combined together to act as a common computing platform, which can perform specific linear computing modulation on any input, is portable, and can be arranged in different spatial positions. The original pattern adopts a segmentation encryption method, so that the original pattern can be selected to have a larger size, thereby accommodating more information; the multiple segments can independently and parallelly perform the encryption and decryption process, which can improve the efficiency of encryption.

[0078] In the case of multiple modulation channels, the input of the metasurface will also change with the processing task, location, and modulation channel. In the face of different tasks, only the input mask changes, which does not affect the design of the metasurface. If you want to successfully decrypt, the electronically encrypted image information needs to be fed into the metasurface in the form of a mask.

[0079] In this embodiment, the metasurface and the mask adopt a separate architecture, and the mask and the metasurface are respectively manufactured on different transparent substrates, which has greater flexibility. In different task scenarios, different original patterns can be selected to prepare different masks, and then the encrypted pattern information in different masks is encoded into the metasurface, and decrypted through the modulation process of the metasurface.

[0080] Example 2

[0081] In this embodiment, an on-chip integrated architecture is provided, in which the mask and the metasurface are integrated onto the same transparent substrate to form an optical modulator, and then the optical modulator is glued and fixed to the CMOS pattern processor to achieve on-chip integration. The CMOS pattern processor is used to collect the output light field of the metasurface and read its light intensity. Since the mask and the metasurface are integrated together, when in use, only a light field of a specific wavelength needs to be incident at a specific position to directly restore the original pattern.

[0082] In this embodiment, the preparation process of the optical modulator includes the following steps:

[0083] Step S1, using ITO glass as a substrate, coating a layer of aluminum film on the surface of the substrate by an ion beam sputtering process and peeling off the shape of the mask;

[0084] Step S2, spin coating a layer of HSQ (Hydrogen Silsesquioxane) on the surface of the mask as an isolation layer;

[0085] Step S3, spin-coating PMMA (Polymethyl methacrylate) photoresist on the surface of the isolation layer, and then making the designed meta-structure surface pattern on the isolation layer by exposure and development;

[0086] Step S4, using atomic layer deposition technology to alternately and cyclically deposit titanium tetrachloride and deionized water on the surface of the isolation layer to form titanium oxide through a gas-solid chemical adsorption reaction, and the number of deposition cycles is set to 2800 to ensure that the exposure pattern is completely filled with titanium oxide;

[0087] Step S5, removing excess titanium oxide on the PMMA photoresist by ion beam etching technology to obtain a titanium oxide nanorod array wrapped by the PMMA photoresist;

[0088] Step S6, using reactive ion beam etching to slowly react the photoresist that wraps the nanorods with oxygen to release the wrapped structure, completing the production of the optical modulator.

[0089] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A metasurface-based encryption method, characterized in that: The steps include: Constructing a proxy model: Based on the inverse design method, a proxy model is constructed to simulate the modulation process of the metasurface on the light field, and learn the mapping relationship between the structural parameters of the metasurface and the optical response under the action of light fields of different wavelengths; Metasurface design: Set an encryption matrix, randomly generate a training matrix with an amplitude value range of 0-1, use the product of the encryption matrix and the training matrix as the input of the metasurface, and minimize the output of the metasurface and the loss of the training matrix as the optimization goal. Optimize the structural parameters of the metasurface under the action of light fields of different wavelengths to obtain the optimal structural parameters of the metasurface required under the action of light fields of each wavelength; Metasurface fabrication: Select the wavelength of the light field and use the optimal structural parameters required under the action of the light field of this wavelength to fabricate the metasurface; Electronic encryption: Provide the original pattern and use the encryption matrix to electronically encrypt the original pattern; Mask design: Encode the encrypted pattern into a mask; Optical decryption: Under the action of a light field of a selected wavelength, a mask is used to transfer the encrypted pattern to the metasurface, and the original pattern is restored after modulation by the metasurface; The proxy model is a fully connected neural network model. The structural parameters of the metasurface and the wavelength of the light field are used as the input of the fully connected neural network model, and the complex amplitude of the output light field of the metasurface is used as the output target of the fully connected neural network model. FDTD is used to simulate the modulation process of the metasurface. The result of FDTD simulation is used as the true value, and the output target of the fully connected neural network model is used as the predicted value. The mean absolute error between the predicted value and the true value is used as the loss function. Stochastic gradient descent is used as the optimization algorithm, and iterative training is performed with the goal of minimizing the loss function until the model is fitted or the maximum number of iterations is reached.

2. The metasurface-based encryption method according to claim 1, characterized in that: The metasurface works in the visible light band and is made of titanium dioxide material. The structural parameters of the metasurface include the size data of each nanounit.

3. The metasurface-based encryption method according to claim 1, characterized in that: The metasurface is composed of multiple rectangular nanounits. The structural parameters include the length and width of each nanounit. The physical model of the fully connected neural network model is expressed as: ； ； ； In the formula, Indicates the i The complex transmission coefficient of each nanometer unit; Represents the first i The spatial coordinates of the nanounits; represents the wavelength of the light field; represents the amplitude of the light field; represents the imaginary part of the light field; Represents the phase of the light field; Respectively represent i The length and width of the nanounit; Represents fully connected neural network model processing; Represents the diffraction coefficient of the light field propagating in free space; Represents the output of the fully connected neural network model; Represents the input of the fully connected neural network model; where: ； In the formula, represents the spatial coordinates of the light field after propagation in free space, .

4. The metasurface-based encryption method according to claim 1, characterized in that: Under the action of each specific wavelength of light field, each encryption matrix has one and only one set of metasurfaces with optimal structural parameters corresponding to it.

5. The metasurface-based encryption method according to claim 1, characterized in that: There are multiple modulation areas on the surface of the metastructure, each modulation area is set to have a light field effect of a different wavelength, and each modulation area is set to have multiple modulation channels, each modulation channel corresponds to an encryption matrix, and is used to perform an optical decryption process matching the encryption matrix under the action of a light field of a specific wavelength in the modulation area.

6. The metasurface-based encryption method according to claim 5, characterized in that: The encryption and decryption process specifically includes the following steps: Electronic encryption: Provide an original pattern, divide the original pattern into multiple sub-patterns, each sub-pattern corresponds to a modulation area, and use the encryption matrix corresponding to any modulation channel in the modulation area to electronically encrypt the sub-pattern; Mask design: Encode the encrypted pattern of each sub-pattern into a mask; Optical decryption: Under the action of the light field of the selected wavelength in each modulation area, the encrypted pattern is transmitted to the metasurface using a mask, and decrypted by the modulation channel in the corresponding modulation area. Finally, the decrypted pattern is spliced ​​to restore the original pattern.

7. The metasurface-based encryption method according to claim 1, characterized in that: The metasurface and mask adopt a separate architecture, and the metasurface and mask are manufactured on different transparent substrates.

8. The metasurface-based encryption method according to claim 7, characterized in that: The metasurface and mask adopt an integrated architecture. The metasurface and mask are integrated onto the same transparent substrate to form an optical modulator, and then the optical modulator is glued and fixed to the CMOS pattern processor to achieve on-chip integration. The CMOS pattern processor is used to collect the output light field of the metasurface and read its light intensity. When in use, a light field of a specific wavelength is incident on a specific modulation area to directly restore the original pattern.

9. The metasurface-based encryption method according to claim 8, characterized in that: The preparation process of the optical modulator includes the following steps: Step S1, using ITO glass as a substrate, coating a layer of aluminum film on the surface of the substrate by an ion beam sputtering process and peeling off the shape of the mask; Step S2, spin coating an isolation layer on the mask surface; Step S3, spin coating photoresist on the surface of the isolation layer, and producing the designed metasurface pattern on the isolation layer by exposure and development; Step S4, using atomic layer deposition technology to alternately and cyclically deposit titanium tetrachloride and deionized water on the surface of the isolation layer, forming titanium oxide through a gas-solid chemical adsorption reaction, and ensuring that the exposure pattern is completely filled with titanium oxide; Step S5, removing excess titanium oxide on the photoresist by ion beam etching technology to obtain a titanium oxide nanorod array wrapped by the photoresist; Step S6, using reactive ion beam etching to slowly react the photoresist that wraps the nanorods with oxygen to release the wrapped structure, completing the manufacture of the optical modulation device.

Images