Information encryption and decryption method for photovoltaic cell cover plate glass

Abstract

The invention discloses an information encryption and decryption method for photovoltaic cell cover plate glass, and relates to the crossing field of a photovoltaic module technology and information security. The method comprises the following steps: acquiring a plaintext image including photovoltaic cell cover plate glass information; generating a mask pattern and a convolution kernel corresponding to the plaintext image through a convolution principle; the mask pattern is displayed on the electrochromic glass; acquiring a mask graphic picture displayed on the shot electrochromic glass, and inputting the mask graphic picture into the neural network model for recognition to obtain a recognition result; the recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification; and carrying out convolution on the identification result and the convolution kernel to obtain a decrypted image. According to the method, the power generation efficiency of the photovoltaic cell is not influenced in the information encryption and decryption process of the photovoltaic cell cover plate glass.

Description

Technical Field

[0001] The present invention relates to the intersection of photovoltaic module technology and information security, and in particular to an information encryption and decryption method for photovoltaic cell cover glass. Background Art

[0002] Currently, in the photovoltaic industry, the cover glass of battery modules, such as tempered glass, coated glass, or electrochromic glass, must not only protect the internal cells from environmental corrosion but also meet key performance requirements such as high light transmittance and weather resistance. With the rapid development of the photovoltaic industry, module production traceability, anti-counterfeiting certification, and full lifecycle management have become particularly important.

[0003] Traditional electrochromic glass encryption methods, such as development based on special ink materials or information encryption based on optical holography, have disadvantages such as the corrosiveness of special ink materials, development conditions that are not suitable for photovoltaic cells, and the large pixel scale required for encryption and decryption methods. As a result, the power generation efficiency of photovoltaic cells is low during the information encryption and decryption process of photovoltaic cell cover glass. Summary of the Invention

[0004] Therefore, it is necessary to provide a method for encrypting and decrypting information on a photovoltaic cell cover glass to address the above technical issues. This method can achieve the goal of not affecting the power generation efficiency of the photovoltaic cell during the encryption and decryption process of the information on the photovoltaic cell cover glass.

[0005] The present invention adopts the following technical solutions: The present invention provides a method for encrypting and decrypting photovoltaic cell cover glass information, comprising: Acquire a plaintext image including photovoltaic cell cover glass information; Generate the mask graphics and convolution kernel corresponding to the plaintext image through the convolution principle; Displaying the mask pattern on the electrochromic glass; A photograph of a mask pattern displayed on the electrochromic glass is obtained and input into a neural network model for recognition, obtaining a recognition result. The neural network model includes a YOLO model and a convolutional neural network connected in series. The recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification. Convolve the recognition result with the convolution kernel to obtain the decrypted image.

[0006] Preferably, generating a mask graphic and a convolution kernel corresponding to the plaintext image by the convolution principle specifically includes: Encode the plaintext image into a pixel size of Plaintext image of Get the pixel size The numerical matrix corresponding to the plaintext image; Random Generation The binary matrix of is used as the mask pattern matrix; the mask pattern is the visualization form of the mask pattern matrix; According to the numerical matrix and the mask pattern matrix, the convolution kernel is obtained by calculation using the convolution theorem.

[0007] Preferably, the convolution kernel is obtained by performing calculations based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: According to the numerical matrix , mask pattern matrix And the mathematical expression formula of the convolution theorem, we get the convolution kernel matrix ; The mathematical expression of the convolution theorem is: ; in, represents the Fourier transform, is the mask pattern matrix, is a numerical matrix; Convolution kernel matrix Perform inverse Fourier transform and get the convolution kernel as: ; in, represents the inverse Fourier transform.

[0008] Preferably, the convolution kernel is obtained by performing calculations based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: based on M 、 and The objective function is constructed by Fourier transform, which is: ; in, represents the regularization parameter, the operator Frobenius 2 norm; Based on numerical matrix , mask pattern matrix Solve the objective function to obtain the convolution kernel matrix The solution is: ; in, Representation matrix The Hermitian transpose of Convolution kernel matrix Perform inverse Fourier transform to obtain the convolution kernel.

[0009] Preferably, the regularization parameter The determination includes: Generate a set of candidate regularization parameters in the logarithmic space List of; For each in the list Solve the frequency domain regularization equation , and obtain the restored image through inverse Fourier transform; Calculate the structural similarity index between the restored image and the plaintext image, and take the image corresponding to the maximum structural similarity index value as , is determined as the regularization parameter.

[0010] Preferably, displaying the mask pattern on the electrochromic glass specifically includes: Design the circuit connecting the electrochromic glass and the microcontroller and connect them in sequence; The mask shape The matrix is ​​stretched to The method further comprises the steps of: writing a row vector list of a pixel size and writing the row vector list into a program of a microcontroller, and running the program of the microcontroller to display a mask pattern on the electrochromic glass.

[0011] Preferably, the mask pattern photo is input into a neural network model for recognition to obtain a recognition result, which specifically includes: Use the YOLO model to identify the image information of electrochromic glass in the mask pattern photo; Converting the recognized image information of the electrochromic glass into a mask pattern binary matrix; Use convolutional neural networks to identify the color of each square in the binary matrix of the masked image; The binary matrix of the mask pattern after color recognition is determined as the recognition result; the color recognition is used to quantify the distribution of the colored area of ​​the electrochromic glass.

[0012] Preferably, the training process of the neural network model specifically includes: Obtain a set of photos of people holding electrochromic glass as the first training set; Get part of the output of the YOLO model and cut and split it as the second training set; Train the YOLO model using the first training set to obtain a trained YOLO model; The convolutional neural network is trained using the second training set to obtain a trained convolutional neural network; The trained YOLO model and the trained convolutional neural network are determined as neural network models.

[0013] The present invention provides a device for encrypting and decrypting information of a photovoltaic cell cover glass, comprising: An acquisition module, used to acquire a plaintext image including information about the photovoltaic cell cover glass; A generation module is used to generate a mask graphic and a convolution kernel corresponding to the plaintext image through the convolution principle; A display module, used for displaying the mask pattern on the electrochromic glass; A recognition module is used to obtain a photograph of a mask pattern displayed on the electrochromic glass and input the mask pattern photograph into a neural network model for recognition to obtain a recognition result; the neural network model includes a YOLO model and a convolutional neural network connected in series; the recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification; The decryption module is used to convolve the recognition result with the convolution kernel to obtain the decrypted image.

[0014] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the encryption and decryption method of the photovoltaic cell cover glass information is implemented.

[0015] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the above-mentioned method for encrypting and decrypting photovoltaic cell cover glass information is implemented.

[0016] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects: the method includes: obtaining a plaintext image including photovoltaic cell cover glass information; generating a mask pattern and a convolution kernel corresponding to the plaintext image. The present invention uses the convolution principle to generate the mask pattern and the convolution kernel of the plaintext image. Convolution is a special mathematical operation that can expand the definition domain of the operation object so that the operation result contains more information. The present invention uses this feature to implicitly store more information of the plaintext image in the convolution kernel. Only a small amount of information needs to be presented to the outside to save almost all the information of the plaintext image, allowing the mask pattern to be kept at a smaller pixel scale, such as The method involves displaying a mask pattern on electrochromic glass, taking a photo of the mask pattern displayed on the electrochromic glass, and inputting the mask pattern photo into a neural network model for recognition, thereby obtaining a recognition result. The neural network model effectively improves recognition accuracy, and convolving the recognition result with a convolution kernel to obtain a decrypted image. This method can achieve the goal of not affecting the power generation efficiency of the photovoltaic cell during the information encryption and decryption process on the photovoltaic cell cover glass. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 A schematic flow chart of a method for encrypting and decrypting information of a photovoltaic cell cover glass provided by the present invention; Figure 2 The target plaintext image provided by the present invention; Figure 3 A schematic diagram of a mask pattern provided by the present invention; Figure 4 A schematic diagram of the convolution kernel matrix provided by the present invention; Figure 5 A schematic diagram of the wiring diagram provided by the present invention; Figure 6 A schematic diagram of the display results provided by the present invention; Figure 7 The target graph to be found by the neural network model provided by the present invention; Figure 8 The identification graph and decoding graph provided by the present invention; Figure 9 A flow chart of the method for encrypting information of photovoltaic cell cover glass provided by the present invention; Figure 10 A schematic diagram of the experimental results provided by the present invention; Figure 11 A schematic diagram of a photovoltaic cell cover glass information encryption and decryption device provided by the present invention; Figure 12 A schematic diagram of a computer device for implementing a method for encrypting and decrypting photovoltaic cell cover glass information provided by the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] Devices such as desktop computers, servers, and laptop computers that can execute the solution of the present invention are described below with the server as the execution subject for the sake of convenience.

[0021] The existing technologies in the photovoltaic industry have the following problems: Limitations of traditional identification technologies. Printed labels or laser engraving: Serial numbers, production dates, and other information can be printed directly on the glass or frame. However, long-term outdoor exposure can easily cause the labels to wear and fade, making it difficult to meet the 25-year or longer service life requirement. QR codes or barcodes: These require a large area (e.g., ≥21×21 pixels), affecting the glass's light transmittance and prone to scanning failure due to surface stains or reflections.

[0022] Balance between light transmittance and power generation efficiency. If the densified pattern covers the cell area, it will block incident light and reduce the module output power (every 1% loss in light transmittance results in approximately a 0.5% power drop, according to IEC 61730).

[0023] Environmental adaptability. Photovoltaic glass must withstand harsh conditions such as ultraviolet rays, high temperature and humidity (e.g., 85°C / 85% RH), and dust. Conventional encryption materials are prone to aging and failure. Relative humidity (RH).

[0024] Invisibility requirements: In scenarios such as building-integrated photovoltaics (BIPV), visible signage can disrupt the aesthetics of the building, necessitating invisible encryption solutions.

[0025] Current encryption technologies are limited. Electrochromic glass has limited applications: Currently, it's primarily used to adjust light transmittance (e.g., in smart windows) and has yet to be used for information encryption. While its voltage-driven color-changing properties enable dynamic displays, they lack integration with encryption algorithms.

[0026] The challenge of miniaturized encryption: Traditional image encryption (such as Fourier transform and chaotic encryption) requires a larger pixel scale to store information, making it difficult to achieve high-capacity data storage in micro arrays such as 5×5 pixels.

[0027] Insufficient decryption robustness: The surface of photovoltaic glass often has interference such as reflection, EVA film glare, and dust obstruction, and the general image recognition algorithm has a high misjudgment rate.

[0028] Industry standardization needs. The global distribution of photovoltaic modules requires compliance with international certifications (such as IEC 61215 and UL 1703). However, existing standards do not cover technical specifications for hidden information storage, resulting in a lack of unified solutions for anti-counterfeiting and traceability.

[0029] The above paragraphs explain the problems that need to be solved by the present invention.

[0030] This invention proposes a method for encrypting and decrypting information on photovoltaic cell cover glass. The core innovations include: The information is compressed to 5×5 pixels through convolution operation, covering only the non-light-receiving area at the edge of the glass, reducing the light transmission loss rate.

[0031] The technical solutions provided by various embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0032] Figure 1 The following is a flow chart of a method for encrypting and decrypting information for photovoltaic cell cover glass according to the present invention, which specifically includes the following steps: S101: Acquire a plaintext image including photovoltaic cell cover glass information.

[0033] The original image containing information such as serial number and production date is obtained as a plaintext image. The plaintext image is required to contain only one dimensional channel, such as a grayscale image. Figure 2 The plaintext image provided by the present invention is Figure 2 The plaintext image shown is Figure A, size Pixels.

[0034] S102: Generate a mask graphic and a convolution kernel corresponding to the plaintext image through the convolution principle.

[0035] In an exemplary embodiment, the mask pattern and convolution kernel corresponding to the plaintext image are generated by the convolution principle, specifically including: encoding the plaintext image into a pixel size of The plaintext image of The numerical matrix corresponding to the plaintext image; randomly generated The binary matrix of is used as the mask pattern matrix; the mask pattern is a visualization form of the mask pattern matrix; according to the numerical matrix and the mask pattern matrix, the convolution theorem is used to calculate and obtain the convolution kernel.

[0036] Specifically, a randomly generated binary matrix of "0 and 1" is used, with each matrix element being either 0 or 1. The upper left, upper right, and lower left corners of the binary matrix are forced to be 1 (corresponding to white), and the lower right corner is forced to be 0 (corresponding to black) to facilitate the positioning of the graphics.

[0037] In one embodiment of the present invention, the present invention provides Figure 3 The mask graphics shown, such as Figure 3 As shown, the mask graphic size is Pixels.

[0038] In an exemplary embodiment, the convolution kernel is obtained by performing calculations based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: , mask pattern matrix And the mathematical expression formula of the convolution theorem, we get the convolution kernel matrix ; The mathematical expression of the convolution theorem is shown in formula (1): (1); in, represents the Fourier transform, is the mask pattern matrix, is a numerical matrix.

[0039] Convolution kernel matrix Perform inverse Fourier transform and obtain the convolution kernel as shown in formula (2): (2); in, represents the inverse Fourier transform.

[0040] In engineering practice, the Fourier transform of the mask matrix often has a condition number that is too large, indicating that it is close to being singular, resulting in instability of the convolution kernel matrix of formula (2). The present invention uses a regularization method to optimize this process.

[0041] In an exemplary embodiment, the convolution kernel is obtained by performing calculations based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: M 、 and The objective function is constructed by Fourier transform, and the objective function is shown in formula (3): (3); in, represents the regularization parameter, the operator Frobenius 2 Norm (F 2 norm).

[0042] Based on numerical matrix , mask pattern matrix Solve the objective function to obtain the convolution kernel matrix The solution is shown in formula (4): (4); in, represent The Hermitian transpose of .

[0043] In an exemplary embodiment, the regularization parameter Confirmation, specifically including: generating a set of candidate regularization parameters in the logarithmic space list; for each Solve the frequency domain regularization equation , and obtain the restored image through inverse Fourier transform; calculate the structural similarity index between the restored image and the plaintext image, and take the corresponding image when the structural similarity index value is the largest , is determined as the regularization parameter.

[0044] Specifically, the logarithmic space range is set according to specific engineering practices, for example, In one embodiment of the present invention, there is provided Figure 4 The convolution kernel matrix shown.

[0045] S103: Displaying the mask pattern on the electrochromic glass.

[0046] In an exemplary embodiment, the mask pattern is displayed through the electrochromic glass, specifically comprising: designing a circuit connecting the electrochromic glass and the microcontroller, and wiring them in sequence; The matrix is ​​stretched to The method further comprises the steps of: writing a row vector list of a pixel size and writing the row vector list into a program of a microcontroller, and running the program of the microcontroller to display a mask pattern on the electrochromic glass.

[0047] Specifically, design the circuit connecting the electrochromic glass and the microcontroller, and connect them in sequence. The wiring diagram is as shown in the reference Figure 5 As shown. Figure 5 Figure (a) shows the planar design of the electrochromic glass screen, which includes the design dimensions and top view appearance. Figure 5 In Figure (b), The squares represent the color-changing glass matrix on the screen, and the characters in the table represent their connections to the corresponding pins on the microcontroller. Figure 5 Direction and design of Figure (b) Figure 1 To.

[0048] Specifically, electrochromic glass is used to display the mask pattern. When a voltage of 3-3.5V is applied to the electrochromic glass in the forward direction, the glass turns blue, and when a voltage of 2-2.5V is applied to the glass in the reverse direction, the glass fades and becomes transparent.

[0049] Specifically, the convolution kernel is stored on the recognition device or in the cloud.

[0050] In one embodiment of the present invention, there is provided Figure 6 The result diagram is shown.

[0051] S104: Obtain a photograph of a mask pattern displayed on the electrochromic glass, input the mask pattern photograph into a neural network model for recognition, and obtain a recognition result; the neural network model includes a YOLO model and a convolutional neural network connected in series; the recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification.

[0052] In an exemplary embodiment, the training process of the neural network model specifically includes: obtaining a set of photos of people holding electrochromic glass as a first training set; obtaining a portion of the output of the YOLO model, and intercepting and segmenting the portion of the output as a second training set; training the YOLO model with the first training set to obtain a trained YOLO model; training the convolutional neural network with the second training set to obtain a trained convolutional neural network; and determining the trained YOLO model and the trained convolutional neural network as a neural network model.

[0053] In one exemplary embodiment, a mask pattern photo is input into a neural network model for recognition, thereby obtaining a recognition result. The method specifically includes: identifying the image information of the electrochromic glass in the mask pattern photo using a YOLO model; converting the identified electrochromic glass image information into a mask pattern binary matrix; identifying the color of each square in the mask pattern binary matrix using a convolutional neural network; and confirming the color-recognized mask pattern binary matrix as the recognition result. Color recognition is used to quantify the distribution of tinted areas in the electrochromic glass.

[0054] Specifically, the present invention provides Figure 7 The neural network model shown identifies the target graph to be sought. Figure 7 The numbers in the box represent the confidence level, which represents the possibility that the object in the box is the target discoloration glass screen.

[0055] Specifically, on the recognition device, the mask pattern photo is input into the neural network model for recognition to obtain the recognition result. The present invention provides the following Figure 8 The identification diagram and decoding diagram shown, Figure 8 Figure (a) is the binary matrix recognized by the neural network model (after calibration).

[0056] S105: Convolve the recognition result with the convolution kernel to obtain a decrypted image.

[0057] Specifically, the recognition result is convolved with the convolution kernel on the recognition device to obtain the decrypted image. The decrypted image is as follows: Figure 8 As shown in Figure (b).

[0058] In an exemplary embodiment, the present invention provides Figure 9 The flowchart of the method for encrypting information of photovoltaic cell cover glass is shown as follows: Figure 9 As shown, Figure 9The target image in the present invention is the plaintext image mentioned in the present invention. First, the target image is processed to obtain a mask pattern and a convolution kernel corresponding to the target image. Second, the mask pattern is displayed on the electrochromic glass by a microcontroller, and the convolution kernel is stored in the microcontroller. Then, a photo of the electrochromic glass is taken by a camera, and the photo is input into the neural network to obtain a recognition result. Finally, the convolution kernel and the recognition result are convolved and decrypted to obtain a decrypted image.

[0059] The present invention proposes a method for encrypting and decrypting information on photovoltaic cell cover glass. Its confidentiality relies on a "password lock" (convolution kernel) rather than a "key" (mask pattern) itself, which makes the entire confidentiality system highly robust to the mask pattern. This allows it to overcome mask recognition errors caused by stains on the surface of electrochromic glass, strong reflections, model errors, etc. in real situations, and still ensures robust decryption. To test its robustness, two indicators, the number of pixels change rate (NPCR) and the structural similarity index (SSIM), are used to measure the impact of pixel changes on the similarity between the image and the reference original image. NPCR refers to the ratio of the number of pixels that undergo transformation to the total pixel quality of the image. Assuming that represents the image pixels obtained after decryption, represents the image pixels of the reference image and the plaintext image, then the SSIM indicators of the two are composed of three components, namely brightness, contrast, and structure. The calculation method of the structural similarity index is shown in formula (5):

[0060] (5); in, SSIM is the structural similarity index, For images x and images y The brightness similarity of For images x and images y The contrast similarity of For images x and images y The structural similarity of , , are all weight parameters.

[0061] When the default value When , SSIM is as shown in formula (6): (6); in, and Respectively and The mean of Representing an image Pixel values ​​and images The covariance between pixel values, and Represent two constants related to image size, For images The variance of pixel values, For images The variance of pixel values.

[0062] The SSIM value range is 0 to 1. When SSIM is equal to 1, it represents the original image. When SSIM is closer to 1, it means that the decrypted image is closer to the plaintext image.

[0063] All pixels of the mask pattern except the corner positioning pixels are randomly inverted, and the NPCR and SSIM values ​​of the mask pattern are calculated. The results are shown in Table 1 and Figure 10 shown.

[0064] Table 1 From Table 1, we can see that even if the mask pattern has 32% error, the SSIM of the decrypted pattern and the original ciphertext remains around 0.6. Figure 10 It can be seen that the information is still recognizable to a certain extent. This shows that the method for encrypting information on a photovoltaic cell cover glass of the present invention is robust against interference from mask patterns, making it effective even in complex real-world environments.

[0065] This method encodes production information, anti-counterfeiting identification, and other data into a micro-mask pattern (e.g., 5×5 pixels) and embeds it into the cover glass surface. Combining convolutional kernel encryption with neural network decryption technology, this method achieves covert storage and high-precision restoration of this information without compromising the transmittance and durability of photovoltaic modules. This method is particularly suitable for traceability management, anti-counterfeiting authentication, and covert data exchange in BIPV within the photovoltaic industry chain.

[0066] When applying the method for encrypting and decrypting information of photovoltaic cell cover glass provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0067] The above is a method for encrypting and decrypting photovoltaic cell cover glass information provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding device for encrypting and decrypting photovoltaic cell cover glass information, such as Figure 11 shown.

[0068] Figure 11A schematic diagram of a photovoltaic cell cover glass information encryption and decryption device provided by the present invention, comprising: The acquisition module 1101 is used to acquire a plain text image including information of the photovoltaic cell cover glass.

[0069] The generating module 1102 is used to generate a mask pattern and a convolution kernel corresponding to the plaintext image through the convolution principle.

[0070] The display module 1103 is used to display the mask pattern on the electrochromic glass.

[0071] Recognition module 1104 is used to obtain a photograph of the mask pattern displayed on the electrochromic glass, input the mask pattern photograph into a neural network model for recognition, and obtain a recognition result; the neural network model includes a YOLO model and a convolutional neural network connected in series; the recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification.

[0072] The decryption module 1105 is used to convolve the recognition result with the convolution kernel to obtain a decrypted image.

[0073] The specific definitions of the photovoltaic cell cover glass information encryption and decryption device can be found in the definitions of the photovoltaic cell cover glass information encryption and decryption method described above and will not be repeated here. The various modules in the aforementioned photovoltaic cell cover glass information encryption and decryption device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the aforementioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the computer device's memory in software form, so that the processor can call and execute the corresponding operations of each of the aforementioned modules.

[0074] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 Provided is a method for encrypting and decrypting photovoltaic cell cover glass information.

[0075] The present invention also provides Figure 12 The structural diagram of the computer equipment shown in FIG. Figure 12 As shown in the figure, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 Provided is a method for encrypting and decrypting photovoltaic cell cover glass information.

[0076] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes in the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0077] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A method for encrypting and decrypting information for photovoltaic cell cover glass, characterized in that: include: Acquire a plaintext image including photovoltaic cell cover glass information; Generate a mask graphic and a convolution kernel corresponding to the plaintext image through the convolution principle; displaying the mask pattern on the electrochromic glass; Obtaining a photograph of a mask pattern displayed on the electrochromic glass, inputting the mask pattern photograph into a neural network model for recognition to obtain a recognition result; the neural network model includes a YOLO model and a convolutional neural network connected in series; the recognition result is binary matrix data generated based on YOLO model positioning and convolutional neural network color classification; The recognition result is convolved with the convolution kernel to obtain a decrypted image.

2. The method according to claim 1, wherein The step of generating a mask graphic and a convolution kernel corresponding to the plaintext image by the convolution principle specifically includes: Encode the plaintext image into a pixel size of Plaintext image of Get the pixel size The numerical matrix corresponding to the plaintext image; Random Generation The binary matrix of is used as the mask pattern matrix; the mask pattern is the visualization form of the mask pattern matrix; The convolution kernel is obtained by performing calculation based on the numerical matrix and the mask pattern matrix using the convolution theorem.

3. The method according to claim 2, wherein The convolution kernel is obtained by performing calculation based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: According to the numerical matrix , the mask pattern matrix And the mathematical expression formula of the convolution theorem, we get the convolution kernel matrix ; The mathematical expression of the convolution theorem is: ； in, represents the Fourier transform, is the mask pattern matrix, is a numerical matrix; For the convolution kernel matrix Perform inverse Fourier transform and get the convolution kernel as: ； in, represents the inverse Fourier transform.

4. The method according to claim 3, wherein The convolution kernel is obtained by performing calculation based on the numerical matrix and the mask pattern matrix using the convolution theorem, specifically including: based on M 、 and The objective function is constructed by Fourier transform, which is: ； in, represents the regularization parameter, the operator Frobenius 2 norm; Based on the numerical matrix , the mask pattern matrix Solve the objective function to obtain the convolution kernel matrix The solution is: ； in, Representation matrix The Hermitian transpose of Convolution kernel matrix Perform inverse Fourier transform to obtain the convolution kernel.

5. The method according to claim 4, wherein The regularization parameter The determination includes: Generate a set of candidate regularization parameters in the logarithmic space List of; For each in the list Solve the frequency domain regularization equation , and obtain the restored image through inverse Fourier transform; Calculate the structural similarity index between the restored image and the plaintext image, and take the corresponding image when the structural similarity index value is the largest , is determined as the regularization parameter.

6. The method according to claim 1, wherein Displaying the mask pattern on the electrochromic glass specifically includes: Design the circuit connecting the electrochromic glass and the microcontroller and connect them in sequence; The mask shape The matrix is ​​stretched to The method comprises the steps of: generating a row vector list of a size, writing the row vector list into a program of a microcontroller, and running the program of the microcontroller to display a mask pattern on the electrochromic glass.

7. The method according to claim 1, wherein Inputting the mask pattern photo into a neural network model for recognition to obtain a recognition result specifically includes: Recognizing image information of the electrochromic glass in the mask pattern photo by the YOLO model; Converting the recognized image information of the electrochromic glass into a mask pattern binary matrix; Identify the color of each square in the binary matrix of the mask graphic through the convolutional neural network; The binary matrix of the mask pattern after color recognition is determined as the recognition result.

8. The method according to claim 7, wherein The training process of the neural network model specifically includes: Obtain a set of photos of people holding electrochromic glass as the first training set; Obtain part of the output of the YOLO model and intercept and split the part of the output as the second training set; Training the YOLO model using the first training set to obtain a trained YOLO model; Training the convolutional neural network using the second training set to obtain a trained convolutional neural network; The trained YOLO model and the trained convolutional neural network are determined as the neural network model.

Images