A unique identification management system for food production

Abstract

The present invention discloses a unique identification management system for food production, belonging to the field of food management. Through a technical closed-loop of artistic design, multi-source data fusion, multi-level anti-counterfeiting collaboration, and blockchain trusted deposit, it constructs an innovative solution with high security, high recognition, and high credibility in the field of food identification, providing strong technical support for brand protection, anti-counterfeiting traceability, and consumer trust establishment in the food industry.

Description

Technical Field

[0001] The present invention relates to the field of food management, and more specifically, to a unique identification management system for food production. Background Art

[0002] In recent years, with the increasing demand of consumers for food safety and traceability transparency, the food industry has gradually deepened its dependence on product unique identification and anti-counterfeiting traceability technologies.

[0003] In the prior art, food identification systems mostly implement data storage and traceability based on traditional barcodes, two-dimensional codes or radio frequency identification (RFID) tags, but their functions are concentrated on basic information recording, with significant limitations. First, the existing identifications lack the ability of artistic design and it is difficult to dynamically integrate elements such as brand culture and regional characteristics into the identifications, resulting in insufficient visual recognition of products and unable to meet the aesthetic value requirements of the high-end market; second, in terms of technology integration, although blockchain technology has been tried for food traceability, its application mostly stays in the data storage stage and it is not very convenient to form a deep collaboration with dynamic identification generation, cross-modal data conversion and multiple anti-counterfeiting mechanisms; at the same time, the existing art design tools rely on manual operations, with low parameterization and automation levels, and it is difficult to efficiently process the quantitative coding of complex graphics, colors and textures. Although artificial intelligence technologies such as generative adversarial networks (GAN) have made breakthroughs in the field of image generation, there is still a lack of mature solutions for cross-modal mapping with production data in industrial scenarios, which limits the personalization and non-replicability of unique identifications.

[0004] In summary, the current food identification system has significant shortcomings in terms of artistry, dynamic generation ability and technology integration. There is an urgent need for a new management system that integrates parametric art design, multi-source data-driven, physical-digital collaborative anti-counterfeiting and blockchain trusted evidence storage to break through the bottleneck of the prior art and achieve the uniqueness, security and traceability of food identifications throughout the entire production-to-consumption link. Summary of the Invention

[0005] 1. Technical Problems to be Solved

[0006] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a unique identification management system for food production, which constructs an innovative solution with high security, high recognition and high credibility in the field of food identification through a technical closed-loop of artistic design, multi-source data fusion, multi-level anti-counterfeiting collaboration and blockchain trusted evidence storage, and provides strong technical support for brand protection, anti-counterfeiting traceability and consumer trust establishment in the food industry.

[0007] 2. Technical Solutions

[0008] To solve the above problems, the present invention adopts the following technical solutions.

[0009] A unique identification management system for food production, comprising:

[0010] An artistic feature parameterization module configured to perform quantifiable encoding on graphics, colors, and textures in a preset artistic element library, output a parameterized data set including a set of vector control points, a color space distribution matrix, and a texture feature vector, and the artistic element library supports importing custom vector graphics through a user interface, automatically performing compliance checks and parameterization conversions, and invoking a regional cultural feature template library according to geographical indication certification requirements;

[0011] A production data acquisition module connected to the sensor network of the food production line to obtain in real time a structured data stream including production timestamps, raw material batch codes, and process control parameters;

[0012] A cross-modal generation engine that receives the parameterized data set and the structured data stream, encodes the production data into artistic style transfer parameters through a pre-trained generative adversarial network model, and fuses them into the parameterized data set to generate a dynamic visual identification containing data fingerprint information;

[0013] A multi-level anti-counterfeiting processing module that performs multi-level anti-counterfeiting processing on the dynamic visual identification, including random microstructure embedding, optical variable layer superposition, and cryptographic hash value calculation;

[0014] A blockchain evidence storage module that stores the feature hash value of the identification and the associated production data in blockchain nodes to establish an immutable traceability chain.

[0015] Further, the artistic feature parameterization module includes a graphics parsing unit, a color quantization unit, and a texture analysis unit. The graphics parsing unit converts geometric patterns into editable vector data through Bezier curve control point extraction and polygon vertex coordinate serialization. The color quantization unit performs three-dimensional discretization encoding of hue-saturation-brightness on the color combination in the artistic elements using the HSV color model. The texture analysis unit calculates the directionality, contrast, and roughness indexes of the texture based on the gray-level co-occurrence matrix to generate a texture feature vector.

[0016] Further, the cross-modal generation engine includes a data-art mapping unit, a style transfer unit, and a dynamic rendering unit. The data-art mapping unit maps the production timestamp, raw material batch code, and process control parameters into the fractal iteration times, polar coordinate basis function parameters, and spiral phase angles respectively through a convolutional neural network. The style transfer unit uses the generator model of the generative adversarial network to fuse the fractal iteration times, polar coordinate basis function parameters, and spiral phase angles into a parameterized data set to generate a unique fractal art pattern. The dynamic rendering unit performs color filling, texture overlay, and dynamic light and shadow rendering on the fractal art pattern according to the preset brand visual specifications, and outputs a vector graphic file of the final logo.

[0017] Further, the data-art mapping unit first converts the raw material batch code into a radius increment sequence in the polar coordinate system through modulo operation, then maps the process control parameters to the iteration termination threshold of the Mandelbrot fractal set after normalization, and generates the initial rotation angle of the spiral line through the millisecond-level value of the production timestamp.

[0018] Further, the multiple anti-counterfeiting processing module includes a physical anti-counterfeiting unit, an optical anti-counterfeiting unit, and a digital anti-counterfeiting unit. The physical anti-counterfeiting unit embeds a microstructure layer containing randomly distributed nano-pits in the logo. The optical anti-counterfeiting unit superimposes an interference film color layer with angle-sensitive characteristics. The digital anti-counterfeiting unit performs an SM hash operation on the vector graphic file of the logo to generate a digital fingerprint, and encrypts the fingerprint and writes it into the invisible coding area at the edge of the logo.

[0019] A method for generating a unique identifier for food production includes the following steps:

[0020] Step 1: Receive a structured data stream and convert it into a set of control parameters through preset rules;

[0021] Step 2: Invoke a pre-trained generative adversarial network model to perform cross-modal fusion on the set of control parameters and the art feature parameterized data set;

[0022] Step 3: Generate a dynamic logo graphic containing fractal geometric features, data fingerprints, and brand visual elements;

[0023] Step 4: Apply physical and digital anti-counterfeiting treatments to the dynamic logo graphic;

[0024] Step 5: Write the logo feature values and associated data into the blockchain deposit contract;

[0025] Step 6: Connect to the food label verification terminal. Use the optical acquisition unit inside the food label verification terminal to capture the physical anti-counterfeiting features of the label through multispectral imaging, and connect to the blockchain node for comparing the matching degree between the on-site collected features and the data stored on the chain;

[0026] Step 7: Use the interactive display unit inside the food label verification terminal to superimpose the production traceability information on the real-time image of the physical label through augmented reality technology. Then, use the anti-counterfeiting verification unit inside the food label verification terminal to detect the random distribution characteristics of the microstructure layer and the angular response curve of the optically variable layer.

[0027] Further, the cross-modal fusion step specifically includes:

[0028] Step 1: Convert the production data stream into a frequency-domain feature vector through Fourier transform;

[0029] Step 2: Use the self-attention mechanism to calculate the correlation weights between the frequency-domain feature vector and the art feature parameterized dataset;

[0030] Step 3: Dynamically weight the art feature parameters according to the correlation weights to generate a fused stylized parameter matrix.

[0031] 3. Beneficial effects

[0032] Compared with the prior art, the advantages of the present invention are as follows:

[0033] (1) In this solution, the art feature parameterization module performs quantifiable encoding on graphics, colors, and textures, and calls the regional cultural feature template library in combination with the geographical indication certification requirements to achieve the dynamic integration of brand cultural elements and regional characteristics. Based on the generative adversarial network (GAN) model of the cross-modal generation engine, production data (such as timestamps, raw material batch codes) is mapped into art style features such as fractal iteration parameters and polar coordinate basis functions to generate a dynamic visual label with both aesthetic value and data uniqueness, solving the problems of traditional labels with single vision and low recognition rate and meeting the needs of the high-end market for artistic labels;

[0034] (2) In this solution, the production data acquisition module obtains the structured data stream of the production line in real time. The cross-modal generation engine dynamically fuses the data stream with art features through technologies such as Fourier transform and self-attention mechanism to generate a unique label containing a data fingerprint. The blockchain evidence storage module binds and stores the label feature hash value with the production data to form an immutable traceability chain, solving the problems of the disconnection between traditional labels and production links and low credibility of traceability information, and realizing the full-link transparent traceability from raw materials to finished products;

[0035] (3) In this solution, the system realizes the automatic coding of artistic elements (such as the extraction of Bezier curve control points and the quantization of the HSV color model) and the real-time generation of dynamic logos through the deep integration of parametric design tools and artificial intelligence models, significantly reducing the manual operation cost. At the same time, technologies such as analog-to-digital conversion and fractal iteration parameter mapping in the data-art mapping unit of the cross-modal generation engine ensure the accurate conversion from production data to artistic features, improving the uniqueness and efficiency of logo generation.

[0036] (4) In this solution, through the multi-spectral imaging and augmented reality technology (AR) of the food label verification terminal, users can capture physical anti-counterfeiting features in real time (such as the distribution of nano-pits and the angular response of interference films), and overlay production traceability information on the physical label image. Combined with the data comparison of blockchain nodes, a "physical-digital" dual-channel verification mechanism is formed, significantly enhancing consumers' trust in product authenticity and traceability information. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a schematic diagram of the system architecture of the present invention;

[0038] Figure 2 is a schematic diagram of the system components and functions of the present invention;

[0039] Figure 3 is a schematic diagram of the steps of the method for generating a unique food production label of the present invention;

[0040] Figure 4 is a mind map of the method for generating a unique food production label of the present invention.

[0041] Description of the reference numerals in the drawings:

[0042] 1. Artistic feature parameterization module; 101. Graphic analysis unit; 102. Color quantization unit; 103. Texture analysis unit;

[0043] 2. Production data collection module;

[0044] 3. Cross-modal generation engine; 301. Data-art mapping unit; 302. Style transfer unit; 303. Dynamic rendering unit;

[0045] 4. Multiple anti-counterfeiting processing module; 401. Physical anti-counterfeiting unit; 402. Optical anti-counterfeiting unit; 403. Digital anti-counterfeiting unit;

[0046] 5. Blockchain evidence storage module. DETAILED DESCRIPTION OF THE INVENTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings of the present invention; obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0048] Embodiment:

[0049] "Xingyao Food Group" is a large comprehensive food enterprise, and its products cover multiple categories such as leisure snacks, seasonings, and frozen foods. The sales network spreads across major supermarkets and online platforms across the country. With the increasingly fierce market competition in the food industry and consumers' growing concern about food safety and quality, the enterprise faces severe challenges. Therefore, "Xingyao Food Group" decides to introduce this unique food production identification management system. The following combines the attached Figures 1 - 4 drawings of the specification to

[0050] I. System implementation steps

[0051] 1. Artistic feature parameterization module (1):

[0052] Graphic analysis unit (101): The brand logo of "Xingyao Food Group" is composed of a unique star pattern and the brand name. The graphic analysis unit uses the Bezier curve control point extraction algorithm to accurately identify the contour control points of the star pattern, and at the same time, through the serialization of polygon vertex coordinates, converts the text contour of the brand name into editable vector data. In order to highlight the regional cultural characteristics of the products, the enterprise imports the vector graphics of auspicious patterns in the local traditional paper-cut art style. The system comprehensively checks the resolution, color mode, file format, size ratio, element composition, etc. of the graphics according to the pre-set strict compliance standards, and completes the parameterization conversion after ensuring compliance. In addition, according to the requirements of geographical indication certification, the system calls the regional cultural feature template library containing local characteristic buildings, folk customs, and natural landscapes, selects representative elements, such as the unique carved patterns of local traditional buildings, and integrates them into the art element library.

[0053] Color quantization unit (102): The product packaging of the enterprise mainly uses bright red and yellow as the main colors, conveying a brand image of vitality and deliciousness. The color quantization unit uses the HSV color model to accurately quantify the main red color tone as hue 30°, saturation 90%, and lightness 85%, and the yellow color is quantified as hue 60°, saturation 80%, and lightness 95%, and records them in detail in the color space distribution matrix. For the auxiliary colors on the packaging, such as green and blue used to distinguish product series, three-dimensional discrete coding is also carried out in detail.

[0054] Texture Analysis Unit (103): On the packaging of casual snacks, textures such as imitation potato chip texture and candy granule texture are often used to enhance the texture. Based on the gray-level co-occurrence matrix, for the imitation potato chip texture, the texture analysis unit accurately calculates its directional angle as 30°, contrast as 0.7, and roughness as 0.5, and generates the corresponding texture feature vector; for the candy granule texture, it calculates the indicators of random directionality, high contrast, and moderate roughness, and generates an accurate texture feature vector.

[0055] 2. Production Data Acquisition Module (2):

[0056] In the production workshop of "Starlight Food Group", a comprehensive and high-precision sensor network is deployed. The time sensor uses atomic clock calibration technology to accurately record the production timestamp of each batch of products, accurate to milliseconds; the raw material batch sensor obtains the raw material batch code by scanning the QR code on the raw material packaging or reading the RFID tag, ensuring that the source of each batch of raw materials is traceable; the process parameter sensor uses high-precision temperature sensors, pressure sensors, flow sensors, etc. to monitor key process control parameters such as frying temperature, seasoning ratio, and packaging weight in real time. For example, when producing a batch of potato chips, the sensor records the production timestamp as 15:45:10.300 milliseconds on May 20, 2023, the raw material batch codes are "POTATO20250701008" (representing a specific batch of potatoes), "OIL20250701003" (representing a specific batch of cooking oil), etc., and the process control parameters are frying temperature 180°C, seasoning ratio salt 0.5%, spices 0.3%, and packaging weight 100 grams ± 2 grams.

[0057] 3. Cross-modal Generation Engine (3):

[0058] Data-Art Mapping Unit (301): Uniquely maps the production timestamp, raw material batch code, and process control parameters. Taking the production timestamp as an example, the last three digits of its millisecond-level value are converted into the initial rotation angle of a spiral through a specific formula. Assuming the last three digits are 250, the initial rotation angle of 25° is generated; for the raw material batch code, it is converted into a radius increment sequence in the polar coordinate system through complex modulo operations and coding mapping algorithms. For example, for the raw material batch code "POTATO20250701008", a set of radius increment values [0.15, 0.25, 0.1] is obtained through calculation; the frying temperature in the process control parameters is mapped to the iteration termination threshold of the Mandelbrot fractal set after normalization. If the frying temperature is 180°C, the iteration termination threshold of 0.85 is obtained after normalization.

[0059] Style transfer unit (302): Using the generator model of the generative adversarial network, it fuses the fractal iteration times, polar coordinate basis function parameters, spiral phase angles, etc. with the artistic feature parameterized dataset. For example, it combines the parameters obtained from the above mapping with the parameterized data such as brand logos, colors, textures, etc. to generate a logo with a unique fractal art pattern, which not only contains the information of the production data but also has a unique artistic style.

[0060] Dynamic rendering unit (303): According to the brand visual specifications preset by the enterprise, it carefully renders the fractal art pattern. First, it fills in the main brand colors red and yellow, then overlays a texture imitating potato chips to make it more in line with the product characteristics, and finally adds a dynamic light and shadow effect to simulate the luster of potato chips under the light, and outputs the vector graphic file of the final logo.

[0061] 4. Multiple anti-counterfeiting processing module (4):

[0062] Physical anti-counterfeiting unit (401): During the logo production process, through advanced nanoimprinting and photolithography technologies, an anti-counterfeiting layer containing randomly distributed nano pits and microstructured lines is embedded on the logo surface. The diameter of these nano pits is between 30 - 80 nanometers, the distribution density is 1200 - 1500 per square millimeter, the width of the microstructured lines is between 50 - 100 nanometers, and the distribution is completely random, making it difficult to replicate.

[0063] Optical anti-counterfeiting unit (402): Using multi-layer vacuum coating and interference technologies, an interference film color layer with angle-sensitive characteristics and a diffractive optical element are superimposed on the logo surface. When observing the logo from different angles, the interference film will show a color gradient from red to yellow, and the diffractive optical element will produce unique pattern changes, and the angle response curves of the color changes and pattern changes are precisely designed and unique.

[0064] Digital anti-counterfeiting unit (403): It performs SM3 hashing operation on the vector graphic file of the logo to generate a digital fingerprint, obtaining a digital fingerprint with a length of 256 bits. Then it uses the AES encryption algorithm to encrypt it, and writes the encrypted fingerprint into the invisible coding area on the edge of the logo. This invisible coding area uses special fluorescent inks and microprinting technologies and can only be revealed under ultraviolet light of a specific wavelength and under a high-power microscope.

[0065] 5. Blockchain evidence storage module (5):

[0066] Store the identified feature hash values together with associated production data, such as production timestamps, raw material batch codes, process control parameters, etc., in blockchain nodes through a dedicated blockchain interface. The blockchain adopts the form of a consortium blockchain, and the nodes include the production department, quality inspection department, logistics department within the enterprise, as well as cooperating third-party regulatory agencies, financial institutions, and industry associations. Each node stores the complete traceability chain information to ensure the immutability and traceability of the data.

[0067] II. Implementation Steps of the Generation Method

[0068] Step 1:

[0069] Receive the production data stream from the sensor network in the production workshop and perform preliminary cleaning and sorting on the data. Convert the production timestamp to a unified time format, such as "YYYY-MM-DD HH:MM:SS.fff"; parse the raw material batch code according to a preset rule to extract key information such as raw material type, batch number, and supplier; perform normalization processing on the process control parameters to make their value range between 0 and 1 for convenient subsequent processing.

[0070] Step 2:

[0071] Call the pre-trained style transfer model and convert the production data stream into a frequency domain feature vector through Fourier transform. Then use the self-attention mechanism to calculate the correlation weights between the frequency domain feature vector and the artistic feature parameterized dataset. For example, for the frequency domain feature vector corresponding to the production timestamp, the correlation weight with the color feature vector is calculated as 0.35, and the correlation weight with the texture feature vector is calculated as 0.25, etc. Finally, dynamically weight the artistic feature parameters according to the correlation weights to generate a fused stylized parameter matrix.

[0072] Step 3:

[0073] Generate a dynamic logo graphic containing fractal geometric features, data fingerprints, and brand visual elements according to the fused stylized parameter matrix. Use a professional graphics rendering engine to convert the fractal iteration times, polar coordinate basis function parameters, etc. into specific graphic shapes and lines, and combine the brand's color and texture elements to generate a logo graphic with a unique visual effect.

[0074] Step 4:

[0075] Apply physical and digital anti-counterfeiting treatments to the dynamic logo graphic in sequence. First, embed a nano-pitted and micro-structured line anti-counterfeiting layer on the graphic surface, then overlay an interference film color layer and a diffractive optical element, and finally generate a digital fingerprint and encrypt it into the invisible coding area.

[0076] Step 5:

[0077] Write the identification feature values and associated data into the blockchain deposit contract according to the format requirements of the blockchain. The contract clearly stipulates the data storage format, access rights, update mechanism, data encryption method, etc., to ensure the secure storage and effective management of data.

[0078] Step Six:

[0079] In the sales terminal and consumer verification scenarios, connect to the food identification verification terminal. Use the optical acquisition unit inside the verification terminal to capture the physical anti-counterfeiting features of the identification through high-resolution multispectral imaging and microstructure detection technologies, such as the distribution of nano-pits, color changes of interference films, patterns of diffractive optical elements, etc. Then transmit the captured feature data to the blockchain node through a secure network connection and compare it with the data stored on the chain to verify the authenticity of the identification.

[0080] Step Seven:

[0081] Use the interactive display unit inside the food identification verification terminal to superimpose the production traceability information onto the real-time image of the physical identification through augmented reality technology. When consumers use the verification terminal to scan the product identification, information such as the production process, raw material sources, quality inspection reports, and nutritional components of the product will be displayed in the form of a 3D model on the screen. At the same time, use the anti-counterfeiting verification unit to detect the random distribution characteristics of the microstructure layer and the angular response curve of the optically variable layer to further verify the anti-counterfeiting performance of the identification.

[0082] By implementing the food production unique identification management system, "Xingyao Food Group" has successfully solved the problems of product anti-counterfeiting and quality traceability. Through the artistic feature parameterization module, the graphics, colors, and textures are quantitatively encoded, and combined with the geographical indication certification requirements, the regional culture feature template library is called to achieve the dynamic integration of brand cultural elements and regional characteristics, significantly enhancing the brand image and market competitiveness. Consumers can easily obtain detailed product information through the verification terminal and be more confident about product quality. Enterprises can also quickly respond to market feedback, optimize the production management process, and improve production efficiency and product quality.

[0083] The above is only the preferred specific implementation manner of the present invention; however, the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its improved concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A food production unique identification management system, characterized in that: include: An art feature parameterization module (1) is configured to perform quantifiable encoding on graphics, colors and textures in a preset art element library, and output a parameterized data set including a vector control point set, a color space distribution matrix and a texture feature vector, wherein the art element library supports importing custom vector graphics through a user interface, and automatically performs compliance checks and parameterization conversions, and calls a regional cultural feature template library according to geographical indication certification requirements; A production data acquisition module (2), connected to a sensor network of a food production line, acquires in real time a structured data stream including production timestamps, raw material batch codes, and process control parameters; A cross-modal generation engine (3) receives the parameterized data set and the structured data stream, encodes the production data into artistic style transfer parameters through a pre-trained generative adversarial network model, and fuses the parameters into the parameterized data set to generate a dynamic visual logo containing data fingerprint information; A multiple anti-counterfeiting processing module (4) performs a multi-level anti-counterfeiting processing including random microstructure embedding, optically variable layer superposition and cryptographic hash value calculation on the dynamic visual mark; The blockchain evidence storage module (5) stores the characteristic hash value of the identifier and the associated production data in the blockchain node to establish an unalterable traceability chain; The cross-modal generation engine (3) comprises a data-art mapping unit (301), a style transfer unit (302) and a dynamic rendering unit (303). The data-art mapping unit (301) maps the production timestamp, the raw material batch code and the process control parameters into the number of fractal iterations, the polar coordinate basis function parameters and the spiral phase angle respectively through a convolutional neural network. The style transfer unit (302) uses a generator model of a generative adversarial network to fuse the number of fractal iterations, the polar coordinate basis function parameters and the spiral phase angle into a parameterized data set to generate a unique fractal art pattern. The dynamic rendering unit (303) performs color filling, texture overlay and dynamic light and shadow rendering on the fractal art pattern according to a preset brand visual specification, and outputs a vector graphics file of a final logo.

2. A food production unique identification management system according to claim 1, characterized in that: The art feature parameterization module (1) comprises a graphics analysis unit (101), a color quantization unit (102) and a texture analysis unit (103). The graphics analysis unit (101) converts a geometric pattern into editable vector data by extracting Bezier curve control points and serializing polygon vertex coordinates. The color quantization unit (102) uses the HSV color model to perform hue-saturation-lightness three-dimensional discrete encoding on the color combination in the art element. The texture analysis unit (103) calculates the directionality, contrast and roughness index of the texture based on the gray level co-occurrence matrix to generate a texture feature vector.

3. A food production unique identification management system according to claim 1, characterized in that: The data-art mapping unit (301) first converts the raw material batch code into a radius increment sequence in a polar coordinate system through a modulus operation, then maps the process control parameter into an iteration termination threshold of the Mandelbrot fractal set after normalization, and then generates an initial rotation angle of the spiral line through the millisecond value of the production timestamp.

4. A food production unique identification management system according to claim 1, characterized in that: The multiple anti-counterfeiting processing module (4) comprises a physical anti-counterfeiting unit (401), an optical anti-counterfeiting unit (402) and a digital anti-counterfeiting unit (403); the physical anti-counterfeiting unit (401) embeds a microstructure layer containing randomly distributed nano-pits in the logo; the optical anti-counterfeiting unit (402) superimposes an interference film color layer with angle-sensitive characteristics; and the digital anti-counterfeiting unit (403) performs an SM3 hash operation on a vector graphics file of the logo to generate a digital fingerprint, and encrypts the fingerprint and writes it into an invisible coding area at the edge of the logo.

5. A method for generating a unique identification for food production, comprising a unique identification management system for food production applied to any one of claims 1 to 4, characterized in that: The following steps are involved: Step 1: Receive structured data stream and convert it into a set of control parameters through preset rules; Step 2: calling a pre-trained generative adversarial network model to cross-modally fuse the control parameter set with the art feature parameterized dataset; Step 3: Generate a dynamic logo graphic containing fractal geometric features, data fingerprints and brand visual elements; Step 4: Apply physical and digital anti-counterfeiting treatment to the dynamic logo graphic; Step 5: Write the identification feature value and associated data into the blockchain evidence storage contract; Step 6: Connect the food label verification terminal, use the optical acquisition unit inside the food label verification terminal to capture the physical anti-counterfeiting features of the label through multi-spectral imaging, and connect it to the blockchain node to compare the matching degree between the on-site acquisition features and the data stored on the chain; Step 7: Using the interactive display unit inside the food label verification terminal, the production traceability information is superimposed on the real-time picture of the physical label through augmented reality technology, and then the anti-counterfeiting verification unit inside the food label verification terminal is used to detect the random distribution characteristics of the microstructure layer and the angle response curve of the optically variable layer.

6. A method for generating a unique identification for food production according to claim 5, characterized in that: The cross-modal fusion step specifically includes: Step 1: Convert the production data stream into a frequency domain feature vector through Fourier transform; Step 2: Use the self-attention mechanism to calculate the correlation weight between the frequency domain feature vector and the art feature parameterized dataset; Step 3: Dynamically weight the artistic feature parameters according to the correlation weights to generate a fused stylized parameter matrix.

Images