Agricultural product anti-counterfeiting packaging method and system based on multi-level structure

Abstract

The present application relates to a kind of agricultural product anti-fake packaging method and system based on multi-level structure.The present application relates to the field of agricultural product packaging anti-fake and product traceability, comprising the three-layer composite packaging body consisting of outer layer encryption two-dimensional code, middle layer rare earth fluorescence physical fingerprint and inner layer food-grade paraffin sealing line, two-dimensional code serial number, fluorescence characteristic vector, sealing line original image and traceability information are encrypted and bound in cloud end in factory end, and form unique index.The terminal adopts gradient three-level verification: scanning code to verify basic information, ultraviolet excitation fluorescence image is compared with packaging body, and RAI algorithm determines sealing and unsealing state, realizes irreversible anti-fake and unsealing detection.The system includes packaging body, cloud verification, terminal verification and enterprise management module, supports AI automatic determination and manual comparison, verification data real-time backflow, can monitor circulation, early warning anomaly, accurate traceability.The present application solves the problem of code separation and unsealing difficult verification, improves the reliability of agricultural product anti-fake and brand control ability.

Description

Technical Field

[0001] This invention relates to the field of agricultural product packaging anti-counterfeiting and product traceability technology, specifically to an agricultural product anti-counterfeiting packaging method and system based on a multi-level structure. Background Technology

[0002] As consumers increasingly demand higher quality, safer, and more authentic agricultural products, the brand protection and market trust system for high-end agricultural products face severe challenges. Counterfeiting, substandard products, and product substitution are rampant, seriously harming the rights and interests of both producers and consumers. Current anti-counterfeiting and traceability technologies applied in this field mainly focus on digital identification and physical structure, but they still suffer from the following inherent flaws, making it difficult to build a closed-loop, reliable anti-counterfeiting system:

[0003] Digital labeling anti-counterfeiting technology:

[0004] The common practice is to print or affix a unique code to the outer surface of the packaging. Consumers scan the code to access a cloud database to verify authenticity and obtain traceability information. The core flaw of this technology lies in the separation of code and product. The label and the packaging itself lack a physical binding relationship. Counterfeiters can mass-produce the same valid code or completely transfer the label from genuine packaging to counterfeit packaging, causing consumers to still see "genuine product" after scanning, creating a regulatory blind spot of "genuine code, fake product".

[0005] Physical structure anti-counterfeiting technology:

[0006] The aim is to provide evidence of tampering through a single act of structural destruction. Its main problems lie in the "blind spot of status verification" and its "ease of counterfeiting." On the one hand, the physical damage is difficult to link with the digital system in real time and objectively, relying on subjective judgment; on the other hand, the externally attached anti-counterfeiting labels themselves can be counterfeited or technically removed and reused, failing to form an irreversible chain of evidence of tampering.

[0007] Patent CN116051125A, "A Method and System for Traceability of Agricultural Products Based on Image Recognition," describes a technology that attempts to embed multiple features into the traceability code to improve anti-counterfeiting performance through step-by-step verification. Its core limitations lie in the high system complexity, reliance on specialized equipment and algorithms for verification, and failure to address the fundamental issue of separating the identification carrier from the product itself. The complex feature extraction and comparison process also raises the verification threshold for consumers, hindering its widespread adoption in retail scenarios.

[0008] Patent CN106600286A describes a reliable traceability system for agricultural product quality and safety based on variable QR codes. This system utilizes variable QR codes to write and query information at each stage of the agricultural product process, emphasizing dynamic information updates and transparency. However, it still fails to overcome the limitations of a weak link between the code and the product. QR codes are easily copied and reposted, making it difficult to prevent packaging recycling or internal item swapping. Furthermore, the system focuses on information traceability rather than physical packaging anti-counterfeiting, failing to provide consumers with direct and reliable evidence of whether the packaging has been opened.

[0009] In summary, existing technologies generally suffer from two major pain points: "separation of code and object" and "difficulty in digitally verifying the opened state." Digital solutions are difficult to bind to physical entities, physical solutions are difficult to link with systems, and complex image recognition solutions face high application barriers and the carrier can still be separated. Therefore, there is an urgent need for an anti-counterfeiting solution that can deeply integrate the unique physical characteristics of the packaging, the irreversible opening mechanism, and the cloud verification system to achieve a closed-loop anti-counterfeiting and traceability system of "one item, one feature; one opening, one record; one scan, one verification." Summary of the Invention

[0010] The purpose of this invention is to provide a method and system for anti-counterfeiting packaging of agricultural products based on a multi-level structure, so as to solve the problems mentioned in the background art.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a method for anti-counterfeiting packaging of agricultural products based on a multi-level structure, comprising the following steps:

[0012] S1. Construct a three-layer composite anti-counterfeiting packaging: from the outside to the inside, an outer packaging layer, a middle packaging layer, and an inner packaging layer are set in sequence; a unique encrypted QR code is printed on the surface of the outer packaging layer; the inner surface of the middle packaging layer is coated with a coating containing micro-nano-level rare earth fluorescent powder to form an unreplicable random fluorescent physical fingerprint; the sealing area of ​​the inner packaging layer is pre-coated with food-grade paraffin wax line.

[0013] S2. Factory-side information binding: Collect fluorescent images of the middle layer packaging and extract feature vectors, collect the original image of the paraffin seal line after the first heat sealing of the inner layer packaging, and encrypt and associate the outer layer QR code serial number, fluorescent feature vector, original image of the seal line and product traceability information and store them in the cloud database to establish a unique binding relationship.

[0014] S3, Terminal Level 3 Verification:

[0015] Level 1: Scan the outer QR code to verify basic product information and initial query status;

[0016] Level 2: After opening the outer packaging, the middle packaging is irradiated with ultraviolet light and a fluorescent image is captured and uploaded. The authenticity of the packaging is confirmed by feature comparison in the cloud.

[0017] Level 3: Remove the inner packaging, take a picture of the paraffin seal line and upload it. The cloud uses the Reliability Assessment Index (RAI) algorithm to determine the sealing status and opening status.

[0018] S3, Enterprise Management Data Feedback: Synchronizes all verification requests and results in real time for dynamic monitoring of product circulation, early warning of abnormal risks, and full-chain traceability.

[0019] Furthermore, in S2, the fluorescence feature comparison adopts the Scale Invariant Feature Transform (SIFT) algorithm to extract the key points and descriptors of the fluorescent spots in the fluorescence image uploaded by the user, and match them with the original feature vector stored in the cloud. When the matching degree is not less than 95%, it is determined that they are the same physical package.

[0020] Furthermore, the reliability assessment index (RAI) in S3 is calculated through the following steps:

[0021] 1. Using the homography matrix calculation based on AKAZE feature points and RANSAC, the user-uploaded sealing line image is perspective aligned with the original image;

[0022] 2. Crop the ROI region and extract its width, sharpness, and color features;

[0023] 3. The RAI value is calculated using the weighted Euclidean distance. The formula is as follows:

[0024] ,

[0025] in , , , , , These represent the relative changes in width, sharpness, and color characteristics, respectively.

[0026] 4. When RAI>0.35, the sealing status is determined to be abnormal, indicating suspected secondary packaging.

[0027] Furthermore, in S2, the factory side uses the AES256 encryption algorithm to encrypt the QR code serial number, fluorescent feature vector, original image of the sealing line, and product traceability information, establishing a unique index of "one bag, one code, one feature, one image".

[0028] Furthermore, the terminal three-level verification in S3 adopts a gradient progressive verification logic. The next level of verification process can only be started after the current level of verification is passed. The request, result, time and terminal geographical location information of each level of verification are all encrypted and stored in the cloud database in real time and written into the full-link traceability file of the corresponding product.

[0029] A multi-level structured anti-counterfeiting packaging system for agricultural products includes:

[0030] The three-layer composite anti-counterfeiting packaging consists of an outer layer, a middle layer, and an inner layer, from the outside in. The outer layer is printed with a unique encrypted QR code. The inner surface of the middle layer is coated with a coating containing micro-nano-level rare earth fluorescent powder. The sealing area of ​​the inner layer is pre-coated with food-grade paraffin wax.

[0031] The cloud-based data verification module is used to store encrypted QR code serial numbers, fluorescent feature vectors, original images of sealing lines, and product traceability information, and to execute fluorescent feature comparison algorithms and reliability assessment index (RAI) algorithms.

[0032] The terminal verification module communicates with the cloud data verification module, allowing consumers to scan the outer QR code, upload fluorescent images and sealing line images, and display the verification results.

[0033] The enterprise management module communicates with the cloud data verification module to obtain verification data and realize product circulation monitoring, anomaly warning and full-chain traceability.

[0034] Furthermore, the outer packaging uses a PET / AL / PE composite film, the middle packaging is a transparent PE self-sealing bag with a self-sealing strip on the top, and the inner packaging uses a biodegradable PLA film.

[0035] Furthermore, the inner packaging surface is printed with a traceability QR code that verifies the encrypted QR code on the outer layer, used for dual unlocking of product traceability information and anti-counterfeiting verification.

[0036] Furthermore, the cloud-based data verification module supports a dual verification mode of manual intuitive comparison and AI automatic judgment, displaying the original image of the sealing line and the user-uploaded image on the same screen, and automatically outputting the feature change amount and judgment result.

[0037] Furthermore, the enterprise management module can view the geographical distribution and verification pass rate of product queries in real time, automatically mark high-frequency abnormal verification areas or multiple unpacking verification requests of a single package, and realize accurate traceability of problematic products to the production line and production shift.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] The agricultural product anti-counterfeiting packaging method and system based on a multi-level structure provided by this invention addresses the two core defects of existing technologies: "separation of code and product" and "difficulty in digitally verifying the unsealed state." By constructing a closed-loop anti-counterfeiting system of "three-layer composite structure + cloud image comparison," significant progress has been made.

[0040] In response to the shortcomings of existing technologies such as QR codes and digital watermarks, which are easily copied or transferred, this invention embeds a unique fluorescent fingerprint into the packaging material and encrypts and binds it with the outer QR code and the inner sealing image at the time of manufacture, forming a three-in-one anti-counterfeiting system of "physical characteristics - digital identity - packaging status", which fundamentally eliminates counterfeiting paths such as label copying and packaging recycling.

[0041] To address the problem that traditional physical anti-counterfeiting measures are difficult to verify online, this invention collects and compares the visual features of the paraffin sealing line and uses the RAI index algorithm to automatically identify morphological changes caused by secondary packaging. This allows the unpacking behavior to be objectively and automatically determined in the cloud, thus creating a closed loop between physical anti-counterfeiting and digital verification.

[0042] This invention provides a multi-level verification process, balancing consumer convenience with corporate credibility. It designs a tiered verification process from QR code scanning and UV light comparison to uploading of sealed images. Consumers can complete the entire verification process using a regular mobile phone, while businesses can obtain tamper-proof verification evidence through the system, greatly enhancing the usability and credibility of the anti-counterfeiting system; empowering businesses to proactively manage and upgrading it into a supply chain management tool. Unlike traditional anti-counterfeiting systems that only verify authenticity, this system collects verification data in real time, supporting businesses to dynamically monitor product flow, identify the risks of cross-selling and counterfeit goods, and achieve a leap from passive anti-counterfeiting to proactive market control and traceability decision-making, improving brand and channel management. It effectively solves two core pain points: "separation of label and physical entity" and "difficulty in digitally and objectively verifying the opened state." Digital solutions are difficult to bind to physical entities, physical solutions are difficult to link with systems, and complex identification solutions face the dilemma of high application thresholds and the carrier still being separable. Attached Figure Description

[0043] Figure 1 This is a structural design drawing of the three-layer composite anti-counterfeiting packaging of the present invention;

[0044] Figure 2 This is an unfolded diagram of the three-layer composite anti-counterfeiting packaging of the present invention;

[0045] Figure 3 This is the system logic flowchart of the present invention;

[0046] Figure 4 This is a flowchart of the system operation method of the present invention.

[0047] In the diagram: 1. Outer packaging; 2. Middle packaging; 3. Inner packaging. Detailed Implementation

[0048] Please see Figure 1 —4. A method for anti-counterfeiting packaging of agricultural products based on a multi-level structure, comprising the following steps:

[0049] S1. Constructing a three-layer composite anti-counterfeiting packaging: From the outside in, an outer packaging 1, a middle packaging 2, and an inner packaging 3 are sequentially arranged. The outer packaging 1 has a unique encrypted QR code printed on its surface, serving as a product protection and information entry layer. The QR code provides consumers with a unique entry point to the cloud verification system, achieving "one item, one code." The inner surface of the middle packaging 2 is coated with a layer containing micro-nano-level rare earth fluorescent powder, forming an unreplicable random fluorescent physical fingerprint, with a built-in unique "physical fingerprint." The fluorescent powder has a random microscopic distribution, presenting an unreplicable fluorescent pattern under ultraviolet light, achieving unique identification of the packaging itself and preventing label transfer. The sealing area of ​​the inner packaging 3 is pre-coated with a food-grade paraffin line, providing irreversible evidence of opening. After the initial controlled heat sealing, a sealing line with specific visual characteristics (gloss, color, edge morphology) is formed; any secondary heating will cause a recognizable visual change. This allows for traceability of packaging authenticity.

[0050] S2. Factory-side information binding: Collect fluorescent images of the middle layer packaging 2 and extract feature vectors; collect the original image of the paraffin seal line after the first heat sealing of the inner layer packaging 3; encrypt and associate the outer layer QR code serial number, fluorescent feature vector, original image of the seal line and product traceability information and store them in the cloud database to establish a unique binding relationship.

[0051] In S2, fluorescence feature comparison uses the Scale Invariant Feature Transform (SIFT) algorithm to extract key points and descriptors of fluorescent spots in user-uploaded fluorescence images and match them with the original feature vectors stored in the cloud. When the matching degree is not less than 95%, it is determined that they are the same physical package.

[0052] In S2, the factory side uses the AES256 encryption algorithm to encrypt the QR code serial number, fluorescent feature vector, original image of the sealing line, and product traceability information, establishing a unique index of "one bag, one code, one feature, one image".

[0053] The system initialization and information binding process includes:

[0054] 1. Image Acquisition:

[0055] Fluorescence image acquisition: Under ultraviolet light in a dark room, acquire fluorescence images of a designated area of ​​the middle layer packaging and extract its microscopic distribution feature vector.

[0056] Sealing image acquisition: Under standard light source, high-definition macro photography is performed on the inner paraffin sealing line after the first heat sealing to fully record its original visual state.

[0057] 2. Information encryption and binding:

[0058] The outer QR code serial number, the middle fluorescent feature vector, and the inner sealing original image, along with traceability information such as product batch and production date, are encrypted with AES-256 and then stored in a cloud database to establish a unique index of "one bag, one code, one feature, and one image".

[0059] The core algorithm of this system consists of two parts: fluorescence feature comparison and sealing line status verification.

[0060] Fluorescence feature matching algorithm:

[0061] 1. Feature Extraction: For user-uploaded ultraviolet light images, scale-invariant feature transform (SIFT) or similar algorithms are used to extract key points and descriptors of fluorescent spots.

[0062] 2. Similarity Calculation: Match the original feature vector stored in the cloud. If the matching degree is higher than the set threshold (e.g., ≥95%), it is determined to be the same physical package.

[0063] S3, Terminal Level 3 Verification:

[0064] Level 1: Scan the outer QR code to verify basic product information and initial query status;

[0065] Level 2: After opening the outer packaging 1, the middle packaging 2 is irradiated with ultraviolet light and a fluorescent image is captured and uploaded. The authenticity of the packaging is confirmed by feature comparison in the cloud.

[0066] Level 3: Remove the inner packaging 3, take a picture of the paraffin sealing line and upload it. The cloud uses the Reliability Assessment Index (RAI) algorithm to determine the sealing status and opening status.

[0067] The Reliability Assessment Index (RAI) in S3 is calculated using the following steps:

[0068] 1. Using the homography matrix calculation based on AKAZE feature points and RANSAC, the user-uploaded sealing line image is perspective aligned with the original image;

[0069] 2. Crop the ROI region and extract its width, sharpness, and color features;

[0070] Width features: The pixel ratio of the sealing line region is calculated by image binarization and pixel statistics.

[0071] Sharpness feature: The degree of blurriness at the edge of the closure is quantified using the Laplacian operator variance.

[0072] Color characteristics: Convert to Lab color space, calculate the mean of the b channel, and evaluate the degree of yellowing of the sealing line.

[0073] 3. The RAI value is calculated using the weighted Euclidean distance. The formula is as follows:

[0074] ,

[0075] in , , , , , These represent the relative changes in width, sharpness, and color characteristics, respectively.

[0076] 4. When RAI>0.35, the sealing status is determined to be abnormal, indicating suspected secondary packaging.

[0077] The S3 terminal three-level verification adopts a gradient progressive verification logic. The next level of verification process can only be started after the current level of verification is passed. The request, result, time and terminal geographical location information of each level of verification are all encrypted and stored in the cloud database in real time and written into the full-link traceability file of the corresponding product.

[0078] S4. Enterprise Management Data Feedback: Synchronizes all verification requests and results in real time for product circulation dynamic monitoring, abnormal risk warning and full-chain traceability.

[0079] All verification requests and results are synchronized to the cloud in real time. The management backend can perform the following functions:

[0080] Dynamic monitoring: View the geographical distribution of product queries and the verification pass rate in real time.

[0081] Intelligent early warning: Automatically marks high-frequency abnormal verification areas or multiple unpacking verification requests of a single package, indicating the risk of counterfeit goods or cross-channel sales.

[0082] End-to-end traceability: Accurately trace problematic products back to specific production lines and shifts, empowering enterprises to upgrade from passive anti-counterfeiting to proactive market control.

[0083] Identification process:

[0084] After receiving the product, users can scan the anti-counterfeiting code on the outer layer to access the cloud and view the fluorescent image of the middle layer and the paraffin seal image of the inner layer. They can also scan the QR code on the inner layer to check its authenticity. After opening the outer packaging, users can shine a UV light on the middle packaging and compare it with the fluorescent image in the cloud. If they are the same, it proves that it has not been tampered with. If they are different, users should be cautious as there may be a switch. Continue to tear open the middle packaging to reach the inner layer. Users need to compare it again with the paraffin seal image in the cloud. If the color and condition are the same, it proves that it has not been opened. If they are different, users should be cautious as there may be a second sealing.

[0085] Image acquisition and comparison system:

[0086] After the initial sealing is completed at the factory, the sealing area of ​​each package is immediately photographed in high definition. The photos are uploaded to a cloud database and linked to the unique QR code of the package.

[0087] The system supports automatic image feature extraction, recording features such as color, gloss, shape, and texture of the paraffin sealing line.

[0088] After purchasing, consumers can scan the QR code on the packaging. The system will not only display product traceability information, but also show the initial sealing image and the current sealing image uploaded by the user side by side for consumers to compare intuitively.

[0089] The system can also integrate AI image comparison algorithms to automatically determine whether the sealing line is abnormal and prompt "suspected secondary packaging".

[0090] Overall Algorithm Goal

[0091] By comparing the original factory sealing image with the current sealing image taken by the user, a RAI (Reliability Assessment Index) is calculated. If the RAI exceeds the threshold (0.35), it is judged as "abnormal / counterfeit".

[0092] Algorithm Flow

[0093] Step 1: Image Registration

[0094] Objective: To eliminate geometric differences such as shooting angle, scaling, and rotation.

[0095] Method: Keypoints and descriptors were extracted using the AKAZE feature detector.

[0096] We use brute-force matching to find the feature correspondence between the two images.

[0097] The top 15% of best matching points are retained, and perspective transformation is performed using RANSAC + homography matrix.

[0098] The current image is warped to the viewpoint of the original image to achieve alignment.

[0099] Steps 2 & 3: ROI cropping + multidimensional feature extraction

[0100] ROI definition: By default, the middle 20% height region of the image is cropped (y∈[0.4h,0.6h]), assuming the sealing line is in this region.

[0101] Extract three types of features (extracting the ROI of the original image and the ROI of the aligned current image separately):

[0102] width The proportion of white pixels after binarization The wider the wax line, the larger the coverage area. Clarity Laplace operator variance The larger the value, the clearer the image; normalize to ≤1 Yellow Mean value of b channel in Lab color space The higher the b value, the more yellow the color.

[0103] Step 4: Difference Quantification and RAI Calculation

[0104] Calculate the relative change (Delta) for each feature:

[0105] Width variation (the wider the better):

[0106] (Expected: Genuine products close to 0, counterfeits > 0.5)

[0107] Decreased clarity (the blurrier the image, the more likely it is to be fake): (Expected: genuine product close to 0, fake product >0.4)

[0108] Color shift (the more yellow, the more fake):

[0109] Then, a weighted RAI index (in Euclidean norm form) is synthesized:

[0110] Wherein the weights are:

[0111] (Most sensitive)

[0112]

[0113]

[0114] Step 5: Determine the result

[0115] If RAI > 0.35 → "status": "FAIL" (suspected forgery)

[0116] Otherwise → "status": "PASS"

[0117] It also returns detailed percentage changes (such as width change +12.34%, blurriness increase of 5.67%, etc.).

[0118] A multi-level structured anti-counterfeiting packaging system for agricultural products includes:

[0119] The three-layer composite anti-counterfeiting packaging consists of an outer packaging 1, a middle packaging 2, and an inner packaging 3, from the outside to the inside. The outer packaging 1 is printed with a unique encrypted QR code. The inner surface of the middle packaging 2 is coated with a coating containing micro-nano-grade rare earth fluorescent powder. The sealing area of ​​the inner packaging 3 is pre-coated with food-grade paraffin wax.

[0120] The cloud-based data verification module is used to store encrypted QR code serial numbers, fluorescent feature vectors, original images of sealing lines, and product traceability information, and to execute fluorescent feature comparison algorithms and reliability assessment index (RAI) algorithms.

[0121] The terminal verification module communicates with the cloud data verification module, allowing consumers to scan the outer QR code, upload fluorescent images and sealing line images, and display the verification results.

[0122] The enterprise management module communicates with the cloud data verification module to obtain verification data and realize product circulation monitoring, anomaly warning and full-chain traceability.

[0123] The outer packaging 1 is made of PET / AL / PE composite film, the middle packaging 2 is a transparent PE self-sealing bag with a self-sealing strip on the top, and the inner packaging 3 is made of biodegradable PLA film.

[0124] The inner packaging layer 3 has a traceability QR code printed on its surface, which verifies the encrypted QR code on the outer layer. This code is used for dual unlocking of product traceability information and anti-counterfeiting verification.

[0125] The cloud-based data verification module supports a dual verification mode of manual comparison and AI automatic judgment. It displays the original image of the sealing line and the image uploaded by the user on the same screen, and automatically outputs the feature change amount and judgment result.

[0126] The enterprise management module allows real-time viewing of the geographical distribution and verification pass rate of product queries, automatically marking high-frequency abnormal verification areas or multiple unpacking verification requests for individual packages, enabling precise traceability of problematic products to the production line and production shift.

[0127] Example 1: A Multi-Level Structure-Based Anti-counterfeiting Packaging Method for Agricultural Products

[0128] This embodiment fully realizes the closed-loop process of anti-counterfeiting packaging for agricultural products, including four core links: three-layer packaging preparation, factory-side information binding, terminal-side full-process identification and verification, and enterprise-side data feedback and control. Simultaneously, the identification process is seamlessly integrated into the entire terminal verification chain. The specific steps are as follows:

[0129] Construction and Preparation of S1 Three-Layer Composite Anti-counterfeiting Packaging

[0130] This step constructs a three-layer composite anti-counterfeiting packaging body with progressively enhanced functions from the outside in, suitable for the vacuum packaging and distribution needs of 5kg high-end rice. The specific structure and preparation parameters are as follows:

[0131] 1. Outer Packaging: Made of food-grade PET / AL / PE composite film, forming a rectangular upright packaging bag with an 8mm wide easy-tear opening at the top. The outer surface is printed with a unique encrypted QR code using variable data printing technology. This QR code is generated using the AES256 encryption algorithm, and the serial number is a 16-digit random code with no duplicates and cannot be cracked in batches. It serves as the sole entry point for consumers to access the cloud verification system, achieving the basic identification of "one item, one code".

[0132] 2. Middle Packaging: A transparent PE self-sealing bag is used, its size adapted to the inner cavity of the outer packaging bag. A raised-and-lower tamper-evident self-sealing strip is installed at the top. The inner surface is uniformly coated with a water-based coating containing micro-nano-level rare earth phosphors using a gravure coating process. The dry film coating contains phosphor particles with a particle size controlled between 500-800nm, randomly dispersed within the coating. This coating has no visible pattern under visible light, but only exhibits randomly distributed fluorescent spot patterns under 365nm ultraviolet light excitation, forming a unique and unreplicable random fluorescent physical fingerprint for each package. The uniqueness error of the fluorescent spot distribution characteristics of a single package's coating does not exceed 0.1%, making it impossible to imitate through printing or replication, thus eliminating the risk of label transfer and packaging tampering at the source.

[0133] 3. Inner Packaging: Made with food-grade biodegradable PLA film, suitable for vacuum packaging of rice. A 5mm wide food-grade fully refined paraffin wax line is pre-coated along the width of the bag opening in the sealing area. The paraffin wax melting point is controlled between 58℃ and 62℃, complying with GB4806.1-2016 "National Food Safety Standard for Food Contact Materials and Products - General Safety Requirements". It is odorless and does not migrate with the rice. Simultaneously, the inner packaging surface is printed with a traceability QR code that uses the same encryption as the outer layer's QR code. The two QR codes can be mutually verified; only after the serial number matches and verification passes can complete product traceability information and anti-counterfeiting verification permissions be unlocked.

[0134] S2 Factory-Side Information Binding and System Initialization

[0135] This step completes the unique encrypted binding of packaging physical characteristics, digital identity, and product traceability information, establishing an immutable anti-counterfeiting and traceability index. The specific process is as follows:

[0136] 1. Standardized image acquisition

[0137] Fluorescence image acquisition: In a darkroom environment, a 10cm×10cm fixed acquisition area of ​​the middle layer packaging was vertically illuminated by a 365nm standard ultraviolet light source, and a clear fluorescence image was obtained by macro photography using a 20-megapixel industrial camera; the key points and descriptors of the fluorescent spots in the image were extracted by using a scale-invariant feature transformation algorithm to generate a 128-dimensional feature vector, which serves as the unique fluorescent physical fingerprint of the packaging.

[0138] Original image acquisition of the sealing line: After the inner packaging completes the vacuum filling of rice, it is heat-sealed in a controlled manner using a precision heat sealer. After cooling to room temperature, the paraffin wax line is evenly impregnated on the surface of the PLA film to form a semi-transparent, clear-edged, pearly-luster fixed sealing line. Under a D65 standard light source, a 24-megapixel industrial macro camera is used to take high-definition pictures of the sealing line area, fully recording the original visual characteristics of the sealing line such as width, edge shape, gloss, and color distribution, and generating the original image of the sealing line.

[0139] 2. Information encryption and unique binding

[0140] The outer QR code serial number, the middle fluorescent feature vector, and the original image of the inner sealing line, along with the full product traceability information, are symmetrically encrypted using the AES256 encryption algorithm and then stored in the cloud blockchain database to establish a unique index of "one bag, one code, one feature, one image," ensuring that the data is tamper-proof and unreverse-proof, thus completing the system initialization.

[0141] S3 Terminal Full-Process Authentication and Three-Level Verification

[0142] This step completes the authentication process, employing a tiered, progressive verification logic. Only after the current level of verification passes can the entry point for the next level of verification and the image data upload channel be unlocked. If any level of verification determines an anomaly, the verification process is immediately terminated, and a risk warning is simultaneously pushed to the terminal and the enterprise management module. Consumers can complete the entire authentication process using a regular smartphone, without the need for specialized equipment. It supports both manual comparison and AI-automated judgment in a dual authentication mode. The specific process is as follows:

[0143] Level 1 Verification: Basic QR Code Scanning Verification and Preliminary Identification

[0144] After purchasing a product, consumers first visually inspect the outer packaging for damage and signs of resealing. They then scan the encrypted QR code on the outer packaging with their smartphones, which redirects them to a cloud verification page. The cloud first verifies the initial query status of the QR code.

[0145] - If this is the first time querying, the page displays basic product traceability information (product name, certification qualifications, production date, production batch), unlocks the next level of verification entry, and prompts consumers for the subsequent verification operation process;

[0146] If this is not the first query, the page will prominently display the product's historical query time, number of queries, and query location, reminding consumers to be wary of the risks of packaging recycling and product tampering.

[0147] If the QR code serial number is not registered in the cloud database, it will be directly identified as a counterfeit product, the verification process will be terminated, and a risk warning will be sent.

[0148] Meanwhile, consumers can scan the traceability QR code on the inner packaging to perform two-way verification with the outer QR code. Once the verification is successful, the complete product traceability file, as well as the original fluorescent image of the middle layer and the original image of the inner sealing line of the packaging, can be unlocked for consumers to preview in advance and complete the initial manual identification.

[0149] Level 2 verification: Fluorescent physical fingerprint comparison and packaging body identification

[0150] Consumers open the outer packaging completely along the easy-tear opening at the top, remove the middle transparent PE self-sealing bag, and proceed to the second-level verification process, which corresponds to the tool-assisted verification step in the identification process. It supports two modes: manual visual identification and AI automatic identification.

[0151] 1. Manual visual identification: Consumers use a portable 365nm ultraviolet flashlight to vertically illuminate the designated collection area on the inner surface of the middle layer of packaging to excite fluorescent spot patterns. The fluorescent spots are then visually compared with the original fluorescent image displayed on the cloud verification page to check whether the distribution, quantity, size, and shape of the fluorescent spots are completely consistent, thus completing the manual identification.

[0152] 2. AI-powered Automatic Intelligent Identification: Consumers click "Intelligent Verification of Fluorescent Features" on the verification page, use their smartphones to capture fluorescent images under ultraviolet light, and upload them to the cloud. The cloud data verification module calls the SIFT algorithm to extract key points and descriptors of fluorescent spots in the uploaded image and performs matching calculations with the original feature vectors stored in the database. When the matching degree is not less than 95%, it is determined to be the same physical packaging, and the page displays "Fluorescent feature matching successful, packaging itself has not been swapped," and unlocks the third-level verification entry. When the matching degree is less than 95%, it is determined that the packaging itself is abnormal, and the page prominently displays "Fluorescent feature mismatch, suspected packaging swap, counterfeit product," terminating the verification process and simultaneously sending a warning message to the enterprise management module.

[0153] Level 3 Verification: Checking the condition of the sealing line and identifying opening behavior.

[0154] Consumers open the self-sealing strip of the middle packaging, remove the inner PLA vacuum packaging bag, and enter the third-level final verification process, which corresponds to the cloud-based final verification stage in the identification process. This stage irreversibly identifies the state of the seal line being broken. It also supports both manual visual identification and AI automatic judgment modes. The specific process is as follows:

[0155] Visual inspection by the consumer: The cloud verification page displays the original image of the packaging seal line and standardized shooting guidelines on the same screen. Consumers can visually inspect the integrity, edge shape, gloss and color of the inner paraffin seal line and compare it with the original image. If the seal line shows abnormalities such as edge diffusion, wax layer melting and accumulation, yellowing and discoloration, or secondary heat sealing marks, it can be directly judged as suspected secondary packaging.

[0156] AI-powered automatic intelligent identification: On the verification page, consumers click "Ultimate Verification of Sealing Status," follow the shooting instructions to take a complete and clear image of the paraffin wax sealing line under uniform natural light, and upload it to the cloud; the cloud data verification module executes the Reliability Assessment Index (RAI) algorithm to complete the following identification steps:

[0157] (1) Image registration: The homography matrix based on AKAZE feature points and RANSAC is used to perform perspective alignment between the user-uploaded sealing line image and the original image, eliminating geometric differences caused by shooting angle, scaling and rotation;

[0158] (2) ROI region cropping: By default, the core region of the sealing line of the middle 20% height area (y∈[0.4h,0.6h]) of the image is cropped to exclude irrelevant background interference;

[0159] (3) Multidimensional feature extraction: For the cropped ROI region, width features, sharpness features, and color features are extracted respectively: the width feature is processed by binarization and color inversion of the image using the Otsu algorithm, and the proportion of white pixels in the sealing line area is statistically analyzed to quantify the width change of the paraffin line; the sharpness feature is calculated by using the Laplacian operator to calculate the image variance and quantify the blurring degree of the sealing line edge. The smaller the value, the more blurred the edge. The value is normalized to ≤1; the color feature is converted to the Lab color space, the mean of the b channel is calculated, and the yellowing degree of the sealing line is evaluated. The higher the b value, the more yellow the color.

[0160] (4) RAI value calculation and determination: The RAI value is calculated using weighted Euclidean distance. The calculation formula is as follows:

[0161] ,

[0162] in , , , , , These represent the relative changes in width, sharpness, and color characteristics, respectively.

[0163] When RAI≤0.35, the sealing status is determined to be normal, and the page will display "The sealing status is normal, it is confirmed to be the first time it has been opened, and the product anti-counterfeiting verification has all passed", and a complete full-chain product traceability file will be displayed.

[0164] When RAI > 0.35, the sealing status is determined to be abnormal. The page will display a prominent warning: "Abnormal sealing status, suspected of secondary packaging or product swapping. Please consume with caution." At the same time, a detailed breakdown of the percentage changes in various characteristics will be provided. The verification process will be terminated and the warning information will be synchronized to the enterprise management module.

[0165] Full record of verification data

[0166] The request time, terminal geographical location, uploaded images, judgment results, and user operation behavior of each level of verification are all synchronously and encryptedly stored in the cloud blockchain database in real time, and synchronously written into the full-chain traceability file of the corresponding product. This ensures that all verification behaviors are traceable and tamper-proof, providing a complete chain of evidence for subsequent brand rights protection and counterfeit product tracing.

[0167] S4 Enterprise Management Terminal Data Feedback and End-to-End Control

[0168] All terminal verification requests and results are synchronized to the enterprise management module in real time, enabling enterprises to upgrade from passive anti-counterfeiting to proactive market control. Specific functionalities are implemented as follows:

[0169] Circulation dynamic monitoring: Enterprises can view the national geographical distribution of product queries, the verification pass rate of each batch of products, and the query frequency of each region in real time through the management backend, so as to accurately grasp the actual circulation trajectory and terminal sales of products;

[0170] Intelligent early warning of abnormal risks: The system automatically marks high-frequency abnormal verification areas, verification requests that exceed the authorized sales area, and multiple opening verification requests for a single package, automatically triggering warnings of counterfeit risk or channel cross-selling, and pushing abnormal details to enterprise management personnel;

[0171] Precise traceability across the entire supply chain: For problematic products with abnormal warnings, a unique index can be used to accurately trace back to the specific production line, production shift, production operator, inspection personnel, logistics carrier, and authorized distributor, achieving closed-loop traceability across the entire supply chain from farm to table, quickly locating the root cause of the problem, and completing risk management;

[0172] Data statistics and analysis: The system automatically generates anti-counterfeiting verification data reports, cross-selling risk reports, and product circulation reports, providing data support for enterprises' brand protection, channel management, production planning, and market decision-making.

[0173] Example 2: Anti-counterfeiting Packaging System for Agricultural Products Based on Multi-level Structure

[0174] This system is used to implement the multi-level structure-based anti-counterfeiting packaging method for agricultural products described in Example 1. It includes a three-layer composite anti-counterfeiting packaging body, a cloud data verification module, a terminal verification module, and an enterprise management module. The modules are interconnected and work together to complete the entire process of anti-counterfeiting verification and traceability control. The specific implementation details are as follows:

[0175] Three-layer composite anti-counterfeiting packaging: from the outside to the inside, it consists of an outer layer, a middle layer, and an inner layer. The specific structure, material parameters, and preparation process are the same as those described in S1 of Example 1.

[0176] Cloud-based data verification module: Deployed on a cloud server cluster and equipped with a blockchain database, it is used to store encrypted QR code serial numbers, fluorescent feature vectors, original images of sealing lines, and product traceability information; it has built-in SIFT fluorescent feature comparison algorithm unit, RAI sealing line status verification algorithm unit, and AES256 encryption and decryption unit, supporting dual verification modes of manual intuitive comparison and AI automatic judgment. It can push the original image of the sealing line and the image uploaded by the user to the terminal verification module on the same screen, and automatically output the feature change amount and judgment result;

[0177] Terminal verification module: This is an H5 webpage / WeChat mini-program based on smartphones, requiring no additional app download. It wirelessly connects to the cloud data verification module and includes a built-in scanning unit, image capture and upload unit, verification result display unit, and standardized operation guidance unit. It allows consumers to scan the outer encrypted QR code, upload fluorescent images and sealing line images, and receive and display verification results and operation guidance from the cloud. It also features a tiered verification permission control unit, which unlocks the operation entry and image upload channel for the next level of verification only after the previous level of verification is completed and the cloud feedback is approved.

[0178] Enterprise Management Module: This module serves as the PC-based management backend and mobile management app, communicating with the cloud-based data verification module. It includes a data monitoring unit, an intelligent early warning unit, a full-chain traceability unit, and a data reporting unit. These units are used to acquire full verification data in real time, enabling dynamic monitoring of product circulation, early warning of abnormal risks, full-chain traceability of problematic products, and data statistical analysis. The specific functions are the same as those described in S4 of Example 1.

Claims

1. A method for anti-counterfeiting packaging of agricultural products based on a multi-level structure, characterized in that, Includes the following steps: S1. Construct a three-layer composite anti-counterfeiting packaging: from the outside to the inside, set an outer packaging (1), a middle packaging (2) and an inner packaging (3); the outer packaging (1) is printed with a unique encrypted QR code; the inner surface of the middle packaging (2) is coated with a coating containing micro-nano-level rare earth fluorescent powder to form an unreplicable random fluorescent physical fingerprint; the sealing area of ​​the inner packaging (3) is pre-coated with food-grade paraffin wax line; S2, Factory-side information binding: Collect the fluorescent image of the middle layer packaging (2) and extract the feature vector, collect the original image of the paraffin sealing line after the first heat sealing of the inner layer packaging (3), encrypt and associate the outer layer QR code serial number, fluorescent feature vector, original image of sealing line with product traceability information and store it in the cloud database to establish a unique binding relationship; S3, Terminal Level 3 Verification: Level 1: Scan the outer QR code to verify basic product information and initial query status; Level 2: After opening the outer packaging (1), the middle packaging (2) is irradiated with ultraviolet light and a fluorescent image is captured and uploaded. The authenticity of the packaging body is confirmed by feature comparison in the cloud. Level 3: Remove the inner packaging (3), take a picture of the paraffin sealing line and upload it. The cloud uses the Reliability Assessment Index (RAI) algorithm to determine the sealing status and unsealing status. S4. Enterprise Management Data Feedback: Synchronizes all verification requests and results in real time for product circulation dynamic monitoring, abnormal risk warning and full-chain traceability.

2. A multi-level structure-based anti-counterfeiting packaging system for agricultural products, employing the method described in claim 1, characterized in that, include: The three-layer composite anti-counterfeiting packaging consists of an outer packaging (1), a middle packaging (2), and an inner packaging (3) from the outside to the inside. The outer packaging (1) is printed with a unique encrypted QR code. The inner surface of the middle packaging (2) is coated with a coating containing micro-nano-grade rare earth fluorescent powder. The sealing area of ​​the inner packaging (3) is pre-coated with food-grade paraffin wax. The cloud-based data verification module is used to store encrypted QR code serial numbers, fluorescent feature vectors, original images of sealing lines, and product traceability information, and to execute fluorescent feature comparison algorithms and reliability assessment index (RAI) algorithms. The terminal verification module communicates with the cloud data verification module, allowing consumers to scan the outer QR code, upload fluorescent images and sealing line images, and display the verification results. The enterprise management module communicates with the cloud data verification module to obtain verification data and realize product circulation monitoring, anomaly warning and full-chain traceability.

3. The method for anti-counterfeiting packaging of agricultural products based on a multi-level structure according to claim 1, characterized in that, In S2, the fluorescence feature comparison uses the Scale Invariant Feature Transform (SIFT) algorithm to extract key points and descriptors of fluorescent spots in the fluorescence image uploaded by the user, and match them with the original feature vector stored in the cloud. When the matching degree is not less than 95%, it is determined that they are the same physical package.

4. The method for anti-counterfeiting packaging of agricultural products based on a multi-level structure according to claim 1, characterized in that, The reliability assessment index (RAI) in S3 is calculated through the following steps:

1. Using the homography matrix calculation based on AKAZE feature points and RANSAC, the user-uploaded sealing line image is perspective aligned with the original image; 2. Crop the ROI region and extract its width, sharpness, and color features; 3. The RAI value is calculated using the weighted Euclidean distance. The formula is as follows: ， in , , , , , These represent the relative changes in width, sharpness, and color characteristics, respectively.

4. When RAI>0.35, the sealing status is determined to be abnormal, indicating suspected secondary packaging.

5. The method for anti-counterfeiting packaging of agricultural products based on a multi-level structure according to claim 1, characterized in that, In S2, the factory side uses the AES256 encryption algorithm to encrypt the QR code serial number, fluorescent feature vector, original image of the sealing line, and product traceability information, and establishes a unique index of "one bag, one code, one feature, one image".

6. The method for anti-counterfeiting packaging of agricultural products based on a multi-level structure according to claim 1, characterized in that, The terminal three-level verification in S3 adopts a gradient progressive verification logic. The next level of verification process can only be started after the current level of verification is passed. The request, result, time and terminal geographical location information of each level of verification are all encrypted and stored in the cloud database in real time and written into the full-link traceability file of the corresponding product.

7. The agricultural product anti-counterfeiting packaging system based on a multi-level structure according to claim 2, characterized in that, The outer packaging (1) is made of PET / AL / PE composite film, the middle packaging (2) is a transparent PE self-sealing bag with a self-sealing strip on the top, and the inner packaging (3) is made of biodegradable PLA film.

8. The agricultural product anti-counterfeiting packaging system based on a multi-level structure according to claim 2, characterized in that, The inner packaging (3) has a traceability QR code printed on its surface that verifies the outer encrypted QR code, which is used for dual unlocking of product traceability information and anti-counterfeiting verification.

9. The agricultural product anti-counterfeiting packaging system based on a multi-level structure according to claim 2, characterized in that, The cloud-based data verification module supports a dual verification mode of manual intuitive comparison and AI automatic judgment. It displays the original image of the sealing line and the image uploaded by the user on the same screen, and automatically outputs the feature change amount and judgment result.

10. The agricultural product anti-counterfeiting packaging system based on a multi-level structure according to claim 2, characterized in that, The enterprise management module allows real-time viewing of the geographical distribution and verification pass rate of product queries, automatically marking high-frequency abnormal verification areas or multiple unpacking verification requests for a single package, enabling precise traceability of problematic products to the production line and production shift.

Images