Safety anti-counterfeiting method based on double anti-counterfeiting codes

Abstract

The invention provides a safety anti-counterfeiting method based on double anti-counterfeiting codes, belongs to the field of anti-counterfeiting technologies, and is used for solving the problems that anti-counterfeiting methods in related technologies are weak in anti-copying, anti-tampering and anti-transfer capabilities and are difficult to cover full life cycle tracing of products. On the basis, generating double codes, associating, binding and storing the double codes and spatio-temporal information, and then performing layered verification; the anti-copying, anti-tampering, anti-transferring and anti-quantum-cracking capabilities of the whole life cycle can be improved, and the verification precision and the scene adaptability are both considered.

Description

Technical Field

[0001] This application relates to the field of anti-counterfeiting technology, and in particular to a secure anti-counterfeiting method based on dual anti-counterfeiting codes. Background Technology

[0002] With the rapid development of the commodity economy, anti-counterfeiting technology plays an increasingly important role in protecting consumer rights and maintaining market order, and is widely used in various fields such as food, medicine, and luxury goods. Among them, QR code anti-counterfeiting has become one of the mainstream anti-counterfeiting methods due to its low cost and ease of identification. It verifies authenticity by setting a QR code on the product and storing product information, combined with encryption algorithms.

[0003] Existing QR code anti-counterfeiting technologies mostly rely on extracting information from single static physical features (such as surface texture or fixed patterns) to generate independent explicit or implicit codes. They achieve dual-code binding only through simple information association (such as storing the location of the implicit code in the explicit code), and the spatiotemporal information often uses timestamps and location data of ordinary precision, without deep integration with product features. Furthermore, the verification process often uses fixed thresholds, failing to consider the impact of product lifecycle, material differences, and environmental changes on features, and lacks effective detection methods against AI-forged features and carrier tampering.

[0004] The existing technologies mentioned above have significant drawbacks: First, feature extraction is easily affected by product deformation and environmental fluctuations, resulting in large feature deviations and weak anti-copying capabilities; second, the dual-code association is loose, the encryption level is low, and there is a lack of quantum attack resistance design, resulting in insufficient anti-tampering capabilities; third, the binding between spatiotemporal information and product features is weak, the chain relationship has poor stability, and the anti-transfer capability is weak; fourth, the verification accuracy and scenario adaptability are low, and it is easy to make misjudgments or omissions, making it difficult to meet the requirements of high-security anti-counterfeiting. Therefore, a new security anti-counterfeiting method is urgently needed to solve the above problems. Summary of the Invention

[0005] This application provides a security anti-counterfeiting method based on dual anti-counterfeiting codes, which can improve the anti-copying, anti-tampering, and anti-transfer capabilities of the anti-counterfeiting method through multi-dimensional feature fusion, deep association of dual codes, chain-based stable binding, and layered precise verification, while taking into account the verification accuracy and scenario adaptability.

[0006] This application provides a secure anti-counterfeiting method based on dual anti-counterfeiting codes. This includes: extracting the physical and spatiotemporal features of the product. The physical features include the product's core invariant features and dynamic response features that change dynamically with the environment. Nonlinear deformation field modeling is used to correct feature deviations caused by product deformation and to calculate deformation confidence. The intensity of environmental change is determined by combining the temperature and humidity changes in the product's environment. The weights of the core invariant features and dynamic response features are dynamically adjusted based on the intensity of environmental change and the deformation confidence. The spatiotemporal features include sub-millimeter spatial coordinates, nanosecond timestamps, and multi-source spatiotemporal data from various nodes in the distribution process during the product's production stage. After fusing and calibrating the multi-source spatiotemporal data, it is fused with the weighted physical features to generate a fused feature vector. Based on the fused feature vector, explicit and implicit QR codes are generated. The explicit QR code adopts a two-layer structure of a base layer and an additional layer. The base layer stores the product's unique identifier, and the additional layer stores the feature evolution prediction factor generated based on the core feature aging model and environmental response coefficient, as well as the hash feature summary of the implicit QR code. The implicit QR code is generated based on Logistic chaotic mapping and adopts a two-layer structure of a base layer and a dynamic layer. The base layer stores... The system generates parameters for the unique identifier of the product, the original hash of its core features, and the explicit QR code. An environmental response code segment, encrypted using TorusLWE lattice cryptography, is dynamically embedded in the layer. This environmental response code segment is generated from real-time environmental parameters and a hidden key. The explicit and implicit QR codes are associated and bound to the product's spatiotemporal information. The timestamps and spatial identifiers of the product's production and circulation are encoded to generate a spatiotemporal combination code. The hash values ​​of the fused feature vector, the explicit QR code, the implicit QR code, and the spatiotemporal combination code are calculated, ensuring that these hash values ​​satisfy algebraic topological homology group constraints and bitwise XOR chain constraints, forming a chain association relationship and storing it in a consortium blockchain with a main-side chain architecture. The generated explicit and implicit QR codes undergo layered verification. The user end verifies the core invariant features and combines AI forgery detection, using a first matching threshold adjusted based on a product lifecycle logarithmic decay model. The professional end verifies all physical features and the chain association relationship, combining damage detection, using a second matching threshold adjusted based on a lifecycle logarithmic decay model and a material correction coefficient. The first matching threshold is less than the second matching threshold.

[0007] By adopting the above technical solutions, the accuracy of feature extraction is improved through nonlinear deformation correction and multi-source spatiotemporal fusion, the security of dual codes is enhanced by combining chaotic mapping and quantum-resistant encryption, the chain stability is enhanced by using homology group constraints and main-side chain evidence storage, and the verification accuracy is improved by layered verification and AI detection. This comprehensively solves the problems of weak anti-copying, anti-tampering, and anti-transfer capabilities and low verification accuracy of existing technologies.

[0008] Furthermore, the method of correcting feature deviations caused by product deformation through nonlinear deformation field modeling includes: constructing a B-spline control mesh on the product surface, collecting displacement vectors of key feature points of the product, and fitting the deformation field matrix of the third-order B-spline basis function using the least squares method; based on the inverse matrix of the deformation field matrix, mapping the coordinates of the deformed product feature points back to the initial reference coordinate system to obtain equivalent initial coordinates and correcting deformation deviations; the deformation field matrix of the third-order B-spline basis function is fitted by collecting displacement vectors of more than 100 key feature points on the product surface using the optical flow method, and the constructed B-spline control mesh is a 3×3×3 three-dimensional control mesh; the equivalent initial coordinates are used to generate fused feature vectors, and the hash value of the fused feature vectors subsequently participates in the algebraic topological homology group constraint of the chain association relationship, wherein the second-order boundary operator image set is used to ensure that the chain association relationship maintains the stability of the overall association structure when the local features of the product undergo small changes due to environmental fluctuations.

[0009] By adopting the above technical solutions, the feature deviations caused by the nonlinear deformation of the product are accurately corrected, the specific implementation method of deformation field modeling and the role of the second-order boundary operator image set are clarified, and the accuracy of the fused feature vectors and the stability of the chain association relationship are further improved.

[0010] Furthermore, the fusion calibration of multi-source spatiotemporal data includes: the multi-source spatiotemporal data includes BeiDou high-precision positioning data, 5G base station triangulation positioning data and inertial navigation data; the extended Kalman filter algorithm is used to filter and denoise the multi-source spatiotemporal data; the optimal spatial coordinates and optimal timestamp are calculated based on the filtered multi-source spatiotemporal data to complete the spatiotemporal data fusion calibration.

[0011] By adopting the above technical solutions, multi-source data complementarity and filtering algorithms are used to eliminate spatiotemporal data noise, improve the accuracy of spatiotemporal information, and provide a high-quality data foundation for deep binding of features, dual codes, and spatiotemporal data.

[0012] Furthermore, the method for generating an implicit QR code based on a Logistic chaotic mapping includes: calculating the mean of the values ​​of each dimension of the fused feature vector, performing a modulo operation on the mean to obtain the initial value of the Logistic chaotic mapping; iterating the initial value a preset number of times and discarding transient values, selecting the iteration result to generate 128-bit encryption parameters for the implicit QR code, and encrypting the fused feature vector using a Torus LWE lattice cipher to generate the implicit QR code.

[0013] By adopting the above technical solution, and utilizing the sensitivity of chaotic mapping and the security of quantum-resistant encryption, the encryption parameters of the hidden QR code are deeply bound to the product features, which greatly improves the anti-copying and anti-cracking capabilities of the hidden QR code.

[0014] Furthermore, the hash value satisfies the algebraic topological homology group constraint by ensuring that the hash value of the fused feature vector, the hash value of the explicit QR code, the hash value of the implicit QR code, and the hash value of the spatiotemporal combination code satisfy the homology group constraint of "fused feature vector hash value - explicit QR code hash value - implicit QR code hash value - spatiotemporal combination code hash value ∈ 2nd order boundary operator image set", thereby ensuring the topological stability of the chain association relationship.

[0015] By adopting the above technical solution, the rigidity of the association between hash values ​​is strengthened through algebraic topological constraints, ensuring that even if local information fluctuates slightly, the overall structure of the chain relationship remains stable, thereby improving the ability to resist tampering.

[0016] Furthermore, the chain-like association stored in the consortium blockchain with a main-side chain architecture includes: the main-side chain architecture includes a main chain and multiple side chains, the side chains are split according to the circulation links and store the chain-like relationship fragments of the corresponding links, a Merkle tree is constructed for the chain-like relationship fragments of each side chain and the root hash is calculated; the root hash and the chain-like relationship of product production and retail terminal nodes are stored in the main chain to form a consortium blockchain with layered evidence storage.

[0017] By adopting the above technical solutions, we can achieve efficient hierarchical evidence storage of chain relationships, reduce the amount of data in the main chain, improve storage and query efficiency, and ensure the integrity of side chain data through Merkle trees, which facilitates full lifecycle traceability.

[0018] Furthermore, the user-side verification combined with AI forgery identification includes: the user-side collecting key dimension data of the product's core invariant features and calculating the entropy value of the feature vector; if the entropy value is lower than the preset lower limit of the natural feature entropy value, it is determined to be a suspected AI-generated forgery feature, triggering in-depth verification by the professional side.

[0019] By adopting the above technical solutions, AI counterfeiting detection is added to the rapid verification on the user end, which can promptly identify false features generated by AI, reduce the risk of counterfeit products passing the initial verification, and improve the comprehensiveness of verification.

[0020] Furthermore, the professional-end verification combined with damage detection includes: professional-end acquisition of the continuity and integrity indicators of the micro-texture on the product surface, calculation of the continuity coefficient of the micro-texture direction angle and the integrity coefficient of the gray value; if the continuity coefficient is lower than a preset continuity threshold or the integrity coefficient is lower than a preset integrity threshold, it is determined that the product carrier has been tampered with or transferred.

[0021] By adopting the above technical solutions, professionals can verify from both feature and carrier dimensions to accurately identify carrier tampering or transfer, further improving the verification accuracy in high-security scenarios.

[0022] Furthermore, the dynamic adjustment of the weights of the core invariant features and the dynamic response features includes: setting the initial weight of the core invariant features to 0.7 and the initial weight of the dynamic response features to 0.3; adjusting the weights based on the normalized value of the intensity of environmental change; when the intensity of environmental change approaches 1, the weight of the core invariant features decreases to 0.2 and the weight of the dynamic response features increases to 0.8; using the deformation confidence as a weight correction factor to perform a secondary correction on the adjusted weights; when the deformation confidence is lower than a preset threshold, increasing the weight ratio of the core invariant features.

[0023] By adopting the above technical solution, the feature weights can be adaptively adjusted according to the environment and deformation state, ensuring that the fused feature vector can balance stability and sensitivity in different scenarios, and improving the adaptability of feature extraction.

[0024] Furthermore, the generation of the environmental response code segment includes: real-time collection of temperature and humidity parameters of the environment in which the product is located, as well as the current nanosecond-level timestamp, to form an environmental parameter set; and encrypting the environmental parameter set using a preset implicit key based on the Torus LWE lattice cryptography algorithm to generate the environmental response code segment.

[0025] By adopting the above technical solution, the environmental response code segment can dynamically change with the real-time environment and has quantum-resistant security properties, ensuring that the copied static code cannot match the real-time environmental parameters and improving the anti-transfer capability.

[0026] In summary, this application has at least the following beneficial effects: 1. It provides a multi-dimensional integrated dual anti-counterfeiting security method, which significantly improves the ability to resist copying, tampering, and transfer; 2. By using nonlinear deformation correction and quantum-resistant encryption, the security of features and dual codes is enhanced, making it suitable for complex application scenarios; 3. By combining layered verification with AI detection, we balance the convenience and accuracy of verification to meet the needs of different security levels.

[0027] It should be understood that the description in the Summary Section is not intended to limit the key or essential features of the embodiments of this application, nor is it intended to restrict the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0028] The above and other features, advantages, and aspects of the embodiments of this application will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A schematic diagram of an exemplary operating environment in which embodiments of this application can be implemented is shown.

[0029] Figure 2A flowchart of a security anti-counterfeiting method based on dual anti-counterfeiting codes is shown in an embodiment of this application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0032] This application provides a secure anti-counterfeiting method based on dual anti-counterfeiting codes, which can improve the ability to resist copying, tampering, and transfer, while taking into account the accuracy of verification and quantum security. It is adaptable to scenarios such as product deformation and complex environments, helps to achieve high-security anti-counterfeiting and ensures the reliability of traceability throughout the entire life cycle.

[0033] Figure 1 A schematic diagram of an exemplary operating environment in which embodiments of this application can be implemented is shown.

[0034] Reference Figure 1 The operating environment includes a full-link macro-hardware system that supports the implementation of "a secure anti-counterfeiting method based on dual anti-counterfeiting codes". This system covers four core links: feature acquisition, data processing, evidence storage and transmission, and verification interaction. The hardware devices in each link realize data interaction and collaboration through wired (such as Ethernet) or wireless (such as 5G, WiFi) communication methods, forming a complete closed loop from raw data acquisition to final anti-counterfeiting verification, ensuring the stable implementation of core functions such as nonlinear deformation correction, multi-source spatiotemporal fusion, quantum-resistant encryption, main and side chain evidence storage, and layered verification.

[0035] The first part of the operating environment is a cluster of feature acquisition and preprocessing equipment. This cluster includes a high-resolution industrial camera (12 megapixels or higher, supporting sub-millimeter spatial resolution), a portable spectrometer, a 3D scanner, a BeiDou / 5G dual-mode positioning module, a nanosecond-level high-precision clock module, a temperature and humidity sensor (accuracy ±0.1℃ / ±1%RH), and a deformation detection sensor. Among them, the high-resolution industrial camera and the portable spectrometer are used to acquire physical features such as microscopic texture and material spectrum of the product surface. The 3D scanner captures the displacement vectors of more than 100 key feature points through optical flow to support nonlinear deformation modeling. The BeiDou / 5G dual-mode positioning module and the nanosecond-level clock module work together to obtain sub-millimeter-level spatial coordinates and nanosecond-level timestamps. The temperature and humidity sensor and the deformation detection sensor collect environmental parameters and product deformation status in real time. All acquisition devices are connected to the subsequent data processing equipment through data interfaces, transmitting the raw feature data to the data processing end for preprocessing.

[0036] The second part of the operating environment is a cluster of data processing and encryption computing devices, mainly composed of high-performance data processing servers and security encryption devices. The high-performance data processing servers are equipped with CPUs with more than 16 cores and GPU acceleration modules, used to run extended Kalman filter algorithms (to achieve multi-source spatiotemporal data fusion and denoising), Logistic chaotic mapping algorithms (to determine the initial value of implicit QR codes), LSTM-Transformer hybrid models (risk prediction), etc., and also support edge-cloud collaborative computing, which can distribute lightweight models to edge devices and receive parameter updates. The security encryption devices include quantum-resistant encryption chips (supporting TorusLWE lattice cryptography) and secure storage modules (encrypted hard disk / trusted execution environment TEE). The former is used to implement quantum-resistant encryption of the implicit QR code dynamic layer at the hardware level, and the latter is used to store sensitive data such as implicit code keys and chaotic mapping parameters. This cluster interacts with feature acquisition devices and blockchain evidence storage devices through communication links, receiving raw data and outputting core data such as processed fused feature vectors and dual code parameters.

[0037] The third part of the operating environment is the blockchain evidence storage and communication equipment cluster, which includes main chain node servers (no fewer than 3, distributed deployment), side chain node servers (deployed separately according to circulation links such as warehousing and transportation), distributed storage devices (such as HDFS distributed file system), and communication transmission equipment. The main chain node servers store the chain relationship of product production / retail terminal nodes and the Merkle root hash of each side chain, and run consortium blockchain consensus mechanisms such as PBFT to ensure data consistency. The side chain node servers store the chain relationship fragments of the corresponding circulation links and build Merkle trees to calculate root hashes. The distributed storage devices are used to retain the feature hashes of each node and the historical versions of the chain relationship. The communication transmission equipment includes IoT gateways and LAN / WAN communication modules. The IoT gateways connect to the sensing devices of circulation nodes and realize low-latency data transmission (latency ≤100ms). The LAN / WAN communication modules ensure the communication stability between blockchain nodes and between the blockchain and other devices. This cluster interfaces with data processing equipment to obtain chain relationship data and provides evidence storage query services to verification and interaction devices.

[0038] The fourth part of the operating environment is the circulation link sensing and edge computing device cluster, which consists of circulation node sensing devices and edge computing terminals. The circulation node sensing devices include RFID card readers and barcode scanners, used to collect information such as circulation node identification and node environmental parameters. The edge computing terminals are embedded servers or industrial tablets equipped with CPUs with more than 4 cores, and deploy lightweight models optimized by federated learning (such as lightweight EKF models and SPONGENT-256 hash algorithms). They can locally compute edge chain fragments and only upload fragment data to the blockchain storage device, reducing the amount of original data transmission. This cluster establishes a connection with the blockchain storage device and data processing device through an IoT gateway to realize the real-time uploading of circulation data and the synchronization of lightweight computing results.

[0039] The fifth part of the operating environment is a hierarchical verification terminal equipment cluster, divided into user-end verification equipment and professional-end verification equipment. The user-end verification equipment consists of smartphones equipped with cameras of 12 megapixels or higher and NFC functionality, supporting explicit QR code scanning and lightweight AI forgery recognition (such as feature entropy value calculation), which can quickly complete core invariant feature matching and suspected forgery warnings. The professional-end verification equipment includes a dedicated terminal integrating a high-precision barcode scanner and a computing terminal equipped with an AI acceleration module. The high-precision barcode scanner is used to collect micro-textures on the product surface to detect the continuity and integrity of the carrier, while the computing terminal runs a full feature matching algorithm and chain-like relationship verification logic. It can also generate AR visualization layers to assist in the presentation of verification results. This cluster connects to the blockchain evidence storage device via wireless communication to query chain-like relationship data to complete the verification judgment.

[0040] The various hardware devices work together: the feature acquisition device provides the raw data foundation, the data processing device completes the core algorithm operation and data encryption, the blockchain storage device ensures the immutability and traceability of the chain relationship, the circulation sensing device realizes the data supplementation throughout the entire life cycle, and the verification terminal device completes the final anti-counterfeiting verification. Together, they constitute the macroscopic equipment environment that supports the efficient and stable operation of "a secure anti-counterfeiting method based on dual anti-counterfeiting codes".

[0041] This application discloses a security anti-counterfeiting method based on dual anti-counterfeiting codes.

[0042] Figure 2 A flowchart of a security anti-counterfeiting method based on dual anti-counterfeiting codes is shown in an embodiment of this application.

[0043] Reference Figure 2 The method specifically includes the following steps: S1: Extract the physical and spatiotemporal characteristics of the product.

[0044] The physical features include the core invariant features of the product and the dynamic response features that change dynamically with the environment. The feature deviation caused by product deformation is corrected by nonlinear deformation field modeling and the deformation confidence is calculated. The intensity of environmental change is determined by the temperature and humidity changes in the environment in which the product is located. The weights of the core invariant features and the dynamic response features are dynamically adjusted according to the intensity of environmental change and the deformation confidence. The spatiotemporal features include sub-millimeter-level spatial coordinates, nanosecond-level timestamps and multi-source spatiotemporal data of each node in the product production stage. After the multi-source spatiotemporal data is fused and calibrated, it is fused with the weighted physical features to generate a fused feature vector.

[0045] In this step, the method of correcting the feature deviation caused by product deformation through nonlinear deformation field modeling includes constructing a B-spline control mesh for the product surface, collecting the displacement vectors of key feature points of the product, and fitting the deformation field matrix of the cubic B-spline basis function using the least squares method; the mathematical expression of the cubic B-spline basis function is as follows: ,in For parameter variables, For node vector values, For zero-order B-spline basis functions (in (The value is 1 within the interval and 0 elsewhere). Deformation field matrix The fitting process is achieved using the least squares method, and the objective function is... ,in For the first The initial coordinate vector of each feature point Let its displacement vector be , Number of feature points Based on the inverse matrix of the deformation field matrix, the coordinates of the deformed product feature points are mapped back to the initial reference coordinate system to obtain equivalent initial coordinates, thus correcting the deformation deviation. The deformation field matrix of the cubic B-spline basis function is fitted using the optical flow method by collecting displacement vectors from more than 100 key feature points on the product surface, and the constructed B-spline control mesh is... A three-dimensional control mesh; the equivalent initial coordinates are used to generate a fused feature vector, the hash value of which is subsequently used to participate in the algebraic topological homology group constraint of the chain association relationship, wherein the second-order boundary operator image set is used to ensure that the chain association relationship maintains the stability of the overall association structure when the local features of the product undergo minor changes due to environmental fluctuations.

[0046] The fusion calibration of multi-source spatiotemporal data includes BeiDou high-precision positioning data, 5G base station triangulation positioning data, and inertial navigation data. An extended Kalman filter algorithm is used to filter and denoise the multi-source spatiotemporal data. The state equation of the extended Kalman filter is: The observation equation is ,in for The state vector at any given time (containing spatial coordinates and timestamp). To control the input, and These are process noise and observation noise, respectively (both following a Gaussian distribution). and These are the state transition function and the observation function, respectively. During the filtering process, prediction is first performed using the state equation. Prior state at time and prior covariance Combined with observed values Calculate Kalman gain (in For the observation matrix, To observe the noise covariance, the posterior state is finally updated. and posterior covariance The optimal spatial coordinates and timestamps are calculated based on the filtered multi-source spatiotemporal data to complete the spatiotemporal data fusion calibration.

[0047] The dynamic adjustment of the weights for core invariant features and dynamic response features includes setting the initial weight of the core invariant features to 0.7 and the initial weight of the dynamic response features to 0.3, and adjusting the weights based on the normalized value of the environmental change intensity. The normalization formula is ,in The change in temperature This represents the change in humidity. and These are the preset maximum temperature change threshold and maximum humidity change threshold, respectively. The weighting adjustment formula is as follows: ,in As the core invariant feature weights, As dynamic response feature weights, when the intensity of environmental change... As the value approaches 1, the weight of the core invariant feature decreases to 0.2, while the weight of the dynamic response feature increases to 0.8; the deformation confidence level is... As a weighting correction factor, the deformation confidence is calculated using the condition number of the deformation field matrix. (in The condition number of the deformation field matrix is ​​denoted as . (The maximum condition number threshold is used as a preset threshold). The adjusted weights are then further corrected using the following formula: , When the deformation confidence level When the weight is below a preset threshold, the weight ratio of the core invariant features is increased.

[0048] When fusing weighted physical features with spatiotemporal features to generate a fused feature vector, the physical feature vector is first... (in For core invariant eigenvectors, (Dynamic response feature vector) and spatiotemporal feature vector (Including optimal spatial coordinates and optimal timestamp) are dimensionally aligned, and then vectors are concatenated to form a fused feature vector. The semicolon indicates a vertical splicing operation.

[0049] S2: Generate explicit and implicit QR codes based on the fused feature vectors.

[0050] The explicit QR code adopts a two-layer structure of a base layer and an additional layer. The base layer stores the product's unique identifier, while the additional layer stores the feature evolution prediction factor generated based on the core feature aging model and environmental response coefficients, as well as the hash feature summary of the implicit QR code. The calculation of the feature evolution prediction factor is based on the core feature aging model, which uses an exponential decay function to describe the change trend of the core feature over time. The mathematical expression is as follows: ,in for The core feature value predicted at time t. The initial core invariant eigenvalues ​​(core invariant features collected from the product manufacturing stage) Extracted from ( ) The core characteristic is the aging coefficient (set according to the material characteristics of the product, such as paper products). Heaven, metal products sky), The duration from product production to the present moment. and These are the environmental response coefficients for temperature and humidity (obtained from a material environmental response database, such as for plastic materials). ), and These are the current environment and the standard environment, respectively. The temperature and humidity differences; the hash feature digest of the hidden QR code is calculated using the SPHINCS+ quantum-resistant hash algorithm, the formula is as follows: ,in This is the base layer data for the hidden QR code. For Torus LWE lattice basis parameters (ring) polynomial degree Modulus The implicit QR code is generated based on Logistic chaotic mapping and adopts a two-layer structure of a base layer and a dynamic layer. The base layer stores the product's unique identifier, the original hash of the core feature, and the generation parameters of the explicit QR code. The original hash of the core feature is the core invariant feature. The hash value is calculated using the following formula: The generation parameters of the explicit QR code include the data length of the base layer and the update cycle of the feature evolution prediction factor of the additional layer; the dynamic layer embeds an environment response code segment encrypted based on Torus LWE lattice cryptography, which is generated by real-time environment parameters and the implicit code key.

[0051] In this step, generating an implicit QR code based on a Logistic chaotic mapping includes calculating the mean of the values ​​of each dimension of the fused feature vector. Include Dimensions (such as) (covering weighted physical and spatiotemporal characteristics), with values ​​for each dimension as follows: The formula for calculating the mean is: The initial value of the Logistic chaotic mapping is obtained by performing a modulo operation on the mean. The modulo operation is as follows: Ensure initial values Then, through a linear transformation, it is mapped to the effective iteration interval of the Logistic mapping [0.1, 0.9], i.e. The iterative formula for the Logistic chaotic mapping is: ,in For control parameters (range of values) Take this place To ensure strong chaotic characteristics, the initial value is iterated a preset number of times (e.g., 1000 times), and transient values ​​are discarded (the results of the first 100 iterations are discarded to avoid the influence of initial value fluctuations). The results of iterations from the 101st to the 1000th are selected, and a 128-bit binary number is extracted from them as the 128-bit encryption parameter of the implicit QR code. The fused feature vector is then encrypted using a Torus LWE lattice cipher. During the encryption process, the fused feature vector... As explicit Hidden key To obtain from discrete Gaussian distribution Sampling dimensional vector Public key From the ring Uniformly distributed sampling on the above, error term from Sampling, ciphertext ,in , This represents the vector inner product operation, generating an implicit QR code.

[0052] The generation of the environmental response code segment includes real-time acquisition of temperature and humidity parameters of the product's environment, as well as the current nanosecond-level timestamp, and temperature parameters. (unit: ), humidity parameters (Unit: %RH), nanosecond-level timestamp (Unit: ns), forming a set of environmental parameters. Based on the TorusLWE lattice cryptography algorithm, the set of environmental parameters is encrypted using a preset implicit key, which is the TorusLWE encryption key mentioned above. During encryption, the environmental parameter set is first... Convert to ring plaintext (Mapping the parameter values ​​to a linear transformation) of (range), then generate ciphertext according to the TorusLWE encryption process. , ,in For the newly sampled public key portion, For the error term of the new sampling, the ciphertext This is the environmental response code segment, a dynamic layer that embeds a hidden QR code.

[0053] S3: Associate and bind the explicit QR code and implicit QR code with the product's spatiotemporal information.

[0054] The product's production and distribution timestamps and spatial identifiers are encoded to generate a spatiotemporal combination code. The product production timestamp is... (Timestamps from the production stage collected by the nanosecond-level high-precision clock module in S1, unit: ns), timestamps of each node in the circulation process are... ( (This refers to the number of distribution nodes, such as warehousing and transportation nodes); the spatial identifier for the production stage is... (S1 Central Asian millimeter-level spatial coordinates, unit: mm), the spatial identifier of the circulation node is as follows (BeiDou at each node) (Dual-mode positioning data). During spatiotemporal combination coding, a spatiotemporal vector is first constructed. Then, a spatiotemporal combination code is generated through Base64 encoding and hash mapping. The formula is ,in For Torus LWE lattice basis parameters (ring) polynomial degree Modulus SPHINCS+ is a quantum-resistant hashing algorithm. It calculates the hash value of the fused feature vector, the hash value of the explicit QR code, the hash value of the implicit QR code, and the hash value of the spatiotemporal combination code, and then fuses the feature vector hash. (The fused feature vector generated for S1), explicit QR code hash ( (Data is the concatenation of the explicit QR code base layer and additional layers), and the implicit QR code hash. ( (Data concatenated from the base layer and dynamic layer of the implicit QR code), spatiotemporal combined code hash. The hash value should satisfy the algebraic topological homology group constraint and the chain constraint of bitwise XOR. The bitwise XOR constraint is: ( (For bitwise XOR operation), ensuring direct association between hash values; forming a chain-like association and storing it in a consortium blockchain with a main-side chain architecture.

[0055] During the product distribution process, the node characteristic information of each distribution node (warehousing, transportation, and distribution node) is acquired in real time. The node characteristic information includes the distribution node identifier ID and the node operation timestamp. (Nanosecond level), node environmental parameters (temperature) ,humidity ) and local feature data of products collected by node devices (e.g., product packaging integrity features collected at warehouse nodes); calculate the hash value of node feature information. ( (For TorusLWE lattice base parameters); hash the node features. Perform a hash overlay operation with the currently stored chained associations to update the chained associations: (For bitwise XOR operation), ensuring that the updated chain relation still satisfies the algebraic topological homology group constraint ( (The text appears to be a series of fragmented sentences and phrases, possibly related to a blockchain or similar system. A coherent translation is not possible without the full context.)

[0056] In this step of the method, the hash value satisfies the algebraic topological homology group constraint, which includes ensuring that the hash value of the fused feature vector, the hash value of the explicit QR code, the hash value of the implicit QR code, and the hash value of the spatiotemporal combination code satisfy the following formula: "Hash value of fused feature vector - hash value of explicit QR code - hash value of implicit QR code - hash value of spatiotemporal combination code". The homology group constraint of the image set of the order boundary operator ensures the topological stability of chain associations. In the algebraic topological homology group, the chain complex is defined as... ,in for Chain group of order (here we take the integer chain group) ), for 1st-order boundary operator; 2nd-order boundary operator Image set Let represent the set of all first-order chains obtained by mapping all second-order chains using boundary operators. Mapping four hash values ​​to elements of a second-order chain: , Then the constraint relationship can be expressed as ,Right now ( (for the kernel of the second-order boundary operator), and ( (Forming a second-order homology group for third-order boundary operators) This constraint ensures the equivalence of chain relationships in terms of topology—the overall association structure remains stable even if local hash values ​​change due to minor perturbations.

[0057] The chain-like relationships are stored in a consortium blockchain with a main-sidechain architecture. This architecture includes a main chain and multiple sidechains. Sidechains are segmented according to circulation stages (e.g., warehousing sidechains, transportation sidechains, distribution sidechains). Each sidechain stores a chain-like relationship fragment corresponding to that stage: the warehousing sidechain stores the chain-like relationships of warehousing nodes. (Storage time-space combination code hash) (Warehouse stage feature vector hashing) and node operation record hashing; the transportation sidechain stores the transportation nodes' hashes. , Etc. Construct a Merkle tree for the chained relation fragments of each sidechain and calculate the root hash. Taking the transportation sidechain as an example, the set of chained relation fragments is as follows: ( (For the number of segments in the transportation process), the Merkle tree construction steps are: 1) Generate leaf nodes like If it is an odd number, supplement. Make the number of nodes even; 3) Generate non-leaf nodes, and sort adjacent leaf nodes according to... The calculation proceeds layer by layer upwards until the root node is generated. (Transportation sidechain Merkle root hash). The root hash and the chained relationship between product production and retail terminal nodes are stored on the main chain. The production node data stored on the main chain includes... (Production phase fusion feature vector hashing) (Explicit QR code hash during production phase) (Hidden QR code hashes during production phase), retail terminal node data includes Simultaneously, it stores the Merkle root hash of each sidechain. This forms a layered evidence storage consortium blockchain—the sidechain stores detailed data at the stage level, while the main chain stores key node data and root hashes. This reduces the storage pressure on the main chain and ensures the integrity of sidechain data through the Merkle root hash.

[0058] S4: Perform layered verification on the generated explicit and implicit QR codes.

[0059] The user-side verification of core invariant features is combined with AI-based forgery detection, employing a first matching threshold adjusted based on a product lifecycle logarithmic decay model. The core of this model is dynamically adjusting the matching threshold based on the product's lifespan from production to verification, adapting to the natural aging of core invariant features over time. Let the initial value of the first matching threshold be... (Based on the feature matching accuracy settings at the time of product delivery, such as) , indicating the initial requirement (The above features match), product production time is (S1 converts nanosecond-level timestamps to days, unit: days), current verification time is (System time during user verification, unit: days), product default lifecycle is (Set according to product type, such as pharmaceuticals) Luxury goods (days), attenuation coefficient is (Controlling the threshold decay rate is based on statistical analysis of a large amount of product aging data, such as...) Then the formula for calculating the first matching threshold is: ,in The threshold is a natural logarithmic function, ensuring that it decays slowly over the lifespan and avoiding misjudgments due to normal aging. Professional verification of all physical features and chain relationships is performed, and combined with damage detection, a second matching threshold is adopted based on the logarithmic decay model of the lifespan and the material correction coefficient. The first matching threshold is less than the second matching threshold.

[0060] In this step, the user-side verification combined with AI forgery detection includes collecting key dimension data of the product's core invariant features from the user-side. The key dimensions of these core invariant features include inflection points of the product's surface texture and material spectral characteristics. A stable dimension (such as) ), forming feature vectors ,in For the first The feature values ​​of each dimension (such as texture inflection point spacing and spectral peak intensity) are used to calculate the entropy value of the feature vector. The entropy value is calculated using the Shannon entropy formula. First, the probability distribution of each feature value is statistically analyzed. (Based on a probability library of natural features of similar products stored locally on the user's device), then calculate the Shannon entropy. ,in Let the entropy be a logarithmic function with base 2. Reflecting the randomness of feature vectors—natural product features have higher entropy values, while AI-generated fake features have lower entropy values ​​due to their strong regularity; if the entropy value is lower than the preset lower limit of natural feature entropy values. (Based on statistical settings of 100,000+ natural product characteristics, such as...) If the result is not found, it is determined to be a suspected AI-generated forgery feature, triggering in-depth verification by professionals.

[0061] Professional-grade verification combined with damage detection includes professionally acquired data on the continuity and integrity of microtextures on the product surface. Microtexture continuity is measured using an angular continuity coefficient, while integrity is measured using a grayscale integrity coefficient. Data is collected from the product surface. Uniformly distributed micro-texture sampling points (e.g.) Record the micro-texture orientation angle of each sampling point. (Angle with the horizontal direction, unit: ) and grayscale value To calculate the continuity coefficient of the microtexture orientation angle, first calculate the variance of the orientation angle. (in (This is the mean of the direction angles), then define the continuity coefficient. ,in The maximum permissible variance is preset (e.g.) (Based on the product's original micro-texture test settings). A value closer to 1 indicates greater continuity in the micro-texture direction, with no obvious breaks or alterations. To calculate the integrity coefficient of the grayscale values, first obtain the standard micro-texture grayscale value vector stored during the product manufacturing stage. Then calculate the total grayscale value deviation. Define an integrity coefficient, where is the maximum grayscale value. The closer the value is to 1, the more consistent the grayscale value is with the original state, indicating no wear or tampering; if the continuity coefficient is lower than the preset continuity threshold... (like (or the integrity coefficient is lower than the preset integrity threshold) (like If so, it is determined that the product carrier has been tampered with or transferred.

[0062] The calculation of the second matching threshold in the professional section requires the addition of a material correction coefficient. Let the material correction coefficient be... (Set according to the material characteristics of the product, such as metal material) Paper material The more stable the material The closer to 1), the initial value of the second matching threshold is (Higher than the initial value of the first matching threshold, such as) The formula for calculating the second matching threshold is: ,make sure This meets the higher precision verification requirements of professional users; when professional users verify all physical features, they need to calculate the real-time acquired full physical feature vector. (Including core invariant features and dynamic response features) and the original physical feature vector stored in the production stage. cosine similarity (where • represents the vector inner product, (for L2 norm), if And the chain association verification passed (i.e.) If the hash value satisfies the constraints of the algebraic topological homology group, then the verification is considered successful.

[0063] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0064] This application utilizes a collaborative approach across the entire technical chain of "feature acquisition - dual-code generation - chained evidence storage - layered verification" to objectively derive a highly robust anti-counterfeiting effect throughout the entire lifecycle. The specific logical derivation is as follows: 1. Feature Acquisition and Preprocessing: Accurate data foundation supports robust anti-interference capabilities The system employs a combination of techniques: nonlinear deformation field modeling (B-spline control mesh + 3rd order basis function fitting), multi-source spatiotemporal fusion (BeiDou / 5G positioning + extended Kalman filtering), and dynamic weight adjustment (environmental change intensity + deformation confidence). Nonlinear deformation modeling corrects complex deformation deviations such as wrinkles and compression through inverse matrix mapping, avoiding misjudgments caused by deformation. Multi-source spatiotemporal fusion eliminates positioning and timestamp errors through filtering and noise reduction, ensuring sub-millimeter spatial accuracy and nanosecond-level temporal accuracy. Dynamic weight adjustment adapts the weights of core invariant features and dynamic response features to environmental fluctuations and deformation reliability, preventing single features from failing in extreme environments. The synergy of these three techniques directly leads to the generation of a fused feature vector possessing the attributes of "deformation independence, spatiotemporal precision, and environmental adaptability," providing a high-quality data foundation for subsequent dual-code generation and chain association, reducing the risk of anti-counterfeiting failure caused by feature interference from the source.

[0065] 2. Dual-code generation stage: Encryption and dynamic design enhance resistance to copying and cracking. The system employs a technical approach of "explicit code dual-layer structure (unique identifier in the base layer + feature evolution prediction factor in the additional layer) + implicit code chaotic mapping + TorusLWE quantum-resistant encryption + dynamic environmental response code segment": The explicit code's feature evolution prediction factor is generated based on a core feature aging model, adapting to the natural changes in features throughout their lifecycle and avoiding matching failures caused by aging; the implicit code uses Logistic chaotic mapping to transform the mean of the fused feature vector into an initial encryption value, deeply binding the encryption parameters to the product features, preventing independent copying; TorusLWE quantum-resistant encryption resists the risk of quantum computing cracking traditional encryption; the dynamic environmental response code segment changes with real-time temperature, humidity, and timestamps, making statically copied code segments unable to match real-time environmental parameters. The synergy of these four elements directly leads to the conclusion that the dual QR code possesses the attributes of "feature binding, quantum-resistant cracking, and dynamic anti-static copying," eliminating the possibility of counterfeit products achieving forgery by copying the QR code from the generation logic.

[0066] 3. Chain-based evidence storage: Constraints and layered design ensure tamper resistance and end-to-end traceability. The system employs a combination of techniques: bitwise XOR hash constraints, algebraic topological homology group constraints, hierarchical main-sidechain data storage (Merkle tree), and dynamic updates in the circulation chain. Bitwise XOR hash constraints ensure a rigid correlation between the hash values ​​of fused features, dual codes, and spatiotemporal combination codes; tampering with a single hash will destroy this correlation. Algebraic topological homology group constraints, through second-order boundary operator image sets, ensure that local feature fluctuations do not affect the overall chain structure stability. Hierarchical main-sidechain data storage stores fragmented data from the circulation process on the sidechain, while key nodes and root hashes are stored on the mainchain, reducing mainchain storage pressure and ensuring sidechain data integrity through the Merkle tree. Dynamic updates in the circulation chain overlay the feature hashes of each node onto the original chain relationship, completing the full lifecycle data. The synergy of these four techniques directly leads to the conclusion that the chain relationship possesses the attributes of "tamper-proof identifiability, structural stability, and end-to-end traceability," preventing chain data from being tampered with or broken at the data storage level, and achieving reliable traceability of products from production to retail.

[0067] 4. Layered Verification Process: Multi-dimensional Detection and Dynamic Thresholds Enable Precise Anti-counterfeiting Capabilities The system employs a technical approach combining "user-side AI counterfeit identification (feature entropy value judgment) + professional-side damage detection (micro-texture continuity / integrity coefficient) + lifecycle logarithmic decay threshold + material correction coefficient." On the user side, feature entropy values ​​distinguish between natural features and AI-generated counterfeit features, quickly intercepting low-level counterfeits. On the professional side, micro-texture coefficients determine whether the carrier has been tampered with or transferred, identifying high-level physical counterfeits. Dynamic thresholds adjust with lifecycle and material characteristics to avoid misjudgments caused by aging or material differences. The hierarchical design of the first / second thresholds balances user-side convenience with high-precision professional-side accuracy. The synergy of these four components directly leads to the following: the verification process possesses the attributes of "full coverage of counterfeit types, accuracy adapted to various scenarios, and low misjudgment rate," accurately distinguishing genuine from counterfeit products at the verification level, avoiding both missed counterfeits and misjudging genuine products due to environmental factors or aging.

[0068] In summary, the end-to-end technical means form a logical closed loop of "accurate data → secure dual codes → stable evidence storage → accurate verification". It can be objectively deduced that this application achieves the core anti-counterfeiting effect of "anti-copying, anti-tampering, anti-transfer, and anti-quantum cracking throughout the entire life cycle", while taking into account the convenience of verification and the adaptability of scenarios, providing reliable technical support for high-security anti-counterfeiting needs.

[0069] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A secure anti-counterfeiting method based on dual anti-counterfeiting codes, characterized in that, include: The product's physical and spatiotemporal features are extracted. The physical features include the product's core invariant features and dynamic response features that change dynamically with the environment. Nonlinear deformation field modeling is used to correct feature deviations caused by product deformation and to calculate deformation confidence. The intensity of environmental change is determined by combining the temperature and humidity changes in the product's environment. The weights of the core invariant features and dynamic response features are dynamically adjusted based on the intensity of environmental change and the deformation confidence. The spatiotemporal features include sub-millimeter-level spatial coordinates, nanosecond-level timestamps, and multi-source spatiotemporal data from various nodes in the product's production process. After fusing and calibrating the multi-source spatiotemporal data, it is fused with the weighted physical features to generate a fused feature vector. Based on the fused feature vector, explicit and implicit QR codes are generated. The explicit QR code adopts a two-layer structure of a base layer and an additional layer. The base layer stores the product's unique identifier, and the additional layer stores the feature evolution prediction factor generated based on the core feature aging model and environmental response coefficient, as well as the hash feature summary of the implicit QR code. The implicit QR code is generated based on Logistic chaotic mapping and adopts a two-layer structure of a base layer and a dynamic layer. The base layer stores the product's unique identifier, the original hash of the core feature, and the generation parameters of the explicit QR code. The dynamic layer embeds an environmental response code segment encrypted based on TorusLWE lattice cryptography. The environmental response code segment is generated by real-time environmental parameters and the implicit code key. The explicit and implicit QR codes are associated and bound with the product's spatiotemporal information. The timestamps and spatial identifiers of the product's production and circulation are encoded to generate a spatiotemporal combination code. The hash values ​​of the fused feature vector, the explicit QR code, the implicit QR code, and the spatiotemporal combination code are calculated to ensure that the hash values ​​satisfy the algebraic topological homology group constraint and the chain constraint relationship of bitwise XOR, forming a chain association relationship and storing it in the consortium chain of the main-side chain architecture. The generated explicit and implicit QR codes are verified in layers. The user side verifies the core invariant features and combines them with AI forgery identification, using a first matching threshold adjusted based on the product lifecycle logarithmic decay model. The professional side verifies all physical features and chain-like relationships, combines them with damage detection, and uses a second matching threshold adjusted based on the lifecycle logarithmic decay model and material correction coefficient. The first matching threshold is less than the second matching threshold.

2. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The method of correcting feature deviations caused by product deformation through nonlinear deformation field modeling includes: Construct a B-spline control mesh on the product surface, collect the displacement vectors of key feature points of the product, and fit the deformation field matrix of the 3rd order B-spline basis function using the least squares method; Based on the inverse matrix of the deformation field matrix, the coordinates of the deformed product feature points are mapped back to the initial reference coordinate system to obtain the equivalent initial coordinates and correct the deformation deviation. The deformation field matrix of the cubic B-spline basis function is fitted by acquiring the displacement vectors of more than 100 key feature points on the product surface using the optical flow method. The constructed B-spline control grid is a 3×3×3 three-dimensional control grid. The equivalent initial coordinates are used to generate the fused feature vector. The hash value of the fused feature vector is subsequently used to participate in the algebraic topological homology group constraint of the chain association relationship. The second-order boundary operator image set is used to ensure that the chain association relationship maintains the stability of the overall association structure when the local features of the product undergo small changes due to environmental fluctuations.

3. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, Fusion calibration of multi-source spatiotemporal data includes: The multi-source spatiotemporal data includes BeiDou high-precision positioning data, 5G base station triangulation positioning data, and inertial navigation data. The extended Kalman filter algorithm is used to filter and denoise the multi-source spatiotemporal data. Optimal spatial coordinates and timestamps are calculated based on the filtered multi-source spatiotemporal data to complete spatiotemporal data fusion calibration.

4. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The generation of implicit QR codes based on Logistic chaotic mapping includes: Calculate the mean of the values ​​of each dimension of the fused feature vector, and perform a modulo operation on the mean to obtain the initial value of the Logistic chaotic mapping; The initial value is iterated a preset number of times and the transient value is discarded. The iteration result is selected to generate a 128-bit encryption parameter for the implicit QR code. The TorusLWE lattice cipher is used to encrypt the fused feature vector to generate the implicit QR code.

5. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The hash value satisfies the following algebraic topological homology group constraints: The hash values ​​of the fused feature vector, the explicit QR code, the implicit QR code, and the spatiotemporal combination code satisfy the homology group constraint of "fused feature vector hash value - explicit QR code hash value - implicit QR code hash value - spatiotemporal combination code hash value ∈ 2nd order boundary operator image set", thus ensuring the topological stability of the chain association relationship.

6. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The chain-like relationships are stored in a consortium blockchain with a main-sidechain architecture, including: The main-sidechain architecture includes a main chain and multiple sidechains. The sidechains are split according to the circulation links and the chain relationship fragments of the corresponding links are stored. A Merkle tree is constructed for the chain relationship fragments of each sidechain and the root hash is calculated. The root hash and the chain relationship between product production and retail terminal nodes are stored on the main chain, forming a hierarchical evidence-gathering consortium chain.

7. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The user-side verification combined with AI forgery detection includes: The user end collects key dimension data of the core invariant features of the product and calculates the entropy value of the feature vector; If the entropy value is lower than the preset lower limit of natural feature entropy value, it is determined to be a suspected AI-generated fake feature, triggering in-depth verification by professionals.

8. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The professional-level verification combined with damage detection includes: The professional end collects the continuity and integrity index of the micro-texture on the product surface, and calculates the continuity coefficient of the micro-texture direction angle and the integrity coefficient of the gray value. If the continuity coefficient is lower than a preset continuity threshold or the integrity coefficient is lower than a preset integrity threshold, it is determined that the product carrier has been tampered with or transferred.

9. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The weights of the dynamically adjusted core invariant features and dynamic response features include: The initial weights of the core invariant features are set to 0.7 and the initial weights of the dynamic response features are set to 0.

3. The weights are adjusted based on the normalized value of the intensity of environmental change. When the intensity of environmental change approaches 1, the weights of the core invariant features are reduced to 0.2 and the weights of the dynamic response features are increased to 0.

8. The deformation confidence level is used as a weight correction factor to make a second correction to the adjusted weights. When the deformation confidence level is lower than the preset threshold, the weight ratio of the core invariant features is increased.

10. The security anti-counterfeiting method based on dual anti-counterfeiting codes according to claim 1, characterized in that, The generation of the environment response code segment includes: Real-time collection of temperature and humidity parameters of the product's environment, as well as the current nanosecond-level timestamp, to form a set of environmental parameters; Based on the Torus LWE lattice cryptography algorithm, the set of environmental parameters is encrypted using a preset implicit key to generate the environmental response code segment.

Images