A supply chain traceability anti-counterfeiting method and system based on rare earth physical fingerprint and fault-tolerant key binding

Abstract

The application discloses a supply chain traceability anti-fake method based on rare earth physical fingerprint and fault-tolerant key binding, and belongs to the technical field of supply chain traceability and anti-fake, which collects inherent and non-reproducible physical fingerprint characteristics of rare earth materials at key process nodes of rare earth raw materials and processed products; high-dimensional feature vectors are obtained through feature extraction and dimension reduction, and a fuzzy extractor (Fuzzy Extractor) in fault-tolerant cryptography is used to perform fault-tolerant processing on the physical fingerprint with measurement noise, to generate stable and reproducible keys and public auxiliary data, and also includes realizing the trinity of "object-code-chain" anti-fake and full-process traceability at the raw material level. Compared with the prior art, the application has the advantages that unique binding can be realized without adding labels on the materials, measurement noise and tampering attacks of the physical fingerprint can be resisted, and the authenticity identification ability and traceability reliability of the rare earth supply chain are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of supply chain traceability and anti-counterfeiting technology, specifically to a supply chain traceability and anti-counterfeiting method and system based on rare earth physical fingerprints and fault-tolerant keys. Background Technology

[0002] Rare earth elements are a collective term for seventeen metallic elements with unique electronic structures. Due to their excellent optical, electrical, and magnetic properties, they are widely used in key areas such as permanent magnet materials, luminescent materials, catalysis, hydrogen storage, polishing, and national defense, earning them the titles of "industrial vitamins" and important strategic resources. Rare earth resources have significant geographical characteristics and strategic value; their mining, separation, and processing are strictly controlled, placing high demands on the traceability and authenticity of the supply chain.

[0003] Existing supply chain traceability and anti-counterfeiting technologies primarily rely on external markings, such as barcodes, QR codes, RFID electronic tags, anti-counterfeiting inks, laser holographic labels, and physicochemical markers attached to products or packaging. A common characteristic of these methods is that the association between the "code" and the "product" is artificially added and separable: the label can be peeled off, copied, transferred, or counterfeited, and the label itself has no inherently unique physical connection to the rare earth material it marks. Once criminals transfer genuine labels to counterfeit, adulterated, or substandard rare earth materials, the traceability system will be unable to identify them, resulting in the problem of "genuine code, counterfeit goods."

[0004] Another type of technology attempts to use physically unclonable functions (PUFs) or inherent characteristics of materials as anti-counterfeiting criteria, such as using random textures or randomly distributed particles on the material surface as "physical fingerprints." However, physical fingerprints of rare earth materials inevitably suffer from measurement noise during collection: feature data collected from the same object at different times, with different devices, and under different environmental conditions are not completely consistent, exhibiting jitter and drift. If the original features are directly used as keys or unique identifiers, repeated measurements from the same source will be judged as inconsistent due to slight differences, leading to genuine products being mistakenly identified as fakes; if the matching threshold is relaxed, different objects will be mistakenly identified as the same, posing a security risk. This makes it difficult to directly combine physical fingerprints with cryptographic systems, limiting their application in high-security traceability scenarios.

[0005] Furthermore, existing traceability systems mostly use centralized databases to store traceability information, which presents problems such as data tampering at a single point, distrust among participating parties, and difficulty in sharing and reconciling data across entities. This makes it difficult to establish a reliable traceability chain covering raw materials to finished products and spanning multiple processing entities. In particular, the physicochemical state of rare earth materials changes during smelting, separation, purification, and processing, resulting in different physical fingerprints for the same batch of materials at different stages, further increasing the difficulty of full-process traceability.

[0006] In summary, existing technologies have the following shortcomings: First, the external labels lack an essential binding with the materials, making them easily transferable and copied, failing to solve the problem of "genuine codes for counterfeit goods"; second, physical fingerprints contain measurement noise, making it difficult to directly and stably bind them to cryptographic keys; third, centralized storage lacks sufficient credibility, making cross-entity traceability difficult to establish; and fourth, there is a lack of effective mechanisms to maintain traceability continuity for rare earth materials as processing procedures evolve. Therefore, there is an urgent need for a rare earth supply chain traceability and anti-counterfeiting method that can establish a physically unique binding at the raw material level, tolerate physical fingerprint measurement noise, and possess tamper-proof and end-to-end traceability capabilities. Summary of the Invention

[0007] The technical problem to be solved by this invention is to overcome the above-mentioned technical defects and provide a supply chain traceability and anti-counterfeiting method and system based on rare earth physical fingerprints and fault-tolerant key binding. It uses the non-replicable physical fingerprints of rare earth materials as the root of trust, uses the fuzzy extractor of fault-tolerant cryptography to solve the problem of key non-repeatability caused by physical fingerprint measurement noise, and combines blockchain to achieve tamper-proof full-process trusted traceability. Thus, it achieves the three-in-one anti-counterfeiting and traceability of "item-code-chain" at the raw material level, fundamentally solving the problems of "genuine code counterfeit goods" and insufficient trust in cross-entity traceability.

[0008] To solve the above-mentioned technical problems, the technical solution provided by this invention is: a supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprints and fault-tolerant keys, comprising the following steps: S1: At the process nodes of rare earth raw materials or rare earth processed products, collect the original data of the physical fingerprint inherent in the object to be traced, and obtain the physical fingerprint feature vector through feature extraction. S2: Use a fuzz extractor to perform a generation algorithm on the physical fingerprint feature vector, and output a stable key and public auxiliary data; S3: Obtain the traceability information of the object to be traced, cryptographically bind the traceability information with the stable key, and generate a digital signature; S4: Write the digital signature, the summary of traceability information, and public auxiliary data into the blockchain or distributed ledger; S5: Re-collect the physical fingerprint of the object to be traced at the verification node to obtain the verification feature vector. Combine the publicly available auxiliary data with the fuzz extractor to execute the regeneration algorithm and regenerate the verification key. When the verification key matches the stable key, the object is determined to be genuine and its full-process traceability information is output.

[0009] Preferably, the physical fingerprint includes at least one of rare earth element fluorescence emission spectra, trace element and isotope ratio combinations, crystal phase structure fingerprints, microscopic morphology textures, and surface random microstructure features.

[0010] Preferably, the feature extraction in S1 includes preprocessing and normalizing the original data to remove baseline drift and environmental noise; Feature point localization and dimensionality reduction are performed on the preprocessed data to obtain real-valued feature vectors of fixed dimensions; The real-valued feature vectors are quantized and encoded to obtain a discrete feature sequence suitable for processing by a fuzzy extractor.

[0011] Preferably, the fuzzy extractor in S2 is constructed based on a security probability and a strong random number extractor, wherein the security probability uses error-correcting codes to tolerate measurement noise of the physical fingerprint; The error correction code is one of BCH code, Reed-Solomon code or LDPC code, and its error correction capability t is set according to the intraclass Hamming distance distribution of repeated measurements of physical fingerprints.

[0012] Preferably, the security profile is constructed using code offsets, including: During the registration phase, codewords for error correction codes are randomly selected, and public auxiliary data is calculated based on the registration feature vector; During the regeneration phase, the codeword containing errors is calculated from the verification feature vector, and the codeword is recovered through error correction decoding to derive the stable key.

[0013] Preferably, the cryptographic binding in S3 involves using a stable key as the seed of the asymmetric signature private key or a symmetric key to sign or calculate the hash value of the traceability information and its preceding traceability records, forming a hash chain traceability record.

[0014] Preferably, the blockchain in S4 is a consortium blockchain, and the public auxiliary data is stored in an off-chain database and its hash value is uploaded to the blockchain.

[0015] Preferably, S5 further includes multi-factor fusion for anti-counterfeiting determination: weighted fusion of the physical fingerprint regeneration key verification result, the spatiotemporal logic consistency verification result of the on-chain traceability record, and the liveness or presence detection result of the object to be traced, outputting the authenticity confidence score, and triggering an early warning and manual re-inspection when the confidence score is lower than the threshold.

[0016] Another aspect of this invention discloses a rare earth processing supply chain traceability and anti-counterfeiting system, comprising: A physical fingerprint acquisition device is used to collect raw physical fingerprint data of rare earth materials; The feature processing and fuzzy extraction module is used to perform feature extraction and the generation and regeneration algorithms of the fuzzy extractor. The password binding and signing module is used to bind the regenerated key with traceability information and generate a signature; The blockchain evidence storage module is used to put signatures, traceability information digests, and publicly available auxiliary data indexes onto the blockchain; The verification terminal is used to re-collect physical fingerprints on-site and complete the authenticity determination and traceability information display.

[0017] The advantages of this invention compared to the prior art are: First, it achieves physical uniqueness binding at the raw material level. This invention uses the inherent, unreplicable, and inseparable physical fingerprint of rare earth materials as the root of trust, essentially binding the "code" and the "object" at the physical level, eliminating the need for tags that can be peeled off or transferred, thus fundamentally solving the problem of "genuine code, counterfeit goods".

[0018] Second, it resolves the contradiction between physical fingerprint noise and key stability. By introducing a fuzzy extractor and error-correcting code from fault-tolerant cryptography, repeated measurements from the same source can reliably regenerate the same key, while measurements from different sources cannot. This allows rare-earth physical fingerprints with measurement noise to be reliably integrated into the cryptographic system for the first time.

[0019] Third, it provides tamper-proof, end-to-end trusted traceability. By leveraging blockchain / distributed ledger and hash chain structures, cross-entity traceability records are immutable, reconcilable, and auditable, establishing a trusted chain covering the entire process from mining source to finished product.

[0020] Fourth, it balances security, privacy, and cost. Publicly disclosing auxiliary data does not reveal the key; the combination of off-chain storage and on-chain evidence hashing balances privacy protection, tamper resistance, and storage costs.

[0021] Fifth, it adapts to the characteristics of rare earth processing. Cross-process fingerprint association mapping and multi-factor fusion judgment enable the method to adapt to the actual situation of rare earth materials evolving with processing, and have good engineering applicability and scalability. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall process of the method of the present invention, showing the relationship between the steps of registration and collection, key generation, signature binding, on-chain evidence storage and verification and traceability; Figure 2 This is a schematic diagram of the processing flow of the fuzzy extractor (generating algorithm Gen and regenerating algorithm Rep) in this invention; Figure 3 This is a schematic diagram of the structure of the rare earth processing supply chain traceability chain (hash chain) of the present invention; Figure 4 This is a block diagram of the composition structure of the traceability and anti-counterfeiting system of the present invention. Detailed Implementation

[0023] The present invention will now be described in further detail with reference to the accompanying drawings.

[0024] Combined with appendix Figure 1-4As shown, a supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprints and fault-tolerant keys includes five main steps: registration and collection, key generation, signature binding, on-chain storage and verification traceability.

[0025] (1) Registration and data collection steps.

[0026] At each process node of rare earth raw materials or processed products, raw data of their inherent physical fingerprints are collected. These physical fingerprints originate from the physicochemical properties of the rare earth materials themselves, specifically including: the fluorescence emission spectra of rare earth ions (such as Eu³⁺, Tb³⁺, Dy³⁺, etc.) under excitation light of a specific wavelength, whose characteristic peak positions, full width at half maximum (FWHM), and relative intensities are determined by the rare earth composition and matrix, exhibiting high origin and batch specificity; the "element-isotope fingerprint" composed of trace element content and rare earth isotope ratios in the rare earth materials, reflecting inherent information about the mineral source and smelting process; the crystal phase structure fingerprint obtained by X-ray diffraction or Raman spectroscopy; and the microscopic morphology and surface random microstructure obtained by microscopic imaging. The collected raw data, after preprocessing, normalization, feature point localization, and dimensionality reduction, yields a fixed-dimensional feature vector, which is then quantized and encoded into a discrete feature sequence W.

[0027] (2) Key generation steps.

[0028] A fuzzy extractor, derived from fault-tolerant cryptography, is used to process the feature sequence W containing measurement noise. The fuzzy extractor consists of a secure sketch and a strong random number extractor. During the registration phase, the generation algorithm Gen(W) is executed, outputting a stable key R and public auxiliary data P. The secure sketch is constructed using the code offset of an error-correcting code (such as BCH, Reed-Solomon, or LDPC codes): a codeword C is randomly selected, and P = W⊕C is calculated. The strong extractor (constructed based on a universal hash or hash function) extracts a uniformly random stable key R from the recovered W. The public auxiliary data P does not reveal information about R and is only used to tolerate noise in the physical fingerprint during the regeneration phase.

[0029] (3) Binding signature steps.

[0030] The traceability information of the object to be traced is obtained, including at least batch identifier, process identifier, timestamp, and location identifier, and may also include extended information such as temperature, equipment number, and operating entity. A stable key R is used as the private key seed for asymmetric signatures or as the symmetric message authentication key to sign the traceability information and the hash values ​​of previous process traceability records, or to calculate message authentication codes. This links adjacent process records in a hash chain, forming a traceable chain covering the entire process from raw materials to finished products, achieving a cryptographic binding of "physical fingerprint—traceability information".

[0031] (4) On-chain evidence storage steps.

[0032] The digital signature, traceability information digest, and storage index of the public auxiliary data P are written into the blockchain or distributed ledger. A consortium blockchain is preferred, with participating parties such as mining, smelting and separation, processing and manufacturing, circulation, and regulation acting as nodes to jointly maintain the ledger. Considering privacy and storage costs, the public auxiliary data P can be stored in an off-chain database, with only its hash value uploaded to the blockchain, ensuring immutability while controlling the scale of on-chain data.

[0033] (5) Verify the traceability steps.

[0034] At the verification node, the physical fingerprint of the object to be traced is re-collected to obtain the verification feature sequence W′. Corresponding public auxiliary data P is obtained from the blockchain or database, and the regeneration algorithm Rep(W′,P) is executed using a fuzzy extractor: W′⊕P is calculated to obtain the erroneous codeword, which is then decoded to recover the codeword C and restore W. Finally, a strong extractor is used to obtain the regeneration key R′. When the difference between W′ and W is within the error correction capability range, R′=R, and the signature reconstructed from R′ matches the on-chain record, indicating it is genuine and outputting full-process traceability information; otherwise, an anomaly is detected and a warning is issued.

[0035] Furthermore, considering the predictable changes in the physical fingerprints of rare earth materials during smelting, separation, purification, and processing, this invention establishes a cross-process transfer mechanism: physical fingerprints are collected before and after key processes, establishing a "pre-process fingerprint—post-process fingerprint" association mapping and uploading it to the blockchain, thereby maintaining the continuity of the traceability chain as the physical fingerprints evolve with processing. Furthermore, the verification process can employ multi-factor fusion judgment, weighting and fusing the physical fingerprint regeneration verification results, the spatiotemporal logical consistency of on-chain records, and the object presence detection results to output a confidence level for authenticity, thus improving robustness.

[0036] Example 1: Anti-counterfeiting of rare earth oxide batches based on fluorescence emission spectroscopy physical fingerprints

[0037] This embodiment uses a batch of europium oxide (Eu2O3) powder produced by a rare earth separation company as the traceable object to illustrate the complete implementation process of the present invention.

[0038] Step 1: Registration and Data Acquisition. At the sampling port at the end of the separation process, a portable fluorescence spectrometer is used to irradiate the powder sample with excitation light at a wavelength of approximately 394 nm to acquire the characteristic fluorescence emission spectrum of Eu³⁺ (the main emission peaks are located around 590 nm and 615 nm). The acquired spectrum is preprocessed with baseline subtraction, intensity normalization, and wavelength calibration. The peak position, relative intensity, and peak shape parameters of several characteristic peaks are extracted. After dimensionality reduction by principal component analysis, a 256-dimensional real feature vector is obtained, which is then encoded into a discrete feature sequence W of length n = 511 bits according to a preset quantization interval.

[0039] Step 2, Key Generation. A security profile is constructed using a BCH(511, k, t) code with error correction capability t. Based on the intra-class Hamming distance distribution obtained from repeated measurements of the same batch of powder under different times and instrument conditions, and the inter-class Hamming distance distribution between different batches, the system selects the error correction capability t, ensuring that the intra-class distance of measurements from the same source has a high probability of being less than t, and the inter-class distance of measurements from different sources has a negligible probability of being less than t. During registration, a BCH codeword C is randomly selected, and the public auxiliary data P = W⊕C is calculated. A strong extractor (e.g., a hash function with random salt s as a parameter) is applied to the recovered W after error correction to obtain a stable key R of length 256 bits. The system records the public auxiliary data P and the salt s.

[0040] Step 3, Bind Signature. Collect the traceability information for this batch, including batch number, process identifier (separate process), timestamp, production line and location identifier, operating entity, etc., and organize it into a structured traceability record M. Derive a signing private key from the stable key R, and digitally sign M and the hash value H_prev of the preceding process record to obtain the signature Sig; H_prev concatenates adjacent process records into a hash chain.

[0041] Step 4, On-chain Evidence Storage. The hash value of M, the signature Sig, and the storage index of the public auxiliary data P (P and salt s are stored in an off-chain database, and their hash values ​​are uploaded to the blockchain) are packaged into a transaction and submitted to the consortium blockchain jointly maintained by the mining, separation, processing, circulation, and regulatory parties. After consensus, the transaction is written into the ledger, forming an immutable traceability record.

[0042] Step 5: Verification and Traceability. During the distribution or regulatory process, verification personnel re-collect fluorescence emission spectra of the same batch of powder and obtain the verification feature sequence W′ following the same procedure. They retrieve the corresponding P and salt s from the off-chain database, calculate W′⊕P to obtain the erroneous codeword, decode it using BCH to recover the codeword C, and then restore W. Finally, they obtain the regenerated key R′ using a strong extractor. If the batch is genuine and the measurement noise is within the error correction range, then R′=R. The signature reconstructed from R′ matches the on-chain Sig, the system determines it to be genuine, and displays the entire traceability information of the batch from its source to the current stage. If the sample has been swapped or adulterated, the difference between W′ and W exceeds the error correction capability, R′≠R, verification fails, and an alert is triggered.

[0043] Example 2: Multimodal anti-counterfeiting combining elemental-isotope fingerprints with microscopic textures

[0044] This embodiment targets sintered NdFeB permanent magnets and employs multimodal physical fingerprinting to enhance uniqueness and resistance to attacks. On one hand, inductively coupled plasma mass spectrometry (ICP-MS) is used to determine the content and isotope ratios of rare earth and trace elements, obtaining an element-isotope fingerprint reflecting the mineral source and smelting process. On the other hand, microscopic imaging is used to collect the grain boundaries and random microscopic textures formed during polishing on the magnet surface. These two types of features are extracted, normalized, and then concatenated and quantized into a discrete feature sequence W. A secure profile is constructed using Reed-Solomon or LDPC codes. The remaining key generation, signature binding, on-chain storage, and verification steps are the same as in Embodiment 1. The combined use of multimodal fingerprints makes it extremely difficult for attackers to simultaneously forge both the elemental isotope composition and the random surface texture of the material, thus significantly improving anti-counterfeiting strength.

[0045] Example 3: Cross-process fingerprint association and end-to-end traceability chain

[0046] Rare earth materials are obtained by smelting and separating oxides from ores, which are then processed into finished products through metal reduction, alloy smelting, and manufacturing. The physical fingerprints undergo predictable evolution between each process. In this embodiment, physical fingerprints are collected before and after each key process, and keys and public auxiliary data are generated for each process. A mapping record of "pre-process fingerprint --- post-process fingerprint" is established and uploaded to the blockchain. Simultaneously, the traceability records of adjacent processes are linked together using a hash chain: the record of process i contains the hash value of the record of process i-1, and is signed by the key generated from the physical fingerprint of process i. This forms a continuous hash chain from the source of the ore to the finished product. Tampering with any record at any stage will cause subsequent hash verification to fail. Verification can be performed on a single process for authenticity authentication, or by tracing back along the hash chain to reconstruct and verify the integrity and spatiotemporal consistency of the entire supply chain.

[0047] Further explanation of the fuzz extractor and parameter selection

[0048] The security and availability of a fuzzy extractor depend on the choice of error correction capability t. Let d_in be the upper quantile of the intra-class Hamming distance distribution for repeated measurements from the same source, and d_out be the lower quantile of the inter-class Hamming distance distribution for dissimilar measurements. t should be chosen to satisfy d_in ≤ t < d_out, thus achieving a balance between the false acceptance rate (FAR) and the false rejection rate (FRR). When a single error-correcting code cannot simultaneously meet the requirements of fault tolerance and security, concatenated codes, list decoding, or multi-block independent error correction followed by fusion can be used to improve performance. A strong extractor is used to compress the error-corrected fingerprint into a uniformly distributed key, ensuring that even if the public auxiliary data P is leaked, an attacker cannot deduce the key R from it, thus satisfying the security definition of a fuzzy extractor.

[0049] System Composition and Deployment Description

[0050] like Figure 4 As shown, the traceability and anti-counterfeiting system implementing the method of this invention includes: a physical fingerprint acquisition device (fluorescence spectrometer, mass spectrometer, microscopic imaging or diffraction equipment, etc.), a feature processing and fuzzy extraction module, a password binding and signature module, a blockchain evidence storage module, and a verification terminal. Each module can be deployed at enterprise production sites, regulatory agencies, and distribution channels, working collaboratively through a consortium blockchain. The acquisition device is responsible for obtaining the original physical fingerprint; the feature processing and fuzzy extraction module completes feature extraction and key generation / regeneration; the password binding and signature module completes the binding and signing of the key and traceability information; the blockchain evidence storage module is responsible for uploading and querying the blockchain; and the verification terminal allows on-site personnel to determine authenticity and display traceability information.

[0051] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

[0052] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0053] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding, characterized in that: Includes the following steps: S1: At the process nodes of rare earth raw materials or rare earth processed products, collect the original data of the physical fingerprint inherent in the object to be traced, and obtain the physical fingerprint feature vector through feature extraction. S2: Use a fuzz extractor to perform a generation algorithm on the physical fingerprint feature vector, and output a stable key and public auxiliary data; S3: Obtain the traceability information of the object to be traced, cryptographically bind the traceability information with the stable key, and generate a digital signature; S4: Write the digital signature, the summary of traceability information, and public auxiliary data into the blockchain or distributed ledger; S5: Re-collect the physical fingerprint of the object to be traced at the verification node to obtain the verification feature vector. Combine the publicly available auxiliary data with the fuzz extractor to execute the regeneration algorithm and regenerate the verification key. When the verification key matches the stable key, the object is determined to be genuine and its full-process traceability information is output.

2. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The physical fingerprint includes at least one of rare earth element fluorescence emission spectra, trace element and isotope ratio combinations, crystal phase structure fingerprints, microscopic morphology textures, and surface random microstructure features.

3. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The feature extraction in S1 includes preprocessing and normalizing the original data, and removing baseline drift and environmental noise; Feature point localization and dimensionality reduction are performed on the preprocessed data to obtain real-valued feature vectors of fixed dimensions; The real-valued feature vectors are quantized and encoded to obtain a discrete feature sequence suitable for processing by a fuzzy extractor.

4. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The fuzzy extractor in S2 is constructed based on a security probability and a strong random number extractor. The security probability uses error-correcting codes to tolerate measurement noise in the physical fingerprint. The error correction code is one of BCH code, Reed-Solomon code or LDPC code, and its error correction capability t is set according to the intraclass Hamming distance distribution of repeated measurements of physical fingerprints.

5. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 4, characterized in that: The security profile is constructed using code offsets and includes: During the registration phase, codewords for error correction codes are randomly selected, and public auxiliary data is calculated based on the registration feature vector; During the regeneration phase, the codeword containing errors is calculated from the verification feature vector, and the codeword is recovered through error correction decoding to derive the stable key.

6. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The cryptographic binding in S3 involves using a stable key as the seed of the asymmetric signature private key or a symmetric key to sign or calculate the hash value of the traceability information and its preceding traceability records, forming a hash chain traceability record.

7. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The blockchain in S4 is a consortium blockchain, and the public auxiliary data is stored in an off-chain database and its hash value is uploaded to the blockchain.

8. The supply chain traceability and anti-counterfeiting method based on rare earth physical fingerprinting and fault-tolerant key binding according to claim 1, characterized in that: The S5 also includes multi-factor fusion for anti-counterfeiting determination: the physical fingerprint regeneration key verification result, the spatiotemporal logic consistency verification result of the on-chain traceability record, and the liveness or presence detection result of the object to be traced are weighted and fused to output the authenticity confidence score. When the confidence score is lower than the threshold, an early warning and manual re-inspection are triggered.

9. A rare earth processing supply chain traceability and anti-counterfeiting system, using the method described in any one of claims 1-8, characterized in that: include: A physical fingerprint acquisition device is used to collect raw physical fingerprint data of rare earth materials; The feature processing and fuzzy extraction module is used to perform feature extraction and the generation and regeneration algorithms of the fuzzy extractor. The password binding and signing module is used to bind the regenerated key with traceability information and generate a signature; The blockchain evidence storage module is used to put signatures, traceability information digests, and publicly available auxiliary data indexes onto the blockchain; The verification terminal is used to re-collect physical fingerprints on-site and complete the authenticity determination and traceability information display.

Images