Automobile parts multi-source fusion tracking method and system based on distributed block chain

Abstract

The application discloses a kind of based on distributed block chain's automobile parts multi-source fusion traceability method and system.First, build blockchain network and traceability platform;Then collect the microcosmic visual characteristic image on the surface of automobile parts, judge texture type and extract features, and combine RFID tag to extract electromagnetic field intensity distribution characteristics, identify electromagnetic fingerprint through neural network;Second, the multi-source characteristic data is bound with production, transportation information identifier and stored in distributed block chain network, using five-layer Merkle tree storage structure to ensure that data is not tampered with;Finally, build the traceability query module of three-layer architecture, and automatically manage data writing and query through smart contract.The application combines multi-feature fusion and blockchain technology, solves the problem of information silos and weak traceability capability in traditional traceability system, realizes the trusted traceability and authenticity identification of automobile parts throughout the life cycle, and improves the data security and query efficiency of supply chain.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of blockchain technology application and Internet of Things traceability, and particularly relates to a multi-source fusion traceability method and system for automobile parts based on a distributed blockchain. BACKGROUND

[0002] Counterfeiting and substandard products of automobile parts have always existed and have been a serious problem. Some unscrupulous businessmen produce counterfeit automobile parts with low-quality materials or processes, which may be very similar to the appearance and logo of the genuine product, but the quality and performance often cannot meet the standard of the genuine product, and it is difficult to meet the requirements of safety and reliability.

[0003] It is difficult to solve the traceability problem of automobile parts simply by relying on one product one code. The object fingerprint identification technology can compare the photographed automobile parts with the product texture image data registered in the cloud by the brand in advance, and complete high-precision identification in a very short time. The metal, fiber, glass, leather or plastic parts used in automobile parts have their special textures that cannot be seen by the naked eye, which are called object fingerprints.

[0004] Currently, there is a traceability identification technology based on texture fingerprint identification technology, and the manual identification factor is still large, and the main identification result still depends on the judgment of experienced appraisers. In automobile parts, hidden electromagnetic fingerprint information can be collected by specific sensors. These sensors usually include magnetoresistance sensors, Hall sensors, etc., which can detect the response of the parts in the magnetic field and convert the response into electrical signals. Subsequently, these electrical signals are received and digitized by a data collector, and by analyzing the preprocessed signals, various features reflecting the electromagnetic characteristics of the automobile parts can be extracted, such as frequency, amplitude, phase, waveform, etc. These features can reflect the true characteristics and potential differences of the parts.

[0005] The openness, security and uniqueness of the blockchain make it naturally meet the needs of reliable traceability. After collecting the key information of the product, the key information is saved on the blockchain, and by maintaining a permanent and unchangeable blockchain data network based on time stamp records, on the one hand, the uniqueness and authenticity of the key information can be established, and on the other hand, by using multi-source data fusion technology, heterogeneous traceability data of each link and stage are analyzed through the traceability label of the automobile parts, hidden electromagnetic fingerprint characteristics and commodity characteristics, and combined to generate a traceable history record of the automobile parts, meeting the various needs of reliable traceability. This technology can also be used for similarity analysis of the appearance and structure of automobile parts, and through similarity analysis, it can determine whether the automobile parts have problems such as infringement of intellectual property rights, effectively protecting the healthy development of the automobile parts industry. SUMMARY

[0006] This invention addresses the problems of information silos and data tampering in existing automotive parts traceability technologies. It proposes a multi-source fusion traceability method and system for automotive parts based on distributed blockchain. By integrating microscopic visual features with RFID electromagnetic fingerprints, production information, and transportation information, and combining blockchain distributed storage and smart contract technology, it achieves reliable recording, tamper-proof traceability, and automated authenticity verification of parts throughout their entire lifecycle, effectively improving supply chain management efficiency.

[0007] The technical solution to achieve the purpose of this invention is:

[0008] A multi-source fusion traceability method for automotive parts based on distributed blockchain, characterized by the following steps:

[0009] Step S1: Construct the blockchain network module:

[0010] The various business entities in the automotive parts supply chain are added to the blockchain as nodes. These business entities include upstream suppliers, manufacturers, downstream distributors, retailers, and customers.

[0011] In the blockchain network, an authentication center is set up, with each supplier, manufacturer, and distributor acting as an independent authentication center to authenticate the authenticity of product information and the rationality of decisions in the automotive parts supply chain.

[0012] A multi-source data fusion and analysis module is constructed as a node to join the blockchain network module. The multi-source data fusion and analysis module is used to receive information collection requests transmitted by various business entities in the supply chain, determine the dataset according to the requests, collect and process the data, and then return the processing results to various business entities in the supply chain.

[0013] Step S2: Construct a multi-source fusion traceability platform module for automotive parts. The multi-source fusion traceability platform module for automotive parts is used to record information of each link in the automotive parts supply chain and uses timestamps to record the time sequence.

[0014] Step S3: Multi-source data collection constitutes the unique identity information of automotive parts:

[0015] Collect microscopic visual feature images of the surface of automotive parts materials and establish a dataset containing microscopic visual feature images of the surface of various automotive parts materials.

[0016] Using a multi-antenna array, the electromagnetic field strength distribution characteristics of hidden electromagnetic fingerprint RFID tags on the outer packaging of automotive parts are collected by an RFID tag reader, and the extracted electromagnetic fingerprint data is analyzed and classified by a neural network algorithm.

[0017] Collect production information, transportation information, inspection information, and import / export information;

[0018] Step S4, Multi-source data fusion: The data collected in step S3 is sent to the multi-source data fusion analysis module, and the multi-source data fusion analysis module is used to process and fuse the data;

[0019] Step S5, Data Storage on the Blockchain: The merged data is written to the blockchain through a smart contract module pre-deployed on the blockchain. The smart contract is a computer protocol that automatically executes the contract terms and is used to realize the automatic verification and storage of data. Each transaction record is packaged into a block, which contains the hash value of the current block, the hash value of the previous block, and timestamp information.

[0020] Step S6, Traceability Query: The user inputs relevant information about the parts through the multi-source fusion traceability platform module for automotive parts. The multi-source fusion traceability platform module for automotive parts searches for matching traceability data on the blockchain according to the query rules in the smart contract module.

[0021] Furthermore, the blockchain network in step S1 adopts a distributed storage structure, which is used to store different types of basic data. Each type of data is mapped to the block body of a corresponding block in the blockchain. The distributed storage data structure consists of the following five parts:

[0022] Data layer: refers to supply chain management related data;

[0023] Categorized Merkle Tree Layer: Data from different categories form different Merkle trees, which do not overlap. This layer is used to store the Merkle trees.

[0024] Classified Merkle tree root layer: This layer stores the roots of various different Merkle trees;

[0025] Directory layer: This layer constructs a directory based on the basic data corresponding to different categories of Merkel roots, and calculates another Merkel root for each category of Merkel root. The number of categories of Merkel roots can be selected according to actual needs.

[0026] Block header layer: Calculate the total Merkle root of all Merkle roots in the directory layer. The block header will contain this root, and includes, but is not limited to, the block header hash of the previous block, the version number of this block, consensus authentication parameters, and timestamp.

[0027] Furthermore, the automotive parts multi-source fusion traceability platform module in step S2 adopts a three-layer architecture design, namely a data application layer, a data management layer, and a data acquisition layer. Among them, the data application layer provides traceability information query services for consumers and suppliers, the data management layer is responsible for managing the data storage, processing, and maintenance required by various business functions, and the data acquisition layer is responsible for collecting and acquiring relevant data during the automotive parts production process. Moreover, in the platform, each business entity only has the permission to use the distributed storage blockchain to store the data in a distributed manner.

[0028] Furthermore, in step S3, the microscopic visual feature images of the collected automotive component material surface are preprocessed, including image segmentation, obtaining the orientation pattern, enhancement and filtering, binarization and thinning; and the texture type is determined based on its two-dimensional frequency domain features, and an appropriate algorithm is selected to extract and analyze the key attribute feature points of the texture features; the texture types include continuous, discontinuous, and contour types. For continuous textures, the Poincaré exponent algorithm is selected; for discontinuous textures, the Gray-Level Co-occurrence Matrix (GLCM) algorithm is selected; and for contour textures, the Freeman chain code and its derivative feature extraction algorithm are used; the extracted texture features are analyzed and matched using the deep learning framework CNN.

[0029] Furthermore, in step S3, the step of collecting the electromagnetic field intensity distribution characteristics of the hidden electromagnetic fingerprint RFID tag includes:

[0030] Step S3-1: Demodulate the RFID tag signal received by the RFID tag reader using IQ modulation to obtain the I-channel and Q-channel signals, and then calculate the amplitude signal by modulus. ,in, , which are the sampling points;

[0031] Step S3-2, for Clustering is performed to obtain the center point vectors corresponding to code symbols 0 and 1. ,in The cluster center point corresponding to code element 0. For the cluster center point corresponding to code 1, the clustering function is expressed as: ,in, Clustering operation functions;

[0032] Step S3-3: Determine based on clustering of silent period signal and EPC signal and The desired signal is obtained through Euclidean distance determination. ,in, , The Euclidean distance decision function;

[0033] Step S3-4: For the amplitude signal Standardization to eliminate amplitude differences:

[0034] ,

[0035] in, Standard amplitude signal;

[0036] Step S3-5: Standard amplitude signal Remove desired signal To calculate noise signals :

[0037] ,

[0038] in, This is a noise signal;

[0039] Step S3-6: For the desired signal Standard amplitude signal Noise signals and the original signal Cross-spectral entropy and related features are calculated by pairwise combination, and cross-spectral entropy features are extracted. Defined as:

[0040] ,

[0041] in, For signal In frequency The probability density function of the energy proportion at that location:

[0042] ,

[0043] The cross-power spectral density, based on the Wiener-Khinchin theorem, is expressed as:

[0044] ,

[0045] in, For signal and The cross-power spectrum reflects the correlation between signals in the frequency domain. for conjugate, Indicates Fourier transform, For signal and The cross-correlation function is used to evaluate the cross-correlation between two signals. The degree of similarity, For time delay variables;

[0046] From the above equation, the cross-spectral entropy can be obtained as:

[0047] ,

[0048] in, For signal and Cross-spectral entropy, as a feature of RFID tags for automotive parts; is the probability density function of the energy proportion of the cross-power spectrum;

[0049] At the same time, the maximum amplitude is introduced. Maximum phase Maximum frequency , and cross-spectral entropy These characteristics are collectively used as features of RFID tags for automotive parts; their expression is:

[0050] .

[0051] Furthermore, in step S3, a neural network algorithm is used to process the multi-source fusion features of the components extracted in steps S3-6, specifically including: maximizing the amplitude of the multi-source fusion features of the components. Maximum phase Maximum frequency , and cross-spectral entropy As input to the neural network, the network is trained using a backpropagation algorithm optimized with momentum terms. After training, the network is used to classify and reason about the fused features of the input, outputting the authenticity of the parts and the corresponding traceability information, and writing the results into a distributed blockchain for evidence storage.

[0052] Furthermore, in step S5, the smart contract module completes decision-making activities and performs distributed ledger recording through decentralized consensus authentication, making the smart contract and product data traceable; in the storage stage, the user calls the smart contract to trace product information up to the chain; the uploaded information will be recorded in the blockchain ledger and a transaction hash address will be returned as the unique identifier of the information in the blockchain.

[0053] Furthermore, each business entity only has the authority to access the distributed storage blockchain. Data storage operations can only be performed after their identity data is encrypted with an encryption key. They have no right to store business-related data that has a mapping relationship with the blockchain. The encryption key is collaboratively managed and backed up by multiple authoritative automotive parts industry organizations.

[0054] When verifying the authenticity of automotive parts, the relevant product data is decrypted using the encryption key provided by the supply chain entity and the official backup key, and the consistency of the two decryption results is compared to determine whether the supply chain entity has tampered with the product data key.

[0055] Furthermore, the automotive parts supply chain is built on an industry-specific network or the Internet, and requires that only business entities in the supply chain join the blockchain system through authorization.

[0056] A multi-source fusion traceability system for automotive parts based on distributed blockchain, applied to a multi-source fusion traceability method for automotive parts based on distributed blockchain, characterized by comprising:

[0057] Blockchain network module: Used to build a blockchain network, adding certification centers, suppliers, manufacturers, distributors, and multi-source data fusion and analysis centers as nodes to achieve information sharing and supervision;

[0058] Automotive Parts Multi-Source Fusion Traceability Platform Module: Used to build an automotive parts multi-source fusion traceability platform;

[0059] Multi-source data acquisition module: used to collect information on automotive parts at various stages, including production time, batch, raw material source, processing technology, transportation route, and sales records;

[0060] Multi-source data fusion and analysis module: used to process and fuse the collected data to form complete traceability information;

[0061] Data on-chain module: Used to write the merged data into the blockchain via smart contracts, ensuring that the data is immutable and traceable;

[0062] Traceability Query Module: This module receives user query requests, searches for matching traceability data on the blockchain according to the query rules in the smart contract, and displays it to the user.

[0063] Smart contract module: used to realize the traceability and automatic collection of automotive parts supply chain information, and to complete decision-making activities and perform distributed ledger through decentralized consensus authentication based on the automatic operation of blockchain.

[0064] Compared with the prior art, the present invention, employing the above technical solution, has the following beneficial effects:

[0065] (1) By integrating the dual physical features of material surface microscopic visual features and RFID electromagnetic fingerprints, a composite identification mechanism of physical fingerprint and digital identity is formed. Microscopic visual features are extracted based on different algorithms adapted to texture type, and electromagnetic fingerprints are analyzed through neural networks to analyze the distribution characteristics of electromagnetic field intensity. After the two are bound with production and transportation information, an uncopyable component identification is formed, which solves the problem of easy counterfeiting and imitation of traditional single tags from the technical principle.

[0066] (2) A blockchain distributed network architecture is adopted, and the data of each link in the supply chain is stored through a five-layer Merkle tree. Each block contains the hash value and timestamp of the previous block to build a chain-like encrypted structure. This ensures that the data cannot be tampered with once it is on the chain. At the same time, through the collaborative authentication mechanism of each business entity as an independent authentication center, the information silos and data tampering risks in the traditional traceability system are solved, and a trusted data chain for the entire supply chain process is established.

[0067] (3) Through the pre-deployed smart contract module, the automatic verification, storage and distributed ledger of data on the chain can be realized, and the real-time on-chain of production, transportation and inspection information can be completed without manual intervention. In the traceability process, the smart contract supports fast retrieval based on unique identifiers and cross-spectral entropy feature comparison. Combined with the collaborative work of the three-layer architecture platform, efficient query and automated authenticity determination are achieved, which greatly reduces the time cost of traditional manual verification.

[0068] (4) By using blockchain timestamp technology and distributed ledger, the information of the entire process of parts from production to consumption is recorded in sequence, supporting bidirectional queries for forward traceability and reverse recall. The three-layer architecture platform further integrates data collection, management and application services, providing a unified information sharing and supervision interface for various business entities, and solving the problems of data asynchrony across links and difficulty in defining responsibilities in traditional systems. Attached Figure Description

[0069] Figure 1 This is a flowchart of the multi-source fusion traceability method for automotive parts based on distributed blockchain proposed in this invention;

[0070] Figure 2 A diagram illustrating the architecture of an automotive parts supply chain based on distributed blockchain.

[0071] Figure 3 This is a diagram illustrating the overall architecture of a multi-source integrated traceability platform for automotive parts.

[0072] Figure 4 This is a diagram of the distributed storage data structure of blocks in a blockchain-based supply chain management system.

[0073] Figure 5 A schematic diagram of a data storage method for certification in the automotive parts supply chain;

[0074] Figure 6 This is a microscopic scanning experiment of genuine and counterfeit automotive ceramic brake pads in Example 1.

[0075] Figure 7 The image shows the microscopic features of the genuine automotive ceramic brake pad product in Example 1.

[0076] Figure 8 The image shows the microscopic features of the pseudo-automotive ceramic brake pad product in Example 1.

[0077] Figure 9 This is an enhanced image of the microscopic features of the genuine automotive ceramic brake pad product in Example 1;

[0078] Figure 10 This is an enhanced image of the microscopic features of the pseudo-automotive ceramic brake pad product in Example 1;

[0079] Figure 11 This is a physical image of the radio frequency identification test equipment in Example 2;

[0080] Figure 12 This is a diagram of the 3×3 multi-label rectangular array experimental setup in Example 2. Detailed Implementation

[0081] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0082] A multi-source fusion traceability method for automotive parts based on distributed blockchain, the process of which is as follows: Figure 1 As shown, the feature is that it includes the following steps:

[0083] Step S1: Construct the blockchain network module:

[0084] The various business entities in the automotive parts supply chain are added to the blockchain as nodes. These business entities include upstream suppliers, manufacturers, downstream distributors, retailers, and customers.

[0085] Establish a certification center in the blockchain network, such as Figure 2 As shown, Figure 2 In this context, AC stands for Certification Center. Each supplier, manufacturer, and distributor in the entire automotive parts supply chain acts as an independent certification center, certifying the authenticity of product information and the rationality of decisions made within the supply chain. By distributing automotive parts supply chain information data within a blockchain system, and allowing authorized entities to access this data, the probability of counterfeit or substandard automotive parts being used as substitutes or "inferior goods being passed off as superior" can be effectively reduced.

[0086] A multi-source data fusion and analysis module is constructed as a node to join the blockchain network module. The multi-source data fusion and analysis module is used to receive information collection requests transmitted by various business entities in the supply chain, determine the dataset according to the requests, collect and process the data, and then return the processing results to various business entities in the supply chain.

[0087] Step S2: Construct a multi-source fusion traceability platform module for automotive parts, such as... Figure 3 As shown, the multi-source fusion traceability platform module for automotive parts is used to record information about each link in the automotive parts supply chain and uses timestamps to record the time sequence.

[0088] Step S3: Multi-source data collection constitutes the unique identity information of automotive parts:

[0089] Collect microscopic visual feature images of the surface of automotive parts materials and establish a dataset containing microscopic visual feature images of the surface of various automotive parts materials.

[0090] Using a multi-antenna array, the electromagnetic field strength distribution characteristics of hidden electromagnetic fingerprint RFID tags on the outer packaging of automotive parts are collected by an RFID tag reader, and the extracted electromagnetic fingerprint data is analyzed and classified by a neural network algorithm.

[0091] Collect production information, transportation information, inspection information, and import / export information;

[0092] Step S4, Multi-source data fusion: The data collected in step S3 is sent to the multi-source data fusion analysis module, and the multi-source data fusion analysis module is used to process and fuse the data;

[0093] Step S5, Data Storage on the Blockchain: The merged data is written to the blockchain through a smart contract module pre-deployed on the blockchain. The smart contract is a computer protocol that automatically executes the contract terms and is used to realize the automatic verification and storage of data. Each transaction record is packaged into a block, which contains the hash value of the current block, the hash value of the previous block, and timestamp information.

[0094] Step S6, Traceability Query: The user inputs relevant information about the parts through the multi-source fusion traceability platform module for automotive parts. The multi-source fusion traceability platform module for automotive parts searches for matching traceability data on the blockchain according to the query rules in the smart contract module.

[0095] Furthermore, the blockchain network in step S1 adopts a distributed storage structure, such as... Figure 3 As shown, the distributed storage structure is used to store different types of basic data, and each type of data is mapped to the block body of the corresponding block in the blockchain; the distributed storage data structure consists of the following five parts:

[0096] Data layer: refers to supply chain management related data;

[0097] Categorized Merkle Tree Layer: Data from different categories form different Merkle trees, which do not overlap. This layer is used to store the Merkle trees.

[0098] Classified Merkle tree root layer: This layer stores the roots of various different Merkle trees;

[0099] Directory layer: This layer constructs a directory based on the basic data corresponding to different categories of Merkel roots, and calculates another Merkel root for each category of Merkel root. The number of categories of Merkel roots can be selected according to actual needs.

[0100] Block header layer: Calculate the total Merkle root of all Merkle roots in the directory layer. The block header will contain this root, and includes, but is not limited to, the block header hash of the previous block, the version number of this block, consensus authentication parameters, and timestamp.

[0101] Furthermore, the automotive parts multi-source fusion traceability platform module in step S2 adopts a three-layer architecture design, namely a data application layer, a data management layer, and a data acquisition layer. Among them, the data application layer provides traceability information query services for consumers and suppliers, the data management layer is responsible for managing the data storage, processing, and maintenance required by various business functions, and the data acquisition layer is responsible for collecting and acquiring relevant data during the automotive parts production process. Moreover, in the platform, each business entity only has the permission to use the distributed storage blockchain to store the data in a distributed manner.

[0102] Furthermore, in step S3, the microscopic visual feature images of the collected automotive component material surface are preprocessed, including image segmentation, obtaining the orientation pattern, enhancement and filtering, binarization and thinning; and the texture type is determined based on its two-dimensional frequency domain features, and an appropriate algorithm is selected to extract and analyze the key attribute feature points of the texture features; the texture types include continuous, discontinuous, and contour types. For continuous textures, the Poincaré exponent algorithm is selected; for discontinuous textures, the Gray-Level Co-occurrence Matrix (GLCM) algorithm is selected; and for contour textures, the Freeman chain code and its derivative feature extraction algorithm are used; the extracted texture features are analyzed and matched using the deep learning framework CNN.

[0103] Microscopic feature images of automotive parts It is stationary in small local regions and can be described by the following model:

[0104] ,

[0105] In the formula, , For respectively exist Local mean and variance of a point in its small neighborhood It is a Gaussian process with zero mean and unit variance.

[0106] For a certain pixel point Size is The neighborhood:

[0107] ,

[0108] In the formula, , They are respectively hour The second moment and second cumulant of a point.

[0109] ,

[0110] In the formula, , They are respectively hour The third moment and third cumulant of a point.

[0111] when In this study, 8×8 blocks were taken from both the foreground and background areas of the microscopic feature image of automotive parts. By comparing the grayscale histograms, it was found that the grayscale values ​​of the foreground area of ​​the microscopic feature image were mainly distributed at the black (grayscale value 0) and white (grayscale value 255) ends, exhibiting a regular distribution with a large variance. In contrast, the grayscale values ​​of the background area of ​​the microscopic feature image of automotive parts were mainly composed of noise, which approximately satisfied a Gaussian distribution. By calculating the third-order cumulant, an appropriate threshold was set based on the actual image grayscale parameters. T ,when At that time, this block is the background area of ​​the microscopic feature image, and this part of the area is cut off.

[0112] Then, the orientation pattern of the microscopic feature image of the automotive parts is obtained. Images representing the microscopic features of automotive parts The grayscale value at that location. Divide the image into sections of size [size missing]. Small blocks, large blocks size set to Among them, for the values ​​of M and N, such that ;Calculate each pixel in a large block based on the Sobel operator gradients on the x and y axes and .

[0113] ,

[0114] Calculation in pixels The direction of the central large block .

[0115] ,

[0116] When large pieces or When the number of zeros reaches a certain level, such as exceeding 80% of the total number of pixels in a large block, It should be set directly to 0 or . This Larger direction As The orientation of the smaller blocks is stored, and the orientation of the next block is calculated iteratively.

[0117] Then, microscopic image enhancement and filtering are performed on the microscopic feature images of automotive parts. A two-dimensional Gabor filter with good directional and frequency selectivity is selected to remove image noise and preserve the texture structure of the microscopic image. Before filtering the microscopic feature images of automotive parts, regions with obvious material property characteristics are normalized by fixing their mean and variance to constants. Normalization can eliminate sensor noise and grayscale value differences caused by different pressures. Here, we use a block-based processing method for the microscopic feature images of automotive parts. Definition For pixels grayscale value at that location , They are blocks The mean and variance, It is a pixel The normalized grayscale value. For the block. The normalized image is defined as containing all pixels in the image as follows: ,

[0118] , These are the expected mean and variance, respectively. Normalization is performed uniformly on all pixels, so it does not change the texture structure of the microscopic feature image of automotive parts.

[0119] Selecting from microscopic feature images of automotive parts The microscopic feature image of the sub-block of a given size, and the mathematical expression of the two-dimensional Gabor function in the spatial domain are as follows: ,

[0120] in, This represents the wavelength, and its value is usually greater than or equal to 2, expressed in pixels, and cannot exceed one-fifth of the input image size. Indicates the direction, with a value ranging from 0 to 360 degrees, specifying the direction of the parallel stripes of the Gabor function. The aspect ratio (spatial aspect ratio) determines the ellipticity of the Gabor function curve; typically, this value is 0.5. When the curve is round, its shape is circular; when... At that time, its shape is continuously stretched along the direction of the parallel stripes. The standard deviation of the Gaussian part cannot be given directly; it is constrained by the center frequency of the filter. x,y Represents the spatial coordinates of an image pixel.

[0121] The feature map of the microscopic feature image of the automotive parts, after image filtering and enhancement, is represented as follows: ,

[0122] A specific component in the frequency domain of a microscopic feature image of automotive parts contains the spatial distribution of the image. This invention, based on the ability of Fourier spectrum to characterize texture features, classifies and determines the texture type (continuous or discontinuous) of the microscopic feature image of automotive parts in the two-dimensional frequency domain. Fourier spectral texture analysis transforms the image from the spatial domain to the frequency domain using Fourier transform, thereby obtaining the frequency distribution of the microscopic feature image of automotive parts. After feature texture classification, based on different texture types, an automatic matching algorithm is used to match a suitable texture feature extraction algorithm for extracting and analyzing key attribute points.

[0123] Transforming the microscopic feature images of automotive parts from the spatial domain to the frequency domain, for For microscopic feature images of automotive parts, if the texture features of these images belong to two-dimensional discrete data, then the process of performing a Fourier transform on them is as follows: ,

[0124] Transforming the microscopic feature images of automotive parts from the spatial domain to the frequency domain, if the texture features of the microscopic feature images of automotive parts belong to two-dimensional continuous data, then the process of performing the Fourier transform is as follows: ,

[0125] Furthermore, in step S3, the step of collecting the electromagnetic field intensity distribution characteristics of the hidden electromagnetic fingerprint RFID tag includes:

[0126] Step S3-1: Demodulate the RFID tag signal received by the RFID tag reader using IQ modulation to obtain the I-channel and Q-channel signals, and then calculate the amplitude signal by modulus. ,in, , which are the sampling points;

[0127] Step S3-2, for Clustering is performed to obtain the center point vectors corresponding to code symbols 0 and 1. ,in The cluster center point corresponding to code element 0. For the cluster center point corresponding to code 1, the clustering function is expressed as: ,in, Clustering operation functions;

[0128] Step S3-3: Determine based on clustering of silent period signal and EPC signal and The desired signal is obtained through Euclidean distance determination. ,in, , The Euclidean distance decision function;

[0129] Step S3-4: For the amplitude signal Standardization to eliminate amplitude differences:

[0130] ,

[0131] in, Standard amplitude signal;

[0132] Step S3-5: Standard amplitude signal Remove desired signal To calculate noise signals :

[0133] ,

[0134] in, This is a noise signal;

[0135] Step S3-6: For the desired signal Standard amplitude signal Noise signals and the original signal Cross-spectral entropy and related features are calculated by pairwise combination, and cross-spectral entropy features are extracted. Defined as:

[0136] ,

[0137] in, For signal In frequency The probability density function of the energy proportion at that location:

[0138] ,

[0139] The cross-power spectral density, based on the Wiener-Khinchin theorem, is expressed as:

[0140] ,

[0141] in, For signal and The cross-power spectrum reflects the correlation between signals in the frequency domain. for conjugate, Indicates Fourier transform, For signal and The cross-correlation function is used to evaluate the cross-correlation between two signals. The degree of similarity, For time delay variables;

[0142] From the above equation, the cross-spectral entropy can be obtained as:

[0143] ,

[0144] in, For signal and Cross-spectral entropy, as a feature of RFID tags for automotive parts; is the probability density function of the energy proportion of the cross-power spectrum;

[0145] At the same time, the maximum amplitude is introduced. Maximum phase Maximum frequency , and cross-spectral entropy These characteristics are collectively used as features of RFID tags for automotive parts; their expression is:

[0146] .

[0147] Furthermore, in step S3, a neural network algorithm is used to process the multi-source fusion features of the components extracted in steps S3-6, specifically including: maximizing the amplitude of the multi-source fusion features of the components. Maximum phase Maximum frequency , and cross-spectral entropy As input to the neural network, the network is trained using a backpropagation algorithm optimized with momentum terms. After training, the network is used to classify and reason about the fused features of the input, outputting the authenticity of the parts and the corresponding traceability information, and writing the results into a distributed blockchain for evidence storage.

[0148] Furthermore, in step S5, the smart contract module completes decision-making activities and performs distributed ledger recording through decentralized consensus authentication, making the smart contract and product data traceable; in the storage stage, the user calls the smart contract to trace product information up to the chain; the uploaded information will be recorded in the blockchain ledger and a transaction hash address will be returned as the unique identifier of the information in the blockchain.

[0149] Furthermore, each business entity only has access to the distributed storage blockchain. Data storage operations can only be performed after their identity data is encrypted using an encryption key. They have no right to store business-related data that has a mapping relationship with the blockchain. The encryption key is collaboratively managed and backed up by multiple authoritative automotive parts industry organizations. Figure 5 As shown;

[0150] When verifying the authenticity of automotive parts, the relevant product data is decrypted using the encryption key provided by the supply chain entity and the official backup key, and the consistency of the two decryption results is compared to determine whether the supply chain entity has tampered with the product data key.

[0151] Furthermore, the automotive parts supply chain is built on an industry-specific network or the Internet, and requires that only business entities in the supply chain join the blockchain system through authorization.

[0152] A multi-source fusion traceability system for automotive parts based on distributed blockchain, applied to a multi-source fusion traceability method for automotive parts based on distributed blockchain, characterized by comprising:

[0153] Blockchain network module: Used to build a blockchain network, adding certification centers, suppliers, manufacturers, distributors, and multi-source data fusion and analysis centers as nodes to achieve information sharing and supervision;

[0154] Automotive Parts Multi-Source Fusion Traceability Platform Module: Used to build an automotive parts multi-source fusion traceability platform;

[0155] Multi-source data acquisition module: used to collect information on automotive parts at various stages, including production time, batch, raw material source, processing technology, transportation route, and sales records;

[0156] Multi-source data fusion and analysis module: used to process and fuse the collected data to form complete traceability information;

[0157] Data on-chain module: Used to write the merged data into the blockchain via smart contracts, ensuring that the data is immutable and traceable;

[0158] Traceability Query Module: This module receives user query requests, searches for matching traceability data on the blockchain according to the query rules in the smart contract, and displays it to the user.

[0159] Smart contract module: used to realize the traceability and automatic collection of automotive parts supply chain information, and to complete decision-making activities and perform distributed ledger through decentralized consensus authentication based on the automatic operation of blockchain.

[0160] [Example 1]

[0161] Example 1 illustrates the experimental process and results of collecting microscopic visual feature images of automotive component material surfaces in step S3. Ceramic brake pads from different manufacturers of the same automotive type were selected as the research object. Microscopic visual feature images of their surfaces were collected as a dataset to verify the effectiveness of the method. The dataset contains 100 microscopic visual feature images of automotive ceramic brake pads; 70 images were selected as training samples, 10 as validation samples, and 20 as test samples.

[0162] The energy, contrast, correlation, and entropy of the gray-level co-occurrence matrix were used as feature parameters to describe the microscopic visual image texture information of automotive parts.

[0163] The image processing of microscopic visual features of automotive parts material surfaces was implemented in Matlab R2018b. The CNN neural network model for recognizing microscopic visual features of automotive parts was implemented using a 64-bit Ubuntu 18.04 operating system, with the environment configured as Python 3.8, PyTorch 1.4.0 combined with the PyCharm compiler.

[0164] like Figure 6 As shown, the microscopic visual feature image processing of automotive component material surfaces was performed using genuine and counterfeit automotive ceramic brake pad experimental materials, and scanned using a KC-X1000 series laser spectral confocal microscope.

[0165] Experimental results are as follows Figure 7 , Figure 8 As shown, a comparison of the microscopic feature images of genuine and counterfeit automotive ceramic brake pads reveals the material differences between them. Subsequent image preprocessing yields the following results: Figure 9 , Figure 10 As shown.

[0166] As can be seen, feature extraction based on a hybrid algorithm of microscopic visual feature extraction and neural network can clearly distinguish the difference between genuine and fake parts.

[0167] [Example 2]

[0168] Example 2 illustrates the experimental process and results of collecting the electromagnetic field intensity distribution characteristics of hidden electromagnetic fingerprint RFID tags on the outer packaging of automotive parts in step S3.

[0169] Seven types of RFID tags commonly found in the market were used, each manufactured by one of three different companies. Before data collection, the same EPC code was written to all 140 tags. Data was then collected from the tags after all tags had the same EPC code written to them. For the automotive parts experiment, genuine and counterfeit automotive brake hoses were selected as the experimental material. These brake hoses were placed inside the tag box for the experiment.

[0170] The radio frequency signal identification testing equipment used in this experiment is the Voyantic Tagformance Pro, such as... Figure 11 As shown. The Voyantic Tagformance Pro is a professional radio frequency identification (RFID) test equipment, primarily used for testing UHF RFID and Near Field Communication (NFC) HF RFID. This equipment is designed for testing tags, cards, and sensor tags, supporting widely used RFID protocols. It can verify and optimize the sensitivity of UHF and HF tags, study the effects of materials, orientation, and distance on tags, and assess tag performance in different applications. The Tagformance Pro is an integrated test equipment, including software, accessories, and a silenced measurement space, forming a complete test system. Its UHF band supports standards such as ISO 18000-6C (EPC C1G2), ISO 18000-6B, and GB / T 29768; its HF band supports standards such as ISO 14443 and ISO 15693.

[0171] The 925MHz standard frequency band was selected for the RFID experiment. Before data acquisition, equipment calibration was performed to ensure data accuracy. During each data acquisition, multiple boxes containing tags indicating genuine and counterfeit automotive brake hoses were placed in a cardboard box on top of the antenna device, forming a 3×3 multi-tag rectangular array. Figure 12 As shown. The remaining tags were not within the magnetic field range of the reading system to reduce the risk of tag collisions. All data acquisition was not conducted in an isolated environment, meaning the environment might include thermal noise, mobile phone noise, wireless network noise, and radio frequency noise. The transmission signal frequency of the radio frequency antenna was recorded, and the sensitivity data of each tag was read.

[0172] An RFID reader is deployed next to an RFID antenna, using an external RFID antenna array to collect information from RFID tags within its reading range in real time. RFID tags are installed on the automotive parts to be sensed, and the data is input into a neural network. Four antennas are evenly deployed on a gantry frame placed indoors, scanning RFID tags within the reading range from different angles. Each antenna can be adjusted in parameters based on commands from the reader, using different transmission powers for tag scanning. Simultaneously, the reflected signal energy value of the RFID tags is accurately collected and used as an identification reference. The RFID antenna array transmits RFID signals, and the RFID reader collects the reflected RFID signal data in real time through the RFID antenna array. Each set of data consists of the RFID signal strength values ​​fed back from the four RFID antennas on the gantry and the arrival time of the feedback RFID signals. Time-frequency division multiplexing technology is used to solve the signal interference problem when multiple tags respond simultaneously in a mixed group of automotive parts.

[0173] The sensitivity data of each tag on the automotive brake hose at this frequency are shown in Table 1.

[0174]

[0175] Table 1 records the experimental data of RFID tag electromagnetic feature acquisition for genuine and counterfeit automotive brake hose samples with different EPC codes under the 925MHz standard test frequency band. The table includes tag name, electronic code (ePC code), sensitivity (dBm) for genuine and counterfeit products, and backscatter power (dBm). Sensitivity reflects the tag's ability to receive radio frequency signals in this frequency band, while backscatter power reflects the intensity of the tag's reflected signal. Comparing the two sets of data for genuine and counterfeit samples shows that genuine tags generally have higher sensitivity than counterfeit tags, while their backscatter power is generally lower. This difference can serve as a key feature parameter for neural network recognition of electromagnetic fingerprints, providing a quantitative basis for the identification of genuine and counterfeit automotive parts.

[0176] Using the data collected in Table 1, four types of features were extracted and input into the neural network model. The final results are shown in Table 2.

[0177]

[0178] As can be seen, the overall accuracy rate reaches 98.2%, indicating that the model's overall discrimination ability is excellent and meets the technical requirements of high-value goods. Further analysis of the category-specific indicators reveals that the recall rate of genuine products (97.5%) is slightly lower than that of counterfeit products (98.8%). This difference may stem from the characteristic attenuation of genuine product signals in complex environments (such as metal interference), leading to some samples being misclassified as counterfeit. Counterfeit products, due to their significantly abnormal electromagnetic characteristics (such as phase shift and spectral entropy change), are more easily identified by the model. The high F1-Score (98.0%), the harmonic mean of precision and recall, confirms that the system has achieved a balance between reducing false alarms and false negatives, highlighting the effectiveness of using the electromagnetic field strength distribution characteristics of multiple RFID tags combined with neural network algorithms in dynamically adjusting the weight mechanism.

[0179] To verify the effectiveness of the algorithm, its performance was compared with that of the traditional support vector machine. The performance comparison is shown in Table 3.

[0180]

[0181] Table 3 shows the model performance comparison results, demonstrating a significant advantage in classification tasks: its accuracy reaches 98.6%, an improvement of 14.2 percentage points compared to the support vector machine (84.4%); its F1 score reaches 0.98, an improvement of 7.2% compared to the support vector machine, representing a relative increase of 7.9%. This result verifies that combining the electromagnetic field strength distribution characteristics of multiple RFID tags with a neural network algorithm effectively enhances the model's ability to represent complex features, significantly improving the accuracy and generalization performance of classification tasks. Although the training time (38.2 seconds) is approximately twice that of the support vector machine (12.7 seconds), this computational cost is an acceptable price to pay in accuracy-sensitive scenarios.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-source fusion traceability method for automotive parts based on distributed blockchain, characterized in that, Includes the following steps: Step S1: Construct the blockchain network module: The various business entities in the automotive parts supply chain are added to the blockchain as nodes. These business entities include upstream suppliers, manufacturers, downstream distributors, retailers, and customers. In the blockchain network, an authentication center is set up, with each supplier, manufacturer, and distributor acting as an independent authentication center to authenticate the authenticity of product information and the rationality of decisions in the automotive parts supply chain. A multi-source data fusion and analysis module is constructed as a node to join the blockchain network module. The multi-source data fusion and analysis module is used to receive information collection requests transmitted by various business entities in the supply chain, determine the dataset according to the requests, collect and process the data, and then return the processing results to various business entities in the supply chain. Step S2: Construct a multi-source fusion traceability platform module for automotive parts. The multi-source fusion traceability platform module for automotive parts is used to record information of each link in the automotive parts supply chain and uses timestamps to record the time sequence. Step S3: Construct a multi-source data acquisition module to form the unique identity information of automotive parts: Collect microscopic visual feature images of the surface of automotive parts materials and establish a dataset containing microscopic visual feature images of the surface of various automotive parts materials. Using a multi-antenna array, the electromagnetic field strength distribution characteristics of hidden electromagnetic fingerprint RFID tags on the outer packaging of automotive parts are collected via an RFID reader. A neural network algorithm is then used to perform feature analysis and classification of the extracted electromagnetic fingerprint data. Specific data collection steps include: Step S3-1: Demodulate the RFID tag signal received by the RFID tag reader using IQ modulation to obtain the I-channel and Q-channel signals, and then calculate the amplitude signal by modulus. ,in, , which are the sampling points; Step S3-2, for Clustering is performed to obtain the center point vectors corresponding to code symbols 0 and 1. ,in The cluster center point corresponding to code element 0. For the cluster center point corresponding to code 1, the clustering function is expressed as: ， in, Clustering operation functions; Step S3-3: Determine based on clustering of silent period signal and EPC signal and The desired signal is obtained through Euclidean distance determination. ,in, , The Euclidean distance decision function; Step S3-4: For the amplitude signal Standardization to eliminate amplitude differences: ， in, Standard amplitude signal; Step S3-5: Standard amplitude signal Remove desired signal To calculate noise signals : ， in, This is a noise signal; Step S3-6: For the desired signal Standard amplitude signal Noise signals and the original signal Cross-spectral entropy and related features are calculated by pairwise combination, and cross-spectral entropy features are extracted. Defined as: ， in, For signal In frequency The probability density function of the energy proportion at that location: ， The cross-power spectral density, based on the Wiener-Khinchin theorem, is expressed as: ， in, For signal and The cross-power spectrum reflects the correlation between signals in the frequency domain. for conjugate, Indicates Fourier transform, For signal and The cross-correlation function is used to evaluate the cross-correlation between two signals. The degree of similarity, For time delay variables; From the above equation, the cross-spectral entropy can be obtained as: ， in, For signal and Cross-spectral entropy, as a feature of RFID tags for automotive parts; is the probability density function of the energy proportion of the cross-power spectrum; At the same time, the maximum amplitude is introduced. Maximum phase Maximum frequency , and cross-spectral entropy These characteristics are collectively used as features of RFID tags for automotive parts; their expression is: ， Simultaneously, it collects production information, transportation information, inspection information, and import / export information; Step S4: Construct a multi-source data fusion analysis module: Send the data collected in step S3 to the multi-source data fusion analysis module, and use the multi-source data fusion analysis module to process and fuse the data; Step S5: Construct the data on-chain module: Write the data fused in step S4 into the blockchain through a smart contract module pre-deployed on the blockchain. The smart contract is a computer protocol that automatically executes the contract terms and is used to realize the automatic verification and storage of data. Each transaction record is packaged into a block, and the block contains the hash value of the current block, the hash value of the previous block, and timestamp information. Step S6: Construct a traceability query module: Users input relevant information about auto parts through the multi-source fusion traceability platform module. The multi-source fusion traceability platform module searches for matching traceability data on the blockchain according to the query rules in the smart contract module.

2. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: The blockchain network in step S1 adopts a distributed storage structure, which is used to store different types of basic data. Each type of data is mapped to the block body of the corresponding block in the blockchain. The distributed storage data structure consists of the following five parts: Data layer: refers to supply chain management related data; Categorized Merkle Tree Layer: Data from different categories form different Merkle trees, which do not overlap. This layer is used to store the Merkle trees. Classified Merkle tree root layer: This layer stores the roots of various different Merkle trees; Directory layer: This layer constructs a directory based on the basic data corresponding to different categories of Merkel roots, and calculates another Merkel root for each category of Merkel root. The number of categories of Merkel roots can be selected according to actual needs. Block header layer: Calculate the total Merkle root of all Merkle roots in the directory layer. The block header will contain this root, and includes, but is not limited to, the block header hash of the previous block, the version number of this block, consensus authentication parameters, and timestamp.

3. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: The automotive parts multi-source fusion traceability platform module in step S2 adopts a three-layer architecture design, namely, a data application layer, a data management layer, and a data acquisition layer. Among them, the data application layer provides traceability information query services for consumers and suppliers, the data management layer is responsible for managing the data storage, processing, and maintenance required by various business functions, and the data acquisition layer is responsible for collecting and acquiring relevant data during the automotive parts production process. In the platform, each business entity only has the permission to use the distributed storage blockchain to store the data in a distributed manner.

4. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: In step S3, the microscopic visual feature images of the collected automotive component material surface are preprocessed, including image segmentation, orientation pattern extraction, enhancement and filtering, binarization and thinning; and the texture type is determined based on its two-dimensional frequency domain features, and an appropriate algorithm is selected to extract and analyze key attribute feature points of the texture features; the texture types include continuous, discontinuous, and contour types. For continuous textures, the Poincaré exponent algorithm is selected; for discontinuous textures, the Gray-Level Co-occurrence Matrix (GLCM) algorithm is selected; and for contour textures, the Freeman chain code and its derivative feature extraction algorithm are used; the extracted texture features are analyzed and matched using the deep learning framework CNN.

5. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: The multi-source fusion features of the components calculated in steps S3-6 are processed using a neural network algorithm, specifically including: maximizing the amplitude of the multi-source fusion features of the components. Maximum phase Maximum frequency , and cross-spectral entropy As input to the neural network, the network is trained using a backpropagation algorithm optimized with momentum terms. After training, the network is used to classify and reason about the fused features of the input, outputting the authenticity of the parts and the corresponding traceability information, and writing the results into a distributed blockchain for evidence storage.

6. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: In step S5, the smart contract module completes decision-making activities and performs distributed ledger recording through decentralized consensus authentication, making the smart contract and product data traceable. In the storage stage, the user calls the smart contract to trace the product information to the blockchain. The uploaded information will be recorded in the blockchain ledger and a transaction hash address will be returned as the unique identifier of the information in the blockchain.

7. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: Each business entity only has the authority to store the distributed blockchain. Its identity data can only be stored after being encrypted with an encryption key. It has no right to store business-related data that has a mapping relationship with the blockchain. The encryption key is jointly managed and backed up by multiple authoritative automotive parts industry organizations. When verifying the authenticity of automotive parts, the relevant product data is decrypted using the encryption key provided by the supply chain entity and the official backup key, and the consistency of the two decryption results is compared to determine whether the supply chain entity has tampered with the product data key.

8. The multi-source fusion traceability method for automotive parts based on distributed blockchain according to claim 1, characterized in that: The automotive parts supply chain is built on an industry-specific network or the Internet, and requires that only business entities in the supply chain join the blockchain system through authorization.

9. A multi-source fusion traceability system for automotive parts based on distributed blockchain, applied to the multi-source fusion traceability method for automotive parts based on distributed blockchain according to any one of claims 1 to 8, characterized in that, include: Blockchain network module: Used to build a blockchain network, adding certification centers, suppliers, manufacturers, distributors, and multi-source data fusion and analysis centers as nodes to achieve information sharing and supervision; Automotive Parts Multi-Source Fusion Traceability Platform Module: Used to build an automotive parts multi-source fusion traceability platform; Multi-source data acquisition module: used to collect information on automotive parts at various stages, including production time, batch, raw material source, processing technology, transportation route, and sales records; Multi-source data fusion and analysis module: used to process and fuse the collected data to form complete traceability information; Data on-chain module: Used to write the merged data into the blockchain via smart contracts, ensuring that the data is immutable and traceable; Traceability Query Module: This module receives user query requests, searches for matching traceability data on the blockchain according to the query rules in the smart contract, and displays it to the user. Smart contract module: used to realize the traceability and automatic collection of automotive parts supply chain information, and to complete decision-making activities and perform distributed ledger through decentralized consensus authentication based on the automatic operation of blockchain.

Images