Anti-counterfeiting traceability system and method based on braille dot matrix star topology structure

Abstract

The invention relates to the technical field of identification coding, in particular to an anti-fake traceability system and method based on a braille dot matrix star topological structure, and the system achieves the combined organization of the length of a connection path and a spatial distribution position through introducing an isolated forest, enables an abnormal state to form a distinguishable isolation result at a structural level, and achieves the anti-fake traceability of the braille dot matrix star topological structure. The dot matrix anomaly does not depend on single verification failure judgment any more, but embodies the anomaly degree through path level difference, so that the refinement level of anomaly recognition is improved, through joint division of isolation states and arrangement positions by the support vector machine, boundary judgment results have clear boundaries, the fuzzy overlapping risk between adjacent abnormal states is reduced, and the accuracy of anomaly recognition is improved. The reproducibility of the abnormal distinguishing result in multiple times of recognition is guaranteed, different abnormal states correspond to stable interval identifiers, the abnormal result does not drift along with the scale change of the dot matrix any more, the consistency of the distinguishing result is improved, and the stability and reliability of anti-fake judgment and traceability records in a complex circulation environment are enhanced.

Description

Technical Field

[0001] This invention relates to the field of identification and coding technology, and in particular to an anti-counterfeiting and traceability system and method based on a Braille dot matrix star topology. Background Technology

[0002] The field of identification coding technology aims to assign unique identifiers to objects and support identification, verification, and association. It focuses on the technical implementation of coding rules, identification carriers, code generation and distribution, code parsing and verification, binding of codes with business data, and recording and tracking of identifiers in the circulation process. Identification carriers in this field are typically based on printable and etchable patterns, including barcodes, QR codes, character sequences, dot matrix structures, and microstructure patterns, which can be combined with tactilely readable graphics to form composite identifiers. Systems in this field typically consist of an encoding end, a labeling end, an identification end, and a data end. The encoding end is responsible for generating unique identifiers and calculating verification information. The labeling end is responsible for mapping the code to an identifiable pattern and forming it on the carrier surface. The identification end is responsible for collecting and parsing the pattern and outputting the coding result. The data end is responsible for binding the code with object information and circulation information and supporting queries.

[0003] A Braille dot matrix star topology-based anti-counterfeiting and traceability system is an identification coding system that uses Braille dot matrix as the identification presentation method and a star topology structure as the dot matrix organization method. It binds code verification with circulation records to achieve anti-counterfeiting verification and traceability. The system assigns each identified object to a unique code and maps the unique code to a dot matrix pattern composed of Braille dots. The dot matrix pattern uses the central dot matrix cluster as the positioning reference and forms multiple radial branch dot matrix clusters around it. Dot matrix areas for positioning, verification, and redundancy recovery are set in the dot matrix pattern, so that the identification end can still complete the positioning and code recovery even when there is rotation, scaling, dirt, or partial loss. The purpose of the system is to form an identification form that can be used by identification devices and touch reading methods, making it difficult for counterfeiters to meet the code verification and topology consistency requirements even if they replicate the dot matrix appearance, thereby achieving anti-copying judgment. At the same time, by binding the code with the records of production, circulation, and sales, the single product identification can be used to trace the source, locate circulation nodes, and support the verification of abnormal circulation.

[0004] Existing identification coding technologies are centered on unique codes and verification rules, typically treating patterns as the carriers of codes. The identification process primarily aims to restore the coded content and verify the verification information. The pattern structure itself lacks the ability to express itself independently. Existing systems mostly use binary results of whether the verification is passed in anomaly identification, which cannot distinguish different anomaly forms. Anomaly states are difficult to analyze further, affecting risk positioning and handling decisions. Circulation records are usually bound by codes as indexes. Once a code is copied, multiple objects may correspond to the same record, causing confusion in the traceability chain and making it difficult to accurately locate anomaly nodes. In practice, when labels are damaged, partially missing, or tampered with, existing technologies rely more on redundant coding and verification, increasing processing costs and resulting in insufficient stability. Due to the lack of the ability to record the evolution of the label structure, existing technologies can often only trace back to the first appearance and last scan position of the code when tracking abnormal circulation paths, failing to reflect the formation stage of the anomaly in the circulation process, thus limiting the application effect of anti-counterfeiting and traceability systems in refined regulatory scenarios. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an anti-counterfeiting and traceability system and method based on a Braille dot matrix star topology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an anti-counterfeiting and traceability system based on a Braille dot matrix star topology includes:

[0007] Dot matrix encoding construction module: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix units, and establish a radial arrangement relationship with the preset reference points in the dot matrix distribution as a reference to form a dot matrix sequence identifier;

[0008] Star topology constraint module: Based on the matrix sequence identifier, compare the connection relationship between the central unit and the branch unit with the stored set of predetermined connection relationships, and compare the branch arrangement order with the stored predetermined arrangement order to obtain the topology constraint result;

[0009] Anomaly Isolation Judgment Module: Based on the lattice sequence identifier and the topological constraint result, according to the isolation path organization rules corresponding to the isolated forest, constructs branch connection paths based on the deviation branch identifier, calculates the branch connection path length and branch spatial distribution position, and generates anomaly isolation identifiers;

[0010] Interval Boundary Discrimination Module: Based on the anomaly isolation identifier, according to the boundary division relationship corresponding to the support vector machine, extract the relative position of the dot matrix sequence identifier in the discrimination boundary, generate the boundary discrimination identifier, and output the boundary discrimination identifier;

[0011] Source tracing coding solidification module: Based on the boundary discrimination identifier, a fixed association relationship is established between the boundary discrimination identifier and the occurrence node, and the record is written into the source tracing link according to the storage writing order to form a source tracing coding record.

[0012] As a further embodiment of the present invention, the dot matrix sequence identifier includes a dot matrix unit identifier, a radial arrangement order mark, and a position identifier corresponding to a preset reference point; the topology constraint result includes a determination result of the connection relationship between the central unit and the branch unit, a determination result of the branch arrangement order, and an inconsistency state mark; the abnormal isolation identifier includes a deviation from the branch identifier, a branch connection path length mark, and a branch spatial distribution position identifier; the boundary discrimination identifier includes an interval number identifier, an arrangement position identifier, and a boundary position mark corresponding to the isolation state; and the traceability coding record includes a boundary discrimination identifier, an occurrence node identifier, and a writing order mark.

[0013] As a further aspect of the present invention, the dot matrix encoding construction module includes:

[0014] Dot matrix position sorting submodule: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix unit, record the corresponding coordinate position of the dot matrix unit on the surface one by one, collect the coordinate values ​​and calculate the adjacent difference, and generate a dot matrix spatial relationship set.

[0015] Radial sequence construction submodule: Based on the set of spatial relationships of the dot matrix, and taking the preset reference point in the dot matrix distribution as the benchmark, sort the relative direction values ​​of the dot matrix units, perform radial grouping and sequence verification, and concatenate the grouping results in sequence to form a dot matrix sequence identifier.

[0016] As a further aspect of the present invention, the star topology constraint module includes:

[0017] Connection relationship verification submodule: Based on the dot matrix sequence identifier, extract the connection items corresponding to the central unit and the branch unit, read the connection pointer number and connection quantity marker, compare the position mapping with the stored set of connection relationships in order, and generate connection relationship comparison results;

[0018] The order comparison submodule: Based on the connection relationship comparison result, it calls the branch order item in the dot matrix sequence identifier, reads the order index number bit by bit, and performs a sequence comparison with the stored predetermined order according to the index position to obtain the topological constraint result.

[0019] As a further aspect of the present invention, the anomaly isolation determination module includes:

[0020] Off-branch identification submodule: Based on the dot matrix sequence identifier and the topological constraint result, read the branch number information recorded in the topological constraint result, extract the corresponding branch index in the order of the number, and check the corresponding branch position in the dot matrix sequence identifier to generate a set of off-branch branches;

[0021] Connection path construction submodule: Based on the set of off-branch nodes, locate the corresponding branch node position in the dot matrix sequence identifier, connect the node identifiers segment by segment along the connection order between nodes, and accumulate the number of connected nodes to form a node order list to obtain the branch connection path data;

[0022] Isolation Identifier Generation Submodule: Based on the branch connection path data, read the number of nodes in the connection path and retrieve the spatial location number of the branch in the dot matrix sequence identifier. According to the correspondence between the isolation path length and the node distribution position in the isolated forest, combine and record the number of nodes and the spatial location number to form an abnormal isolation identifier corresponding to the branch.

[0023] As a further aspect of the present invention, the isolated forest constructs an input item set composed of path length values ​​based on the number of nodes recorded in the branch connection path data. The input item set is randomly dimensioned and randomly split points are set to generate multiple sets of isolated path structures. The path level values ​​corresponding to the input items are recorded in each isolated path structure, and the path level values ​​are collected to form an isolated level record.

[0024] As a further aspect of the present invention, the interval boundary discrimination module includes:

[0025] Isolation Identifier Reading Submodule: Based on the abnormal isolation identifier, the branch number field is read item by item and rearranged in order, and the isolation status value is extracted and bound to the number one by one to form an index entry set that corresponds to the dot matrix sequence identifier, thus obtaining the isolation index set;

[0026] Boundary location extraction submodule: Based on the isolation index set, read the isolation state value and arrangement number corresponding to each index, map the isolation state value to the interval to which the state value belongs according to the boundary division relationship corresponding to the support vector machine and classify the position, record the interval number corresponding to each index, and generate a boundary location set;

[0027] Discrimination identifier generation submodule: Based on the boundary position record set, read the arrangement number and segment number in the record one by one, combine them into an identifier unit sequence in a predetermined order, and write the identifier unit sequence into a unified number position to form a boundary discrimination identifier.

[0028] As a further aspect of the present invention, the support vector machine constructs an input item set composed of state values ​​and arrangement numbers by recording isolation state values ​​and corresponding arrangement numbers in the isolation index set, performs boundary function calculation on the input item set, determines the positional relationship of each input item relative to the separation boundary during the calculation process, and maps the positional relationship to interval numbers according to a preset interval division rule to form an interval mapping record corresponding to each index.

[0029] As a further aspect of the present invention, the traceability coding and solidification module includes:

[0030] Relationship Construction Submodule: Based on the boundary discrimination identifier, read the number field in the boundary discrimination identifier and extract the corresponding occurrence node identifier, align the number field and node identifier in order, write the number item and node item accordingly, mark the order position and confirm the item, and establish a relationship record;

[0031] Link writing and solidification submodule: Based on the association record, read the stored writing order and extract the corresponding sequence number, retrieve the record item according to the sequence number and write it to the link position, synchronously write the sequence identifier to each record and complete the position registration to form a traceability coding record.

[0032] A method for anti-counterfeiting and traceability based on a Braille dot matrix star topology, wherein the method is executed based on the aforementioned anti-counterfeiting and traceability system based on a Braille dot matrix star topology, includes the following steps:

[0033] S1: Based on the dot matrix distribution on the surface of the identified object, collect the spatial location data of the dot matrix units, and perform radial sorting of the dot matrix units with a preset reference point as a benchmark to form a set of dot matrix identifiers with sequential relationships, thus obtaining the dot matrix sequence identifier.

[0034] S2: Based on the dot matrix sequence identifier, read the connection relationship information between the central unit and the branch unit, compare it with the predetermined connection relationship set, and check the branch arrangement order to obtain the topological constraint result corresponding to the dot matrix structure.

[0035] S3: Based on the dot matrix sequence identifier and the topological constraint result, identify the deviation branch identifier, and combine the branch connection path length information with the branch spatial distribution position to generate an abnormal isolation identifier;

[0036] S4: Based on the anomaly isolation identifier, according to the boundary division relationship corresponding to the support vector machine, perform interval mapping on the content corresponding to the anomaly isolation identifier, determine the position of the dot matrix sequence identifier in the boundary division interval, and form and output the boundary discrimination identifier;

[0037] S5: Based on the boundary discrimination identifier, the boundary discrimination identifier is associated with the occurrence node by number, and written into the traceability link according to the storage writing order. The writing order is marked to form a traceability coding record.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] 1. In this invention, by introducing isolated forests to jointly organize the connection path length and spatial distribution location, abnormal states form distinguishable isolation results at the structural level. The point anomaly no longer relies on a single verification failure judgment, but reflects the degree of anomaly by the difference in path level, thereby improving the precision of anomaly identification.

[0040] 2. In this invention, by using a support vector machine to jointly divide the isolation state and the arrangement position, the boundary discrimination result has a clear boundary, reducing the risk of fuzzy overlap between adjacent abnormal states and ensuring the reproducibility of the abnormality discrimination result in multiple identifications;

[0041] 3. In this invention, by making stable interval identifiers correspond to different abnormal states, the abnormal results no longer drift with changes in the dot matrix size, thereby improving the consistency of the discrimination results and enhancing the stability and reliability of anti-counterfeiting judgment and traceability records in complex circulation environments. Attached Figure Description

[0042] Figure 1 This is a system flowchart of the present invention;

[0043] Figure 2 This is a schematic diagram of the method steps of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0045] Example 1

[0046] Please see Figure 1 The present invention provides a technical solution: an anti-counterfeiting and traceability system based on a Braille dot matrix star topology structure.

[0047] Dot matrix encoding construction module: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix units, and establish a radial arrangement relationship with the preset reference points in the dot matrix distribution as a reference to form a dot matrix sequence identifier;

[0048] Star topology constraint module: Based on the lattice sequence identifier, it compares the connection relationship between the central unit and the branch unit with the stored set of predetermined connection relationships, and compares the branch arrangement order with the stored predetermined arrangement order to obtain the topology constraint result;

[0049] Anomaly Isolation Detection Module: Based on the lattice sequence identifier and topological constraint results, according to the isolation path organization rules corresponding to the isolated forest, branch connection paths are constructed based on the deviation branch identifier, the length of the branch connection path and the spatial distribution position of the branch are calculated, and anomaly isolation identifiers are generated.

[0050] The boundary discrimination module: Based on the anomaly isolation marker, according to the boundary division relationship corresponding to the support vector machine, it extracts the relative position of the dot matrix sequence marker in the discrimination boundary, generates the boundary discrimination marker, and outputs the boundary discrimination marker;

[0051] Source tracing coding solidification module: Based on the boundary discrimination identifier, a fixed association relationship is established between the boundary discrimination identifier and the occurrence node, and the record is written into the source tracing link in the order of storage to form a source tracing coding record.

[0052] The dot matrix sequence identifier includes the dot matrix unit identifier, the radial arrangement order mark, and the position mark corresponding to the preset reference point. The topology constraint result includes the determination result of the connection relationship between the central unit and the branch unit, the determination result of the branch arrangement order, and the inconsistency state mark. The abnormal isolation identifier includes the deviation branch identifier, the branch connection path length mark, and the branch spatial distribution position mark. The boundary discrimination identifier includes the interval number identifier, the arrangement position identifier, and the boundary position mark corresponding to the isolation state. The traceability coding record includes the boundary discrimination identifier, the occurrence node identifier, and the writing order mark.

[0053] The dot matrix encoding construction module includes:

[0054] Dot matrix position sorting submodule: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix unit, record the corresponding coordinate position of the dot matrix unit on the surface one by one, collect the coordinate values ​​and calculate the adjacent difference, and generate a dot matrix spatial relationship set.

[0055] Radial sequence construction submodule: Based on the spatial relationship set of the dot matrix, and taking the preset reference point in the dot matrix distribution as the benchmark, sort the relative direction values ​​of the dot matrix units, perform radial grouping and sequence verification, and concatenate the grouping results to form a dot matrix sequence identifier;

[0056] The dot matrix location organization submodule: Based on the dot matrix distribution formed on the surface of the identified object, the binary dot matrix image is processed for region labeling using the eight-neighbor connectivity labeling method. The input image resolution is set to 1200 pixels by 1200 pixels, and the pixel values ​​are limited to 0 and 255. The connectivity determination adopts the 8-direction adjacency rule. The starting value of the label number is set to 1 and increments in the discovery order. For each numbered region, the pixel coordinates within the region are collected and the geometric center coordinates are calculated. The horizontal coordinate of the geometric center is retained to 6 decimal places and recorded. The vertical coordinate of the geometric center is retained to 6 decimal places and recorded. The number, horizontal coordinate, and vertical coordinate are written into the coordinate record table and sorted in ascending order by number. The difference is calculated for adjacent numbered records. The difference includes the horizontal difference and the vertical difference. The horizontal difference is the difference between the horizontal coordinate values ​​of two adjacent records and is retained to 6 decimal places. The vertical difference is the difference between the vertical coordinate values ​​of two adjacent records and is retained to 6 decimal places. The unit is millimeters. The number, horizontal coordinate, vertical coordinate, horizontal difference, and vertical difference are collected into a set of record entries to generate a dot matrix spatial relationship set.

[0057] The radial sequence construction submodule: Based on the point space relationship set, a configuration-driven center point locking method is used to determine the preset reference point. The preset reference point rules are pre-written into the configuration file. The configuration items include a threshold of 6 adjacency numbers, a lower limit of 4 mm for the adjacency distance range, an upper limit of 6 mm for the adjacency distance range, and center positioning coordinates set to 50 mm horizontally and 50 mm vertically. During execution, the number of adjacencies for each record is counted, and records that meet the threshold are filtered to form a candidate set. The record with the smallest difference from the center positioning coordinates in the candidate set is selected as the reference point. For the remaining records, directional segment numbers are generated one by one. The directional segment division is set to 8 segments with a segment width of 45 degrees. A radial distance value is generated for each record. The radial distance is obtained by looking up a table with a millimeter scale and retains 6 decimal places. Records are organized in ascending order by directional segment number and in ascending order by radial distance value within the same directional segment. The organization results are concatenated in sequence as a sequence field and written into the identifier record to form a point sequence identifier.

[0058] The star topology constraint module includes:

[0059] Connection relationship verification submodule: Based on the dot matrix sequence identifier, extract the corresponding connection items of the central unit and the branch unit, read the connection pointer number and connection quantity marker, compare the position mapping with the stored set of connection relationships in order, and generate connection relationship comparison results;

[0060] The sorting order comparison submodule: Based on the connection relationship comparison results, it calls the branch sorting order item in the dot matrix sequence identifier, reads the sorting index number bit by bit, and performs a sequential comparison with the stored predetermined sorting order according to the index position to obtain the topological constraint results;

[0061] The connection relationship verification submodule uses an adjacency list sequential comparison algorithm based on the dot matrix sequence identifier to verify the connection items between the central unit and the branch unit. A list of connection items is pre-written into the dot matrix sequence identifier. The list fields include connection start number, connection end number, and connection sequence number. The connection start number ranges from 1 to 64, the connection end number ranges from 1 to 64, and the connection sequence number ranges from 1 to 16. During execution, the connection start number and connection end number are read sequentially in ascending order of connection sequence number. A fixed-length sequence mapping comparison method is used to align the read connection items with a predetermined set of connection relationships stored in the configuration file. The predetermined set of connection relationships is pre-set in text format, with each line recording one connection relationship and fields separated by semicolons. The field order is connection start number, connection end number, and connection sequence number. During the verification process, numerical equality is determined according to the sequence number, and a determination flag is recorded. The determination flag takes values ​​of 0 and 1. After all connection items are verified, the determination flag sequence is compiled in the original order to generate the connection relationship comparison result.

[0062] The order alignment submodule, based on the connection relationship alignment results, uses a sequence index bit-by-bit verification method to check the branch order. A branch order index list is pre-written into the dot matrix sequence identifier. The index list consists of an integer sequence with a limited length of 8 items, each with a value ranging from 0 to 7. During execution, the index number is read bit-by-bit according to the list index positions 0 to 7, and a predetermined order table stored in the configuration file is read synchronously. The predetermined order table is written in a fixed array format with consistent length. A numerical comparison is performed between the current position index number and the corresponding position number in the predetermined order, generating an alignment mark. The alignment mark takes values ​​of 0 and 1. The alignment mark is written into the result sequence according to the index position order. After completing the verification of all index positions, the result sequence is aggregated and recorded to generate the topological constraint result.

[0063] The anomaly isolation determination module includes:

[0064] Off-branch identification submodule: Based on the dot matrix sequence identifier and topological constraint results, read the branch number information recorded in the topological constraint results, extract the corresponding branch index in the order of the numbers, and check the corresponding branch position in the dot matrix sequence identifier to generate a set of off-branch branches;

[0065] Connection path construction submodule: Based on the off-branch set, locate the corresponding branch node position in the dot matrix sequence identifier, concatenate the node identifiers segment by segment along the connection order between nodes, and accumulate the number of concatenated nodes to form a node order list, thus obtaining the branch connection path data;

[0066] Isolation Identifier Generation Submodule: Based on branch connection path data, read the number of nodes in the connection path and retrieve the spatial location number of the branch in the dot matrix sequence identifier. According to the correspondence between the isolation path length and the node distribution location in the isolated forest, combine and record the number of nodes and the spatial location number to form an abnormal isolation identifier corresponding to the branch.

[0067] The deviation branch identification submodule: Based on the dot matrix sequence identifier and topological constraint results, it uses a sequential consistency comparison algorithm to verify the branch numbers. During execution, it first reads the branch number sequence from the topological constraint results. The branch numbers are recorded in the form of consecutive integers, with the starting number set to 1 and the maximum number of numbers set to 8. They are written into the comparison list in ascending order of the numbers. At the same time, it reads the branch position sequence from the dot matrix sequence identifier. Each item in the branch position sequence corresponds to a branch starting node number. It performs an item-by-item position comparison operation on the two sets of sequences. The comparison rule is set as follows: if the number values ​​are equal and the sequence index positions are consistent, it is considered a consistent state. If the number values ​​are unequal and the index positions are inconsistent, it is considered a deviation state. The number corresponding to the item determined to be in a deviation state is written into the deviation record table. The deviation record table is rearranged in ascending order of the numbers, and the rearranged number sequence is written into the set field to generate a deviation branch set.

[0068] The connection path construction submodule: Based on the set of deviation branches, it uses a sequential chain traversal method to construct the branch connection path. During execution, it reads the numbers in the set of deviation branches one by one and locates the starting node position corresponding to the number in the dot matrix sequence identifier. The node position is represented by the sequence index value. It performs item-by-item traversal operation according to the connection order recorded in the dot matrix sequence identifier. The traversal rule is set to read the node number in the direction of increasing sequence index. During the reading process, it writes the node number into the path record list one by one until the branch termination mark number is -1. It stops traversing when the number of nodes in the path record list is accumulated and counted. The initial value of the number of nodes is set to 1 and increases with the number of traversed nodes. The node number sequence and the node number value are written into the path record entry. A path record entry is generated independently for each deviation branch. All path record entries are collected to form the branch connection path data.

[0069] The isolation marker generation submodule: Based on branch connection path data, it processes the path data using the isolated forest path hierarchy calculation method. During execution, the number of nodes in each path record is used as input, with the number of input items matching the number of deviation branches. The node number range is limited to 2 to 32. An isolation tree structure is constructed, with 100 isolation trees. The node number input item in each isolation tree is written in random order, and a random split operation is performed at each level. The split threshold is set as an integer between the minimum and maximum number of nodes. The hierarchy value of each input item when reaching the terminal node in the isolation tree is recorded. The arithmetic mean of the hierarchy values ​​of the same input item in the 100 isolation trees is calculated, and the average hierarchy value is retained to two decimal places. Simultaneously, the spatial position number corresponding to the branch in the dot matrix sequence identifier is read, with the spatial position number range limited to 0 to 63. The average hierarchy value and the spatial position number are combined sequentially and written into the record entry to form an abnormal isolation marker.

[0070] Isolation Forest constructs an input item set consisting of path length values ​​based on the number of nodes recorded in the branch connection path data. It randomly selects dimensions and sets random split points for the input item set to generate multiple isolated path structures and records the path level values ​​corresponding to the input items in each isolated path structure. The path level values ​​are then aggregated to form an isolation level record.

[0071] An isolated forest, according to the formula:

[0072]

[0073] in: To improve the isolation status value, The number of nodes is taken from the branch connection path data as input, and 2 is the base of the exponentiation operation. This refers to the path hierarchy weight coefficient. Spatial bias weighting coefficient, For missing penalty weighting coefficients, The topology penalty weight coefficient, The isolation tree index is set to a value ranging from 1 to 100. For input items In the The isolation path level value in the isolation tree. For the isolation tree sequence number Summation operator from 1 to 100 This is used as an average coefficient to convert the summation result into the average path level value of 100 isolated trees. The branch spatial locations are numbered with values ​​ranging from 0 to 63. 32 is the center reference value for the location number and is used to calculate the relative center offset. This represents the absolute offset of the location number relative to the center reference value, with 32 as the normalized denominator used to scale the absolute offset to the range of 0 to 1. For the natural logarithm function operator, The missing percentage is the ratio of the number of missing lattice elements to the number of lattice elements that should be present. This represents the degree of topological inconsistency, and its value is the percentage of numbers with a value of 0 in the connection relationship comparison marker and the arrangement order comparison marker in the topological constraint results. To stabilize the term and to avoid a denominator of 0, For harmonic numbers and represent The calculation results The sum index of the harmonic number is calculated and takes values ​​ranging from 1 to 1. , The sample size of the input item set is defined as the number of input items corresponding to the number of deviation branches. This is the noise penalty factor. The noise figure is identified by summing the standard deviations of the lateral and longitudinal differences in the spatial relationship set of the lattice.

[0074] Execution process: First, the dot matrix encoding construction module extracts the branch connection path data corresponding to the off-branch from the Braille dot matrix star topology and writes the number of nodes of each off-branch into the input item. The number of input items is recorded to obtain the sample size. The anomaly isolation judgment module constructs 100 isolation trees and assigns them sequence numbers. Perform a random partitioning and isolation process on trees numbered 1 to 100, and record the relationship between each isolation tree and the input item. Path hierarchy value generated and to Dimension accumulation and multiply by The average path level term is obtained, and then the spatial location number corresponding to the off-branch is read from the lattice sequence identifier. And calculate the spatial bias term. The missing proportion is obtained from the missing data of the matrix. And calculate the missing penalty item. The topology constraint results output by the star topology constraint module statistically determine the degree of topology inconsistency. And calculate the topology penalty term. The noise figure is calculated from the spatial relation set of the lattice. Combined with noise penalty coefficient Formation of noise suppression term At the same time, based on the sample size Calculate the normalization term Then, multiply the average path level term, spatial bias term, missing value penalty term, and topology penalty term by their respective weighting coefficients. The terms β, γ, and δ are summed to form the numerator of the exponent, and the product of the normalization term and the noise suppression term is used as the denominator of the exponent, with the result negative to form the exponent value. Finally, the improved isolation state value is obtained by performing an exponentiation operation with base 2. and will The isolation status value, serving as an anomaly isolation identifier, is written into the traceability coding and solidification module and established with the occurrence node to form a verifiable isolation level record, supporting anti-counterfeiting verification and traceability.

[0075] The interval boundary discrimination module includes:

[0076] Isolation Identifier Reading Submodule: Based on the abnormal isolation identifier, the branch number field is read item by item and rearranged in order. The isolation status value is extracted item by item and bound to the number to form an index entry set that corresponds to the dot matrix sequence identifier, thus obtaining the isolation index set.

[0077] Boundary location extraction submodule: Based on the isolation index set, read the isolation state value and arrangement number corresponding to each index, map the isolation state value to the interval to which the state value belongs according to the boundary division relationship corresponding to the support vector machine, classify the position, record the interval number corresponding to each index, and generate a boundary location set;

[0078] Discrimination identifier generation submodule: Based on the boundary location record set, read the permutation number and segment number in the record item by item, combine them into an identifier unit sequence in a predetermined order, and write the identifier unit sequence into a unified number position to form a boundary discrimination identifier;

[0079] The isolation identifier reading submodule: Based on the abnormal isolation identifier, it uses a sequential index rearrangement algorithm to read the branch number field. During execution, it reads the corresponding branch number of each record in the abnormal isolation identifier in the storage order. The branch number value range is set from 1 to 8. The read number is written to a temporary sequence cache and rearranged in order using a fixed-length array. The rearrangement rules are pre-written into the configuration file. The configuration content includes setting the standard branch order list to 1, 2, 3, 4, 5, 6, 7, 8. During execution, the positions of the numbers in the cache are aligned in order, and a placeholder mark with a value of 0 is written for the positions of the non-existent numbers. After the number rearrangement is completed, the isolation status value recorded in the abnormal isolation identifier is read one by one. The isolation status value is a floating-point number with 2 decimal places. The isolation status value is bound one by one with the corresponding rearranged branch number and written into the index record entry. The index record entries are linearly arranged according to the rearranged number order to form a continuous index sequence. The entire index sequence is written into the set field to generate an isolation index set.

[0080] Boundary Location Extraction Submodule: Based on the isolation index set, a support vector machine boundary partitioning algorithm is used to perform interval mapping processing on the isolation state values. During execution, the isolation state values ​​and corresponding permutation numbers are read one by one from the isolation index set. The permutation numbers are set to a range of 0 to 63. The isolation state values ​​and permutation numbers are combined to form a two-dimensional input record, and the preset support vector machine model parameters are loaded. The model parameters are written to the model file through offline training. The parameter content includes the kernel function type set to linear kernel, the penalty parameter value set to 1.0, and the interval tolerance parameter value set to 0.001. During the execution process, boundary judgment calculation is performed on each input record. The calculation process is completed by linear weighted summation and comparison with a threshold. The judgment result is mapped to the interval number. The interval number is preset to 4 and takes the values ​​0, 1, 2, and 3. The interval number corresponding to each input record is written to the position record table, and the position record table is sorted and collected according to the permutation number order to generate the boundary location set.

[0081] The discrimination identifier generation submodule generates discrimination identifiers based on the boundary position set using a sequential concatenation encoding method. During execution, it reads the permutation number and corresponding interval number from the boundary position set one by one. The permutation number is used as the position index and arranged in ascending order. The interval number is written as the content value into the encoding buffer. During the writing process, a fixed-length encoding method is used, with the encoding bit width set to 2 decimal digits. Zero-padding is performed on numbers with insufficient bit width. After all permutation numbers are written, the contents of the encoding buffer are concatenated sequentially according to the permutation order to form a continuous identifier unit sequence. The identifier unit sequence is written to a unified number position and the writing length is locked at 128 bits. The identifier writing registration is completed, forming the boundary discrimination identifier.

[0082] Support Vector Machine (SVM) constructs an input item set consisting of state values ​​and permutation numbers by recording the isolation state values ​​and corresponding permutation numbers in the isolation index set. It performs boundary function calculation on the input item set and determines the positional relationship of each input item relative to the separation boundary during the calculation process. The positional relationship is mapped to interval numbers according to the preset interval division rules, forming interval mapping records corresponding to each index.

[0083] Support Vector Machine, according to the formula:

[0084]

[0085] in: To improve the classification output notation and to characterize the positional relationship of input items relative to the separating hyperplane, The input vector to be discriminated consists of isolation state values, permutation numbers, spatial bias, and topological consistency factors. (sign) The sign function is used to map the output of the decision function to -1. Support vector index with values ​​ranging from 1 to , The number of support vectors is written to the offline trained model file and loaded during inference. For the first Each support vector corresponds to a Lagrange multiplier, which is read from the model parameter file. For the first Each support vector corresponds to a class label and takes values ​​of 1 and -1. For support vector indexing From 1 to Perform a weighted summation operator. This is a linear kernel inner product operator used to calculate the similarity between the input vector and the extended feature vectors of the support vectors. The bias term is read from the model parameter file. The isolation status value is a floating-point number derived from the isolation status record in the isolation index set and rounded to two decimal places. The mean of the isolation states is obtained statistically from the training samples and written into the model parameter file. The isolation state standard deviation is obtained statistically from the training samples and written into the model parameter file. The isolation state weighting coefficient ranges from 0.35 to 0.65. To arrange the numbers and take values ​​from 0 to 63, The mean of the permuted numbers is obtained statistically from the training samples and written into the model parameter file. The standard deviation of the numbering is obtained by statistical analysis of the training samples and written into the model parameter file. The weighting coefficients for the numbering are set and range from 0.10 to 0.30. The spatial location is assigned a number ranging from 0 to 63 and is read from the lattice sequence identifier. 32 is the spatial location center reference value and is used to generate the offset normalization quantity. This represents the absolute offset of the spatial location relative to the central reference value. This is a spatial offset normalization value used to scale the offset to the range of 0 to 1. This is the spatial bias weighting coefficient, with values ​​ranging from 0.05 to 0.25. This is the topology consistency factor, and its value is the percentage of values ​​of 1 in the connection relationship comparison marker and the sort order comparison marker in the topology constraint results, rounded to 4 decimal places. This is the topology consistency weighting coefficient, with a value ranging from 0.05 to 0.30. For the first Each support vector expands the feature vector and is read from the model parameter file. For the first Each support vector corresponds to a standardized component of the isolated state and is read from the model parameter file. For the first Each support vector corresponds to a numbered, standardized component, which is read from the model parameter file. For the first Each support vector corresponds to a spatial bias component, which is read from the model parameter file. For the first Each support vector corresponds to a topological consistency component and is read from the model parameter file;

[0086] Execution process: First, the interval boundary discrimination module reads the isolation status values ​​one by one from the isolation index set. With arrangement number The spatial location number corresponding to the index is read from the dot matrix sequence identifier. The star topology constraint module outputs the topology constraint results, which are obtained by comparing the statistical connection relationships with the arrangement order, comparing the number of 1 values ​​in the labels, and dividing by the total number of labels to obtain the topology consistency factor. Load the support vector machine model parameter file and read the number of support vectors. And read the index of each support vector. Corresponding Lagrange multipliers With category tags And read the bias term And read the extended feature support vectors Then read the mean of the isolation state from the model parameter file. Standard deviation of isolation status And calculate the standardized isolation state. And read the average of the permutation numbers. Standard deviation of permutation number And calculate the standardized permutation number. Computational space bias normalization And thus the weighting coefficients , The verification process includes fixing the kernel function type to a linear kernel and the penalty parameter to 1.0 in the historical labeled sample set. The value ranges from 0.35 to 0.65, with a step size of 0.05. The value ranges from 0.10 to 0.30, with a step size of 0.05. The value ranges from 0.05 to 0.25, with a step size of 0.05. Calculated and filtered Valid combinations falling between 0.05 and 0.30 are selected, and for each valid combination, the classification output of all samples is calculated. The interval number consistency score is used as the objective function, and the combination with the highest score is fixed as the final weight coefficient and written into the model parameter file. Subsequently, a weighted input vector is constructed. For support vector indexes From 1 to Calculate the weighted input vector item by item and the first... Support vectors expand feature vectors inner product and multiplied by Perform cumulative summation and add the bias term. The decision function output value is obtained, and finally the sign function is executed on the decision function output value. Obtain the classification output symbol According to the preset interval division rules, the classification output symbols are mapped to interval numbers and written into the interval mapping record. An index is established and bound with the permutation number so that the traceability coding solidification module can write it into the traceability link to complete anti-counterfeiting verification and abnormal traceability location.

[0087] The traceability coding and solidification module includes:

[0088] The association relationship construction submodule reads the number field from the boundary discrimination identifier and extracts the corresponding occurrence node identifier. It then aligns the number field with the node identifier, writes the numbered items and node items accordingly, marks the order position, confirms the items, and establishes the association relationship record.

[0089] Link writing and solidification submodule: Based on the association relationship record, read the storage writing order and extract the corresponding sequence number, retrieve the record item according to the sequence number and write it to the link position, synchronously write the sequence identifier to each record and complete the position registration to form a traceability coding record;

[0090] The association construction submodule: Based on the boundary discrimination identifier, it uses a sequential alignment mapping algorithm to bind the number field and the occurrence node identifier. During execution, it reads the number field from the boundary discrimination identifier in byte order. The number field length is set to 128 bits, and each 2-digit decimal number is used as an item number unit. At the same time, it reads the occurrence node identifier from the node registration table. The occurrence node identifier is recorded in integer form and the value range is 1 to 999. The number unit sequence and the occurrence node identifier sequence are sequentially aligned according to the index position. The alignment rule is set to perform one-to-one writing when the index positions are the same. During the writing process, a sequential position value is assigned to each pair of number units and node identifiers. The starting value of the sequential position is set to 1 and increments according to the writing order. The number unit, node identifier, and sequential position are collected and written into the relationship table record item. After performing integrity verification on all record items, the writing state is locked, and the association relationship record is generated.

[0091] Link writing and solidification submodule: Based on the associated records, the sequence number-driven chain writing method is used to perform record solidification processing. During execution, the writing sequence number list is read from the sequence control table. The sequence number is in integer form and arranged in ascending order. The associated record items are retrieved one by one according to the sequence number and written to the specified location in the traceability link storage area. The link storage area adopts a linear address structure and uses the sequence number as the address offset. During the writing process, a sequence identifier field is synchronously written for each record. The value of the sequence identifier field is consistent with the sequence number. After the writing is completed, the location registration operation is performed, the sequence number and link address are written to the index table and the entry status is locked to form a traceability coded record.

[0092] 10. A method for anti-counterfeiting and traceability based on a Braille dot matrix star topology, characterized in that, when executed according to any one of claims 1-9, the method comprises the following steps:

[0093] S1: Based on the dot matrix distribution on the surface of the identified object, collect the spatial location data of the dot matrix units, and perform radial sorting of the dot matrix units with a preset reference point as a benchmark to form a set of dot matrix identifiers with sequential relationships, thus obtaining the dot matrix sequence identifier.

[0094] S2: Based on the dot matrix sequence identifier, read the connection relationship information between the central unit and the branch unit, compare it with the established connection relationship set, and check the branch arrangement order to obtain the topological constraint result corresponding to the dot matrix structure.

[0095] S3: Based on the dot matrix sequence identifier and topological constraint results, identify the deviation branch identifier, and combine the branch connection path length information with the branch spatial distribution position to generate an abnormal isolation identifier;

[0096] S4: Based on the anomaly isolation identifier, according to the boundary division relationship corresponding to the support vector machine, the content corresponding to the anomaly isolation identifier is mapped to the interval to determine the position of the dot matrix sequence identifier in the boundary division interval, and the boundary discrimination identifier is formed and output.

[0097] S5: Based on the boundary discrimination identifier, the boundary discrimination identifier is associated with the occurrence node by number, and written into the traceability link according to the storage writing order. The writing order is marked to form a traceability coding record.

[0098] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A counterfeit-proof and traceability system based on a Braille dot matrix star topology, characterized in that, The system includes: Dot matrix encoding construction module: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix units, and establish a radial arrangement relationship with the preset reference points in the dot matrix distribution as a reference to form a dot matrix sequence identifier; Star topology constraint module: Based on the matrix sequence identifier, compare the connection relationship between the central unit and the branch unit with the stored set of predetermined connection relationships, and compare the branch arrangement order with the stored predetermined arrangement order to obtain the topology constraint result; Anomaly Isolation Judgment Module: Based on the lattice sequence identifier and the topological constraint result, according to the isolation path organization rules corresponding to the isolated forest, constructs branch connection paths based on the deviation branch identifier, calculates the branch connection path length and branch spatial distribution position, and generates anomaly isolation identifiers; Interval Boundary Discrimination Module: Based on the anomaly isolation identifier, according to the boundary division relationship corresponding to the support vector machine, extract the relative position of the dot matrix sequence identifier in the discrimination boundary, generate the boundary discrimination identifier, and output the boundary discrimination identifier; Source tracing coding solidification module: Based on the boundary discrimination identifier, a fixed association relationship is established between the boundary discrimination identifier and the occurrence node, and the record is written into the source tracing link according to the storage writing order to form a source tracing coding record.

2. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The dot matrix sequence identifier includes a dot matrix unit identifier, a radial arrangement order marker, and a position identifier corresponding to a preset reference point. The topology constraint result includes a determination result of the connection relationship between the central unit and the branch unit, a determination result of the branch arrangement order, and an inconsistency state marker. The abnormal isolation identifier includes a deviation branch identifier, a branch connection path length marker, and a branch spatial distribution position identifier. The boundary discrimination identifier includes an interval number identifier, an arrangement position identifier, and a boundary position marker corresponding to the isolation state. The traceability coding record includes a boundary discrimination identifier, an occurrence node identifier, and a writing order marker.

3. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The dot matrix encoding construction module includes: Dot matrix position sorting submodule: Based on the dot matrix distribution formed on the surface of the identified object, read the spatial position relationship of the dot matrix unit, record the corresponding coordinate position of the dot matrix unit on the surface one by one, collect the coordinate values ​​and calculate the adjacent difference, and generate a dot matrix spatial relationship set. Radial sequence construction submodule: Based on the set of spatial relationships of the dot matrix, and taking the preset reference point in the dot matrix distribution as the benchmark, sort the relative direction values ​​of the dot matrix units, perform radial grouping and sequence verification, and concatenate the grouping results in sequence to form a dot matrix sequence identifier.

4. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The star topology constraint module includes: Connection relationship verification submodule: Based on the dot matrix sequence identifier, extract the connection items corresponding to the central unit and the branch unit, read the connection pointer number and connection quantity marker, compare the position mapping with the stored set of connection relationships in order, and generate connection relationship comparison results; The order comparison submodule: Based on the connection relationship comparison result, it calls the branch order item in the dot matrix sequence identifier, reads the order index number bit by bit, and performs a sequence comparison with the stored predetermined order according to the index position to obtain the topological constraint result.

5. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The anomaly isolation determination module includes: Off-branch identification submodule: Based on the dot matrix sequence identifier and the topological constraint result, read the branch number information recorded in the topological constraint result, extract the corresponding branch index in the order of the number, and check the corresponding branch position in the dot matrix sequence identifier to generate a set of off-branch branches; Connection path construction submodule: Based on the set of off-branch nodes, locate the corresponding branch node position in the dot matrix sequence identifier, connect the node identifiers segment by segment along the connection order between nodes, and accumulate the number of connected nodes to form a node order list to obtain the branch connection path data; Isolation Identifier Generation Submodule: Based on the branch connection path data, read the number of nodes in the connection path and retrieve the spatial location number of the branch in the dot matrix sequence identifier. According to the correspondence between the isolation path length and the node distribution position in the isolated forest, combine and record the number of nodes and the spatial location number to form an abnormal isolation identifier corresponding to the branch.

6. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The isolated forest constructs an input item set composed of path length values ​​based on the number of nodes recorded in the branch connection path data. Random dimension selection and random splitting point setting are performed on the input item set to generate multiple sets of isolated path structures. The path level values ​​corresponding to the input items are recorded in each isolated path structure, and the path level values ​​are collected to form an isolated level record.

7. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The interval boundary discrimination module includes: Isolation Identifier Reading Submodule: Based on the abnormal isolation identifier, the branch number field is read item by item and rearranged in order, and the isolation status value is extracted and bound to the number one by one to form an index entry set that corresponds to the dot matrix sequence identifier, thus obtaining the isolation index set; Boundary location extraction submodule: Based on the isolation index set, read the isolation state value and arrangement number corresponding to each index, map the isolation state value to the interval to which the state value belongs according to the boundary division relationship corresponding to the support vector machine and classify the position, record the interval number corresponding to each index, and generate a boundary location set; Discrimination identifier generation submodule: Based on the boundary position record set, read the arrangement number and segment number in the record one by one, combine them into an identifier unit sequence in a predetermined order, and write the identifier unit sequence into a unified number position to form a boundary discrimination identifier.

8. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The support vector machine constructs an input item set consisting of state values ​​and arrangement numbers by recording the isolation state values ​​and corresponding arrangement numbers in the isolation index set. It performs boundary function calculation on the input item set and determines the positional relationship of each input item relative to the separation boundary during the calculation process. The positional relationship is mapped to interval numbers according to the preset interval division rules to form interval mapping records corresponding to each index.

9. The anti-counterfeiting and traceability system based on a Braille dot matrix star topology according to claim 1, characterized in that, The traceability coding and solidification module includes: Relationship Construction Submodule: Based on the boundary discrimination identifier, read the number field in the boundary discrimination identifier and extract the corresponding occurrence node identifier, align the number field and node identifier in order, write the number item and node item accordingly, mark the order position and confirm the item, and establish a relationship record; Link writing and solidification submodule: Based on the association record, read the stored writing order and extract the corresponding sequence number, retrieve the record item according to the sequence number and write it to the link position, synchronously write the sequence identifier to each record and complete the position registration to form a traceability coding record.

10. A method for anti-counterfeiting and traceability based on a Braille dot matrix star topology, characterized in that, The anti-counterfeiting and traceability system based on the Braille dot matrix star topology structure according to any one of claims 1-9 includes the following steps: S1: Based on the dot matrix distribution on the surface of the identified object, collect the spatial location data of the dot matrix units, and perform radial sorting of the dot matrix units with a preset reference point as a benchmark to form a set of dot matrix identifiers with sequential relationships, thus obtaining the dot matrix sequence identifier. S2: Based on the dot matrix sequence identifier, read the connection relationship information between the central unit and the branch unit, compare it with the predetermined connection relationship set, and check the branch arrangement order to obtain the topological constraint result corresponding to the dot matrix structure. S3: Based on the dot matrix sequence identifier and the topological constraint result, identify the deviation branch identifier, and combine the branch connection path length information with the branch spatial distribution position to generate an abnormal isolation identifier; S4: Based on the anomaly isolation identifier, according to the boundary division relationship corresponding to the support vector machine, perform interval mapping on the content corresponding to the anomaly isolation identifier, determine the position of the dot matrix sequence identifier in the boundary division interval, and form and output the boundary discrimination identifier; S5: Based on the boundary discrimination identifier, the boundary discrimination identifier is associated with the occurrence node by number, and written into the traceability link according to the storage writing order. The writing order is marked to form a traceability coding record.

Images