A batch information tracing system and method for effervescent and sustained-release health products

Abstract

A batch information traceability system and method for effervescent and sustained-release health products, belonging to the field of product traceability technology, aims to solve the problem of inaccurately locating defective sub-batches. The batch information traceability method for effervescent and sustained-release health products includes: obtaining the manufacturing environment sequence of the effervescent and sustained-release health products; extracting the transient anomaly node set from the manufacturing environment sequence; generating a microscopic state fingerprint of the effervescent and sustained-release health products based on the transient anomaly node set; constructing a sustained-release coating forming feature map; performing dimensionality reduction and fusion of the microscopic state fingerprint and the sustained-release coating forming feature map to generate batch quality mapping primitives; injecting the batch quality mapping primitives into the redundant fields of the basic batch code to generate a composite traceability code matrix; and outputting the composite traceability code matrix to the surface printing device of the corresponding product batch. This invention improves the accuracy of batch traceability for effervescent and sustained-release health products and the efficiency of quality defect location.

Description

Technical Field

[0001] This invention relates to the field of product traceability technology, specifically to a batch information traceability system and method for effervescent and sustained-release health products. Background Technology

[0002] Batch traceability for health products is a key technological means of recording product identity information at the production, warehousing, logistics, and sales stages. By assigning corresponding one-dimensional or two-dimensional codes to the packaging, the production date, material source, and corresponding production team information of the batch of products can be recorded. Due to their special physicochemical properties, the quality of effervescent and sustained-release health products is highly susceptible to fluctuations in the production microenvironment. If the relative humidity of the environment surges briefly during the tableting stage of effervescent tablets, it can easily cause the tablets to absorb moisture, affecting the final foaming effect. During the coating stage of sustained-release products, even a slight imbalance in the ratio of spray flow rate to inlet air temperature can lead to uneven coating thickness, thereby altering the release curve of the drug or nutrient.

[0003] Existing batch traceability systems mainly remain at the level of macro data recording; batch codes can only correspond to static identifiers within a continuous production cycle and cannot be linked to the transient micro-environmental physical state generated during the manufacturing stage; when subsequent market feedback indicates that a certain batch of products has abnormal dissolution or effervescence failure, the lack of traceability information linking to fluctuations in the micro-environment of pharmaceutical equipment often leads to the inability to accurately locate specific defective sub-batch, thereby forcing companies to implement large-scale blind recall measures.

[0004] To address the above issues, a batch information traceability system and method for effervescent and sustained-release health products is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a batch information traceability system and method for effervescent and sustained-release health products. By using this device, the problems in the above-mentioned background are solved: batch traceability of effervescent and sustained-release health products only records macroscopic static information, cannot be linked to transient micro-environmental fluctuations in production, cannot accurately locate defective sub-batch when quality is abnormal, and can only be recalled on a large scale.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for batch information traceability of effervescent and sustained-release health products, the method comprising: A manufacturing environment sequence for effervescent and sustained-release health products is obtained; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than the data dimension of the first temperature and humidity matrix. Extract the set of transient mutation nodes from the manufacturing environment sequence; generate the microscopic state fingerprint of the effervescent and sustained-release health products based on the set of transient mutation nodes; Obtain the coating spray flow rate sequence and the inlet air temperature sequence during the coating stage; calculate the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence to construct a slow-release coating forming feature map; the sampling period of the inlet air temperature sequence is greater than the sampling period of the coating spray flow rate sequence; The microscopic fingerprint and the slow-release coating forming feature map are dimensionality-reduced and fused to generate batch quality mapping primitives. The batch quality mapping primitive is injected into the redundant field of the basic batch code to generate a composite traceability code matrix, and the composite traceability code matrix is ​​output to the surface printing device of the corresponding product batch.

[0007] Further, the extraction of the transient variable node set in the manufacturing environment sequence includes: Calculate the first difference between adjacent time slices in the first temperature and humidity matrix and the second difference between adjacent time slices in the second temperature and humidity matrix; the weighting coefficient of the second difference is greater than the weighting coefficient of the first difference; Set a basic mutation threshold; If the first difference is greater than the basic anomaly threshold, the corresponding time slice is marked as a first-level abnormal node; If the second difference is greater than the basic anomaly threshold, the corresponding time slice is marked as a second-level abnormal node; The first-level abnormal nodes and the second-level abnormal nodes are aggregated to generate the transient abnormal node set.

[0008] Further, generating the microscopic state fingerprint of the effervescent and sustained-release health products based on the transient mutation node set includes: Extract the first environmental time sequence identifier of the time when the first-level abnormal node occurs; Extract the second environmental time sequence identifier of the time when the second-level anomaly node occurs; The locality-sensitive hashing algorithm is used to perform feature mapping between the first environment time sequence identifier and the second environment time sequence identifier; The numerical values ​​after feature mapping are concatenated to output a one-dimensional hexadecimal microstate fingerprint; the character length of the microstate fingerprint is less than the original byte length of the transient mutation node set.

[0009] Further, calculating the ratio trajectory of the coating spray flow rate sequence to the inlet air temperature sequence includes: The coating spray flow rate sequence and the inlet air temperature sequence are time-stamp aligned. Extract the coating spray flow rate and inlet air temperature values ​​at the same timestamp; Calculate the quotient of the coating spray flow rate value divided by the inlet air temperature value to obtain a discrete ratio point set; The discrete ratio point set is smoothly fitted using a spline interpolation function to output a continuous ratio trajectory.

[0010] Furthermore, the construction of the sustained-release coating molding feature map includes: Calculate the set of first derivatives of the ratio trajectory; Extract the zero-crossing points from the set of first derivatives; The peak and trough regions corresponding to the zero-crossing points are transformed into two-dimensional thermodynamic images to construct the characteristic map of the slow-release coating formation.

[0011] Further, the step of dimensionality reduction and fusion of the microscopic fingerprint and the sustained-release coating molding feature map to generate batch quality mapping primitives includes: The core spectral vector of the sustained-release coating forming feature map was extracted using principal component analysis. The microscopic fingerprint is converted into a digital array; Perform a tensor product operation between the digital array and the core graph vector; The tensor product operation result is binarized to generate the batch quality mapping primitive; the data sparsity of the batch quality mapping primitive is greater than the data sparsity of the core graph vector.

[0012] Further, the step of injecting the batch quality mapping primitive into the redundant field of the basic batch code to generate a composite traceability code matrix includes: Parse the data structure of the basic batch code; Locate the unoccupied blank validation area in the data structure and set the blank validation area as the redundant field; The batch quality mapping primitive is written into the redundant field using an error correction coding algorithm; According to the preset QR code array rules, the basic batch code after the data is written is recombined to generate the composite traceability code matrix.

[0013] Furthermore, after outputting the composite traceability code matrix to the surface printing device for the corresponding product batch, the method further includes: Receive a traceability parsing request sent by a barcode scanning terminal; the traceability parsing request carries the scanned composite traceability code matrix to be parsed; Extract and separate the plaintext batch information and hidden basic data from the composite traceability code matrix to be parsed; the read permission level of the plaintext batch information is lower than the read permission level of the hidden basic data.

[0014] Furthermore, extracting and separating the plaintext batch information and hidden primitive data from the composite traceability code matrix to be parsed also includes: The hidden primitive data is restored by using a preset reverse decoding tree to reconstruct the target microstate fingerprint; The target micro-state fingerprint is matched with a preset risk knowledge base for similarity matching; If the similarity reaches the risk threshold, a warning pop-up message will be generated for the individual package. Record the time when the transient anomaly node corresponding to the target microstate fingerprint that triggers the warning pop-up command occurs; Lock the edge sub-batch product sequence that is at the same processing station during the time of occurrence; An interception and isolation command is issued to the edge sub-batch product sequence; the number of the edge sub-batch product sequence is less than the total number of products covered by the plaintext batch information.

[0015] This invention also discloses another technical solution: a batch information traceability system for effervescent and sustained-release health products, the system comprising: The manufacturing environment sequence acquisition module acquires the manufacturing environment sequence of effervescent and sustained-release health products; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than that of the first temperature and humidity matrix. The microstate fingerprint generation module extracts the transient anomaly node set from the manufacturing environment sequence; and generates the microstate fingerprint of the effervescent and sustained-release health products based on the transient anomaly node set. The slow-release coating feature map construction module acquires the coating spray flow rate sequence and the inlet air temperature sequence during the coating stage; calculates the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence, and constructs a slow-release coating forming feature map; the sampling period of the inlet air temperature sequence is greater than the sampling period of the coating spray flow rate sequence; The quality mapping primitive fusion module performs dimensionality reduction fusion of the micro-state fingerprint and the slow-release coating forming feature map to generate batch quality mapping primitives. The composite traceability code generation and output module injects the batch quality mapping primitive into the redundant field of the basic batch code to generate a composite traceability code matrix, and outputs the composite traceability code matrix to the surface printing device of the corresponding product batch.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention upgrades the traditional static batch identification into a dynamic traceability carrier that is deeply bound to the transient microenvironment fluctuations in production by simultaneously collecting multi-dimensional manufacturing environment sequences during the tableting and coating stages. It fully associates the physicochemical state of effervescent and sustained-release health products with the production microenvironment parameters, fundamentally filling the technical gap where batch information is disconnected from the production microenvironment.

[0017] 2. To address the issue that existing traceability systems cannot accurately pinpoint defective sub-batches when there are micro-environmental fluctuations in pharmaceutical equipment or when quality anomalies occur, a dual feature system is constructed, combining micro-state fingerprints and sustained-release coating molding characteristic maps. This system enables precise marking and location of transient anomalies and process ratio deviations during the production process. When quality anomalies occur, defective sub-batches can be directly identified, replacing the large-scale blind recall model and reducing recall losses and market operation costs for enterprises.

[0018] 3. To address the issues of moisture absorption failure caused by humidity fluctuations during the tableting stage of effervescent tablets and uneven coating caused by spray and temperature imbalances during the coating stage of sustained-release products, this study collects environmental data from the tableting and coating stages in a differentiated manner and assigns weights to distinguish between them. This allows for early detection of transient changes in the microenvironment and risks of process imbalances, providing early warnings of typical quality problems such as moisture absorption during tableting and uneven coating thickness, thereby improving the stability and pass rate of finished effervescent and sustained-release health products.

[0019] 4. This invention injects microscopic state fingerprints and process feature maps into redundant fields of basic batch codes through dimensionality reduction fusion, achieving implicit quality feature embedding without altering the appearance and basic structure of existing batch codes. It is compatible with existing packaging printing and barcode scanning equipment, requiring no additional hardware and exhibiting strong adaptability for on-site deployment. Through hierarchical permission design, it separates plaintext batch information from implicit quality data, catering to multiple scenarios such as basic queries by ordinary consumers, internal quality control by enterprises, and precise traceability by regulatory authorities, thereby improving the security and practicality of traceability data. Attached Figure Description

[0020] Figure 1 This is a flowchart outlining the steps of the present invention. Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0021] 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.

[0022] Example 1: As Figure 1 As shown in the figure, this embodiment provides a method for batch information traceability of effervescent and sustained-release health products. The method includes: Step S101: Obtain the manufacturing environment sequence of effervescent and sustained-release health products; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than that of the first temperature and humidity matrix.

[0023] Step S102: Extract the transient mutation node set from the manufacturing environment sequence; generate the microscopic state fingerprint of effervescent and sustained-release health products based on the transient mutation node set.

[0024] Step S103: Obtain the coating spray flow rate sequence and the inlet air temperature sequence during the coating stage; calculate the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence, and construct a characteristic map of the slow-release coating formation; the sampling period of the inlet air temperature sequence is greater than the sampling period of the coating spray flow rate sequence.

[0025] Step S104: Dimensionally reduce and fuse the microscopic fingerprint and the slow-release coating forming feature map to generate batch quality mapping primitives.

[0026] Step S105: Inject the batch quality mapping primitive into the redundant field of the basic batch code to generate a composite traceability code matrix, and output the composite traceability code matrix to the surface printing device of the corresponding product batch.

[0027] In some embodiments, a manufacturing environment sequence of effervescent and sustained-release health products is obtained; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than that of the first temperature and humidity matrix; a set of transient variation nodes is extracted from the manufacturing environment sequence; a microscopic state fingerprint of the effervescent and sustained-release health products is generated based on the set of transient variation nodes; a coating spray flow rate sequence and an inlet air temperature sequence are obtained for the coating stage; the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence is calculated to construct a sustained-release coating forming feature map; the sampling period of the inlet air temperature sequence is greater than that of the coating spray flow rate sequence; the microscopic state fingerprint and the sustained-release coating forming feature map are dimensionality-reduced and fused to generate a batch quality mapping primitive; the batch quality mapping primitive is injected into the redundant fields of the basic batch code to generate a composite traceability code matrix, and the composite traceability code matrix is ​​output to the surface printing device of the corresponding product batch.

[0028] Specifically, environmental data continuously transmitted by temperature and humidity sensing devices deployed at the tableting production station is collected; temperature and humidity values ​​within a continuous time period during the tableting stage are organized into a fixed-dimensional data array according to chronological order; environmental data continuously transmitted by various environmental sensing devices deployed at the coating production station is collected; temperature, humidity, and spatial location data within a continuous time period during the coating stage are integrated into a higher-dimensional data array according to chronological order; the number of environmental elements to be monitored in the coating production stage is greater than that in the tableting production stage; the number of dimensions in the higher-dimensional data array is always higher than that in the fixed-dimensional data array; all environmental data values ​​at adjacent time points in the manufacturing environment sequence are compared segment by segment; time points where environmental data values ​​suddenly change within a short period of time are selected; all selected time points with sudden changes are summarized into a fixed data set; production time information corresponding to all time points in the data set is extracted; and the extracted production time information undergoes feature preservation and data compression processing.

[0029] The compressed feature data is concatenated into a single-dimensional character identifier; the flow rate of the spray delivery equipment during the coating stage is collected; the flow rate values ​​within a continuous time period are arranged into a continuous data sequence according to time sequence; the temperature values ​​of the air inlet equipment during the coating stage are collected; the temperature values ​​within a continuous time period are arranged into a continuous data sequence according to time sequence; the rate of change of the air inlet temperature is lower than the rate of change of the spray flow rate; the data acquisition interval for the air inlet temperature is set to be longer than the data acquisition interval for the spray flow rate; data content with completely identical time nodes in the two sets of data sequences is filtered out; isolated data content with mismatched time nodes in the two sets of data sequences is removed; the spray flow rate value and the air inlet temperature value at the same time node are divided.

[0030] All division results are arranged chronologically as scattered numerical points; uniform transitional numerical content is added between the scattered numerical points; all added numerical points are connected into a continuous curve without breaks or drastic fluctuations; the rate of change of the continuous curve is calculated point by point; key points where the rate of change of numerical change reverses are selected; peak areas formed by continuously rising numerical values ​​and trough areas formed by continuously falling numerical values ​​are delineated around the key points; the numerical values ​​in the peak and trough areas are converted into different shades of color content; the color content is arranged according to spatial location to form a two-dimensional planar image; invalid background information and interfering data content in the two-dimensional planar image are removed; continuous core data that can represent the overall characteristics of the two-dimensional planar image is extracted; single-dimensional character identifiers are converted into continuous numerical arrangement content bit by bit; the continuous numerical arrangement content is fully integrated with the continuous core data of the two-dimensional planar image.

[0031] Values ​​greater than a fixed threshold in the fused data are uniformly set to one fixed value; values ​​less than or equal to the fixed threshold in the fused data are uniformly set to another fixed value; thus obtaining simplified core quality characteristic data after removing redundant information; disassembling the internal data layout structure of the product's conventional batch code; determining the function and storage location of each data segment in the conventional batch code; locating blank storage areas within the conventional batch code that do not carry any production information; converting the simplified core quality characteristic data into a format that can be written to the code; filling the converted format content into the blank storage area; rearranging and combining the batch codes after filling in the data according to the nationally unified QR code graphic layout rules; forming new QR code graphic data; transmitting the new QR code graphic data to the product's outer packaging dedicated printing equipment; and completing the traceability code printing operation on the corresponding batch of product outer packaging.

[0032] The manufacturing environment sequence is a continuous data set recording the physical environment of the workshop throughout the entire production process of effervescent and sustained-release health products; the first temperature and humidity matrix is ​​a fixed-dimensional data array recording temperature and humidity values ​​throughout the tableting stage; the second temperature and humidity matrix is ​​a higher-dimensional data array recording multiple environmental parameters throughout the coating stage; the transient anomaly node set is a summary set of time points in the manufacturing environment sequence where environmental data values ​​suddenly change; the microscopic state fingerprint is a single-dimensional character identifier representing sudden environmental changes during product production; the coating spray flow rate sequence is a continuous data sequence recording the change of spray flow rate values ​​over time during the coating stage; the inlet air temperature sequence is a continuous data sequence recording the change of inlet air temperature values ​​over time during the coating stage; the ratio trajectory is a continuous smooth curve fitted after calculating the division between spray flow rate values ​​and inlet air temperature values; sustained-release coating... The layer forming feature map is a two-dimensional planar image formed by converting the numerical features of the coating process curve; the batch quality mapping primitive is a simplified core quality feature data formed by fusing environmental change features and coating process features; the basic batch code is the original batch code data that records the product's routine production information; the redundant field is the blank storage area within the basic batch code that does not carry information; the composite traceability code matrix is ​​a QR code graphic data formed by rearranging the core quality feature data; multi-dimensional collection of production environment data and process data can comprehensively cover the core elements affecting product quality; unified data time series benchmarks and dimensional specifications can eliminate traceability deviations caused by data differences; the fusion of environmental features and process features and the embedding of batch codes can achieve implicit and accurate traceability of product quality; avoiding blind recall of entire batches of products caused by a single environmental anomaly can reduce the company's production and recall costs.

[0033] In some embodiments, extracting a set of transient anomaly nodes from a manufacturing environment sequence includes: calculating a first difference between adjacent time slices in a first temperature and humidity matrix and a second difference between adjacent time slices in a second temperature and humidity matrix; the weighting coefficient of the second difference is greater than the weighting coefficient of the first difference; setting a basic anomaly threshold; if the first difference is greater than the basic anomaly threshold, marking the corresponding time slice as a first-level anomaly node; if the second difference is greater than the basic anomaly threshold, marking the corresponding time slice as a second-level anomaly node; and aggregating the first-level anomaly nodes and the second-level anomaly nodes to generate a set of transient anomaly nodes.

[0034] Extract temperature and humidity values ​​from two adjacent time points in the fixed-dimensional data array during the tableting stage; subtract the value of the previous time point from the value of the later time point to obtain the variation range of environmental data between adjacent time points during the tableting stage. Extract all environmental parameter values ​​from two adjacent time points in the higher-dimensional data array during the coating stage; subtract the value of the previous time point from the value of the later time point to obtain the variation range of environmental data between adjacent time points during the coating stage. The impact of environmental changes during the coating stage on the final product quality is greater than that during the tableting stage; assign a higher impact weight to the variation range of environmental data during the coating stage; pre-set... Define the maximum permissible value for normal fluctuations in environmental parameters; this value serves as a fixed threshold for judging whether environmental data is abnormal; compare the fluctuation range of environmental data during the tableting stage with the fixed threshold; when the fluctuation range exceeds the fixed threshold, mark this time point as the time point of environmental abnormality during the tableting stage; compare the fluctuation range of environmental data during the coating stage with the fixed threshold; when the fluctuation range exceeds the fixed threshold, mark this time point as the time point of environmental abnormality during the coating stage; summarize and integrate all the time points of environmental abnormality during the tableting stage and the coating stage; form a set of time points of sudden changes in environmental data.

[0035] The first difference is the magnitude of environmental data variation at adjacent time points in the fixed-dimensional data array during the tableting stage; the second difference is the magnitude of environmental data variation at adjacent time points in the higher-dimensional data array during the coating stage; the basic anomaly threshold is the maximum allowable value for normal fluctuations in environmental parameters; the first-level abnormal node is the time point when environmental data in the tableting stage exceeds the normal fluctuation range; the second-level abnormal node is the time point when environmental data in the coating stage exceeds the normal fluctuation range; distinguishing the impact weights of environmental changes in different production stages can accurately locate core abnormal links; and hierarchical marking of environmental abnormal nodes can improve the accuracy of subsequent feature processing.

[0036] In some embodiments, generating a microstate fingerprint of effervescent and sustained-release health products based on a transient anomaly node set includes: extracting a first environmental time sequence identifier at the time of occurrence of a first-level anomaly node; extracting a second environmental time sequence identifier at the time of occurrence of a second-level anomaly node; performing feature mapping on the first environmental time sequence identifier and the second environmental time sequence identifier using a locality-sensitive hashing algorithm; concatenating the feature-mapped values ​​to output a one-dimensional hexadecimal microstate fingerprint; the character length of the microstate fingerprint is less than the original byte length of the transient anomaly node set.

[0037] Extract production time information corresponding to environmental anomalies during the tableting stage; convert the production time information into fixed-format digital identifiers; extract production time information corresponding to environmental anomalies during the coating stage; convert the production time information into fixed-format digital identifiers; input the two types of fixed-format digital identifiers into the data processing program; the data processing program retains the core feature content of the digital identifiers; the data processing program converts the core feature content into fixed-length feature data content; concatenate the two types of feature data sequentially according to the order of tableting stage first, then coating stage; form single-dimensional character data content; the data processing program removes redundant information from the original data; the final total length of the character data is shorter than the original data length of the set of environmental anomaly nodes.

[0038] Among them, the first environmental time sequence identifier is a fixed-format digital time identifier corresponding to the environmental anomaly time node in the tableting stage; the second environmental time sequence identifier is a fixed-format digital time identifier corresponding to the environmental anomaly time node in the coating stage; the micro-state fingerprint is a single-dimensional character data content formed by splicing feature data; compressing discrete environmental anomaly time information can reduce the space occupied by data storage and transmission; retaining core features can ensure the effectiveness of subsequent data fusion.

[0039] In some embodiments, calculating the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence includes: performing timestamp alignment processing on the coating spray flow rate sequence and the inlet air temperature sequence; extracting the coating spray flow rate value and the inlet air temperature value at the same timestamp; calculating the quotient of the coating spray flow rate value divided by the inlet air temperature value to obtain a discrete ratio point set; and using a spline interpolation function to smoothly fit the discrete ratio point set to output a continuous ratio trajectory.

[0040] Retrieve continuous data sequences of coating spray flow rate and inlet air temperature; filter data from both sets of data sequences where the time nodes are completely consistent; remove isolated data from the two sets of data sequences where the time nodes do not match; extract the spray flow rate and inlet air temperature values ​​corresponding to each matching time node; use the spray flow rate value as the dividend; use the inlet air temperature value as the divisor; perform division calculations to obtain the corresponding results; arrange all calculation results in chronological order; form a set of dispersed numerical points; supplement the dispersed numerical points with uniformly transitioning numerical content; connect all numerical points into a continuous curve without breaks or drastic fluctuations.

[0041] Timestamp alignment is an operation to filter data with consistent time nodes in two sets of data sequences; discrete ratio point set is a set of scattered numerical points formed by dividing spray flow rate and inlet air temperature; ratio trajectory is a continuous smooth curve formed after supplementing transitional values; unified process data time series benchmark can amplify the micro-characteristics of process ratio imbalance; smooth fitting curve can intuitively present the process fluctuation law of slow-release coating formation.

[0042] In some embodiments, constructing a feature map of sustained-release coating formation includes: calculating the set of first derivatives of the ratio trajectory; extracting zero-crossing points from the set of first derivatives; performing image conversion on the peak and trough regions corresponding to the zero-crossing points to obtain a two-dimensional thermodynamic image, and constructing a feature map of sustained-release coating formation.

[0043] The process involves: calculating the rate of change of values ​​for a continuous smooth curve point by point; obtaining the rate of change data for all points on the curve; identifying key points where the rate of change changes from positive to negative; identifying key points where the rate of change changes from negative to positive; delineating peak regions formed by continuously rising values ​​around key points; delineating trough regions formed by continuously falling values ​​around key points; converting the values ​​of peak and trough regions into different shades of color; arranging the color information according to spatial location to form a two-dimensional planar image; the set of first derivatives represents the rate of change data for all points on the continuous smooth curve; zero-crossing points are key points where the rate of change of values ​​on the curve reverses; a two-dimensional thermodynamic image is a two-dimensional planar image formed by transforming the characteristics of the numerical region; a slow-release coating forming characteristic map is a two-dimensional planar image formed by transforming the characteristics of the coating process curve; converting abstract process curves into visual images can intuitively represent process stability; and a standardized image carrier can be adapted for subsequent cross-modal data fusion operations.

[0044] In some embodiments, the micro-state fingerprint and the sustained-release coating forming feature map are dimensionality-reduced and fused to generate batch quality mapping primitives, including: extracting the core map vector of the sustained-release coating forming feature map using principal component analysis; converting the micro-state fingerprint into a digital array; performing a tensor product operation between the digital array and the core map vector; binarizing the tensor product operation result to generate batch quality mapping primitives; the data sparsity of the batch quality mapping primitives is greater than the data sparsity of the core map vector.

[0045] The process involves: removing invalid background information and interfering data from a 2D image; extracting continuous core data that represents the overall characteristics of the 2D image; converting single-dimensional character identifiers into continuous numerical arrangements; fully fusing the continuous numerical arrangements with the continuous core data of the 2D image; uniformly setting values ​​greater than a fixed threshold in the fused data to one fixed value; uniformly setting values ​​less than or equal to a fixed threshold in the fused data to another fixed value; obtaining simplified data containing only two values; eliminating a large amount of redundant information during data fusion and numerical conversion; ensuring that the proportion of effective information in the final simplified data is higher than that in the continuous core data of the image; the core map vector is the continuous core data of the image after removing invalid information; the numerical array is the continuous numerical arrangement formed by character identifier conversion; the batch quality mapping primitive is the simplified core quality feature data formed after data fusion and numerical conversion; cross-modal fusion of environmental and process features can improve the stability of quality feature data; and the highly simplified data content can adapt to the writing requirements of batch coding blank areas.

[0046] In some embodiments, injecting batch quality mapping primitives into redundant fields of basic batch codes to generate a composite traceability code matrix includes: parsing the data structure of the basic batch code; locating unoccupied blank check areas in the data structure and setting the blank check areas as redundant fields; using an error correction coding algorithm to write batch quality mapping primitives into the redundant fields; and reorganizing the basic batch codes after writing the data according to preset QR code array rules to generate a composite traceability code matrix.

[0047] Disassemble the internal data layout structure of the product's standard batch code; determine the function and storage location of each data segment in the standard batch code; locate the unused blank storage area within the standard batch code; set this blank storage area as the dedicated writing location for core quality feature data; convert the simplified core quality feature data into writable binary data content; automatically correct data deviations during the writing process; ensure the integrity and accuracy of the written data; rearrange and combine the batch code data after writing the core quality feature data according to a unified QR code graphic layout rule; and form new QR code graphic data.

[0048] The blank verification area is an unused blank storage area within the basic batch code; the composite traceability code matrix is ​​a QR code graphic data formed by rearranging the core quality feature data; seamless embedding of quality feature data can retain the original code's recognition rules and appearance; and it can reduce the production transformation costs of enterprises without upgrading existing equipment.

[0049] In some embodiments, after outputting the composite traceability code matrix to the surface printing device of the corresponding product batch, the method further includes: receiving a traceability parsing request sent by a barcode scanning terminal; the traceability parsing request carries the scanned composite traceability code matrix to be parsed; extracting and separating plaintext batch information and hidden basic data from the composite traceability code matrix to be parsed; the reading permission level of the plaintext batch information is lower than the reading permission level of the hidden basic data.

[0050] The system opens a data receiving port; continuously receives product traceability information query commands transmitted by various barcode scanning devices; continuously receives QR code graphic data collected by various barcode scanning devices; according to the encoding and parsing rules; extracts publicly verifiable routine production information from the QR code graphic data; extracts encrypted core quality feature data from the QR code graphic data; sets data access permissions; ordinary barcode scanning devices can only read publicly verifiable routine production information; only authorized dedicated devices can read encrypted core quality feature data; the traceability parsing request is a product traceability information query command sent by the barcode scanning device; the composite traceability code matrix to be parsed is the QR code graphic data to be decoded collected by the barcode scanning device; the plaintext batch information is publicly verifiable routine production information in the QR code; the hidden primitive data is the encrypted core quality feature data in the QR code; hierarchical permission management can meet the dual needs of ordinary consumers for querying and enterprise supervision and traceability.

[0051] In some embodiments, extracting and separating plaintext batch information and hidden primitive data from the composite traceability code matrix to be parsed further includes: restoring the hidden primitive data through a preset reverse decoding tree to reconstruct the target microstate fingerprint; matching the target microstate fingerprint with a preset risk knowledge base for similarity; and generating a warning pop-up command for the single package if the similarity reaches a risk threshold. The transient anomaly node corresponding to the target microstate fingerprint that triggers the warning pop-up command is recorded at its occurrence time; the edge sub-batch product sequences located at the same processing station within the occurrence time are locked; an interception and isolation command is issued to the edge sub-batch product sequences; the number of edge sub-batch product sequences is less than the total number of products covered by the plaintext batch information.

[0052] According to the pre-set reverse data parsing rules, the encrypted core quality characteristic data is restored to the original environmental anomaly character identifier; the database content of the environmental anomaly characteristics of historical non-conforming products is retrieved; the restored environmental anomaly character identifier is compared with all characteristic data in the database one by one; the matching degree value between the environmental anomaly character identifier and the database characteristic data is calculated; when the matching degree value reaches a fixed threshold, a single product risk warning instruction is sent to the barcode scanning device; the production time node corresponding to the environmental anomaly character identifier that triggered the risk warning instruction is recorded; the data content of the workshop production management system is queried; all products produced in the same tableting station or coating station within the production time node are locked; control instructions are sent to the product sorting equipment and warehouse management equipment; the interception and isolation operation of this part of the products is executed; this part of the products are small batches of products produced within the environmental anomaly time node; the number of products is much lower than the total number of products covered by the regular batch code.

[0053] The reverse decoding tree is the reverse data parsing rule for restoring the core quality feature data of encryption; the target micro-state fingerprint is the restored original environmental anomaly character identifier; the risk knowledge base is a database storing the environmental anomaly characteristics of historical non-conforming products; the early warning pop-up instruction is a single product risk warning instruction received by the barcode scanning device; the edge sub-batch product sequence is a small batch of potentially defective products produced within the time node of environmental anomalies; the interception and isolation instruction is a product isolation control instruction executed by product sorting and warehousing management equipment; precise screening of micro-quality risks can prevent defective products from entering the consumption stage; fine-grained interception of defective products can significantly reduce the economic losses of enterprises.

[0054] Example 2: Refer to Figure 2 Based on Example 1, this embodiment also provides a batch information traceability system for effervescent and sustained-release health products, including: The manufacturing environment sequence acquisition module acquires the manufacturing environment sequence of effervescent and sustained-release health products; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than that of the first temperature and humidity matrix.

[0055] The microstate fingerprint generation module extracts the transient anomaly node set from the manufacturing environment sequence; and generates the microstate fingerprint of effervescent and sustained-release health products based on the transient anomaly node set.

[0056] The slow-release coating feature map construction module obtains the coating spray flow rate sequence and inlet air temperature sequence during the coating stage; calculates the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence, and constructs the slow-release coating forming feature map; the sampling period of the inlet air temperature sequence is longer than the sampling period of the coating spray flow rate sequence.

[0057] The quality mapping primitive fusion module performs dimensionality reduction fusion of micro-state fingerprints and slow-release coating forming feature maps to generate batch quality mapping primitives.

[0058] The composite traceability code generation and output module injects batch quality mapping primitives into the redundant fields of the basic batch code to generate a composite traceability code matrix, and outputs the composite traceability code matrix to the surface printing device of the corresponding product batch.

[0059] The manufacturing environment sequence acquisition module continuously collects environmental sensing equipment data from the tableting station and coating station.

[0060] The manufacturing environment sequence acquisition module constructs temperature and humidity data arrays of different dimensions; the micro-state fingerprint generation module filters time nodes of sudden changes in environmental data; the micro-state fingerprint generation module converts node information into single-dimensional character identifiers; the slow-release coating feature map construction module collects spray flow rate and inlet air temperature data during the coating stage; the slow-release coating feature map construction module calculates the process ratio trajectory and converts it into a visualized two-dimensional planar image; the quality mapping primitive fusion module fuses the character identifiers with the two-dimensional planar image; the quality mapping primitive fusion module generates simplified core quality feature data; the composite traceability code generation and output module writes the core quality feature data into the blank area of ​​the batch code; the composite traceability code generation and output module generates new QR code graphic data and transmits it to the printing equipment. The manufacturing environment sequence acquisition module is a hardware and software combination unit for collecting environmental data throughout the entire production process; the micro-state fingerprint generation module is a processing unit for processing abnormal environmental data and generating character identifiers; the slow-release coating feature map construction module is a processing unit for processing coating process data and generating visualized images; the quality mapping primitive fusion module is a processing unit for fusing environmental features and process features; and the composite traceability code generation and output module is a control unit for generating traceability codes and transmitting them to the printing equipment.

[0061] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0062] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for batch information traceability of effervescent and sustained-release health products, characterized in that, The method includes: S101, Obtain the manufacturing environment sequence of effervescent and sustained-release health products; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than the data dimension of the first temperature and humidity matrix. S102, Extract the set of transient mutation nodes from the manufacturing environment sequence; Generate the microscopic state fingerprint of the effervescent and sustained-release health products based on the set of transient mutation nodes; S103, obtain the coating spray flow rate sequence and the inlet air temperature sequence of the coating stage; calculate the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence, and construct a slow-release coating forming feature map; the sampling period of the inlet air temperature sequence is greater than the sampling period of the coating spray flow rate sequence; S104, the microscopic fingerprint and the slow-release coating forming feature map are dimensionality-reduced and fused to generate batch quality mapping primitives; S105, the batch quality mapping primitive is injected into the redundant field of the basic batch code to generate a composite traceability code matrix, and the composite traceability code matrix is ​​output to the surface printing device of the corresponding product batch.

2. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The extraction of the transient variable node set from the manufacturing environment sequence includes: Calculate the first difference between adjacent time slices in the first temperature and humidity matrix and the second difference between adjacent time slices in the second temperature and humidity matrix; the weighting coefficient of the second difference is greater than the weighting coefficient of the first difference; Set a basic mutation threshold; If the first difference is greater than the basic anomaly threshold, the corresponding time slice is marked as a first-level abnormal node; If the second difference is greater than the basic anomaly threshold, the corresponding time slice is marked as a second-level abnormal node; The first-level abnormal nodes and the second-level abnormal nodes are aggregated to generate the transient abnormal node set.

3. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The step of generating the microscopic state fingerprint of the effervescent and sustained-release health products based on the transient mutation node set includes: Extract the first environmental time sequence identifier of the time when the first-level anomaly node occurs; Extract the second environmental time sequence identifier of the time when the second-level anomaly node occurs; The locality-sensitive hashing algorithm is used to perform feature mapping between the first environment time sequence identifier and the second environment time sequence identifier; The numerical values ​​after feature mapping are concatenated to output a one-dimensional hexadecimal microstate fingerprint; the character length of the microstate fingerprint is less than the original byte length of the transient mutation node set.

4. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The calculation of the ratio trajectory between the coating spray flow rate sequence and the inlet air temperature sequence includes: The coating spray flow rate sequence and the inlet air temperature sequence are time-stamp aligned. Extract the coating spray flow rate and inlet air temperature values ​​at the same timestamp; Calculate the quotient of the coating spray flow rate value divided by the inlet air temperature value to obtain a discrete ratio point set; The discrete ratio point set is smoothly fitted using a spline interpolation function to output a continuous ratio trajectory.

5. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The construction of the sustained-release coating molding feature map includes: Calculate the set of first derivatives of the ratio trajectory; Extract the zero-crossing points from the set of first derivatives; The peak and trough regions corresponding to the zero-crossing points are transformed into two-dimensional thermodynamic images to construct the characteristic map of the slow-release coating formation.

6. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The step of dimensionality reduction and fusion of the microscopic fingerprint and the sustained-release coating molding feature map to generate batch quality mapping primitives includes: The core spectral vector of the sustained-release coating forming feature map was extracted using principal component analysis. The microscopic fingerprint is converted into a digital array; Perform a tensor product operation between the digital array and the core graph vector; The tensor product operation result is binarized to generate the batch quality mapping primitive; the data sparsity of the batch quality mapping primitive is greater than the data sparsity of the core graph vector.

7. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, The step of injecting the batch quality mapping primitive into the redundant fields of the basic batch code to generate a composite traceability code matrix includes: Parse the data structure of the basic batch code; Locate the unoccupied blank validation area in the data structure and set the blank validation area as the redundant field; The batch quality mapping primitive is written into the redundant field using an error correction coding algorithm; According to the preset QR code array rules, the basic batch code after the data is written is recombined to generate the composite traceability code matrix.

8. The method for batch information traceability of effervescent and sustained-release health products according to claim 1, characterized in that, After outputting the composite traceability code matrix to the surface printing device for the corresponding product batch, the method further includes: Receive a traceability parsing request sent by a barcode scanning terminal; the traceability parsing request carries the scanned composite traceability code matrix to be parsed; Extract and separate the plaintext batch information and hidden basic data from the composite traceability code matrix to be parsed; the read permission level of the plaintext batch information is lower than the read permission level of the hidden basic data.

9. A method for batch information traceability of effervescent and sustained-release health products according to claim 8, characterized in that, Extracting and separating the plaintext batch information and hidden primitive data from the composite traceability code matrix to be parsed also includes: The hidden primitive data is restored by using a preset reverse decoding tree to reconstruct the target microstate fingerprint; The target micro-state fingerprint is matched with a preset risk knowledge base for similarity matching; If the similarity reaches the risk threshold, a warning pop-up message will be generated for the individual package. Record the time when the transient anomaly node corresponding to the target microstate fingerprint that triggers the warning pop-up command occurs; Lock the edge sub-batch product sequence that is at the same processing station during the time of occurrence; An interception and isolation command is issued to the edge sub-batch product sequence; the number of the edge sub-batch product sequence is less than the total number of products covered by the plaintext batch information.

10. A batch information traceability system for effervescent and sustained-release health products, characterized in that, A method for tracing batch information of effervescent and sustained-release health products according to any one of claims 1-9, characterized in that the system comprises: The manufacturing environment sequence acquisition module acquires the manufacturing environment sequence of effervescent and sustained-release health products; the manufacturing environment sequence includes a first temperature and humidity matrix for the tableting stage and a second temperature and humidity matrix for the coating stage; the data dimension of the second temperature and humidity matrix is ​​greater than that of the first temperature and humidity matrix. The microstate fingerprint generation module extracts the transient anomaly node set from the manufacturing environment sequence; and generates the microstate fingerprint of the effervescent and sustained-release health products based on the transient anomaly node set. The slow-release coating feature map construction module acquires the coating spray flow rate sequence and the inlet air temperature sequence during the coating stage; calculates the ratio trajectory of the coating spray flow rate sequence and the inlet air temperature sequence, and constructs a slow-release coating forming feature map; the sampling period of the inlet air temperature sequence is greater than the sampling period of the coating spray flow rate sequence; The quality mapping primitive fusion module performs dimensionality reduction fusion of the micro-state fingerprint and the slow-release coating forming feature map to generate batch quality mapping primitives. The composite traceability code generation and output module injects the batch quality mapping primitive into the redundant field of the basic batch code to generate a composite traceability code matrix, and outputs the composite traceability code matrix to the surface printing device of the corresponding product batch.

Images