Drug identification information intelligent association method and system based on multi-source data fusion

Abstract

The invention provides a drug identification information intelligent association method and system based on multi-source data fusion. According to the method, firstly, multi-source heterogeneous data generated in the medicine circulation process is obtained, then key information elements printed on the surface of a medicine package are recognized, meanwhile, a unique identifier bound with the medicine identity is extracted, and then a dynamic tracking link of a medicine transportation path is established; then cross-modal association matching is carried out on the key information element, the unique identifier and the dynamic tracking link to generate a dynamic mapping relation, finally the consistency of the dynamic mapping relation is verified, and the verified dynamic mapping relation is synchronously updated to a supervision terminal of a medicine circulation link. According to the technical scheme provided by the invention, the identification precision and timeliness of risks such as package tampering and path forgery are improved, the abnormal omission ratio caused by data acquisition errors or artificial tampering is reduced, and the reliability of cross-modal data collaborative verification is enhanced.

Description

Technical Field

[0001] The present application relates to the field of information association technology, and in particular to a method and system for intelligently associating drug identification information based on multi-source data fusion. Background Art

[0002] During the drug circulation process, the key information printed on the surface of the drug packaging (such as production batch number, expiration date), the unique identifier of the electronic supervision code, and the spatiotemporal trajectory data of the logistics node are the core basis for verifying the authenticity of drugs and tracking the circulation path.

[0003] Existing technologies currently use blockchain to bind drug electronic supervision codes and the spatiotemporal information of logistics nodes into data blocks, using the tamper-proof feature to store key events, and verifying timestamp continuity through smart contract rules to trigger abnormality alerts. However, this solution only relies on electronic supervision codes and logistics data, and does not integrate the printed information in the packaging image, resulting in the inability to identify counterfeit behaviors that tamper with the packaging but retain the electronic supervision code. At the same time, its preset rules cannot adapt to dynamic changes in transportation routes (such as detours and transit), and it is easy to misjudge compliant routes as abnormal. In addition, there is a lack of multi-source collaborative verification mechanism between electronic supervision codes, logistics trajectories and visual information, making it difficult to achieve full-link closed-loop supervision through cross-modal data (such as consistency between electronic codes and packaging batch numbers). Summary of the Invention

[0004] The present application provides a method and system for intelligent association of drug identification information based on multi-source data fusion, which is used to solve the problems in the existing technology that reliance on a single data dimension leads to the inability to identify tampered packaging information, preset rules cannot adapt to dynamic changes in transportation routes and cause misjudgments, and multi-source data fragmentation makes it difficult to achieve cross-modal collaborative verification.

[0005] In a first aspect, the present application provides a method for intelligently associating drug identification information based on multi-source data fusion, comprising: Acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes; Based on the geometric distribution characteristics of the image data, the system identifies key information elements printed on the surface of the drug packaging, and parses the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug's identity. Based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of logistics nodes, a dynamic tracking link of the drug transportation route is established; Cross-modal association matching of the key information elements, unique identifiers, and node events in the dynamic tracking link is performed to generate a dynamic mapping relationship between drug identification information and different data sources; Through dynamic mapping relationships, the consistency of image data, text data and spatiotemporal trajectory data in the drug circulation process can be verified in real time, and the verified dynamic mapping relationships can be synchronously updated to the regulatory terminals in the drug circulation link.

[0006] Optionally, multi-source heterogeneous data generated during the drug circulation process is obtained, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes, including: Capturing image data of the surface of the drug packaging using a preset scanning device, wherein the scanning device dynamically adjusts the scanning resolution according to the physical size of the drug packaging and adaptively adjusts the image contrast based on the ambient light intensity; The initial text data of the drug electronic supervision code is read by a code scanning device, and the missing characters in the initial text data are checked for integrity based on the encoding rules of the electronic supervision code. When the position of the missing characters is in a predefined non-key field, the original content of the valid field in the text data is retained to obtain the target text data; The recording device of the logistics node captures the time information and position coordinates of the drugs passing through each logistics node during transportation, and arranges the time information and position coordinates into a position change sequence with continuous timestamps according to the transportation order of the drugs to form spatiotemporal trajectory data.

[0007] Optionally, based on the geometric distribution characteristics in the image data, key information elements printed on the surface of the drug packaging are identified, and predefined encrypted fields in the electronic supervision code are parsed to extract a unique identifier bound to the drug identity, including: Analyze the geometric distribution characteristics of the image data, determine the boundary range of the text area based on the layout direction of the printed content on the drug packaging surface, and segment the text area containing the drug name and production batch number from the image data into independent key information elements; Based on the key information elements, the text data of the electronic supervision code is decomposed into fields according to predefined encryption rules. According to the arrangement order and length constraints of the encrypted fields after decomposition, character fragments of a fixed number of bits are extracted from the first field and the last field in the arrangement order, and the character fragments are spliced into a unique identifier bound to the identity of the drug.

[0008] Optionally, a dynamic tracking link for the drug transportation route is established based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics nodes, including: Extract the timestamp-continuous location coordinates of the logistics nodes from the spatiotemporal trajectory data, and generate the initial links of the transportation path according to the increasing order of the timestamps of the location coordinates; Detecting whether the position coordinate change between adjacent logistics nodes in the initial link exceeds a preset distance threshold, retaining the current logistics node when it does not exceed the preset distance threshold, and connecting the retained logistics nodes in series to form an initial dynamic tracking link based on timestamp continuity; When the change in position coordinates exceeds the preset distance threshold, the abnormal logistics node with the largest timestamp interval among the adjacent logistics nodes is eliminated, and the remaining logistics nodes are reconnected in series to generate a target dynamic tracking link.

[0009] Optionally, cross-modal association matching is performed on the key information elements, the unique identifier, and the node events in the dynamic tracking link to generate a dynamic mapping relationship between drug identification information in different data sources, including: Extract the timestamps and location coordinates corresponding to the drug's flow through each logistics node from the dynamic tracking link, match the timestamps with the production date in the key information element, and generate a correlation relationship in the time dimension; Binding the unique identifier to the logistics node where the drug first appears in the dynamic tracking link to generate an association relationship in the location dimension; When there is a timestamp or location coordinate conflict between the association relationship in the time dimension and the association relationship in the location dimension, the conflicting timestamps or conflicting location coordinates are logically replaced based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information between different data sources.

[0010] Optionally, when there is a conflict between timestamps or location coordinates in the association relationship in the time dimension and the association relationship in the location dimension, logical replacement of the conflicting timestamps or conflicting location coordinates is performed based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information of different data sources, including: Detect whether the difference between the timestamp of the logistics node where the drug flows through the preset location in the association relationship on the time dimension and the production date in the key information element exceeds a preset circulation cycle threshold, and determine it as a timestamp conflict node when it exceeds the preset circulation cycle threshold; Detect whether the transportation direction between the logistics node location coordinates where the drug first appears in the association relationship on the location dimension and the subsequent logistics node location coordinates in the dynamic tracking link violates the preset geographical path constraints, and determine the node as a location coordinate conflict node if the preset geographical path constraints are violated; For a timestamp conflict node, extract the timestamps of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link, calculate the average time interval between adjacent logistics nodes, and replace the conflicting timestamp in the timestamp conflict node with a corrected timestamp obtained by adding the production date to the average time interval. For nodes with conflicting position coordinates, the position coordinates of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link are extracted. According to the position change direction between the adjacent logistics nodes, a continuous direction of the transportation path is generated. The conflicting position coordinates of the node with conflicting position coordinates are replaced with the intermediate coordinates calculated based on the continuous direction. Rebind the corrected timestamp and intermediate coordinates to the corresponding logistics nodes to generate a dynamic mapping relationship after conflict elimination.

[0011] Optionally, for a node with conflicting position coordinates, the position coordinates of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link are extracted, and a continuous direction of the transportation path is generated according to the position change direction between the adjacent logistics nodes. The conflicting position coordinates of the node with conflicting position coordinates are replaced with intermediate coordinates calculated based on the continuous direction, including: Connecting the position coordinates of the adjacent previous logistics node and the position coordinates of the adjacent next logistics node to form a straight path, and determining the moving direction vector of the transportation path based on the coordinate difference between the starting point and the end point of the straight path; Dividing orthographic projection areas on both sides of the straight path according to the angular range of the moving direction vector; Projecting the original coordinates of the position coordinate conflict node vertically onto the straight path, and detecting whether the projection point of the vertical projection falls within the orthographic projection area; if not, translating the coordinates of the conflicting position of the position coordinate conflict node along the moving direction vector to the nearest projection point on the straight path; When the projection point falls within the positive projection area, the position coordinates of the adjacent previous logistics node are used as the starting point, and the position coordinates are extended along the moving direction vector to the position coordinates of the next logistics node, and the conflicting position coordinates of the position coordinate conflicting node are replaced with the intermediate coordinates between the starting point and the end point divided according to the preset distance ratio.

[0012] In a second aspect, the present application provides a drug identification information intelligent association system based on multi-source data fusion, comprising: An acquisition module is used to acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes; The recognition module is used to identify key information elements printed on the surface of the drug packaging based on the geometric distribution characteristics of the image data, and to parse the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug identity; Establish a module for establishing a dynamic tracking link of the drug transportation route based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics nodes; A matching module, configured to perform cross-modal correlation matching between the key information elements, the unique identifier, and the node events in the dynamic tracking link, and generate a dynamic mapping relationship between drug identification information and different data sources; The verification module is used for dynamic mapping relationships, real-time verification of the consistency of image data, text data and spatiotemporal trajectory data during the drug circulation process, and synchronously updates the verified dynamic mapping relationships to the regulatory terminals in the drug circulation link.

[0013] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for intelligent association of drug identification information based on multi-source data fusion as described in the first aspect above.

[0014] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a method for intelligent association of drug identification information based on multi-source data fusion as described in the first aspect.

[0015] The embodiment of the present application builds a dynamic tracking link and generates a cross-modal mapping relationship by integrating multi-source heterogeneous data of drug packaging images, electronic supervision codes and logistics trajectories, thereby solving the problem of tampering identification lag caused by traditional single data verification. It utilizes the geometric characteristics of image data to extract printed information, parse unique identifiers of encrypted fields, and establish dynamic links of spatiotemporal trajectories to achieve cross-validation and real-time consistency detection of multi-source data. At the same time, it is synchronously updated to the supervision terminal through dynamic mapping relationships, forming a closed-loop supervision mechanism in the drug circulation process, effectively improving the accuracy and timeliness of identifying risks such as packaging tampering and path forgery.

[0016] Furthermore, based on weight 1, through the dual association of time dimension (matching production date with logistics node timestamp) and location dimension (unique identifier with first node binding), combined with the context logic of adjacent nodes in the dynamic tracking link, the conflicting timestamps or location coordinates are replaced, which solves the misjudgment problem caused by data asynchrony or tampering in the existing technology. For example, the conflicting timestamps are corrected based on the average of the time intervals of adjacent nodes, and the intermediate coordinates are calculated according to the continuous direction of the transportation path. This ensures that the corrected mapping relationship not only conforms to the physical space continuity of the logistics path, but also can adapt to the dynamic changes of the transportation path (such as detours and transfers), significantly reducing the abnormal missed detection rate caused by data collection errors or human tampering, and at the same time enhancing the reliability of cross-modal data collaborative verification.

[0017] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] Figure 1 A flowchart of a method for intelligently associating drug identification information based on multi-source data fusion provided by the present application is shown; Figure 2 The present invention provides a schematic diagram of a drug identification information intelligent association system based on multi-source data fusion; Figure 3 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0021] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0022] In order to solve the problem of delayed tampering identification caused by traditional drug identification verification relying on a single data source, the real-time and accuracy of multi-source data collaborative verification should be improved.

[0023] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0024] Figure 1 The present invention provides a flowchart of a method for intelligently associating drug identification information based on multi-source data fusion, as shown in FIG. Figure 1 As shown, the method includes: Step 101: Acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes.

[0025] In this step, the multi-source heterogeneous data includes: The image data of the surface of pharmaceutical packaging refers to the visual information of the surface of pharmaceutical packaging collected by scanning equipment, including the original pixel set of printed text (such as production batch number, expiration date), patterns and anti-counterfeiting marks.

[0026] The text data of the electronic drug supervision code refers to the encrypted string read by the scanning device, which contains a predefined field structure (such as manufacturer code, production batch), and some fields are generated by an asymmetric encryption algorithm.

[0027] The spatiotemporal trajectory data of logistics nodes refers to a serialized data set that records the timestamps (accurate to seconds) and geographic location coordinates (latitude and longitude) of the logistics nodes (such as warehouses and transfer stations) that the drugs pass through during transportation, and is arranged in chronological order.

[0028] In this embodiment, a pre-set scanning device first captures images of the drug packaging surface. The scanning device dynamically adjusts resolution based on package size (e.g., increasing resolution to capture fine details for large packages) and automatically adjusts contrast based on ambient light intensity (e.g., reducing exposure in bright light environments) to ensure clear and recognizable printed text. Secondly, after reading the original text data of the electronic supervision code, the scanning device performs an integrity check on missing characters based on its encoding rules (e.g., segment length, check digit position). If the missing characters are in non-critical fields (e.g., padding fields), the original content of valid fields (e.g., manufacturer code, production batch) is retained to generate de-redundant target text data. Finally, a recording device at a logistics node captures the timestamps and location coordinates of each node as the drug passes through. These are arranged into a time-stamped sequence of position changes in the order of transportation. Isolated nodes caused by device delays or signal loss (e.g., nodes with time intervals exceeding 30 minutes) are eliminated to generate spatiotemporal trajectory data.

[0029] For example, during the circulation process of drugs produced by a pharmaceutical factory, the packaging image is first captured by a scanning device deployed at the warehouse exit: when the packaging size is large, the scanning device increases the resolution to 600dpi to capture the small-font production batch number, and automatically reduces the contrast in strong light environments to avoid overexposure of printed content.

[0030] Subsequently, when the scanning device read the electronic supervision code, it was found that a field was missing 3 characters. After verification, it was confirmed that the missing position was a reserved fill-in field (non-critical field). The original content of the manufacturer code "A01" and production batch "2023-09" was retained to generate the target text data.

[0031] During the logistics transportation process, the recording device captures the spatiotemporal trajectory data of a certain drug, which includes a timestamp sequence (such as 09:00:00, 09:15:00, and 09:45:00) and corresponding coordinates (such as longitude 116.4→116.5→116.7). Among them, the interval between 09:15:00 and 09:45:00 exceeds 30 minutes. The system determines it as an abnormal node and removes it, generating the corrected spatiotemporal trajectory data (09:00:00→116.4, 09:45:00→116.7).

[0032] Step 102: Based on the geometric distribution characteristics in the image data, identify the key information elements printed on the surface of the drug packaging, and parse the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug identity.

[0033] In this step, the geometric distribution characteristics refer to the layout rules of the printed content (such as text, barcode) on the surface of the pharmaceutical packaging, including the direction of the text area (horizontal / vertical arrangement), the spacing between adjacent elements, and the alignment.

[0034] Key information elements refer to independent areas separated by geometric distribution characteristics, containing printed content such as drug name, production batch number, expiration date, etc., and have fixed semantic labels (such as "production batch number: 2023-09").

[0035] Predefined encryption fields refer to encryption segments divided according to rules in the electronic supervision code text data. For example, the first field is the manufacturer code (plain text), the middle field is the production batch (encrypted), and the last field is the check code (hash value).

[0036] In this embodiment, the image data is first analyzed for its geometric distribution characteristics. Based on the horizontal or vertical orientation of the printed content, the boundaries of the text area are determined (for example, horizontally aligned text areas are segmented using a horizontal projection histogram). The areas containing the drug name and production batch number are then segmented into independent key information elements. Secondly, based on the semantic tags of the key information elements (such as the production batch number), the text data of the electronic supervision code is field-decomposed according to predefined encryption rules. Based on field length and position constraints (for example, the first field is fixed at 3 digits, the last field is fixed at 5 digits), fixed-digit character segments are extracted from the first and last fields (for example, the first to third digits of the first field, the last 3 to 5 digits of the last field). These segments are then sequentially concatenated to generate a unique identifier. Finally, the unique identifier is logically bound to the production batch number within the key information element (for example, the unique identifier prefix is consistent with the production batch number), ensuring the unique correspondence between the identifier and the drug's identity.

[0037] For example, following the case of a pharmaceutical factory in step 101, after scanning the image of its drug packaging, the system recognizes the horizontally arranged text area and segments out the key information elements containing the production batch number "2023-09"; at the same time, the electronic supervision code text data "A01#5X8@2023-09#Z2Y" is decomposed into the manufacturer code "A01", the encrypted field "5X8", the production batch "2023-09" and the check code "Z2Y" according to the rules, and extracts the first field "A01" and the last 3 digits "2Y" of the last field, and splices them together to generate a unique identifier "A012Y".

[0038] When the production batch number "2023-09" matches the unique identifier prefix "A01", it is confirmed that the identifier is bound to the drug identity, providing a data basis for cross-modal association in the subsequent dynamic tracking link.

[0039] Step 103: Establish a dynamic tracking link for the drug transportation route based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics node.

[0040] In this step, the location change sequence refers to a set of geographical location coordinates of logistics nodes arranged in chronological order, reflecting the spatial movement trend of the drug transportation route (such as from warehouse A → transfer station B → pharmacy C).

[0041] Timestamp continuity means that the timestamps recorded at logistics nodes increase in order of transportation and are spaced reasonably, avoiding abnormal nodes with reversed time or extremely long intervals.

[0042] A dynamic tracking link refers to a transportation path formed by a series of logistics nodes with continuous timestamps and reasonable position changes. Abnormal nodes can be dynamically corrected to maintain the logical consistency of the path.

[0043] In this embodiment, first, the timestamp-continuous position coordinates of the logistics nodes are extracted from the spatiotemporal trajectory data, and the initial links are generated in ascending order of timestamps (for example, the nodes are arranged in timestamp order as A→B→C).

[0044] Secondly, the change in the position coordinates of adjacent nodes in the initial link (for example, the distance difference between nodes A and B) is detected. If the change does not exceed the preset distance threshold (such as a reasonable transportation radius), the node is retained and connected in series as the initial dynamic tracking link; if it exceeds the threshold, the abnormal node with the largest timestamp interval among adjacent nodes is eliminated (for example, the timestamp interval between node B and A and C is too long).

[0045] Finally, the remaining nodes are reconnected according to the timestamp continuity to generate a revised target dynamic tracking link (for example, after removing B, the link becomes A→C), and the rationality of the transportation direction is verified (for example, whether A→C meets the geographical path constraints).

[0046] For example, taking the drug case from step 102, its initial link includes logistics node A (09:00:00, longitude 116.4), node B (09:15:00, longitude 116.5), and node C (09:45:00, longitude 116.7). Detection reveals that the distance between nodes B and C exceeds a threshold (e.g., a reasonable threshold is 50 kilometers, but the actual distance is 120 kilometers), and that the timestamp interval between nodes B (30 minutes between B and C) is much greater than the interval between nodes A and B (15 minutes). The system then identifies node B as an abnormal node and removes it, generating a revised target dynamic tracking link A→C.

[0047] The corrected link verifies the rationality of the transportation direction (the longitude increase from A to C is consistent with the transportation trend from west to east), providing an accurate path basis for the cross-modal association in the subsequent step 104.

[0048] Step 104 : Cross-modal association matching is performed on the key information elements, the unique identifier, and the node events in the dynamic tracking link to generate a dynamic mapping relationship between drug identification information and different data sources.

[0049] In this step, the node event refers to the combination of the timestamp and location coordinates of the logistics node in the dynamic tracking link, which represents the spatiotemporal state of the drug during transportation (for example, a node event is "09:00:00, longitude 116.4").

[0050] Cross-modal association matching refers to mapping the semantic information of different data sources (images, texts, spatiotemporal trajectories) through logical rules, such as matching production batch numbers with logistics node timestamps, and binding unique identifiers with first-appearance nodes.

[0051] The dynamic mapping relationship refers to a cross-data source mapping table generated through association matching, which records the consistent association of drug identification information in different data sources (such as the production batch number "2023-09" corresponds to the logistics node timestamp "09:00:00").

[0052] In this embodiment, the node events (timestamps and location coordinates) of each logistics node are first extracted from the dynamic tracking link, and the timestamp is matched with the production date in the key information element to generate a time dimension association relationship (for example, whether the difference between the production date "2023-09-01" and the node timestamp "09:00:00" is within a reasonable transportation cycle). Secondly, the unique identifier is bound to the location coordinates of the logistics node where the drug first appears in the dynamic tracking link to generate a location dimension association relationship (for example, the unique identifier "A012Y" first appears at the coordinate position of node A). When there is a conflict in the association relationship of the time dimension or location dimension (for example, the timestamp of a node is earlier than the production date, or the location of the first node appears inconsistent with the transportation direction of the subsequent node), the conflicting timestamps or location coordinates are logically replaced based on the association relationship of adjacent nodes in the dynamic tracking link (such as the timestamp of the previous node and the location of the next node) to form a corrected dynamic mapping relationship.

[0053] For example, taking the drug case from step 103, its key information elements include the production batch number "2023-09" and the unique identifier "A012Y." The dynamic tracking link is node A (09:00:00, longitude 116.4) → node C (09:45:00, longitude 116.7). The system matches the time corresponding to the production batch number (production date 2023-09-01) with the timestamp of node A (09:00:00), creating a time-dimensional relationship. It also binds the unique identifier "A012Y" to the location coordinates of node A, creating a location-dimensional relationship. If it is subsequently detected that the timestamp of a node (such as 09:30:00) is earlier than the production date, the average time interval (22.5 minutes) of its adjacent nodes (such as the previous node 09:00:00 and the next node 09:45:00) is extracted, and the conflicting timestamp is corrected to a reasonable time after "production date + interval average" (2023-09-01 09:22:30), and the dynamic mapping relationship is updated.

[0054] Step 105 , through the dynamic mapping relationship, the consistency of the image data, text data and spatiotemporal trajectory data in the drug circulation process is verified in real time, and the verified dynamic mapping relationship is synchronously updated to the supervision terminal of the drug circulation link.

[0055] In this step, real-time verification is performed based on the association rules in the dynamic mapping relationship (such as matching timestamps with production dates and continuous trends in location coordinates) to perform instant consistency checks on newly added logistics nodes or data changes, ensuring that images, texts, and spatiotemporal data are always logically consistent during circulation.

[0056] The supervision terminal refers to the data receiving and display platform deployed in various entities in the drug circulation process (such as pharmaceutical factories, logistics companies, and pharmacies), which is used to receive verified dynamic mapping relationships and trigger alarms or approval processes.

[0057] In this embodiment, the time dimension association (matching the production date with the logistics node timestamp) and the location dimension association (binding the unique identifier to the first-appearing node) in the dynamic mapping relationship are first used to verify the timestamp and location coordinates of the newly added logistics node: if the timestamp of a node is earlier than the production date or the location coordinates violate the continuous direction of the transportation path, it is determined to be a data anomaly; secondly, when an anomaly is detected, the conflicting data is logically replaced based on the contextual relationship of adjacent nodes in the dynamic tracking link (such as the mean time interval and the location direction) to generate a corrected dynamic mapping relationship; finally, the corrected dynamic mapping relationship is pushed to the supervision terminal through the preset data synchronization interface, such as the logistics provider's transportation monitoring system or the pharmacy's warehousing review platform, to trigger the exception handling process (such as suspension of warehousing, manual review).

[0058] For example, in the case of a drug from step 104, its dynamic mapping relationship already associates the production batch number "2023-09", the unique identifier "A012Y", and the dynamic tracking link A→C (timestamps 09:00:00→09:45:00). When the timestamp of the newly added logistics node D is "2023-08-31 08:30:00" (which is earlier than the production date 2023-09-01), the system identifies a time conflict anomaly.

[0059] Based on the average time interval between adjacent nodes A and C (45 minutes), the timestamp of node D is corrected to "2023-09-01 09:22:30" (production date + interval average). After updating the dynamic mapping relationship, it is pushed to the pharmacy's supervision terminal through the API interface, triggering a "production date discrepancy" alarm and suspending the warehousing of this batch of drugs. The abnormal status will be resolved after manual review.

[0060] In order to improve the acquisition quality of image and text data and solve the problem of information distortion caused by environmental interference or data loss, in some embodiments, according to step 101, multi-source heterogeneous data generated during the drug circulation process is obtained, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of the logistics node, including: Step 201 : collecting image data of the surface of the drug package by a preset scanning device, wherein the scanning device dynamically adjusts the scanning resolution according to the physical size of the drug package and adaptively adjusts the image contrast based on the ambient light intensity.

[0061] In this step, dynamic resolution adjustment based on physical size refers to the scanner automatically matching the resolution to the actual size of the drug packaging (e.g., large box versus small bottle), ensuring that printed details (such as small batch numbers) on packages of varying sizes are clearly legible. Adaptive contrast adjustment based on ambient light intensity refers to the scanner automatically adjusting the image's brightness and darkness based on ambient light intensity (e.g., strong or weak light) to avoid blurring or missing printed content due to overexposure or underexposure.

[0062] In this embodiment, a scanning device is first used to capture images of the pharmaceutical packaging. The required resolution is calculated based on the physical dimensions of the packaging (such as length, width, and height). For large-sized packages, the resolution is increased to capture fine text, while for small-sized packages, the resolution is reduced to avoid data redundancy. Secondly, the built-in photosensor of the scanning device detects the ambient light intensity in real time. If the light is too strong, the contrast is reduced and shadow compensation is increased. If the light is insufficient, the contrast is increased and the brightness is enhanced to ensure that the boundaries of the printed text area (such as the production batch number and expiration date) are clear. Finally, the adjusted image data is saved in a preset format and bound to the unique identifier of the pharmaceutical product, providing high-quality input for the identification of key information elements in subsequent steps.

[0063] Step 203: The time information and position coordinates of the drugs passing through each logistics node during transportation are captured by the recording device of the logistics node, and the time information and position coordinates are arranged into a position change sequence with continuous timestamps according to the transportation order of the drugs to form spatiotemporal trajectory data.

[0064] In this step, the recording device at the logistics node refers to positioning and timing equipment (such as GPS modules and RFID readers) deployed at transportation nodes (such as warehouses and transfer stations). It is used to record the time and location of the drug passing through the node. The time-stamped continuous position change sequence refers to a collection of logistics node timestamps and location coordinates arranged in the order of drug transportation. The timestamps must be increasing, and the position changes must conform to the physical continuity of the transportation path.

[0065] In this embodiment, a recording device is first used to capture in real time the timestamps (accurate to the second) and location coordinates (e.g., longitude and latitude) of the drug as it passes through each logistics node, and these are bound to the drug's unique identifier. Secondly, the timestamps and location coordinates of the nodes are arranged into an initial sequence according to the order of transportation. The time interval between adjacent nodes is checked to see if it exceeds a preset threshold (e.g., the upper limit of reasonable transportation time) or if the location coordinates suddenly change (e.g., if the reasonable transportation radius is exceeded). If an abnormal node is detected (e.g., an excessively long time interval or a sudden change in location), the node is removed and a corrected continuous sequence is generated based on the timestamps and location coordinates of the previous and next nodes. Finally, the corrected timestamps and location coordinates are sequentially integrated into spatiotemporal trajectory data and associated with the drug's packaging image data and electronic supervision code text data to form a complete circulation traceability record.

[0066] To resolve the issue of separation between package printing information and the electronic supervision code and improve the ability to accurately extract key information elements and unique identifiers, in some embodiments, according to step 102, based on the geometric distribution characteristics of the image data, key information elements printed on the surface of the drug packaging are identified, and predefined encrypted fields in the electronic supervision code are parsed to extract the unique identifier bound to the drug identity, including: Step 301: perform geometric distribution characteristic analysis on the image data, determine the boundary range of the text area according to the layout direction of the printed content on the drug packaging surface, and segment the text area containing the drug name and production batch number from the image data into independent key information elements.

[0067] In this step, geometric distribution analysis involves analyzing the overall layout structure of the text area by identifying the arrangement patterns of printed content on the pharmaceutical packaging surface (e.g., horizontal or vertical alignment, character spacing consistency). The text area boundary refers to the pixel range of a rectangular or polygonal area defined by the layout direction (e.g., horizontally aligned text areas extend along the horizontal axis), encompassing specific semantic content (e.g., drug name, production batch number).

[0068] In this embodiment, the geometric distribution characteristics of the image data are first analyzed: according to the horizontal or vertical arrangement direction of the printed content, the pixel density distribution is statistically analyzed along the main axis (horizontal or vertical direction), and a projection histogram is generated to locate the starting and ending boundaries of the text area (for example, the peak interval in the horizontal projection histogram corresponds to the horizontal text area). Secondly, based on the boundary range, the image is divided into several independent areas, and the semantic label areas containing preset keywords (such as "batch number" and "expiration date") are screened out, and the corresponding pixel blocks are extracted as key information elements. Finally, the background noise (such as decorative patterns or non-text areas) is removed through the boundary detection method, retaining a clear image of the key information element, providing structured visual input for the subsequent electronic supervision code analysis.

[0069] Step 302: Based on the key information elements, the text data of the electronic supervision code is decomposed into fields according to predefined encryption rules. According to the arrangement order and length constraints of the encrypted fields after decomposition, character fragments of a fixed number of bits are extracted from the first field and the last field in the arrangement order, and the character fragments are spliced into a unique identifier bound to the drug identity.

[0070] In this step, the predefined encryption rules refer to the delimiter definition, encryption segment length and position constraints of the fields in the electronic supervision code text data, for example, the first field is the manufacturer code (plain text), the middle field is the encrypted production batch, and the last field is the check code.

[0071] A character segment of a fixed number of bits refers to a character segment intercepted from a specific field according to encryption rules, such as extracting the first 3 digits of the manufacturer code from the first field and extracting the last 2 digits of the check code from the last field.

[0072] In this embodiment, the field decomposition rules of the electronic supervision code are first determined based on the production batch number in the key information element (such as "2023-09"): the text data is split into multiple fields according to the predefined separator (such as "@" or "#"), and the length of each field is verified to meet the constraints (for example, the first field is fixed to 3 digits and the last field is fixed to 4 digits). Secondly, according to the order of the fields, a fixed number of character fragments are extracted from the first field (such as the first 3 digits of the first field "A01"), and a fixed number of character fragments are extracted from the last field (such as the last 2 digits of the last field "2Y"), and the two fragments are spliced in sequence to generate a unique identifier (such as "A012Y"). Finally, the unique identifier is logically bound to the production batch number in the key information element (for example, the unique identifier includes the production batch number prefix) to ensure its unique correspondence with the drug identity, providing a data basis for subsequent cross-modal association.

[0073] In order to solve the problem of tracking link breakage caused by interference from abnormal nodes in the logistics path and improve the reliability of dynamic correction of the transportation path, in some embodiments, according to step 103, a dynamic tracking link of the drug transportation path is established based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics nodes, including: Step 401 extracts the timestamp-consistent location coordinates of logistics nodes from the spatiotemporal trajectory data. These coordinates are arranged in ascending timestamp order to generate the initial links of the transportation route. In this step, timestamp-consistent location coordinates refer to logistics nodes whose timestamps are strictly increasing according to the drug transportation sequence and are spaced appropriately (e.g., without reversed chronological order or excessively long intervals). Furthermore, the location coordinates maintain spatial continuity along the transportation route (e.g., without sudden geographic changes). The initial links, formed by sequentially connecting timestamp-consistent location coordinates in chronological order, include any anomalous nodes (e.g., nodes with sudden changes in time or location).

[0074] In this example, the timestamps and location coordinates of all logistics nodes are first extracted from the spatiotemporal trajectory data and arranged into an ordered sequence in ascending timestamp order (e.g., from earliest to latest). Next, the time intervals between adjacent nodes are verified based on timestamp continuity to ensure they are within a reasonable transport time range. If they are, the node is retained; otherwise, it is marked as pending verification. Finally, the retained nodes are concatenated in timestamp order to generate an initial link, and the location coordinates of the pending nodes are recorded. This provides the raw data foundation for removing abnormal nodes and correcting paths in subsequent steps.

[0075] Step 402, detect whether the change in position coordinates between adjacent logistics nodes in the initial link exceeds a preset distance threshold. When it does not exceed the preset distance threshold, retain the current logistics node, and concatenate the retained logistics nodes into an initial dynamic tracking link based on timestamp continuity.

[0076] In this step, the location coordinate change refers to the actual distance difference between the geographic coordinates of adjacent logistics nodes. This is calculated by the difference in longitude and latitude, reflecting the transportation span of the drug between adjacent nodes. The preset distance threshold is the maximum allowable distance difference calculated based on the reasonable transportation speed and time interval for drug transportation scenarios. It is used to determine whether the node location is abnormal (for example, the maximum distance a truck can travel in 1 hour is 100 kilometers).

[0077] In this embodiment, the location coordinates of adjacent logistics nodes are first extracted from the initial link, and the actual distance difference between them is calculated (for example, the straight-line distance from node A to node B). Next, the calculated distance difference is compared with a preset distance threshold. If it does not exceed the threshold, the node is determined to be within a reasonable transport span and retained. If it exceeds the threshold, the node is marked as an abnormal node and temporarily stored. Finally, the retained nodes are concatenated in ascending timestamp order to form the initial dynamic tracking link (for example, A→B→C), and the abnormal node information is passed to subsequent steps for removal or correction.

[0078] Step 403: When the change in position coordinates exceeds the preset distance threshold, the abnormal logistics node with the largest timestamp interval among the adjacent logistics nodes is eliminated, and the remaining logistics nodes are reconnected in series to generate a target dynamic tracking link.

[0079] In this step, the abnormal logistics node with the largest timestamp interval refers to the node with the largest timestamp difference among adjacent logistics nodes, where the time difference significantly exceeds the reasonable transportation time (for example, the distance between a node and the previous node is too long, far exceeding the average transportation speed). The target dynamic tracking link refers to the corrected path formed by reconnecting the remaining logistics nodes based on timestamp continuity after removing the abnormal nodes, ensuring the spatiotemporal logical rationality of the transportation path.

[0080] In this embodiment, the first step is to check whether the change in the location coordinates between adjacent logistics nodes exceeds a preset distance threshold (for example, the distance difference between node A and node B exceeds 100 kilometers). If so, it is determined to be an abnormal span. Secondly, the timestamp intervals between the adjacent nodes in the abnormal span are calculated (for example, the interval between nodes A and B is 15 minutes, and the interval between nodes B and C is 30 minutes). The node with the largest timestamp interval (for example, node B) is selected as the abnormal node. Finally, the abnormal node (node B) is removed, and the remaining nodes are reconnected in ascending timestamp order (for example, A→C). The revised path is verified to ensure that it conforms to the geographic transportation direction (for example, increasing longitude), generating the target dynamic tracking link.

[0081] In order to resolve verification contradictions caused by cross-modal data association conflicts and improve the consistent mapping capabilities of time and location dimensions, in some embodiments, according to step 104, the key information elements, unique identifiers, and node events in the dynamic tracking link are cross-modally associated and matched to generate a dynamic mapping relationship between drug identification information in different data sources, including: Step 501: extract the timestamps and location coordinates corresponding to the drug flowing through each logistics node from the dynamic tracking link, match the timestamps with the production date in the key information element, and generate an association relationship in the time dimension.

[0082] In this step, the temporal relationship refers to the logical mapping between the drug's production date and the timestamps at the logistics node. For example, for a drug with a production date of "2023-09-01," its logistics node timestamp should be later than that date and meet the preset circulation cycle threshold (e.g., a transportation cycle of no more than 30 days). Production date matching verifies whether the drug's circulation path complies with the time logic from production to transportation by comparing the sequence and reasonableness of the intervals between the production date and logistics node timestamps.

[0083] In this embodiment, the timestamps and location coordinates of each logistics node are first extracted from the dynamic tracking link and arranged in ascending order of timestamps into a node event sequence. Secondly, the timestamp of each node is matched with the production date in the key information element: if the timestamp is later than the production date and the difference is within the circulation cycle threshold (for example, transportation within 30 days after production), a time dimension association relationship is generated; if the timestamp is earlier than the production date or the difference exceeds the threshold, it is marked as a time conflict node. Finally, based on the context time interval of adjacent nodes (for example, the average transportation time of the previous node and the next node), the timestamps of the conflicting nodes are logically corrected (for example, replaced with "production date + average transportation time") to generate an association relationship that conforms to the time logic.

[0084] Step 502: Bind the unique identifier to the logistics node where the drug first appears in the dynamic tracking link, generating a location-based association. In this step, the first-appearing logistics node refers to the earliest logistics node recorded in timestamp order in the dynamic tracking link. Its location coordinates represent the starting point of the drug's distribution path. This location-based association, which binds the unique identifier to the location coordinates of the first logistics node, is used to verify that the transportation direction of subsequent nodes complies with geographic path constraints (e.g., extending from the starting point in a reasonable direction).

[0085] In this embodiment, the logistics node with the earliest timestamp is first extracted from the dynamic tracking link (for example, the timestamp of node A is 09:00:00), and its location coordinates (such as longitude 116.4) are bound to a unique identifier (such as "A012Y") to generate an association relationship of the location dimension (such as "A012Y→longitude 116.4"). Secondly, based on this association relationship, it is verified whether the direction of change of the location coordinates of subsequent logistics nodes is reasonable (for example, increasing from longitude 116.4 to 116.7, which is consistent with the transportation direction from west to east). If the location coordinates of subsequent nodes violate the direction constraint (such as a sudden decrease in longitude), a location conflict alarm is triggered, and the corrected coordinates are calculated based on the positions of adjacent nodes to update the association relationship.

[0086] Step 503: When there is a timestamp or location coordinate conflict between the association relationship in the time dimension and the association relationship in the location dimension, the conflicting timestamp or conflicting location coordinate is logically replaced based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information between different data sources.

[0087] In this step, a timestamp conflict occurs when a logistics node's timestamp precedes the production date or exceeds the circulation cycle threshold, resulting in a conflicting relationship in the time dimension. A location coordinate conflict occurs when a logistics node's location coordinate violates the transport direction constraint (e.g., reverse jump) established by the initial node binding, resulting in a conflicting relationship in the location dimension. Logical replacement involves revising conflicting data to eliminate the conflict based on the contextual relationships between adjacent nodes in the dynamic tracking link (e.g., mean time interval and location direction).

[0088] In this embodiment, first, the timestamp conflict in the time dimension association relationship (such as the timestamp of a node is earlier than the production date) or the coordinate conflict in the position dimension association relationship (such as the reverse jump in the transportation direction) is detected. For timestamp conflicts, the timestamps of the previous node and the next node adjacent to the conflicting node are extracted, the average time interval between the two is calculated, and the conflicting timestamp is replaced with a corrected timestamp obtained by adding the production date to the average (for example, the production date is D, the interval average is T, and the corrected timestamp is D+T). For position coordinate conflicts, the position coordinates of the previous node and the next node adjacent to the conflicting node are extracted, and the intermediate coordinates based on the continuous direction of the transportation path are generated (for example, the coordinates of the previous node are (X1, Y1), the coordinates of the next node are (X2, Y2), and the intermediate coordinates are ((X1+X2) / 2, (Y1+Y2) / 2)), and the conflicting coordinates are replaced. Finally, the corrected timestamps or position coordinates are re-bound to the nodes to generate a dynamic mapping relationship after the conflict is eliminated.

[0089] In order to resolve the misjudgment problem caused by abnormal timestamps or location coordinates and improve the efficiency of conflict resolution based on the logistics link context logic, in some embodiments, according to step 503, when there is a timestamp or location coordinate conflict in the association relationship in the time dimension and the association relationship in the location dimension, the conflicting timestamps or conflicting location coordinates are logically replaced based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information of different data sources, including: Step 601, detect whether the difference between the timestamp of the logistics node where the drug flows through the preset position in the association relationship on the time dimension and the production date in the key information element exceeds the preset circulation cycle threshold. When it exceeds the preset circulation cycle threshold, it is determined to be a timestamp conflict node.

[0090] In this step, the preset circulation cycle threshold refers to the maximum allowable time span from the production date to the entry of the drug into the circulation link. For example, the drug must be transported within 30 days after production, and exceeding the deadline is considered an abnormality.

[0091] The logistics nodes at the preset locations refer to the key nodes in the drug distribution path where timestamp verification is mandatory (such as the first-stop warehouse and regional distribution center).

[0092] In this embodiment, the timestamp of the pre-set logistics node (e.g., the timestamp of the first warehouse node) is first extracted from the dynamic tracking link. The time difference is then calculated with the production date in the key information element. Next, the time difference is compared with a preset circulation cycle threshold. If the difference exceeds the threshold (e.g., if the production date is D and the node timestamp is D+35 days), the node is identified as a timestamp conflict node. Finally, the timestamp and difference exceeding threshold information of the conflicting node are recorded to provide a basis for conflict location during subsequent logical replacement steps.

[0093] Step 602: Detect whether the transportation direction between the logistics node location coordinates where the drug first appears in the association relationship on the location dimension and the subsequent logistics node location coordinates in the dynamic tracking link violates the preset geographical path constraints. If the preset geographical path constraints are violated, it is determined to be a location coordinate conflict node.

[0094] In this step, the preset geographic path constraints refer to the geographic direction rules that the drug transportation route must follow (such as from west to east or from south to north), or prohibit sudden changes in direction (such as a sudden decrease in longitude). The subsequent logistics node location coordinates refer to the location coordinates of all logistics nodes recorded after the initial node in the dynamic tracking link.

[0095] In this embodiment, the location coordinates of the initial and subsequent nodes are first extracted from the dynamic tracking link and arranged in timestamp order to form a path sequence. Next, the transport direction between adjacent nodes is calculated (for example, the sign of the longitude difference indicates east / west) and compared with the preset geographic path constraints. If the direction suddenly changes (for example, from east to west), the node is determined to have a location coordinate conflict. Finally, the location coordinates and direction deviation information of the conflicting nodes are recorded to provide a basis for conflict location determination during subsequent logical replacement.

[0096] Step 603: For the timestamp conflict node, extract the timestamps of the adjacent previous logistics node and the next logistics node in the dynamic tracking link of the logistics node, calculate the average time interval between adjacent logistics nodes, and replace the conflicting timestamp in the timestamp conflict node with the corrected timestamp of the production date plus the average time interval.

[0097] In this step, the average time interval refers to the arithmetic mean of the timestamp differences between the preceding and succeeding nodes of the conflicting node. This is used to calculate a corrected timestamp that is consistent with reasonable logistics transportation times. The corrected timestamp is a logically reasonable timestamp generated by superimposing the average time interval on the production date. This is used to eliminate time dimension conflicts.

[0098] In this embodiment, the timestamps of the adjacent previous and next nodes of the timestamp conflicting node are first extracted, and the time interval difference between the two is calculated (for example, the timestamp of the next node minus the timestamp of the previous node). The absolute value of the difference is taken as the time interval. Secondly, the average time interval of the adjacent nodes is calculated (for example, the interval from the previous node to the conflicting node is T1, and the interval from the conflicting node to the next node is T2, and the average is (T1+T2) / 2). This average is then added to the production date to generate a revised timestamp (for example, if the production date is D, the revised timestamp is D+average). Finally, the original timestamp of the conflicting node is replaced with the revised timestamp, and the time dimension association in the dynamic mapping relationship is updated to ensure the temporal logical continuity of the transportation path.

[0099] Step 604: For the node with conflicting position coordinates, extract the position coordinates of the adjacent previous logistics node and the next logistics node in the dynamic tracking link of the logistics node, generate a continuous direction of the transportation path according to the position change direction between the adjacent logistics nodes, and replace the conflicting position coordinates of the node with the intermediate coordinates calculated based on the continuous direction.

[0100] In this step, the continuous direction refers to the direction of the transport route generated based on the coordinate trends of the adjacent previous and next nodes (e.g., increasing longitude, decreasing latitude), ensuring that the route conforms to geospatial logic. Intermediate coordinates refer to reasonable location coordinates calculated based on the continuous direction and are used to replace conflicting coordinates to ensure that the transport route is spatially continuous and smooth.

[0101] In this embodiment, the position coordinates of the preceding and succeeding nodes of the node with conflicting position coordinates are first extracted, and the transport direction vector between them is calculated (for example, a positive longitude difference indicates eastward, and a negative longitude difference indicates westward). Next, a continuous direction is generated based on the transport direction vector (for example, an eastward transport path should maintain increasing longitude). The original coordinates of the conflicting node are vertically projected onto the continuous direction path. If the projected point meets the preset geographic path constraints (for example, the projected point is located on the line connecting the adjacent nodes), the projected point coordinates are replaced. If the projected point deviates from the path constraints, the intermediate coordinates are calculated based on the position change trends of the adjacent nodes (for example, the average of the coordinates of the preceding and succeeding nodes is taken). Finally, the conflicting coordinates are replaced with the intermediate coordinates to ensure the continuity and directional rationality of the transport path.

[0102] Step 605: Rebind the corrected timestamp and the intermediate coordinates to the corresponding logistics nodes to generate a dynamic mapping relationship after conflict elimination.

[0103] In this step, the dynamic mapping relationship after conflict elimination refers to the cross-modal data mapping table formed after the corrected timestamp and location coordinates are re-associated with the logistics node, reflecting the spatiotemporal logical consistency of the drug circulation path.

[0104] In this embodiment, the timestamp corrected in step 603 and the intermediate coordinates corrected in step 604 are first bound to the corresponding logistics nodes, replacing the original conflicting data. Next, based on the corrected node data, the time dimension association (matching the production date with the corrected timestamp) and the location dimension association (binding the unique identifier to the corrected coordinates) are regenerated to form a new dynamic mapping relationship. Finally, the updated dynamic mapping relationship is pushed to the monitoring terminal, triggering the data synchronization and alarm release process, completing the closed-loop monitoring process.

[0105] In order to solve the problem of unreasonable coordinates caused by sudden changes in the transportation path and improve the physical consistency of the calculation of the continuous direction of the geographical path, in some embodiments, according to step 604, for the node with conflicting position coordinates, the position coordinates of the adjacent previous logistics node and the next logistics node of the logistics node in the dynamic tracking link are extracted, and the continuous direction of the transportation path is generated according to the position change direction between the adjacent logistics nodes. The conflicting position coordinates of the node with conflicting position coordinates are replaced with the intermediate coordinates calculated based on the continuous direction, including: Step 701: Connect the position coordinates of the adjacent previous logistics node and the position coordinates of the adjacent next logistics node to form a straight path, and determine the moving direction vector of the transportation path based on the coordinate difference between the starting point and the end point of the straight path.

[0106] In this step, the straight path refers to a virtual line formed by connecting the position coordinates of the adjacent previous and next nodes, representing the theoretical shortest transportation path between the two points. The movement direction vector refers to a vector generated based on the coordinate difference between the starting and ending points of the straight path, which is used to describe the transportation direction (such as east or north) and movement trend.

[0107] In this example, the coordinates of the adjacent previous and next nodes are first extracted, and the two points are connected to form a straight line path. Next, the longitude and latitude differences between the starting point and the end point are calculated to generate a movement direction vector, and its angular range is calculated to indicate the appropriate direction of the transportation path.

[0108] Step 702: Divide the orthographic projection areas on both sides of the straight path according to the angular range of the moving direction vector.

[0109] In this step, the orthographic projection area refers to the reasonable transportation path range delineated on both sides of the straight path based on the angle range of the moving direction vector, which is used to constrain the boundary of the conflict coordinate correction.

[0110] In this embodiment, the reasonable range of the transport path is first determined based on the angular range of the movement direction vector. Next, a predetermined width (e.g., 10% of the path length) is expanded along both sides of the linear path to form an orthographic projection area. This ensures that the corrected coordinates fall within this area to avoid deviation from the actual transport direction.

[0111] Step 703, vertically project the original coordinates of the position coordinate conflict node onto the straight path, and at the same time detect whether the projection point of the vertical projection falls within the orthographic projection area. When it does not fall within the orthographic projection area, translate the coordinates of the conflict position of the position coordinate conflict node along the moving direction vector to the nearest projection point on the straight path.

[0112] In this step, vertical projection refers to projecting the original coordinates of the conflict node onto the path in a direction perpendicular to the straight path to generate theoretical reasonable coordinates.

[0113] In this embodiment, the original coordinates of the conflicting node (e.g., longitude 116.3, latitude 40.1) are first projected perpendicularly onto the linear path A→C to calculate the coordinates of the projected point. Next, the projection point is checked to see if it lies within the orthographic projection area. If not (e.g., the projection point deviates from the path extension range), the conflicting coordinates are translated along the movement direction vector to the nearest projection point on the linear path. If so, the process proceeds to step 704.

[0114] Step 704: When the projection point falls within the positive projection area, the position coordinates of the adjacent previous logistics node are used as the starting point, and the position coordinates are extended along the moving direction vector to the position coordinates of the next logistics node, and the conflicting position coordinates of the position coordinate conflicting node are replaced with the intermediate coordinates between the starting point and the end point divided according to the preset distance ratio.

[0115] In this step, the preset distance ratio segmentation refers to determining the position of the intermediate coordinates according to a reasonable segmentation rule of the transportation route (such as equal division or time ratio).

[0116] In this embodiment, if the projection point is located within the orthographic projection area, the adjacent previous node is used as the starting point, and the moving direction vector is extended to the next node. The path is divided according to a preset distance ratio (for example, 1 / 2 of the path length), and the intermediate coordinates (such as longitude 116.55, latitude 40.35) are generated to replace the original coordinates of the conflicting node.

[0117] Figure 2 The present invention provides a schematic diagram of a drug identification information intelligent association system based on multi-source data fusion, as shown in FIG. Figure 2 As shown, the system includes: An acquisition module is used to acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes; The recognition module is used to identify key information elements printed on the surface of the drug packaging based on the geometric distribution characteristics of the image data, and to parse the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug identity; Establish a module for establishing a dynamic tracking link of the drug transportation route based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics nodes; A matching module, configured to perform cross-modal correlation matching between the key information elements, the unique identifier, and the node events in the dynamic tracking link, and generate a dynamic mapping relationship between drug identification information and different data sources; The verification module is used for dynamic mapping relationships, real-time verification of the consistency of image data, text data and spatiotemporal trajectory data during the drug circulation process, and synchronously updates the verified dynamic mapping relationships to the regulatory terminals in the drug circulation link.

[0118] Figure 2 The intelligent association system of drug identification information based on multi-source data fusion can be executed Figure 1 The implementation principle and technical effects of the method for intelligently associating drug identification information based on multi-source data fusion described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the intelligent association system for drug identification information based on multi-source data fusion in the above embodiment has been described in detail in the relevant embodiments of the method and will not be elaborated on here.

[0119] In one possible design, Figure 2 The drug identification information intelligent association system based on multi-source data fusion of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0120] The processing component 32 is used for the above Figure 1 The embodiment provides an intelligent association method for drug identification information based on multi-source data fusion.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for intelligent association of drug identification information based on multi-source data fusion, characterized in that: include: Acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes; Based on the geometric distribution characteristics of the image data, the system identifies key information elements printed on the surface of the drug packaging, and parses the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug's identity. Based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of logistics nodes, a dynamic tracking link of the drug transportation route is established; Cross-modal association matching of the key information elements, unique identifiers, and node events in the dynamic tracking link is performed to generate a dynamic mapping relationship between drug identification information and different data sources; Through dynamic mapping relationships, the consistency of image data, text data and spatiotemporal trajectory data in the drug circulation process can be verified in real time, and the verified dynamic mapping relationships can be synchronously updated to the regulatory terminals in the drug circulation link.

2. The method according to claim 1, characterized in that Acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes, including: Capturing image data of the surface of the drug packaging using a preset scanning device, wherein the scanning device dynamically adjusts the scanning resolution according to the physical size of the drug packaging and adaptively adjusts the image contrast based on the ambient light intensity; The initial text data of the drug electronic supervision code is read by a code scanning device, and the missing characters in the initial text data are checked for integrity based on the encoding rules of the electronic supervision code. When the position of the missing characters is in a predefined non-key field, the original content of the valid field in the text data is retained to obtain the target text data; The recording device of the logistics node captures the time information and position coordinates of the drugs passing through each logistics node during transportation, and arranges the time information and position coordinates into a position change sequence with continuous timestamps according to the transportation order of the drugs to form spatiotemporal trajectory data.

3. The method according to claim 1, characterized in that Based on the geometric distribution characteristics of the image data, the system identifies key information elements printed on the surface of the drug packaging, parses the predefined encrypted fields in the electronic supervision code, and extracts the unique identifier bound to the drug identity, including: Analyze the geometric distribution characteristics of the image data, determine the boundary range of the text area based on the layout direction of the printed content on the drug packaging surface, and segment the text area containing the drug name and production batch number from the image data into independent key information elements; Based on the key information elements, the text data of the electronic supervision code is decomposed into fields according to predefined encryption rules. According to the arrangement order and length constraints of the encrypted fields after decomposition, character fragments of a fixed number of bits are extracted from the first field and the last field in the arrangement order, and the character fragments are spliced into a unique identifier bound to the identity of the drug.

4. The method according to claim 1, wherein Based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of logistics nodes, a dynamic tracking link for drug transportation routes is established, including: Extract the timestamp-continuous location coordinates of the logistics nodes from the spatiotemporal trajectory data, and generate the initial links of the transportation path according to the increasing order of the timestamps of the location coordinates; Detecting whether the position coordinate change between adjacent logistics nodes in the initial link exceeds a preset distance threshold, retaining the current logistics node when it does not exceed the preset distance threshold, and connecting the retained logistics nodes in series to form an initial dynamic tracking link based on timestamp continuity; When the change in position coordinates exceeds the preset distance threshold, the abnormal logistics node with the largest timestamp interval among the adjacent logistics nodes is eliminated, and the remaining logistics nodes are reconnected in series to generate a target dynamic tracking link.

5. The method according to claim 1, wherein Cross-modal association and matching of the key information elements, unique identifiers, and node events in the dynamic tracking link are performed to generate a dynamic mapping relationship between drug identification information in different data sources, including: Extract the timestamps and location coordinates corresponding to the drug's flow through each logistics node from the dynamic tracking link, match the timestamps with the production date in the key information element, and generate a correlation relationship in the time dimension; Binding the unique identifier to the logistics node where the drug first appears in the dynamic tracking link to generate an association relationship in the location dimension; When there is a timestamp or location coordinate conflict between the association relationship in the time dimension and the association relationship in the location dimension, the conflicting timestamps or conflicting location coordinates are logically replaced based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information between different data sources.

6. The method according to claim 5, characterized in that When there is a conflict between timestamps or location coordinates in the association relationship on the time dimension and the association relationship on the location dimension, the conflicting timestamps or location coordinates are logically replaced based on the association relationship of adjacent logistics nodes in the dynamic tracking link to obtain a dynamic mapping relationship between drug identification information in different data sources, including: Detect whether the difference between the timestamp of the logistics node where the drug flows through the preset location in the association relationship on the time dimension and the production date in the key information element exceeds a preset circulation cycle threshold, and determine it as a timestamp conflict node when it exceeds the preset circulation cycle threshold; Detect whether the transportation direction between the logistics node location coordinates where the drug first appears in the association relationship on the location dimension and the subsequent logistics node location coordinates in the dynamic tracking link violates the preset geographical path constraints, and determine the node as a location coordinate conflict node if the preset geographical path constraints are violated; For a timestamp conflict node, extract the timestamps of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link, calculate the average time interval between adjacent logistics nodes, and replace the conflicting timestamp in the timestamp conflict node with a corrected timestamp obtained by adding the production date to the average time interval. For nodes with conflicting position coordinates, the position coordinates of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link are extracted. According to the position change direction between the adjacent logistics nodes, a continuous direction of the transportation path is generated. The conflicting position coordinates of the node with conflicting position coordinates are replaced with the intermediate coordinates calculated based on the continuous direction. Rebind the corrected timestamp and intermediate coordinates to the corresponding logistics nodes to generate a dynamic mapping relationship after conflict elimination.

7. The method according to claim 6, characterized in that For nodes with conflicting position coordinates, the position coordinates of the adjacent previous and next logistics nodes of the logistics node in the dynamic tracking link are extracted, and a continuous direction of the transportation path is generated according to the position change direction between the adjacent logistics nodes. The conflicting position coordinates of the node with conflicting position coordinates are replaced with intermediate coordinates calculated based on the continuous direction, including: Connecting the position coordinates of the adjacent previous logistics node and the position coordinates of the adjacent next logistics node to form a straight path, and determining the moving direction vector of the transportation path based on the coordinate difference between the starting point and the end point of the straight path; Dividing orthographic projection areas on both sides of the straight path according to the angular range of the moving direction vector; Projecting the original coordinates of the position coordinate conflict node vertically onto the straight path, and detecting whether the projection point of the vertical projection falls within the orthographic projection area; if not, translating the coordinates of the conflicting position of the position coordinate conflict node along the moving direction vector to the nearest projection point on the straight path; When the projection point falls within the positive projection area, the position coordinates of the adjacent previous logistics node are used as the starting point, and the position coordinates are extended along the moving direction vector to the position coordinates of the next logistics node, and the conflicting position coordinates of the position coordinate conflicting node are replaced with the intermediate coordinates between the starting point and the end point divided according to the preset distance ratio.

8. An intelligent association system for drug identification information based on multi-source data fusion, characterized in that: include: An acquisition module is used to acquire multi-source heterogeneous data generated during the drug circulation process, wherein the multi-source heterogeneous data includes image data of the drug packaging surface, text data of the drug electronic supervision code, and spatiotemporal trajectory data of logistics nodes; The recognition module is used to identify key information elements printed on the surface of the drug packaging based on the geometric distribution characteristics of the image data, and to parse the predefined encrypted fields in the electronic supervision code to extract the unique identifier bound to the drug identity; Establish a module for establishing a dynamic tracking link of the drug transportation route based on the position change sequence and timestamp continuity contained in the spatiotemporal trajectory data of the logistics nodes; A matching module, configured to perform cross-modal correlation matching between the key information elements, the unique identifier, and the node events in the dynamic tracking link, and generate a dynamic mapping relationship between drug identification information and different data sources; The verification module is used for dynamic mapping relationships, real-time verification of the consistency of image data, text data and spatiotemporal trajectory data during the drug circulation process, and synchronously updates the verified dynamic mapping relationships to the regulatory terminals in the drug circulation link.

9. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for intelligent association of drug identification information based on multi-source data fusion as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the method for intelligently associating drug identification information based on multi-source data fusion as described in any one of claims 1 to 7 is implemented.

Images