A medicine storage management method and device, a terminal device and a storage medium

Abstract

The application discloses a medicine storage management method and device, a terminal equipment and a storage medium. The method comprises the following steps: searching for a shelf center vector matched with a medicine-in vector, and placing the medicine-in on a shelf corresponding to the shelf center vector; the shelf center vector is a center vector of a vector set formed by all medicine vectors on the corresponding shelf; obtaining a point cloud diagram corresponding to the sub-medicine warehouse at two time points before and after medicine-in, or a point cloud diagram corresponding to the sub-medicine warehouse at two time points before and after medicine-out, and obtaining a residual heat map through an iterative closest point algorithm; and checking whether the corresponding shelf medicine and medicine order information with a difference value greater than a preset difference threshold in the residual heat map are matched. The application realizes the intelligentization, precision and high efficiency of medicine storage management, solves the pain points in traditional medicine storage, and provides reliable technical support for the whole life cycle quality management of medicines.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medicine management, and in particular to a medicine storage management method and device, a terminal device and a storage medium. BACKGROUND

[0002] With the rapid development of the medical circulation system, the types of medicines are increasing, and higher requirements for the safety, compliance and efficiency of warehouse management are put forward. Traditional medicine warehouse management mainly relies on manual operation and barcode scanning, which has the following problems: first, the storage area allocation is extensive: only according to the temperature control condition to divide the large area, lack of comprehensive consideration of the properties of medicines (such as volume, turnover rate, price, etc.), leading to high-frequency medicines far away from the entrance and exit, low space utilization; second, the warehouse check relies on manual operation: after the medicine is put on the shelf or taken out of the warehouse, whether it is accurately placed in the target position, usually there is no automatic verification mechanism, and the wrong or missing placement can not be found in time. SUMMARY

[0003] The embodiments of the present application provide a medicine storage management method, device, terminal device and storage medium, which realizes intelligent allocation of medicine into the warehouse, dynamic optimization of storage location and automatic verification of operation behavior, and improves the safety, accuracy and traceability of medicine storage.

[0004] To achieve the above-mentioned purpose, the first aspect of the embodiments of the present application provides a medicine storage management method, comprising:

[0005] obtaining information of the incoming medicine according to the unique identification code of the incoming medicine;

[0006] selecting a sub-medicine warehouse meeting the storage condition according to the information of the incoming medicine;

[0007] vectorizing the information of the incoming medicine to obtain an incoming medicine vector;

[0008] finding a shelf center vector matching the incoming medicine vector in the point cloud graph corresponding to the sub-medicine warehouse, and placing the incoming medicine on the shelf corresponding to the shelf center vector; the shelf center vector is a center vector of a vector set formed by all medicine vectors on the corresponding shelf;

[0009] obtaining the point cloud graph corresponding to the sub-medicine warehouse at two time points before and after the incoming, or the point cloud graph corresponding to the sub-medicine warehouse at two time points before and after the outgoing, and obtaining a residual heat map through an iterative closest point algorithm;

[0010] checking whether the corresponding shelf medicine and the medicine order information in the residual heat map with a difference value greater than a preset difference threshold match.

[0011] In a possible implementation manner of the first aspect, before the selecting a storage warehouse meeting the storage condition according to the information of the incoming medicine, the method further comprises:

[0012] The main drug warehouse is divided into an ambient temperature zone drug warehouse, a cool zone drug warehouse, a cold storage zone drug warehouse, a variable temperature zone drug warehouse, and a return zone drug warehouse.

[0013] K-means clustering is performed on the drug vectors of all drugs in each drug warehouse, and the number of cluster centers is equal to the number of shelves in the corresponding drug warehouse.

[0014] After clustering, the center vector of each cluster is used as the initial shelf center vector, and the medicines corresponding to the clusters are placed on the same shelf;

[0015] K-means clustering is performed after each receipt or shipment to update the center vector of each shelf.

[0016] In one possible implementation of the first aspect, the step of vectorizing the purchased drug information to obtain a purchased drug vector specifically includes:

[0017] The drug name, production date, expiration date, quantity, volume, and selling price of the purchased drugs are obtained from the purchased drug information, and a multi-dimensional purchased drug vector is generated.

[0018] In one possible implementation of the first aspect, the step of finding the shelf center vector that matches the incoming drug vector in the point cloud map corresponding to the sub-drug warehouse, and placing the incoming drug on the shelf corresponding to the shelf center vector, specifically includes:

[0019] In the point cloud map corresponding to the sub-drug warehouse, each shelf plane is bound to a shelf center vector;

[0020] Calculate the cosine distance between the incoming medicine vector and each of the shelf center vectors, take the shelf center vector with the smallest cosine distance value as the matching shelf center vector, and place the incoming medicine on the shelf corresponding to the shelf center vector.

[0021] In one possible implementation of the first aspect, obtaining the residual heatmap through the iterative nearest point algorithm specifically includes:

[0022] The iterative nearest point algorithm is used to register the first point cloud map of the previous time step to the second point cloud map of the next time step.

[0023] After registration, the shortest distance from each point in the first point cloud map to the second point cloud map is calculated;

[0024] If the nearest distance between all points in a connected region is greater than or equal to the mean of the magnitudes of all drug vectors in the connected region, the connected region is marked in red.

[0025] In one possible implementation of the first aspect, marking the connected region in red includes:

[0026] If the nearest distance between all points within the connected region is less than the mean of the magnitudes of all drug vectors within the connected region, and the nearest distance between all points within the connected region is greater than or equal to a distance threshold, the connected region is marked in yellow; the distance threshold is the weighted sum of shelf area, drug volume, and drug quantity.

[0027] If the nearest distance between all points within the connected region is less than a distance threshold, the connected region is marked in blue.

[0028] In one possible implementation of the first aspect, checking whether the shelf medicines and medicine order information corresponding to the difference values ​​in the residual heatmap that are greater than a preset difference threshold match specifically includes:

[0029] Check whether the names and quantities of the drugs on the shelves and in the drug order information corresponding to the red connected areas in the residual heatmap are consistent. If they are consistent, randomly select several drug samples to check the production date, expiration date, and drug packaging.

[0030] A second aspect of this application provides a pharmaceutical storage management device, comprising:

[0031] The acquisition module is used to obtain information about the purchased medicines based on their unique identifiers.

[0032] The selection module is used to select a sub-drug warehouse that meets the storage conditions based on the purchased drug information;

[0033] The vector module is used to vectorize the purchased drug information to obtain a purchased drug vector;

[0034] The matching module is used to find the shelf center vector that matches the vector of the incoming medicine in the point cloud map corresponding to the sub-medicine warehouse, and place the incoming medicine on the shelf corresponding to the shelf center vector; the shelf center vector is the center vector of the vector set formed by all medicine vectors on the corresponding shelf;

[0035] The heatmap module is used to obtain the point cloud map corresponding to the sub-drug warehouse at two times before and after the goods are received or at two times before and after the goods are shipped, and to obtain the residual heatmap by iterative nearest point algorithm.

[0036] The inspection module is used to check whether the corresponding shelf medicines and medicine order information for those whose difference values ​​in the residual heatmap are greater than a preset difference threshold match.

[0037] A third aspect of this application provides a terminal device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the drug storage management method described above.

[0038] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the pharmaceutical storage management method described above.

[0039] Compared to existing technologies, this invention first accurately obtains drug information through unique identifiers and transforms this information into multi-dimensional vector representations, making drug characteristics quantifiable and calculable. Based on the similarity matching between drug vectors and shelf center vectors, it automatically determines the optimal storage location, achieving intelligent placement of similar drugs. It then uses point cloud technology to create a three-dimensional spatial model of the warehouse, comparing point cloud data from different periods using the Iterative Closest Point (ICP) algorithm to accurately calculate spatial changes and generate an intuitive residual heatmap. Finally, based on preset thresholds, the residual heatmap is analyzed to automatically identify areas of abnormal change and intelligently match and verify them with order information.

[0040] In summary, by vectorizing medicines and matching them with shelf-center vectors, we ensure that medicines are intelligently placed according to their characteristics, solving the problems of "not being able to find goods" or "difficulty in detecting misplaced goods" in traditional warehousing, and significantly improving the accuracy of medicine location. The strategy of storing similar medicines together not only facilitates picking operations but also makes the warehouse space layout more rational and improves space utilization. The residual heat map technology transforms traditional comprehensive inventory counting into targeted verification, reducing anomaly detection time from several hours to a few minutes, thus greatly improving inventory reconciliation efficiency. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating a pharmaceutical warehousing management method according to an embodiment of the present invention;

[0042] Figure 2 This is a schematic diagram of the structure of a pharmaceutical storage management device according to an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] To resolve the above issues, please refer to [link / reference]. Figure 1An embodiment of the present invention provides a method for pharmaceutical warehouse management, comprising:

[0045] S10. Obtain information about the purchased medicines based on their unique identification codes.

[0046] S11. Select a sub-drug warehouse that meets the storage conditions based on the purchased drug information.

[0047] S12. Vectorize the purchased drug information to obtain the purchased drug vector.

[0048] S13. In the point cloud map corresponding to the sub-drug warehouse, find the shelf center vector that matches the incoming drug vector, and place the incoming drug on the shelf corresponding to the shelf center vector; the shelf center vector is the center vector of the vector set formed by all drug vectors on the corresponding shelf.

[0049] S14. Obtain the point cloud map corresponding to the sub-drug warehouse at two time points before and after the purchase or at two time points before and after the shipment, and obtain the residual heat map by iterative nearest point algorithm.

[0050] S15. Check whether the corresponding shelf medicines and medicine order information for those with a difference value greater than a preset difference threshold in the residual heat map match.

[0051] S10-S11 is the process of precise information acquisition and storage matching: First, detailed drug information is automatically obtained through a unique identifier. Then, based on the drug's storage requirements (such as temperature, humidity, etc.), it is intelligently matched to a suitable sub-warehouse area. This solves the problem of inconsistent storage conditions caused by manual judgment in traditional warehousing. Automated storage condition matching ensures that drugs are always stored in a suitable environment, effectively reducing drug quality problems caused by improper storage and complying with GSP (Good Supply Practice) requirements.

[0052] S12-S13 is the process of drug vectorization and intelligent placement: The multi-dimensional attributes of drugs (name, production date, expiration date, quantity, volume, price, etc.) are transformed into standardized vectors. By calculating the similarity between the vector of a newly arrived drug and the center vector of each shelf (the average of all drug vectors on the shelf), the optimal storage location is automatically determined. This vectorization matching mechanism ensures that drugs are intelligently placed according to their characteristics, improving drug location accuracy and solving the pain point of "not being able to find the goods" in traditional warehousing. This process implements an intelligent placement strategy of "centralized storage of similar drugs," which can further utilize algorithms based on cosine distance to ensure matching accuracy. The centralized storage strategy of similar drugs makes the warehouse layout more scientific and rational, reducing unnecessary movement and search time, thereby improving picking efficiency.

[0053] S14-S15 pertain to the spatial perception and anomaly detection process: 3D scanning technology is used to acquire warehouse point cloud data. The iterative nearest neighbor (ICP) algorithm is then used to accurately register point clouds from different periods, generating an intuitive residual heatmap. The system automatically identifies areas of abnormal change (differences exceeding preset thresholds) and intelligently compares and verifies them with order information, achieving a shift from "comprehensive inventory" to "partial verification." Furthermore, point cloud technology and heatmaps provide a visual representation of warehouse status, enabling a shift from "experience-driven" to "data-driven" approaches, and providing reliable technical support for the entire lifecycle quality management of pharmaceuticals.

[0054] The above-described method and implementation methods achieve intelligent, precise, and efficient drug storage management through technological innovation, providing a brand-new solution for drug safety and quality management, and have significant economic value and social benefits.

[0055] For example, before selecting a storage warehouse that meets the storage conditions based on the purchased drug information, the process includes:

[0056] The main drug warehouse is divided into an ambient temperature zone drug warehouse, a cool zone drug warehouse, a cold storage zone drug warehouse, a variable temperature zone drug warehouse, and a return zone drug warehouse.

[0057] K-means clustering is performed on the drug vectors of all drugs in each drug warehouse, and the number of cluster centers is equal to the number of shelves in the corresponding drug warehouse.

[0058] After clustering, the center vector of each cluster is used as the initial shelf center vector, and the medicines corresponding to the clusters are placed on the same shelf;

[0059] K-means clustering is performed after each receipt or shipment to update the center vector of each shelf.

[0060] Warehouse zoning can be referenced as follows:

[0061] Ambient temperature zone (20-30℃): 1500㎡, including 30 shelves;

[0062] Cool area (8-20℃): 1200㎡, including 25 shelves;

[0063] Cold storage area (2-8℃): 1000㎡, including 20 shelves;

[0064] Variable temperature zone: 800㎡, including 15 shelves;

[0065] Returns area: 500㎡, including 10 shelves;

[0066] Taking the room temperature zone as an example, K-means clustering is performed on existing medicines, where K = 30 (number of shelves).

[0067] Clustering features include: drug category, frequency of use, volume, and expiration date.

[0068] Clustering results:

[0069] Shelves 1-5: Cardiovascular medications

[0070] Shelves 6-10: Antibiotics (including cephalosporins, etc.)

[0071] Shelves 11-15: Digestive system medications

[0072] ...(Other categories follow the same pattern)

[0073] The center vector of each shelf is calculated and stored. For example, the center vector of shelf number 7 (antibiotics) is:

[0074] [0.85,20230315,20250228,85,0.0028,12.5] (representing: antibiotic category identifier, average production date, average expiration date, average quantity, average volume, and average price, respectively)

[0075] The main warehouse is divided into five functional areas: ambient temperature zone, cool zone, refrigerated zone, variable temperature zone, and return zone, strictly adhering to Good Supply Practice (GSP) management standards for pharmaceuticals. Building upon this, an innovative K-means clustering algorithm is introduced to group drug vectors within each area according to similarity, with the number of cluster centers strictly corresponding to the number of physical shelves. This design ensures that the algorithm output is directly mapped to physical storage space, avoiding the mismatch between traditional clustering methods and actual shelf space. Furthermore, the K-means clustering algorithm can automatically identify the inherent relationships between drug characteristics, achieving more accurate shelf allocation than manual classification.

[0076] The central vector of each cluster is used as the "shelf central vector," and medicines within the same cluster are grouped together on the same shelf, forming a scientific initial layout. More importantly, this method establishes a dynamic update mechanism—K-means clustering is re-performed after each receipt or shipment, updating the shelf central vector in real time. This continuous optimization ensures that shelf allocation is always in an optimal state, adapting to real-time inventory changes. The dynamic clustering mechanism enables the system to automatically adjust shelf allocation strategies according to inventory changes, improving space utilization. Especially during seasonal fluctuations in medicine prices, the system can automatically rebalance the load on each shelf, avoiding situations where some shelves are overcrowded while others are idle.

[0077] In summary, by strictly binding the K-means clustering algorithm to the number of physical shelves, the problem of mapping algorithm output to physical space is creatively solved, which can also improve the practicality, adaptability and management efficiency of the warehouse management system.

[0078] For example, the step of vectorizing the purchased drug information to obtain a purchased drug vector specifically includes:

[0079] The drug name, production date, expiration date, quantity, volume, and selling price of the purchased drugs are obtained from the purchased drug information, and a multi-dimensional purchased drug vector is generated.

[0080] Suppose a new batch of "enteric-coated aspirin tablets" has arrived, and its details are as follows:

[0081] Drug Name: Enteric-coated Aspirin Tablets;

[0082] Production date: 2023-06-15;

[0083] Valid until: June 14, 2025;

[0084] Quantity of medicine: 200 boxes;

[0085] Drug volume: 0.0025 m3 per box;

[0086] Price of medicine: 18.5 yuan / box;

[0087] A drug category knowledge base can be established to map drug names to standard categories. By analyzing drug names using NLP technology, active ingredients and dosage forms can be extracted. For example, "Aspirin Enteric-coated Tablets" has "Aspirin" as its active ingredient and belongs to the "Antiplatelet Drugs" category. In the predefined category vector space, the corresponding code value for "Antiplatelet Drugs" is 0.92.

[0088] Convert the date to days from today: 0 days (assuming today is 2023-06-15). However, directly using the number of days still has a dimension problem, so further standardization is needed: Set a reference range: 0-730 days (2 years); Standardization formula: Standardized value = (current number of days - minimum value) / (maximum value - minimum value); Standardized value: (0-0) / (730-0) = 0.00;

[0089] Calculate the remaining validity period in days: 730 days (2 years). Standardize the data by setting a reference range of 0-1095 days (3 years). The standardized value is: (730-0) / (1095-0)=0.67.

[0090] Considering that the quantity of medicines in the warehouse usually ranges from 1 to 1000 boxes, the standardized formula is: (200-1) / (1000-1)=0.20;

[0091] Original volume: 0.0025m 3 / box, considering that the volume of the drug is usually in the range of 0.001-0.01m 3Standardized value: (0.0025-0.001) / (0.01-0.001)=0.17;

[0092] Considering that the price range of drugs is usually between 5 and 100 yuan, the standardized value is: (18.5-5) / (100-5) = 0.14. After the above processing, the 6-dimensional drug vector for aspirin enteric-coated tablets is: [0.92, 0.00, 0.67, 0.20, 0.17, 0.14]. Each dimension of this vector represents: drug category code (antiplatelet drug), standardized value of production date (latest production), standardized value of expiration date (2 years remaining), standardized value of drug quantity (200 boxes), standardized value of drug volume (small volume), and standardized value of drug price (medium price).

[0093] For example, the step of finding the shelf center vector that matches the incoming drug vector in the point cloud map corresponding to the sub-drug warehouse, and placing the incoming drug on the shelf corresponding to the shelf center vector, specifically includes:

[0094] In the point cloud map corresponding to the sub-drug warehouse, each shelf plane is bound to a shelf center vector;

[0095] Calculate the cosine distance between the incoming medicine vector and each of the shelf center vectors, take the shelf center vector with the smallest cosine distance value as the matching shelf center vector, and place the incoming medicine on the shelf corresponding to the shelf center vector.

[0096] First, calculate the center vector of each shelf (based on the currently stored medicines), and then spatially bind these vectors to the shelf plane, for example:

[0097] Shelf #1 (Cardiovascular Category): [0.85, 0.25, 0.60, 0.35, 0.20, 0.15]

[0098] Shelf #2 (Respiratory System Products): [0.30, 0.40, 0.55, 0.60, 0.25, 0.20]

[0099] Shelf #3 (Antibiotics): [0.88, 0.35, 0.52, 0.45, 0.22, 0.18]

[0100] Shelf #4 (Digestive System): [0.25, 0.50, 0.48, 0.55, 0.28, 0.22]

[0101] ...(Other shelves are similar)

[0102] As mentioned earlier, the 6-dimensional vector of amoxicillin, after standardization, is: A = [0.90, 0.00, 0.70, 0.30, 0.21, 0.17]. Calculations show that its cosine distance to shelf number 3 (antibiotics) is the smallest, at 0.058.

[0103] -Shelf #1 (Cardiovascular): Cosine distance = 0.421;

[0104] -Shelf #2 (Respiratory System): Cosine distance = 0.385;

[0105] -Shelf #3 (Antibiotics): Cosine distance = 0.058;

[0106] -Shelf #4 (Digestive System): Cosine distance = 0.412;

[0107] -Shelf #5 (Vitamins): Cosine distance = 0.523;

[0108] The staff placed amoxicillin in the designated location on shelf number 3, updated the drug information on shelf number 3, and recalculated the center vector of shelf number 3.

[0109] New center vector = (original vector × original number of drugs + new drug vector) / (original number of drugs + 1) = [0.882, 0.32, 0.53, 0.46, 0.22, 0.18].

[0110] This embodiment achieves seamless integration of physical and data spaces by binding the shelf plane to the center vector. The use of cosine distance enables the system to accurately capture the pattern similarity of drug characteristics, rather than simply comparing a single attribute, significantly improving matching accuracy. Furthermore, updating the shelf center vector after each inventory entry allows the system to automatically adjust the matching strategy according to inventory changes, always maintaining optimal performance.

[0111] For example, obtaining the residual heatmap through the iterative nearest point algorithm specifically includes:

[0112] The iterative nearest point algorithm is used to register the first point cloud map of the previous time step to the second point cloud map of the next time step.

[0113] After registration, the shortest distance from each point in the first point cloud map to the second point cloud map is calculated;

[0114] If the nearest distance between all points in a connected region is greater than or equal to the mean of the magnitudes of all drug vectors in the connected region, the connected region is marked in red.

[0115] For example, marking the connected region in red includes:

[0116] If the nearest distance between all points within the connected region is less than the mean of the magnitudes of all drug vectors within the connected region, and the nearest distance between all points within the connected region is greater than or equal to a distance threshold, the connected region is marked in yellow; the distance threshold is the weighted sum of shelf area, drug volume, and drug quantity.

[0117] If the nearest distance between all points within the connected region is less than a distance threshold, the connected region is marked in blue.

[0118] Suppose a 150㎡ ambient temperature warehouse in a pharmaceutical distribution center contains 10 shelves. The system automatically collects point cloud data every 2 hours with a resolution of 2cm. Two point cloud scans were performed this afternoon at 2:00 PM (T1 time) and 4:00 PM (T2 time). The system needs to analyze the warehouse changes over these 2 hours using the Iterative Closest Point (ICP) algorithm.

[0119] ICP registration process: Coarse alignment is performed based on the fixed structure of the warehouse (walls, columns, etc.), followed by iterative optimization: 1st iteration: average distance error 0.15m; 2nd iteration: average distance error 0.08m; 3rd iteration: average distance error 0.03m; 4th iteration: average distance error 0.01m (reaching the convergence threshold, iteration stops); the two point clouds are successfully aligned with a registration error of <2cm, and the nearest corresponding point in the T2 point cloud is found for each point in the T1 point cloud.

[0120] It should be noted that the above process can use a region growing algorithm to identify connected regions in the point cloud.

[0121] Red area: Distance greater than the size of the medicine indicates that medicine has been significantly moved or is missing, and must be dealt with immediately; Yellow area: Distance within the expected inbound / outbound range, the system automatically matches and verifies the order; Blue area: Minor changes, which are normal scanning errors and should be avoided to prevent false alarms.

[0122] The distance threshold can be calculated as follows:

[0123] Distance threshold = shelf area × w1 + drug volume × w2 + drug quantity × w3. Example of actual calculation:

[0124] For standard pharmaceutical products (such as amoxicillin): shelf space: 1.2m² 2 ×0.3=0.36; Drug volume: 0.0028m³ 3 ×0.5=0.0014; Quantity of medicines: 150 boxes ×0.2=30; Dimensionless processing: (0.36+0.0014+30) / 1000=0.030014m≈0.03m. The weighted calculation of the distance threshold takes into account the combined effects of shelf space, medicines, and quantity, making the system's judgment more accurate and reasonable.

[0125] This embodiment utilizes a residual heatmap method based on point cloud technology and the ICP algorithm to transform traditional passive warehouse management into proactive intelligent monitoring. This not only improves inventory accuracy but also provides strong protection for drug quality and safety. Especially given the increasingly stringent requirements of GSP (Good Supply Practice) for pharmaceuticals, this technology can effectively reduce compliance risks and has significant industry application value.

[0126] For example, checking whether the shelf-side medicines and medicine order information corresponding to differences greater than a preset difference threshold in the residual heatmap match specifically includes:

[0127] Check whether the names and quantities of the drugs on the shelves and in the drug order information corresponding to the red connected areas in the residual heatmap are consistent. If they are consistent, randomly select several drug samples to check the production date, expiration date, and drug packaging.

[0128] This embodiment employs a three-tiered verification system: Level 1: Comparison of drug name and quantity (highly automated, rapid screening); Level 2: Inspection of production date and expiration date (targeted sampling to ensure quality); Level 3: Inspection of packaging condition (assessing potential quality risks). This allows for a rapid name and quantity comparison to quickly eliminate obvious anomalies, followed by targeted inspection of factors that may affect drug quality. The results of the Level 1 verification then determine whether further in-depth inspection is necessary. This verification mechanism, through name-quantity comparison and quality sampling, ensures both efficiency and quality, achieving a balance between efficiency and safety.

[0129] The above embodiments, based on the anomaly verification method using residual heatmaps, transform traditional passive warehouse management into proactive quality control, which not only improves inventory accuracy but also provides strong protection for drug quality and safety.

[0130] Compared to existing technologies, this invention first accurately obtains drug information through unique identifiers and transforms this information into multi-dimensional vector representations, making drug characteristics quantifiable and calculable. Based on the similarity matching between drug vectors and shelf center vectors, the optimal storage location is automatically determined, achieving intelligent placement of similar drugs. Point cloud technology is used to create a three-dimensional spatial model of the warehouse, and the iterative nearest point (ICP) algorithm is used to compare point cloud data from different periods to accurately calculate spatial changes and generate an intuitive residual heatmap. Then, based on preset thresholds, the residual heatmap is analyzed to automatically identify abnormal change areas and intelligently match and verify them with order information.

[0131] In summary, by vectorizing medicines and matching them with shelf-center vectors, we ensure that medicines are intelligently placed according to their characteristics, solving the problems of "not being able to find goods" or "difficulty in detecting misplaced goods" in traditional warehousing, and significantly improving the accuracy of medicine location. The strategy of storing similar medicines together not only facilitates picking operations but also makes the warehouse space layout more rational and improves space utilization. The residual heat map technology transforms traditional comprehensive inventory counting into targeted verification, reducing anomaly detection time from several hours to a few minutes, thus greatly improving inventory reconciliation efficiency.

[0132] Please see Figure 2 One embodiment of this application provides a drug storage management device, including an acquisition module 20, a selection module 21, a vector module 22, a matching module 23, a heat map module 24, and an inspection module 25.

[0133] Module 20 is used to obtain information about the purchased medicines based on their unique identifiers.

[0134] Selection module 21 is used to select a sub-drug warehouse that meets the storage conditions based on the purchased drug information;

[0135] Vector module 22 is used to vectorize the purchased drug information to obtain a purchased drug vector;

[0136] Matching module 23 is used to find the shelf center vector that matches the incoming drug vector in the point cloud map corresponding to the sub-drug warehouse, and place the incoming drug on the shelf corresponding to the shelf center vector; the shelf center vector is the center vector of the vector set formed by all drug vectors on the corresponding shelf;

[0137] Heatmap module 24 is used to obtain point cloud maps of the sub-drug warehouses at two times before and after the goods are received or at two times before and after the goods are shipped, and to obtain residual heatmaps through iterative nearest point algorithm.

[0138] The inspection module 25 is used to check whether the corresponding shelf medicines and medicine order information for those with a difference value greater than a preset difference threshold in the residual heat map match.

[0139] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the drug storage management device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0140] Compared to existing technologies, this invention first accurately obtains drug information through unique identifiers and transforms this information into multi-dimensional vector representations, making drug characteristics quantifiable and calculable. Based on the similarity matching between drug vectors and shelf center vectors, it automatically determines the optimal storage location, achieving intelligent placement of similar drugs. It then uses point cloud technology to create a three-dimensional spatial model of the warehouse, comparing point cloud data from different periods using the Iterative Closest Point (ICP) algorithm to accurately calculate spatial changes and generate an intuitive residual heatmap. Finally, based on preset thresholds, the residual heatmap is analyzed to automatically identify areas of abnormal change and intelligently match and verify them with order information.

[0141] In summary, by vectorizing medicines and matching them with shelf-center vectors, we ensure that medicines are intelligently placed according to their characteristics, solving the problems of "not being able to find goods" or "difficulty in detecting misplaced goods" in traditional warehousing, and significantly improving the accuracy of medicine location. The strategy of storing similar medicines together not only facilitates picking operations but also makes the warehouse space layout more rational and improves space utilization. The residual heat map technology transforms traditional comprehensive inventory counting into targeted verification, reducing anomaly detection time from several hours to a few minutes, thus greatly improving inventory reconciliation efficiency.

[0142] One embodiment of this application provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the drug storage management method described above.

[0143] One embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the pharmaceutical storage management method described above.

[0144] The computer device may be a smartphone, tablet, desktop computer, or cloud server, among other computing devices. This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the figures are merely examples of computer devices and do not constitute a limitation on the computer device. It may include more or fewer components than illustrated, or a combination of certain components, or different components, such as input / output devices, network access devices, etc.

[0145] The processor referred to can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0146] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard drive or RAM. In other embodiments, the memory may be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory may include both internal and external storage units of the computer device. The memory is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory can also be used to temporarily store data that has been output or will be output.

[0147] This application provides a computer program product that, when run on a computer device, enables the computer device to execute the steps described in the various method embodiments above.

[0148] In the several embodiments provided in this application, it will be understood that each block in the flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the figures. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved.

[0149] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0150] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for pharmaceutical warehousing management, characterized in that, include: Obtain information about the purchased medicines based on their unique identification codes; The main pharmaceutical warehouse is divided into ambient temperature zone, cool zone, refrigerated zone, variable temperature zone, and return zone. K-means clustering is performed on the drug vectors of all drugs in each warehouse, with the number of cluster centers equal to the number of shelves in the corresponding warehouse. After clustering, the center vector of each cluster is used as the initial shelf center vector, and the drugs corresponding to each cluster are placed on the same shelf. K-means clustering is performed after each receipt or shipment to update the center vectors of each shelf. Select a sub-drug warehouse that meets the storage conditions based on the purchased drug information; The purchased drug information is vectorized to obtain the purchased drug vector; In the point cloud map corresponding to the sub-pharmacy warehouse, the shelf center vector that matches the incoming medicine vector is found, and the incoming medicine is placed on the shelf corresponding to the shelf center vector. Specifically, this includes: in the point cloud map corresponding to the sub-pharmacy warehouse, binding each shelf plane to a shelf center vector; calculating the cosine distance between the incoming medicine vector and each shelf center vector, taking the shelf center vector with the smallest cosine distance value as the matching shelf center vector, and placing the incoming medicine on the shelf corresponding to the shelf center vector; the shelf center vector is the center vector of the vector set formed by all medicine vectors on the corresponding shelf. Obtain point cloud maps of the sub-pharmacy warehouse at two different times before and after receiving goods, or at two different times before and after shipping goods. Obtain a residual heatmap using an iterative nearest-point algorithm. Specifically, this includes: registering the first point cloud map from the previous time step to the second point cloud map at the next time step using the iterative nearest-point algorithm; after registration, calculating the nearest distance from each point in the first point cloud map to the second point cloud map; if the nearest distance of each point within a connected region is greater than or equal to the mean of the magnitudes of all drug vectors within the connected region, the connected region is marked in red; if the nearest distance of each point within a connected region is less than the mean of the magnitudes of all drug vectors within the connected region, and the nearest distance of each point within the connected region is greater than or equal to a distance threshold, the connected region is marked in yellow; the distance threshold is the weighted sum of shelf area, drug volume, and drug quantity; if the nearest distance of each point within a connected region is less than the distance threshold, the connected region is marked in blue. Check whether the shelf medicines and medicine order information corresponding to the difference values ​​greater than the preset difference threshold in the residual heat map match. Specifically, this includes checking whether the names and quantities of medicines in the shelf medicines and medicine order information corresponding to the red connected areas in the residual heat map are consistent. If they are consistent, randomly select several medicine samples to check the production date, expiration date and medicine packaging.

2. The pharmaceutical warehousing management method as described in claim 1, characterized in that, The step of vectorizing the purchased drug information to obtain a purchased drug vector specifically includes: The drug name, production date, expiration date, quantity, volume, and selling price of the purchased drugs are obtained from the purchased drug information, and a multi-dimensional purchased drug vector is generated.

3. A pharmaceutical storage management device, characterized in that, include: The acquisition module is used to obtain information about the purchased medicines based on their unique identifiers. It is also used to divide the total drug warehouse into ambient temperature drug warehouse, cool area drug warehouse, refrigerated drug warehouse, variable temperature drug warehouse, and return area drug warehouse; to perform K-means clustering on the drug vectors of all drugs in each drug warehouse, the number of cluster centers is equal to the number of shelves in the corresponding drug warehouse; after clustering, the center vector of each cluster is used as the initial shelf center vector, and the drugs corresponding to the cluster are placed on the same shelf; K-means clustering is performed after each receipt or each shipment to update the center vectors of each shelf; The selection module is used to select a sub-drug warehouse that meets the storage conditions based on the purchased drug information; The vector module is used to vectorize the purchased drug information to obtain a purchased drug vector; The matching module is used to find the shelf center vector that matches the incoming drug vector in the point cloud map corresponding to the sub-drug warehouse, and place the incoming drug on the shelf corresponding to the shelf center vector. Specifically, it includes: binding each shelf plane to a shelf center vector in the point cloud map corresponding to the sub-drug warehouse; calculating the cosine distance between the incoming drug vector and each shelf center vector, taking the shelf center vector with the smallest cosine distance value as the matching shelf center vector, and placing the incoming drug on the shelf corresponding to the shelf center vector; the shelf center vector is the center vector of the vector set formed by all drug vectors on the corresponding shelf. The heatmap module is used to obtain point cloud maps corresponding to the sub-pharmacy warehouse at two different times before and after receiving goods, or at two different times before and after shipping goods. A residual heatmap is obtained through an iterative nearest-point algorithm. Specifically, this includes: registering the first point cloud map from the previous time step to the second point cloud map at the next time step using the iterative nearest-point algorithm; after registration, calculating the nearest distance from each point in the first point cloud map to the second point cloud map; if the nearest distance of each point within a connected region is greater than or equal to the mean of the magnitudes of all drug vectors within the connected region, the connected region is marked in red; if the nearest distance of each point within a connected region is less than the mean of the magnitudes of all drug vectors within the connected region, and the nearest distance of each point within the connected region is greater than or equal to a distance threshold, the connected region is marked in yellow; the distance threshold is a weighted sum of shelf area, drug volume, and drug quantity; if the nearest distance of each point within a connected region is less than the distance threshold, the connected region is marked in blue. The inspection module is used to check whether the shelf medicines and medicine order information corresponding to the difference values ​​in the residual heat map that are greater than a preset difference threshold match. Specifically, it includes checking whether the names and quantities of medicines in the shelf medicines and medicine order information corresponding to the red connected areas in the residual heat map are consistent. If they are consistent, several medicine samples are randomly selected to check the production date, expiration date and medicine packaging.

4. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the pharmaceutical storage management method as described in any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the pharmaceutical storage management method as described in any one of claims 1 to 2.

Images