Granary warehouse-in and warehouse-out management method and system based on account data matching

Abstract

The invention discloses a granary warehouse-in and warehouse-out management method and system based on account data matching, and the method comprises the steps: obtaining a warehouse-in and warehouse-out type and a target grain variety according to warehouse-in and warehouse-out instruction information; according to the warehouse-in and warehouse-out type and the target grain variety, generating object label information; determining a target storage area by combining a pre-trained account-reality matching model according to the inventory state information and the warehouse-in and warehouse-out instruction information of the current warehouse; matching the real object label information with the accounting information of the target storage area to generate a matching result; triggering a warehouse-in or warehouse-out operation instruction according to a matching result; after warehouse-in and warehouse-out operation is completed, inventory state information is updated; according to the obtained regular inventory checking instruction, triggering an inventory checking process, and obtaining actual inventory checking data; comparing the actual inventory data with the account data to obtain a difference analysis result; and triggering a difference processing instruction according to the difference analysis result. According to the invention, the account and real data matching accuracy and the granary warehouse-in and warehouse-out efficiency in granary warehouse-in and warehouse-out management are improved

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of warehouse in-out management, in particular to a warehouse in-out management method and system based on account-real data matching. BACKGROUND

[0002] Traditional warehouse in-out management relies on manual recording and checking, which is prone to errors and low in efficiency. In recent years, with the development of Internet of Things technology, informationized identification information such as RFID has been applied to warehouse management to improve the automation degree of information collection.

[0003] At present, warehouse in-out management usually adopts manual entry of account data, then reads real object information through a bar code or an RFID tag, and finally manually checks whether the two are consistent; however, manual checking is prone to errors such as batch number transcription errors and variety identification confusion, which leads to disconnection between account data and actual inventory, and in the scenario of large data volume in-out, the efficiency is low, which affects the in-out efficiency.

[0004] Therefore, how to improve the account-real data matching accuracy in warehouse in-out management has practical application value and significance. SUMMARY

[0005] In order to improve the account-real data matching accuracy in warehouse in-out management and improve the warehouse in-out efficiency, the application provides a warehouse in-out management method and system based on account-real data matching.

[0006] In the first aspect, the application achieves the purpose by adopting the following technical scheme: A warehouse in-out management method based on account-real data matching, comprising: obtaining in-out instruction information, and obtaining an in-out type and a target grain variety according to the in-out instruction information; generating real object label information according to the in-out type and the target grain variety, wherein the real object label information comprises variety identification, batch number, origin information and storage time; obtaining inventory state information of a current warehouse, and determining a target storage area by combining a pre-trained account-real matching model according to the inventory state information and the in-out instruction information; matching the real object label information with account information of the target storage area to generate a matching result; triggering an in-out operation instruction according to the matching result; updating the inventory state information and generating operation record information after completing the in-out operation; obtaining a regular inventory instruction, triggering an inventory checking process according to the regular inventory instruction, and obtaining actual inventory checking data; Compare the actual inventory data with the book data to obtain a difference analysis result; According to the difference analysis result, a difference processing instruction is triggered.

[0007] By adopting the above technical solution, in order to improve the consistency of book data and actual data, the present application precisely matches the physical label information containing variety identification, batch number, origin information and storage time with the book information of the target storage area, avoids the mistakes and omissions of manual recording, ensures that the book data and the fields (such as batch, variety) of the physical label completely correspond, and through the combination of the pre-trained book and actual matching model, when determining the target storage area, not only the inventory capacity is considered, but also the origin, variety and batch time in the physical label are comprehensively evaluated, so that the storage area allocation meets the requirements of grain characteristics and actual environment, which is conducive to reducing the mixed storage rate of the same variety and different batches, and improving the grain storage environment compliance rate; according to the matching result, the warehousing or delivery operation instruction is directly generated, which can reduce the manual intervention link, and the automatic generation of operation records is conducive to traceability and responsibility definition, and improves the automation and intelligent level of the grain warehouse operation; further, through the comparison of the actual inventory data and the book data obtained by the regular inventory process, the difference analysis result is generated, which can quickly locate the quantity difference, variety difference or quality difference, and trigger the targeted processing instruction, for example, the quantity difference automatically starts the approval process to correct the ledger, the variety difference automatically pushes the area adjustment task, and the quality difference automatically isolates the abnormal grain, which is conducive to reducing the inventory loss rate; through the real-time matching of the physical label information and the book information, the automatic archiving of the operation records and the process recording of the difference processing, the present application forms a full-link data closed loop from warehousing to delivery, which not only improves the accuracy of book and actual data matching in the grain warehouse operation, but also improves the grain warehouse operation efficiency.

[0008] In a preferred example of the present application, the matching of the physical label information with the book information of the target storage area generates a matching result, which specifically includes: extracting the current storage state and capacity information of the target storage area; checking whether the variety in the physical label information is consistent with the variety allowed to be stored in the target storage area; checking the batch in the physical label information with the storage quantity and position of different batches of the same variety in the target storage area; verifying whether the quality grade in the physical label information meets the quality requirements of the target storage area; comprehensively evaluating the matching result to generate a suggestion of allowing warehousing, rejecting warehousing or adjusting the storage position.

[0009] By adopting the technical scheme, multi-dimensional data verification of variety consistency, batch storage quantity verification, and quality grade matching is adopted, fine management and control of storage area allocation are realized, traditional manual experience judgment is converted into rule-based automatic decision, and problems such as mixed storage of different batches of the same variety and illegal storage of low-quality grain are effectively avoided.

[0010] In a preferred example of the present application: the difference processing instruction is triggered according to the difference analysis result, specifically including: analyzing the difference type of the difference analysis result, the difference type including quantity difference, variety difference and quality difference; generating a difference processing strategy according to the difference type; The difference processing strategy is generated according to the difference type, including: for quantity difference, triggering a surplus or loss processing flow; for variety difference, triggering a variety exchange or storage location adjustment flow; for quality difference, triggering a quality detection or isolated storage flow; record the difference processing process and result, and update the accounting information.

[0011] By adopting the technical scheme, a classification processing framework based on difference type is established, and the inventory problem is quickly responded through differentiated strategies. For quantity difference (such as surplus / loss), the system automatically triggers the financial approval and account correction process, and the processing time efficiency is shortened from 2-3 days of traditional manual inventory to real-time completion; for variety difference, the system recommends the optimal exchange scheme (such as migrating the misplaced batch of corn to the designated area) through intelligent algorithm, and for quality difference, the system triggers the isolation and detection linkage mechanism to facilitate timely discovery and processing of abnormal grain.

[0012] In a preferred example of the present application: the generation step of the pre-trained account-real matching model includes: Collecting grain warehouse physical data and pre-processing the account data of the accounting system; determining a data matching strategy based on the pre-processed physical data and account data; analyze the difference and correlation mode between the physical data and the account data, and determine the influence weight of each data on inventory management; divide the different categories of physical data into first-class basic attribute data and second-class dynamic change data, and establish the correlation between the data to obtain data correlation information; Based on the data correlation information, the physical label information of the first-class basic attribute data and the warehouse position information of the second-class dynamic change data are obtained, and the data matching strategy is input to obtain the inventory state fusion result; obtain the synchronization record time and update frequency between the first type of basic attribute data and the second type of dynamic change data; input the synchronization record time, update frequency and inventory state fusion result into a preset account-real matching model for model optimization to obtain the pre-trained account-real matching model.

[0013] By adopting the above technical solutions, a high-precision inventory state prediction model is constructed by adopting data hierarchical fusion and dynamic optimization. The account-real matching model adopts a dual-track input architecture of static basic attributes and dynamic data. Static attributes such as varieties and batches are separated and processed from dynamic data such as storage positions and environmental parameters. The weight distribution is dynamically adjusted according to parameters such as synchronization record time and update frequency. High-frequency variable data (such as warehouse entry and exit records) are calibrated in real time to improve the management efficiency of warehouse entry and exit.

[0014] In a preferred example of the present application, the data matching strategy is determined based on the preprocessed physical data and account data, including: Based on the preprocessed physical data and account data, the characteristic information of each data is obtained, including data type, source, importance, and update period. The importance level of each data in inventory management is determined. The physical label information of each data is associated with the importance level to generate a data matching strategy suitable for the current scenario.

[0015] By adopting the above technical solutions, the data importance priority of each data is automatically divided through characteristic analysis (such as data type and update period) of physical label data (such as quality grade) and account data (such as storage location). For example, the quality grade of grain close to the shelf life is set as the highest priority, and the matching strategy requires it to be stored in an independent fresh-keeping area. For low-frequency variable batch data, a relaxed matching rule is adopted to improve inventory turnover efficiency.

[0016] In a preferred example of the present application, the data correlation information is obtained, specifically including: The first type of basic attribute data and the second type of dynamic change data are input into a preliminary account-real matching model. According to the characteristic information and importance level of each data, the dependency relationship and consistency between different data are determined. Based on the preset inventory management accuracy rate standard, key data is selected and marked. According to the key data and corresponding characteristic information, synchronization record time and update frequency of the first type of basic attribute data and the second type of dynamic change data, the influence weight of each data on inventory management is calculated. The correlation between each data is established to form detailed data correlation information.

[0017] By adopting the above technical solution, based on the screening and weight calculation of key data, the deep correlation patterns between grain warehouse management data are analyzed and determined. First, key fields are screened based on data characteristics (such as the timeliness of production dates). Then, the impact weight of each data on inventory management is calculated by combining historical difference cases (such as the weight of production location on storage area selection is 0.6, and the weight of batch on batch conflict is 0.8). Finally, a cross-data category correlation network is established to provide an accurate data foundation for the optimization of the account-to-physical matching model.

[0018] In a preferred embodiment of this application, the step of inputting the fusion result of the synchronization recording time, update frequency, and inventory status into a preset account-to-physical matching model for model optimization includes: Adjust the parameter settings in the inventory matching model according to the determined influence weights, optimize the inventory status fusion process, and obtain the optimized inventory status fusion result; The results of the fusion of the synchronization recording time, update frequency, and optimized inventory status are input into the data matching strategy to generate feature mapping information for each type of data. Using the feature mapping information, based on the characteristic information and influence weight of each type of data, and combined with the synchronization recording time and update frequency, the model optimization parameters are calculated; Based on the optimized parameters of the model, the final account-to-actual matching model is obtained.

[0019] By adopting the above technical solution, real-time calibration and adaptive adjustment of model parameters were achieved. By introducing temporal characteristics such as synchronization recording time and update frequency, the system dynamically adjusts the weights of model parameters: assigning high-sensitivity parameters to real-time updated inbound and outbound data, and employing a low-frequency update strategy for historical ledger data. For example, when a sudden increase in the inbound and outbound frequency of a batch of grain is detected, the model automatically increases the calculation frequency for its storage location matching to ensure real-time updates of storage locations; while reducing the computational load for long-term static reserve grain.

[0020] Secondly, the objective of this invention is achieved through the following technical solution: A grain warehouse inbound and outbound management system based on matching account and physical inventory data, the system comprising: The instruction acquisition module is used to acquire inbound / outbound instruction information, and to acquire the inbound / outbound type and target grain variety based on the inbound / outbound instruction information. The label generation module is used to generate physical label information based on the inbound / outbound type and the target grain variety, wherein the physical label information includes variety identifier, batch number, place of origin information and inbound time; The inventory status monitoring module is used to obtain the current inventory status information of the warehouse; A target area determination module is configured to determine a target storage area according to the inventory status information and the warehouse-in / out instruction information in combination with a pre-trained account-physical matching model; An account-physical matching module is configured to match the physical label information with the account information of the target storage area to generate a matching result; An operation control module is configured to trigger a warehouse-in or warehouse-out operation instruction according to the matching result and update the inventory status information after completing the warehouse-in / out operation to generate operation record information; An inventory management module is configured to trigger an inventory checking process in response to a periodic inventory checking instruction, obtain actual inventory checking data, and compare the actual inventory checking data with account data to obtain a difference analysis result; A difference processing module is configured to trigger a difference processing instruction according to the difference analysis result.

[0021] In a third aspect, the application aims to achieve the above technical solutions. A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above-mentioned grain warehouse-in / out management method based on account-physical data matching when executing the computer program.

[0022] In a fourth aspect, the application aims to achieve the above technical solutions. A computer-readable storage medium stores a computer program, and the computer program implements the steps of the above-mentioned grain warehouse-in / out management method based on account-physical data matching when executed by a processor.

[0023] In summary, the present application has at least one of the following beneficial technical effects: 1. By generating physical label information containing variety identification, batch number, origin and warehouse-in time, the identity of the grain product throughout its life cycle is digitized. The physical label information is synchronized in real time with the electronic account of the accounting system, solving the problem of easy errors and omissions in traditional manual recording (such as batch number transcription errors and variety confusion), and ensuring the initial consistency of account-physical data from the source. The present application forms a full-link data closed loop from warehouse-in to warehouse-out, not only improving the accuracy of account-physical data matching in grain warehouse-in / out management, but also improving the efficiency of grain warehouse-in / out. 2. The physical label information is matched and evaluated with the account information of the target storage area in multiple dimensions, including variety consistency, batch matching, quality level verification and capacity adaptation, etc. key factors, so as to realize precise control of warehouse-in operation. The present application can effectively avoid the risk of incorrect warehouse-in and mixed storage, and improve the safety and standardization of grain management. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1is a flowchart of a grain warehouse in-out management method based on account-real data matching in an embodiment of the present application; Figure 2 is a flowchart of a grain warehouse in-out management method based on account-real data matching in an embodiment of the present application before step S3; Figure 3 is a device schematic diagram in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The present application will be further described in detail below with reference to the accompanying drawings.

[0026] In an embodiment, as shown in the figure, Figure 1 The present application discloses a grain warehouse in-out management method based on account-real data matching, which specifically comprises the following steps: S1: Obtain in-out instruction information, and obtain in-out type and target grain variety according to the in-out instruction information.

[0027] In the embodiment, the in-out instruction information refers to an electronic signal or data instruction triggering the grain warehouse to perform in-out operation, including operation type, target grain variety, quantity and other key information, and the data format can be structured text, bar code scanning result or RFID signal. The data is derived from a user terminal (PC / mobile terminal), an automatic device (AGV trolley, conveyor belt sensor) or a third-party system (ERP / MES). The grain variety information includes wheat, japonica rice, corn and other variety information.

[0028] Specifically, step S1 comprises: S11: Receive an in-out request from a user terminal or an automatic sensor.

[0029] In the embodiment, the in-out request is received by an RFID handheld terminal, a bar code scanner or an integrated sensor node. For example, in the in scenario, when a transport vehicle arrives at the entrance of the grain warehouse, the vehicle-mounted RFID tag is automatically recognized by the access control system, triggering the in request. For another example, in the out scenario, the grain warehouse management system receives an out order sent by a downstream processing plant through an ERP system, and parses the grain variety (such as “2024 Heilongjiang japonica rice”) and quantity demand in the order.

[0030] S12: Analyze the in-out request and extract the in-out type, which includes in and out.

[0031] In this embodiment, inbound refers to storing newly purchased or produced food grains in the warehouse; outbound refers to delivering food grains in the warehouse to downstream users or processing links, which is automatically identified by keywords (such as "INBOUND" and "OUTBOUND") in the instruction or operation targets (storage area / picking area); the instruction text of the "inbound and outbound request" is parsed using natural language processing (NLP) technology, or the key fields in the structured data are directly extracted. For example: The instruction "INBOUND-20241015-WHEAT-500T" is parsed as: Inbound type: inbound (INBOUND); Target grain species: wheat (WHEAT); Target quantity: 500 tons.

[0032] S13: Identify target grain species information, which is obtained by scanning physical labels or manual input.

[0033] In this embodiment, the target grain species information refers to the specific category of food grains involved in this time's outbound or inbound operation, which is identified by a combination of variety name + origin + quality grade, such as "Heilongjiang japonica rice - first grade". Manual input is used when there is no label or the label is damaged, and the HMI (human-machine interface) is manually entered.

[0034] S14: Associate the inbound and outbound types and grain species information to generate inbound and outbound instruction information.

[0035] Specifically, users can initiate or accept inbound and outbound instructions through user terminals, such as fixed terminals of PC systems in warehouse management offices and mobile terminals of handheld PDAs.

[0036] Further, Internet of Things devices are also deployed in the grain storage environment to automatically collect inbound and outbound related data, including RFID readers, barcode scanners, weight sensors, and location sensors. The RFID reader identifies the electronic tag attached to the food items; the barcode scanner parses one-dimensional / two-dimensional barcode information; the weight sensor weighs the load of the transport vehicle (such as the ground weight data); and the location sensor is used to track the real-time location of the AGV trolley or pallet.

[0037] S2: Generate physical label information according to the inbound and outbound types and target grain species, which includes variety identification, batch number, origin information, and inbound time.

[0038] In this embodiment, step S2 includes: S21: Generate a unique identification code based on the grain species, origin information, and batch information.

[0039] In this embodiment, a unique identification code is generated using a hash function (such as SHA-256) or a segment splicing rule, the variety name (such as "Japonica rice"), the origin code (such as "Harbin, Heilongjiang" defined in "GB / T 16830-2023"), and the batch number (such as "20240315-001") are generated using a hash function (such as SHA-256) or a segment splicing rule to generate a unique code (such as NH-RICE-20240315-001) and stored in the database of the accounting system.

[0040] S22: Encode the origin information, production date, shelf life information, and quality grade information into the physical label to generate the physical label information.

[0041] In this embodiment, the origin information, production date, shelf life, and quality grade, etc. key data are encoded into machine-readable physical labels such as two-dimensional codes, barcodes, and RFID chips.

[0042] Specifically, the origin information in the encoding field is the country / region code (such as "CN"), the administrative division code (such as "150100" representing Hohhot, Inner Mongolia Autonomous Region), the production date uses UTC timestamp (such as 2024-03-15T08:00:00Z), the shelf life uses the form of "numerical value + unit", such as 12M represents 12 months), and the quality grade is divided according to the national standard code (such as "GB 1351-2008 Level 1"). The two-dimensional code in the encoding format uses QR Code format, supports mixed encoding of Chinese characters and numbers; the RFID chip can be stored in EPC (Electronic Product Code) format, compatible with Gen2 protocol.

[0043] S23: Attach the physical label information in the form of RFID tag, two-dimensional code or bar code to the grain commodity.

[0044] In this embodiment, the encoding information is carried by physical carriers such as RFID tags, two-dimensional codes or barcodes to realize the physical binding of grain commodities and digital information. The attachment forms include direct printing on the surface of the packaging bag, wrapping and fixing by heat shrink film, and pasting on the side wall of the tray or container.

[0045] S24: Real-time synchronization of physical label information and electronic account in the accounting system Specifically, through software interface or middleware technology, the physical label information and the electronic account in the accounting system are kept dynamically consistent. The synchronization methods include active push and passive pull, where active push means that the label is immediately sent to WMS through HTTP API or MQTT protocol after generation; passive pull means that new label data is extracted from the database by ETL tool at specified period (such as every hour).

[0046] S3: Obtain the inventory state information of the current warehouse, and determine the target storage area according to the inventory state information and the warehouse-in and warehouse-out instruction information in combination with a pre-trained account-real matching model.

[0047] In this embodiment, the inventory state information is obtained by real-time collection of warehouse data through a wireless sensor network, including the capacity of each storage area in the granary and the inventory distribution, the inventory distribution being the storage data of each type of grain in the warehouse, including dynamic information such as position, quantity, variety, batch, quality grade, etc. The pre-trained account-real matching model is a pre-trained machine learning model, which can comprehensively consider the inventory state, warehouse-in and warehouse-out instructions and business rules, dynamically recommend or automatically assign the optimal storage area and storage strategy suggestion. The pre-trained account-real matching model outputs the number of the target storage area, such as "A area 3 row 5 column", and the storage strategy suggestion, such as "need to adjust the existing storage layout to accommodate the new batch".

[0048] Specifically, the inventory state is the real-time capacity, environmental parameters (temperature, humidity) and stored grain variety information of each storage area in the granary; the warehouse-in and warehouse-out instructions include the target grain variety, quantity, quality grade and storage time limit; the business rules include centralized storage of the same variety, preferential storage of near-expiration grain, and isolation storage of high-risk varieties, and the specific business rules are customized as needed.

[0049] For example, in the scenario of wheat warehouse-in: The warehouse management system receives the warehouse-in instruction (variety: strong gluten wheat, quantity: 800 tons, quality grade: first class); The account-real matching model queries the inventory state and finds that the remaining capacity of A area is 500 tons and the environmental parameters meet the requirements, and the remaining capacity of B area is 300 tons but the humidity is too high; According to the business rule "first-class wheat needs to be stored in a humidity ≤20% area", the model recommends A area as the main storage area, and the remaining 300 tons is temporarily stored in C area; The warehouse management system generates the instruction: "allocate 800 tons of strong gluten wheat to A area (500 tons) and C area (300 tons)".

[0050] S4: Match the physical label information with the account information of the target storage area to generate a matching result.

[0051] In this embodiment, matching the physical label information with the account information means comparing the variety, batch, quality grade, etc. in the physical label with the electronic account of the target storage area at the field level to verify the storage legality. The matching result includes the matching state and the suggested measures, wherein the matching state includes allowing warehouse-in / out, rejecting, and manual review; the suggested measures include adjusting the storage location, isolation treatment, and supplementing the quality inspection report.

[0052] Specifically, step S4 includes: S41: Extract the current storage state and capacity information of the target storage area.

[0053] In this embodiment, the current storage state and capacity information reflect the real-time physical carrying capacity and storage content of the target storage area, including dynamic data such as remaining capacity, stored varieties, and storage duration, and static data such as area size, weight limit, and environmental parameters (temperature and humidity), to provide spatial and resource constraints for storage decision-making; real-time data can be collected through RFID readers and weight sensors deployed in the warehouse, or static parameters of the storage area (such as a maximum weight of 500 tons and an area of 200 square meters) can be retrieved by the WMS (Warehouse Management System).

[0054] S42: Check whether the variety in the physical label information is consistent with the allowed storage variety in the target storage area.

[0055] In this embodiment, the variety consistency check ensures that the physical label variety matches the allowed storage variety in the target area. The warehouse management system database has a pre-set allowed storage list and conflict detection rules, where the allowed storage list can be divided based on grain characteristics (such as moisture absorption and pest risk), and the conflict detection rules mean that mutually exclusive varieties (such as indica rice and glutinous rice) cannot be stored in the same area.

[0056] S43: Check the batch in the physical label information and the storage quantity and position of different batches of the same variety in the target storage area.

[0057] In this embodiment, the warehouse management system also performs batch storage compatibility checks to assess the rationality of storage allocation for different batches of the same variety. In actual applications, storage strategies and capacity limits can be set, with storage strategies divided by batch number and capacity limits set to a single batch occupying no more than a threshold of the area capacity (e.g., 30%).

[0058] S44: Verify whether the quality grade in the physical label information meets the quality requirements of the target storage area.

[0059] In this embodiment, it is determined whether the physical quality grade meets the storage standards of the target area. In actual applications, the storage environment of the grain warehouse and the storage requirements that can be met need to be further divided, such as grain warehouse A area meeting the constant temperature requirement, which is a constant temperature warehouse; grain warehouse C area is a general warehouse and cannot meet the constant temperature storage requirement; the "grain storage and grain warehouse environment" level-environment mapping rules are set in the warehouse management system, such as first-class grain requiring a constant temperature warehouse and second-class grain being stored in a general warehouse. Real-time verification of quality grade compliance is performed based on environmental sensor data to ensure the storage quality and safety of high-value grain.

[0060] S45: Comprehensive evaluation of the matching result to generate a suggestion of allowing warehouse entry, rejecting warehouse entry, or adjusting the storage location.

[0061] In this embodiment, a storage decision suggestion is generated through multi-dimensional verification results. The application uses a weighted comprehensive algorithm to calculate the matching degree of varieties, batches, and quality grades by weighting. The decision priority is: safety compliance > storage efficiency > cost control, to realize the automation and scientization of storage decision, and improve the quality control level of grain storage.

[0062] S5: According to the matching result, trigger the warehouse-in or warehouse-out operation instruction.

[0063] In this embodiment, according to the matching result generated in step S4 (such as allowing warehouse-in / warehouse-out, adjusting storage location, rejecting operation, etc.), the system automatically generates corresponding device control instructions to drive automatic devices (such as AGV, stacker, conveying line) to perform physical operations, and synchronously updates the accounting system.

[0064] For example, the matching result is "allowing warehouse-in", and the warehouse management system generates a warehouse-in instruction according to the "target storage area" (such as A area 3 row 5 column) and "storage strategy" (such as "preferentially storing in the area close to the warehouse-out port") in the matching result, and dispatches devices such as AGV to call the AGV trolley to carry the goods from the unloading area to the target storage position; start the stacker control, such as the stacker placing the goods to the specified shelf layer according to the instruction; RFID binding refers to automatically scanning the storage position RFID tag and the goods tag after the goods are placed in the storage position, to establish the physical and digital association.

[0065] S6: After completing the warehouse-in or warehouse-out operation, update the inventory state information and generate operation record information.

[0066] In this embodiment, after the warehouse-in or warehouse-out operation is completed, the warehouse management system and the accounting system automatically trigger the inventory data update process, and generate electronic records containing operation details, to ensure that the accounting system and the physical inventory are real-time synchronized. The warehouse management system and the accounting system of the application are associated and connected, and can perform real-time data update.

[0067] For example, the inventory update after the warehouse-in operation is completed is as follows: After the AGV trolley completes the goods carrying, the actual storage position (such as "A area 3 row 5 column"), the storage quantity (such as "50 tons"), and the goods state (such as "perfect and intact") are collected through the on-board sensor; the storage position RFID tag is scanned to confirm that the goods are in place.

[0068] At this time, the warehouse management system calls the WMS interface to perform the inventory increment operation and update the storage area capacity information, and pushes the updated inventory data to the accounting system through the API interface to synchronize the financial accounts.

[0069] S7: Obtain a regular inventory instruction, and trigger an inventory process according to the regular inventory instruction to obtain actual inventory data.

[0070] In this embodiment, the inventory check instruction is automatically generated or received by external input according to preset check rules or external trigger conditions, and the periodic or temporary inventory check process is started. The preset check rules include timing trigger rules and event trigger rules, wherein the timing trigger refers to automatically generating the instruction based on the preset period (such as the first day of each month, the end of each quarter); the event trigger refers to triggering when the inventory fluctuation exceeds the threshold (such as ±10% of the single warehouse inventory), or when the abnormal in-out warehouse alarm occurs, and the trigger check rule based on the environment temperature and humidity can also be set; the external trigger condition refers to manual triggering, that is, the management personnel manually issues the instruction through the management terminal. The instruction content of the periodic check instruction includes the check range (whole warehouse / zone / single product), the check time window (such as the night operation period), and the priority (emergency / regular).

[0071] Specifically, the inventory check process refers to coordinating the hardware devices and software systems to execute the automated or semi-automated inventory data collection and checking process according to the instruction content, and the checking process includes the applied checking devices (such as the RFID scanning array of the fixed reader deployed at the warehouse entrance and exit or the shelf channel, the mobile terminal of the PDA or tablet computer, or the unmanned drone / AGV carrying the RFID reader) and the operation details, such as zone scanning and concurrent processing. The concurrent processing refers to parallel scanning of multiple devices.

[0072] The actual check data refers to the complete physical inventory data set formed by integrating the data collected by the devices and manually entered, and the comparison and analysis with the original data of the accounting system, including unstructured data and structured data.

[0073] S8: Comparing the actual check data with the book data to obtain the difference analysis result.

[0074] In this embodiment, the comparison of the actual check data and the book data includes quantity difference detection, variety difference detection, quality grade difference detection, and storage location difference detection, wherein the quantity difference detection difference value = actual inventory - book inventory, the difference rate (the ratio of the difference value to the book inventory) is obtained, and the difference is determined: if the difference rate exceeds the threshold (such as ±1%), an alarm is triggered; in this example, the difference rate is 0.4%<1%, which is determined to be within the allowable error range.

[0075] The variety difference detection, for example, the physical variety (corn) is inconsistent with the book variety (wheat); a variety mixing storage alarm is triggered. At the same time, if the mixing storage violates the business rules (such as "single variety special storage"), it is determined as a serious difference.

[0076] The quality grade difference detection, for example, actual inventory data, is that the moisture content of the "first-class wheat" sample in the C area of the grain depot is 13.5%; and the account data is that the quality standard of "first-class wheat" in the C area of the grain depot is moisture ≤12.5%, and the quality comparison between the two is: the actual moisture exceeds the standard (13.5%>12.5%); then the quality degradation alarm is triggered. If the quality difference affects the storage safety, it is determined that it needs to be isolated and treated.

[0077] Specifically, the difference type is classified according to the preset difference determination rule. For example, the determination rule includes: when the difference type is quantity difference, the determination rule is actual inventory ≠ account inventory ± allowable error, such as a certain grain product ± 0.5% within the allowable error. When the difference type is variety difference, the determination rule is that the physical variety does not belong to the account allowed storage list at all, such as storing corn in a japonica rice warehouse. When the difference type is quality difference, the determination rule is that the actual quality grade < the account registered grade. When the difference type is location difference, the determination rule is that the physical storage location ≠ the account record location.

[0078] Specifically, the difference analysis result also includes different priority evaluation contents of different levels, wherein the high priority refers to the quality difference and the safety violation, the medium priority refers to the quantity difference and the location deviation, and the low priority refers to the label falling off and the temporary transfer.

[0079] S9: Triggering a difference processing instruction according to the difference analysis result.

[0080] In this embodiment, step S9 includes: S91: analyzing the difference type of the difference analysis result, the difference type including quantity difference, variety difference and quality difference.

[0081] In this embodiment, the processing strategy is automatically matched based on a machine learning algorithm and a preset business rule (such as "variety mixed storage is prohibited", and the specific business rule clauses can be customized as needed), wherein the machine learning algorithm is used to optimize the strategy through historical data, such as high-frequency problems being processed first.

[0082] S92: generating a difference processing strategy according to the difference type.

[0083] According to the difference type, the difference processing strategy is generated, including: S921: for quantity difference, triggering a surplus or loss processing flow.

[0084] In this embodiment, the quantity difference is automatically accounted for, and the adjustment process of the financial and inventory records is triggered.

[0085] For example, when the actual inventory > book value, trigger the process of surplus, otherwise trigger the process of loss. The process of surplus triggers the process of financial audit, generates an "approval sheet for surplus", and after manual approval and verification, updates the inventory record to make the book value = actual value. The process of loss includes starting the material accountability process (such as checking loading and unloading records), generating a restocking work order or insurance claim application.

[0086] S922: For variety differences, trigger the process of variety replacement or storage location adjustment.

[0087] In this embodiment, the variety mixed storage problem is physically isolated or the location is adjusted to ensure storage compliance.

[0088] For example, A area misstores "corn" (should store "wheat"), at this time the AGV can be operated to carry "corn" to the designated storage position in C area; update the WMS record to: change the original storage position (A-03-05) to the new storage position (C-05-03).

[0089] S923: For quality differences, trigger the process of quality detection or isolated storage.

[0090] In this embodiment, the substandard quality grain is isolated or rechecked to prevent inferior products from flowing into the next link.

[0091] Specifically, the isolated storage process, such as the actual moisture content of grain exceeding the standard (such as "14%>12.5%"), operates the AGV to transfer the problem batch to the "inspection area" and triggers the laboratory reinspection process.

[0092] S924: Record the difference processing process and results, and update the accounting information.

[0093] In this embodiment, through the block chain storage technology and database update, it is ensured that the difference processing process is traceable and the accounting data is real-time synchronized.

[0094] In an embodiment, as shown in Figure 2 Before step S3, the generation step of the pre-trained account-inventory matching model includes: S301: Collect grain physical data and pre-process the account data of the accounting system.

[0095] In the embodiment, the grain warehouse physical data is collected by an RFID reader to collect physical tag information. The account data is historical storage records exported from a WMS system, such as storage location, storage time, and warehouse in-out flow. The preprocessing operation includes data cleaning, format unification, and data alignment. The format unification refers to aligning the string format (such as WHEAT-2024XJ001) of the RFID tag with the database field of the account. The data alignment refers to associating the physical data with the account record based on the timestamp or batch number (such as matching the “2024-XJ-001 batch” in the warehouse in-out record on March 15, 2024).

[0096] S302: Determine a data matching strategy based on the preprocessed physical data and the account data.

[0097] In the embodiment, a matching rule for the physical data and the account data is formulated, and a matching logic of a key field is defined. The matching logic includes exact matching and fuzzy matching. The matching fields include mandatory matching fields and optional matching fields. The mandatory matching fields include variety, batch number, and quality grade. The optional matching fields include origin (such as allowing different districts in the same city) and production date (such as allowing a ±3-day error).

[0098] Specifically, step S302 includes: S3021: Based on the preprocessed physical data and the account data, obtain characteristic information of each data, including data type, source, importance, and update period.

[0099] In the embodiment, the data type includes structured data and unstructured data. The structured data is, for example, RFID tag information (variety, batch number, and quality grade), and the unstructured data is, for example, sensor logs (temperature and humidity curve, and device running status). The data source is, for example, physical data from an RFID reader and a two-dimensional code scanning device, and account data from a WMS system and an ERP system (i.e., an accounting system). The importance is classified, for example, a high-priority quality grade, a storage location, and high-priority data that directly affect food safety; a medium-priority variety and batch number, and medium-priority data that affect storage strategy; and a low-priority origin and production date, and low-priority data that serve as auxiliary references. The update period is, for example, a fixed-time update and a real-time update.

[0100] S3022: Analyze the role of each data in inventory management to determine the importance level of each data.

[0101] In this embodiment, the role evaluation dimension of each data includes business impact, economic value and compliance risk. The business impact refers to whether data anomalies cause storage violations (such as mixed storage and over-storage); the economic value refers to the economic loss caused by data errors (such as misjudging high-quality grain as ordinary grain); and the compliance risk refers to the risk level of violating the Grain Circulation Management Regulations. The scoring method adopts a machine learning model to predict the data failure probability (using a scoring method of 1 to 5 points) by training a random forest classifier, and to perform grade division, for example, the score of key data is ≥4, such as quality grade and storage location; the score of important data is 3≤score<4, such as variety and batch number; and the score of auxiliary data is <3, such as place of production and production date.

[0102] S3023: Associate the physical label information of each data with the importance level to generate a data matching strategy suitable for the current scene.

[0103] In this embodiment, the data importance level is bound with the physical label information to form an executable matching rule library.

[0104] Specifically, the rule mapping of the matching rule library includes: key data is required to be matched by 100%, such as directly rejecting if the quality grade is inconsistent; important data is allowed to have certain errors, such as requiring manual audit for mixed storage of the same variety and different batches; and auxiliary data is only for reference (such as minor differences in place of production do not affect storage allocation, and different places in the same city).

[0105] The data matching strategy adopts a weighted matching and dynamic adjustment manner, wherein the weighted matching gives higher matching weight to high-priority data, and the dynamic adjustment refers to adjusting the strategy strictness according to the inventory pressure (such as storage saturation).

[0106] For example, the current needs to handle the "2024-XJ-003 batch" wheat into the warehouse. The warehouse management system detects that the target storage area has stored "2024-XJ-001 batch" wheat (same variety), and generates an adjustment instruction according to the data matching strategy "allowing mixed storage of the same variety and different batches", and if the quality grade is not matched, directly triggers rejection and alarm.

[0107] S303: Analyze the difference and correlation mode between physical data and account data to determine the influence weight of each data on inventory management.

[0108] In this embodiment, the difference degree between physical data and account data is quantified, and the causal relationship between data is mined, such as the influence of storage location on grain loss. The difference analysis includes calculation of statistical indicators and abnormality detection, wherein the statistical indicators include variety consistency rate, batch matching rate and quality grade deviation rate; and the abnormality detection refers to identifying abnormal data points (such as sudden increase of moisture content of a batch).

[0109] Specifically, the analysis of the association mode includes correlation analysis and association rule analysis, wherein the correlation analysis includes analyzing the relationship between the storage temperature and the grain mold rate by a Pearson coefficient, and the association rule is, for example, "if the storage quantity of the same batch > 500 tons, then the mold risk is increased by 20%", and the association rule can be defined according to the importance level of the data and the physical label information.

[0110] The distribution of the influence weight is based on the importance of the business: quality level (weight 0.4), storage location (weight 0.3), and batch consistency (weight 0.3).

[0111] For example, the storage data of "2024-XJ-001 batch" is analyzed, and it is found that the mold rate of the same batch of grain stored in the A area is 2%, and the mold rate of the same batch of grain stored in the B area is 5%; it is determined that the "storage area" and the "mold rate" are strongly correlated (correlation coefficient 0.75); and the matching strategy is adjusted to preferentially allocate to the A area with more stable temperature and humidity.

[0112] S304: The different categories of physical data are divided into first category basic attribute data and second category dynamic change data, and the association relationship between the data is established to obtain data association information.

[0113] In this embodiment, the physical data and the dynamic data are classified, and the logical association between the cross-category data is established, such as the binding relationship between the variety and the warehouse position. The first category basic attribute data includes variety, origin, quality level, and has static attributes; the second category dynamic change data includes warehouse position, inventory quantity, and environmental parameters, and has real-time change attributes.

[0114] Specifically, the data association information is obtained, and specifically includes: S3041: The first category basic attribute data and the second category dynamic change data are input into a preliminary account-physical matching model.

[0115] In this embodiment, the basic attributes are coded as One-Hot Vector, and the dynamic data is normalized to the interval [0, 1].

[0116] For example, the newly entered "2024-XJ-002 batch" corn needs to be matched with a storage area; the basic attributes {variety: corn, quality level: first class} and the dynamic data {warehouse position: B area idle warehouse, environmental parameters: {temperature: 22℃, humidity: 60%}} are input into the model; the model outputs a preliminary matching score: B area matching degree 85%.

[0117] S3042: According to the characteristic information and the importance level of each data, the dependency relationship and the consistency between different data are judged.

[0118] In this embodiment, the logical association and conflict points between different data are determined based on data characteristics and importance level. The dependency relationship includes positive association and reverse association. The positive association refers to the storage location and the temperature and humidity requirement (e.g., constant temperature warehouse stores only one level of grain). The reverse association refers to the batch number and the production date (e.g., batch 2024-XJ-001 corresponds to the production date 2024-03-10).

[0119] Specifically, the consistency check adopts a rule engine and conflict detection. The rule engine is, for example, if the storage amount of the same batch is greater than 500 tons, then the storage needs to be divided into warehouses. The conflict detection is, for example, if the storage location of the physical label information is inconsistent with the storage location of the account record, an alarm is triggered.

[0120] S3043: Based on the preset inventory management accuracy rate standard, the key data is selected and marked.

[0121] In this embodiment, based on the preset inventory management accuracy rate standard (e.g., 98% matching rate), the core data that plays a decisive role in inventory management is screened.

[0122] Specifically, the data importance is scored by a warehouse expert (1-5 points). For example, when the data type is quality grade, the importance score is 5 points; when the data type is quality grade, the importance score is 4 points; when the data type is variety, the importance score is 3 points; and when the data type is origin, the importance score is 2 points. A machine learning model trained based on a random forest model is used to enable the machine learning model to predict the influence probability of data missing on inventory error, and the features with a contribution degree TOP 30% are reserved as key data.

[0123] S3044: According to the key data and the corresponding characteristic information, the synchronization record time and the update frequency of the first type of basic attribute data and the second type of dynamic change data, the influence weight of each data on inventory management is calculated.

[0124] In this embodiment, the contribution degree of each data to inventory management is quantified in combination with the data synchronization frequency, update time and importance level.

[0125] Specifically, the weight calculation formula is as follows Wherein, i is the data variable index; 、 、 is the weight coefficient, which is determined by cross-validation; is the key degree of the ith type of data in inventory management, which is defined by business rules or expert experience; is the update frequency of the data, i.e., the number of times the data is refreshed per unit time. The higher the update frequency (the smaller the denominator), the larger the value, and the higher the weight contribution; The data synchronization time difference (the smaller the weight is higher); Threshold is the synchronization time threshold, which is the maximum allowed synchronization time difference.

[0126] For example, the importance level of the quality level = 5, the update frequency = once a day, and the synchronization time difference = 5 seconds, then the weight = 0.45; the importance level of the storage location = 4, the update frequency = once every 2 hours, and the synchronization time difference = 30 seconds, then the weight = 0.35.

[0127] S3045: Establish the association between the data, and form detailed data association information.

[0128] In this embodiment, the key data is integrated with the weight to construct an association network representing the logical relationship between the data.

[0129] For example, the association rule base for establishing the association relationship is: Rule 1: Quality level = first level, must be stored in constant temperature warehouse; Rule 2: Different batches of the same variety, stored separately and the distance ≥ 5 meters.

[0130] After using Neo4j graph database to represent the entity relationship, a JSON format association rule set is generated to obtain the data association information.

[0131] S305: Based on the data association information, obtain the physical label information of the first type of basic attribute data and the storage location information of the second type of dynamic change data, and input the data matching strategy to obtain the inventory state fusion result.

[0132] In this embodiment, the basic attribute and dynamic data are integrated to generate a fusion feature vector representing the current inventory state.

[0133] Specifically, the data fusion method includes: encoding discrete attributes (variety) into one-hot vector, normalizing continuous attributes (temperature), and then weighted sum (such as fusion value = 0.6 × quality level + 0.4 × temperature), the output result is like inventory state vector [0.8, 0.3, 0.9] (representing quality, storage compliance, and environmental suitability respectively).

[0134] S306: Obtain the synchronization record time and update frequency between the first type of basic attribute data and the second type of dynamic change data.

[0135] In this embodiment, the update time stamp and frequency of the physical data and the account data are recorded to quantify the data timeliness. The physical data update frequency is, for example, polling once every 5 minutes through the sensor; the account data update frequency is, for example, triggering real-time update for warehouse-in and warehouse-out operations. The timeliness evaluation is that if the time difference between the two is ≤ 1 minute, it is determined to be synchronized, otherwise an alarm is triggered (such as "data delay needs manual check").

[0136] For example, the physical label update time adopts the RFID write time (2025-03-15 14:30:00); the account book update time adopts the WMS system record time (2025-03-15 14:31:00).

[0137] S307: input the synchronization record time, the update frequency and the inventory state fusion result into a preset account-physical matching model for model optimization to obtain a pre-trained account-physical matching model.

[0138] In this embodiment, step S307 comprises: S3071: adjust the parameter setting in the account-physical matching model according to the determined influence weight, optimize the inventory state fusion process, and obtain an optimized inventory state fusion result.

[0139] In this embodiment, the model parameters are dynamically adjusted through the influence weight, the multi-source data fusion process is optimized, and the accuracy of the inventory state representation is improved. The initial model is trained using historical data, and the initial parameters are set. The parameter update rule is: the sensitivity of the model to different data is adjusted according to the influence weight (such as quality grade weight 0.4 and storage location weight 0.3).

[0140] Example formula: Wherein, is the adjusted data sensitivity; is the data sensitivity before adjustment; is the learning rate, such as 0.01; is the influence weight of the ith type of data; is the loss function gradient.

[0141] S3072: input the synchronization record time, the update frequency and the optimized inventory state fusion result into the data matching strategy together to generate feature mapping information of each type of data.

[0142] In this embodiment, the synchronization record time, the update frequency and the optimized inventory state fusion result are combined to generate the mapping relationship between the data features and the inventory state.

[0143] Specifically, the feature mapping logic of the static features is: the variety mapping is one-hot encoding, such as mapping to [1, 0, 0,...]; the dynamic features such as temperature are normalized, such as mapping to [0.75].

[0144] For example, a feature mapping table of various data is constructed, for example: the data type is quality grade, and the feature value is level one, then the mapping is [1, 0, 0]. The data type is storage location, and the feature value is A area, then the mapping result is [0.3, 0.7]. The data type is environmental parameter, and the feature value is temperature 18℃, then the mapping result is 0.75.

[0145] For example, the feature map of "2024-XJ-006 batch" corn needs to be generated; in the case of mapping the physical label information to [1, 0, 0] (variety = japonica rice) and mapping the warehouse environment to [0.3, 0.7, 0.75] (A area, storage capacity 300 tons, temperature 18°C), the generated comprehensive feature vector is [1, 0, 0, 0.3, 0.7, 0.75].

[0146] S3073: Using the feature mapping information, based on the characteristic information of each data and its influence weight, the model optimization parameters are calculated based on the synchronous recording time and the update frequency.

[0147] In this embodiment, the types of model optimization parameters include learning rate update according to data update frequency (high-frequency data reduces learning rate) and regularization coefficient update according to data importance level adjustment (key data increases regularization strength).

[0148] Specifically, the update calculation formula of the learning rate is: wherein, is the initial learning rate, which is 0.01 in this embodiment; is the importance level coefficient (for example, the importance level of first-class grain is =0.8).

[0149] Example calculation: the quality level of first-class grain (assuming the importance level is 5, =5 seconds, Threshold=10 seconds), then =0.006.

[0150] The calculation of the regularization coefficient needs to combine the data importance level k and the basic regularization strength ( ), and the specific formula design is: wherein, k is the data importance level, taking values from 1 to 5, 1 is auxiliary data, and 5 is key data; is the basic regularization strength, which is 0.01 by default; is the enhancement factor of the importance level, taking values from 0.3 to 0.7, controlling the rate of k growth.

[0151] S3074: Obtain the final account-real match model according to the model optimization parameters.

[0152] In this embodiment, the cross-validation (such as K-fold=5) is used to test the model generalization ability; the verification indicators are accuracy, recall rate and F1-score.

[0153] Specifically, the model optimization is a parameter adjustment using a gradient descent algorithm to optimize the matching strategy weight, and adjusting the learning rate, such as an initial value of 0.01, decaying by 0.1 per iteration.

[0154] Specifically, the training process of the account-real matching model includes training and optimization based on the input data of the fusion result vector, the synchronization record time, the update frequency, the loss function of the cross-entropy loss function (measuring the deviation of the predicted storage area and the actual area), and the optimization target of minimizing the prediction error (such as the training set accuracy reaching 95%).

[0155] It should be understood that the sequence numbers of the steps in the above embodiments do not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0156] In an embodiment, a grain warehouse in-out management system based on account-real data matching is provided, which corresponds to the grain warehouse in-out management method based on account-real data matching in the above embodiment.

[0157] A grain warehouse in-out management system based on account-real data matching includes an instruction acquisition module, a label generation module, an inventory state monitoring module, a target area determination module, an account-real matching module, an operation control module, an inventory management module, and a difference processing module. The detailed description of each functional module is as follows: The instruction acquisition module is configured to acquire in-out warehouse instruction information, and acquire the in-out warehouse type and target grain species according to the in-out warehouse instruction information; The label generation module is configured to generate physical label information according to the in-out warehouse type and target grain species, wherein the physical label information includes species identification, batch number, origin information, and storage time; The inventory state monitoring module is configured to acquire the inventory state information of the current warehouse; The target area determination module is configured to determine the target storage area according to the inventory state information and the in-out warehouse instruction information, in combination with the pre-trained account-real matching model; The account-real matching module is configured to match the physical label information with the account information of the target storage area, and generate a matching result; The operation control module is configured to trigger the in-out warehouse operation instruction according to the matching result, and update the inventory state information after completing the in-out warehouse operation to generate operation record information; The inventory management module is configured to trigger the inventory checking process in response to the periodic inventory checking instruction, acquire the actual inventory checking data, and compare it with the account data to obtain the difference analysis result; The difference processing module is configured to trigger the difference processing instruction according to the difference analysis result.

[0158] The specific limitation of the grain warehouse in-out management system based on account-real data matching can refer to the limitation of the grain warehouse in-out management method based on account-real data matching, which will not be repeated here. Each module in the grain warehouse in-out management system based on account-real data matching can be realized by software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0159] In one embodiment, a computer device, which can be a server, has an internal structure as shown in Figure 3 The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is used to store physical label information, inventory status information, and operation record information, etc. The network interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement a grain warehouse in-out management method based on account-real data matching.

[0160] In one embodiment, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: S1: Obtain in-out instruction information, and obtain the in-out type and target grain variety according to the in-out instruction information; S2: Generate physical label information according to the in-out type and target grain variety, wherein the physical label information includes variety identification, batch number, origin information, and in-out time; S3: Obtain the inventory status information of the current warehouse, and determine the target storage area according to the inventory status information and the in-out instruction information in combination with the pre-trained account-real matching model; S4: Match the physical label information with the account information of the target storage area to generate a matching result; S5: Trigger the in-out operation instruction according to the matching result; S6: Update the inventory status information and generate operation record information after completing the in-out operation; S7: Obtain a regular inventory instruction, trigger an inventory process according to the regular inventory instruction, and obtain actual inventory data; S8: Compare the actual inventory data with the book data to obtain a difference analysis result; S9: Trigger a difference processing instruction according to the difference analysis result.

[0161] In one embodiment, a computer readable storage medium is provided, and a computer program is stored on the computer readable storage medium. The computer program is executed by a processor to implement the following steps: S1: Obtain warehouse entry and exit instruction information, and obtain a warehouse entry and exit type and a target grain variety according to the warehouse entry and exit instruction information; S2: Generate physical label information according to the warehouse entry and exit type and the target grain variety, wherein the physical label information includes variety identification, batch number, origin information and warehouse entry time; S3: Obtain inventory state information of a current warehouse, and determine a target storage area according to the inventory state information and the warehouse entry and exit instruction information in combination with a pre-trained book and reality matching model; S4: Match the physical label information with bookkeeping information of the target storage area to generate a matching result; S5: Trigger a warehouse entry or exit operation instruction according to the matching result; S6: Update the inventory state information and generate operation record information after completing the warehouse entry and exit operation; S7: Obtain a regular inventory instruction, trigger an inventory checking process according to the regular inventory instruction, and obtain actual inventory data; S8: Compare the actual inventory data with the book data to obtain a difference analysis result; S9: Trigger a difference processing instruction according to the difference analysis result.

[0162] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0163] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of functional units and modules is exemplified. In actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the above-described functions.

[0164] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for grain warehouse entry and exit management based on matching account and physical data, characterized in that, include: Obtain inbound / outbound instruction information, and based on the inbound / outbound instruction information, obtain the inbound / outbound type and target grain variety; Based on the inbound / outbound type and the target grain variety, generate physical label information, wherein the physical label information includes variety identifier, batch number, place of origin information and inbound time; Obtain the current inventory status information of the warehouse, and determine the target storage area based on the inventory status information and the inbound / outbound instruction information, combined with a pre-trained account-to-physical matching model. The physical tag information is matched with the accounting information of the target storage area to generate a matching result; Based on the matching results, trigger an inbound or outbound operation command; After completing the inbound and outbound operations, update the inventory status information and generate operation record information; Obtain a periodic inventory count instruction, and trigger the inventory count process based on the periodic inventory count instruction to obtain the actual inventory count data; Compare the actual inventory data with the book data to obtain the discrepancy analysis results; Based on the difference analysis results, a difference processing instruction is triggered.

2. The method for grain warehouse entry and exit management based on matching account and physical data as described in claim 1, characterized in that, The step of matching the physical tag information with the accounting information in the target storage area to generate a matching result specifically includes: Extract the current storage status and capacity information of the target storage area; Verify that the variety in the physical label information matches the variety allowed to be stored in the target storage area; Check the batch information in the physical label and the storage quantity and location of different batches of the same variety in the target storage area; Verify whether the quality level in the physical label information meets the quality requirements of the target storage area; The system comprehensively evaluates the matching results and generates recommendations to allow entry into the database, reject entry, or adjust the storage location.

3. The method for grain warehouse entry and exit management based on matching account and physical data as described in claim 1, characterized in that, The step of triggering a difference processing instruction based on the difference analysis results specifically includes: The difference analysis results are analyzed to determine the types of differences, including quantitative differences, varietal differences, and quality differences. Generate a difference handling strategy based on the difference type; The step of generating a difference processing strategy based on the difference type includes: For quantity discrepancies, trigger the inventory surplus or shortage handling process; For differences in varieties, trigger the process of variety exchange or adjustment of storage location; For quality discrepancies, trigger quality inspection or isolation storage procedures; Record the discrepancy processing procedure and results, and update accounting information.

4. The method for grain warehouse entry and exit management based on matching account and physical data as described in claim 1, characterized in that, The steps for generating the pre-trained account-to-actual matching model include: Collect physical data of grain warehouses and preprocess it with the ledger data in the accounting system; Determine the data matching strategy based on preprocessed physical data and ledger data; Analyze the differences and correlation patterns between physical data and ledger data to determine the impact weight of each data type on inventory management; Different categories of physical data are divided into the first category of basic attribute data and the second category of dynamic change data, and the relationship between the data is established to obtain data relationship information; Based on data association information, obtain the physical tag information of the first type of basic attribute data and the warehouse location information of the second type of dynamic change data, and input the data matching strategy to obtain the inventory status fusion result; Obtain the synchronization recording time and update frequency between the first type of basic attribute data and the second type of dynamically changing data; The results of the fusion of the synchronization recording time, update frequency and inventory status are input into the preset account-to-physical matching model for model optimization, thereby obtaining the account-to-physical matching model.

5. A method for grain warehouse entry and exit management based on matching account and physical data as described in claim 4, characterized in that, The data matching strategy determined based on preprocessed physical data and ledger data includes: Based on preprocessed physical data and ledger data, characteristic information of each type of data is obtained, including data type, source, importance, and update cycle; Analyze the role of each data point in inventory management and determine the importance level of each data point; Associating the physical label information of each data type with its importance level generates a data matching strategy suitable for the current scenario.

6. A method for grain warehouse entry and exit management based on matching account and physical data according to claim 4, characterized in that, The obtained data association information specifically includes: Input the first type of basic attribute data and the second type of dynamic change data into the preliminary account-to-physical matching model; Based on the characteristics and importance level of each type of data, determine the dependencies and consistency between different data; Based on preset inventory management accuracy standards, select and mark key data; Based on the key data and corresponding characteristic information, the synchronization recording time and update frequency of the first type of basic attribute data and the second type of dynamic change data, calculate the impact weight of each type of data on inventory management. Establish relationships between various data points to generate detailed data association information.

7. A method for grain warehouse entry and exit management based on matching account and physical data as described in claim 6, characterized in that, The step of inputting the fusion result of the synchronization recording time, update frequency, and inventory status into a preset account-to-physical matching model for model optimization includes: Adjust the parameter settings in the inventory matching model according to the determined influence weights, optimize the inventory status fusion process, and obtain the optimized inventory status fusion result; The results of the fusion of the synchronization recording time, update frequency, and optimized inventory status are input into the data matching strategy to generate feature mapping information for each type of data. Using the feature mapping information, based on the characteristic information and influence weight of each type of data, and combined with the synchronization recording time and update frequency, the model optimization parameters are calculated; Based on the optimized parameters of the model, the final account-to-actual matching model is obtained.

8. A grain warehouse inbound and outbound management system based on matching account and physical data, characterized in that, The system includes: The instruction acquisition module is used to acquire inbound / outbound instruction information, and to acquire the inbound / outbound type and target grain variety based on the inbound / outbound instruction information. The label generation module is used to generate physical label information based on the inbound / outbound type and the target grain variety, wherein the physical label information includes variety identifier, batch number, place of origin information and inbound time; The inventory status monitoring module is used to obtain the current inventory status information of the warehouse; The target area determination module is used to determine the target storage area based on the inventory status information and the inbound / outbound instruction information, combined with a pre-trained account-to-physical matching model. The physical item matching module is used to match the physical item tag information with the accounting information of the target storage area to generate a matching result; The operation control module is used to trigger inbound or outbound operation instructions based on the matching results, and update the inventory status information and generate operation record information after the inbound or outbound operation is completed. The inventory management module is used to respond to periodic inventory count instructions, trigger the inventory count process, obtain actual inventory count data, and compare it with the book data to obtain the discrepancy analysis results. The difference processing module is used to trigger difference processing instructions based on the difference analysis results.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the grain warehouse entry and exit management method based on matching account and physical data as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the grain warehouse entry and exit management method based on the matching of account and physical data as described in any one of claims 1 to 7.

Images