Intelligent management cabinet combining dual-drive algorithm and OCR (Optical Character Recognition) and control method

Abstract

The invention belongs to the field of storage cabinet management, and particularly relates to an intelligent management cabinet combining a dual-drive algorithm and OCR recognition and a control method, the intelligent management cabinet comprises an electronic display screen, a screen cabinet key cabinet, a tool storage cabinet and an information management and control system; the information management and control system integrates an information storage subsystem and a management cabinet control subsystem, the information storage subsystem is used for automatically filing and storing label information of screen cabinet keys and tools, and the management cabinet control subsystem realizes intelligent control of a management cabinet through a work ticket generation module, an identification control module and an alarm module; the work ticket generation module is used for generating a work ticket containing tool information; the recognition control module recognizes the work ticket by using an OCR technology and controls the access of the tools through a built-in algorithm; the alarm module monitors the tool returning condition in real time; the tool management efficiency and safety can be effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of storage cabinet management, and in particular relates to an intelligent management cabinet and a control method combining a dual-drive algorithm and OCR recognition. Background Art

[0002] The computer room of Yandu Central Control Station consists of key parts such as the main computer room, battery room, UPS power room and communication room. There are nearly 60 cabinets in the computer room, equipped with line patrol meters, screwdrivers, network cable pliers, lighting pens, emergency lights, cable ties, diagonal pliers and many other tools to meet the needs of daily management and maintenance. However, due to the large number of cabinet keys and tools and their frequent use, the staff spends a lot of time looking for the required items, affecting work efficiency. In addition, multiple cabinet keys and multiple tools may be required for work. If the corresponding items cannot be found in time, the work progress will be seriously affected. In addition, the staff may mistakenly enter the non-working area due to taking the wrong cabinet key during work, which not only affects the safety of the system and equipment, but also may bring network security risks. After the work is completed, the cabinet keys and tools lack a unified storage location, and random placement may cause the loss of items.

[0003] For example, the patent with authorization announcement number CN104778574B discloses an intelligent management system and an intelligent management cabinet thereof, wherein the intelligent management system includes: an intelligent control unit, a real-time communication unit, a data acquisition unit and a query and analysis unit; the intelligent management cabinet includes a cabinet body, a plurality of intelligent drawers are arranged in the cabinet body, and the cabinet body is also provided with an intelligent identity recognition device for identifying the identity of the user and an intelligent electric lock with an automatic pop-up mechanism; the intelligent identity recognition device is a fingerprint recognition device or an ID card recognition device or a user name and password recognition device.

[0004] The above prior art has the following problems: the prior art still has many loopholes in the access and management of machine room repair tools, which makes the storage of tools chaotic and affects work efficiency; in order to solve this problem, the present invention provides an intelligent management cabinet and control method combining dual-drive algorithm and OCR recognition. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes an intelligent management cabinet combining a dual-drive algorithm and OCR recognition, wherein the intelligent management cabinet includes an electronic display screen, a screen cabinet key cabinet, a tool storage cabinet and an information management and control system; the information management and control system integrates an information storage subsystem and a management cabinet control subsystem, the information storage subsystem is used to automatically file and save label information of screen cabinet keys and tools, and the management cabinet control subsystem realizes intelligent control of the management cabinet through a work ticket generation module, an identification control module and an alarm module; the work ticket generation module is used to generate a work ticket containing tool information; the identification control module utilizes OCR technology to identify the work ticket, and controls the access of tools through a built-in algorithm; the alarm module monitors the return of tools in real time; the present invention can effectively improve the efficiency and safety of tool management.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent management cabinet combining a dual-drive algorithm and OCR recognition, comprising: a management cabinet body, the management cabinet body comprising a tool storage cabinet and an information management and control system;

[0008] Tool storage cabinets are used for tool storage and monitoring;

[0009] The information management and control system includes an information storage subsystem and a management cabinet control subsystem; the information storage subsystem includes an information repository and an information update module; the information repository includes a tool information sub-repository; the tool information sub-repository automatically files and saves the label information of tools through natural language algorithms and graph databases, and updates the label information corresponding to the tools newly stored in the tool storage cabinet to the tool information sub-repository through the information update module, and deletes the label information corresponding to the obsolete tools from the information repository;

[0010] The management cabinet control subsystem includes a work ticket generation module, an identification control module and an alarm module; the work ticket generation module includes an intelligent matching unit, a work ticket generation unit and a work ticket printing unit;

[0011] The intelligent matching unit is used to match the tool information sub-library through the configured matching model according to the real-time input user demands and maintenance parameters, and obtain the tool location information and label information corresponding to the user demands; the generation and printing unit is used to generate the maintenance process corresponding to the user demands and the safety measures list of the maintenance process through the text generation algorithm according to the acquired tool attribute information and user demands and repair parameters; the work ticket printing unit is used to generate the work ticket corresponding to the user demands through the preset work ticket template according to the acquired tool location information, repair process and the safety measures list of the repair process, and print out the work ticket entity;

[0012] The identification and control module is used to identify and control the storage and retrieval process of tools based on the obtained work ticket and the built-in deep learning algorithm; the alarm module is used to monitor and alarm the storage process of used tools in real time.

[0013] Specifically, the recognition control module includes an information recognition unit and a control instruction unit; the information recognition unit is used to perform text recognition on the work ticket entity through a device with a built-in OCR text recognition model to obtain the work ticket text information; the control instruction unit is used to generate the switch instruction of the user's tool storage cabinet through the PID algorithm according to the location information and tool attribute information of the tool in the acquired work ticket text information, and mark the storage status of the tool corresponding to the user's request stored in the tool information sub-library as "in use";

[0014] The alarm module includes an RFID identification unit and an alarm unit; the RFID identification unit identifies and verifies the high-density RFID tags on the stored tools after use through the RFID equipment configured in the corresponding tool storage cabinet. If the corresponding stored tools belong to the tool storage cabinet where the tools are currently stored, the tool storage status corresponding to the user's request stored in the tool information sub-library is converted for the second time. If it does not belong, the corresponding identification result is transmitted to the alarm unit; the alarm unit is used to give real-time alarm reminders on the tool storage process through the configured voice prompt algorithm based on the obtained identification results.

[0015] Specifically, the label information of the tool includes the identification number of the tool storage cabinet, the physical attribute information and the functional attribute information of the tool; the steps of automatically filing and saving the label information of the tool by the tool information sub-library include:

[0016] A1. Obtain label information corresponding to tools and instruments, annotate the obtained label information using the ternary method, and obtain a tool and instrument information data set;

[0017] A2. Use the pre-trained Bert model to build an information extraction model, input the acquired information into the information extraction model, and obtain the label information keyword triple set corresponding to each tool;

[0018] A3. Extract tool function attribute keywords from the tool corresponding label information keyword triple set, and use the function attribute keywords as classification labels to perform secondary labeling on each tool to obtain a tool classification data set;

[0019] A4. Construct a tool classification model based on a support vector machine, input the tool classification data set into the tool classification model for training, obtain a trained tool classification model, and output the tool classification result;

[0020] A5. Use the tool classification result as the first-level node of the tool information sub-library, the tool name attribute as the second-level node, the tool storage cabinet identification number corresponding to the tool as the third-level node, the physical attribute information and functional attribute information as the fourth-level node, construct the tool storage status as the fifth-level node, and set the initial storage status to "unused";

[0021] A6. The hierarchical nodes constructed based on A5 save the tag information keyword triple set corresponding to each tool under the corresponding node, and construct a tool information sub-library with hierarchical nodes.

[0022] Specifically, the step of automatically filing and saving the label information of the tool in the tool information sub-library further includes:

[0023] A7. The trained tool classification model is built into the tool information sub-library. When a new tool is stored in the tool storage cabinet, the tool classification model is used to update the label information corresponding to the new tool to the corresponding node.

[0024] A8. Obtain tool service life information and tool damage image information, annotate the tool damage image information with damage type and damage score, and obtain a tool damage detection data set;

[0025] A9. Use the random forest algorithm to build a damage detection model, input the tool damage detection data set into the damage detection model for training, and obtain a trained damage detection model;

[0026] A10. Set the elimination score threshold, embed the damage detection model into the tool information sub-library, and obtain the damage type and damage score of the corresponding tool through the built-in damage detection model based on the tool service life information and tool damage image information obtained in real time;

[0027] A11. If the damage score of the corresponding tool obtained is greater than the elimination score threshold, an alarm will be issued to the corresponding tool, and the corresponding tool label information will be deleted from the tool information sub-library.

[0028] Specifically, the steps of generating a maintenance process corresponding to the user's demands and a list of safety measures for the maintenance process include:

[0029] B1. Obtain user demand information, and extract the corresponding maintenance parameter keyword information in the user demand information through the information extraction model built into the tool information sub-library;

[0030] B2. Use the SENet algorithm to build a matching model, embed the matching model into the tool information sub-library interface, match the maintenance parameter keyword information obtained with the tool label information stored in the tool information sub-library through the matching model embedded in the interface, and obtain the tool matching score corresponding to the user's request;

[0031] B3. Set a matching score threshold. If the tool matching score is greater than the matching score threshold, obtain the corresponding tool location information and label information. If the tool matching score is less than or equal to the matching score threshold, issue an alarm prompt that there is no corresponding tool.

[0032] B4. Input the acquired tool label information and the maintenance parameter keyword information corresponding to the user's request into the text generation algorithm to obtain the maintenance process corresponding to the user's request and the safety measures list of the maintenance process;

[0033] B5. Input the maintenance process corresponding to the user's demands and the corresponding safety measures list of the maintenance process into the preset maintenance process template to obtain the electronic version of the maintenance process and the corresponding safety measures list.

[0034] Specifically, the step of generating a switch instruction for the tool storage cabinet requested by the user includes:

[0035] C1. Obtain historical work ticket image data and corresponding text data, and preprocess the acquired data;

[0036] C2. Build a lightweight OCR text recognition model, input the preprocessed work ticket image data and the corresponding text data into the OCR text recognition model for training, and obtain a trained OCR text recognition model;

[0037] C3. The trained OCR text recognition model is built into the text recognition device configured in the management cabinet, and the text recognition device is used to extract text from the real-time printed work ticket entity to obtain the text information in the work ticket entity;

[0038] C4. Use the information extraction model to extract the identification number and location information of the corresponding tool storage cabinet from the text information in the work ticket entity, input the identification number and location information of the tool storage cabinet into the PID algorithm, and generate the corresponding tool storage cabinet switch control instruction;

[0039] C5. By generating the corresponding tool storage cabinet switch control instruction, the corresponding storage cabinet on the management cabinet is controlled to pop open. When the corresponding tool storage cabinet pops open, the storage status of the corresponding tool stored in the tool information sub-library is converted to "in use".

[0040] Specifically, the steps of real-time monitoring and alarming of the storage process of the used tools include:

[0041] D1. After the corresponding tools are used, use the work ticket to open the corresponding tool storage cabinet through C3-C5, put the tools into the corresponding tool storage cabinet, and align the high-density RFID tag on the tools with the configured RFID device;

[0042] D2. Identify the high-density RFID tags placed on the tools through the configured RFID device, obtain the corresponding tag information, and compare the obtained tag information with the tool tag information stored in the tool information sub-library, and obtain the comparison result;

[0043] D3. According to the comparison result, if the tools put in belong to the tools storage cabinet at the current moment, the configured voice prompt algorithm is used to prompt that the tools are stored correctly, and the tool storage cabinet door can be closed normally;

[0044] D4. According to the comparison result, if the tools placed in the cabinet do not belong to the tools stored in the cabinet at the current moment, the configured voice prompt algorithm will prompt that the tools are stored incorrectly, and the corresponding tool storage cabinet door will be prohibited from closing. If the corresponding tool storage cabinet door is closed by mistake, the tool storage cabinet door where the tools are stored incorrectly will automatically pop open.

[0045] A management cabinet control method combining dual-drive algorithm and OCR recognition, the steps comprising:

[0046] S1. Obtain tool attribute information and the location information of the corresponding storage cabinet, construct a tool information sub-library through natural language algorithm and graph database, and update the tool information sub-library in real time through the feedback model;

[0047] S2. Obtain user demands and maintenance parameters, and match them with the tool information sub-library through the configured matching model to obtain the tool location information and label information corresponding to the user demands;

[0048] S3. Generate the corresponding repair process and the safety measures list of the corresponding repair process through the text generation algorithm according to the tool attribute information, user demands and maintenance parameters, and use the preset work ticket template to generate the work ticket entity corresponding to the user demands;

[0049] S4. Based on the acquired work ticket entity, the text information of the work ticket is obtained by performing text recognition on the work ticket entity through the OCR text recognition model, and the tool position information and tool attribute information in the work ticket text information are used to generate the switch command of the tool storage cabinet requested by the user through the PID algorithm, and at the same time, the storage status of the corresponding tool is converted;

[0050] S5. When the tools are put back into the corresponding tool storage cabinet after use, the high-density RFID tags on the tools are identified and verified through the built-in RFID device in the tool storage cabinet. If they are, the storage status of the tools is converted again. If they are not, an alarm is given through the configured voice prompt algorithm.

[0051] A computer-readable storage medium, characterized in that computer instructions are stored thereon, and when the computer instructions are executed, a management cabinet control method combining a dual-drive algorithm and OCR recognition is executed.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] In view of the deficiencies in the prior art, the present invention realizes the automatic management and real-time updating of tool information through an integrated information management and control system, utilizing natural language processing technology and graph database, thereby ensuring the accuracy and timeliness of the information; the intelligent matching unit can quickly match suitable tools according to the actual needs and maintenance parameters of the user, and generate a work ticket containing maintenance procedures and safety measures through a work ticket generation and printing unit, thereby greatly simplifying the workflow and improving operational efficiency; the recognition control module reads the work ticket information through OCR technology, and controls the switch of the tool storage cabinet in conjunction with the PID algorithm to ensure that the tools are correctly accessed; in addition, the alarm module monitors the return of tools in real time through RFID technology, and immediately issues an alarm through a voice prompt algorithm once an abnormality is found, thereby effectively preventing the tools from being misplaced or lost, thereby ensuring the safety of operation; in summary, the present invention not only significantly improves the level of intelligence in tool management, but also greatly improves work efficiency and reduces potential safety hazards. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a structural diagram of an intelligent management cabinet combining a dual-drive algorithm and OCR recognition according to Embodiment 1 of the present invention;

[0055] Figure 2 This is a flow chart of a Bert model of an intelligent management cabinet combining a dual-drive algorithm and OCR recognition in Example 1 of the present invention;

[0056] Figure 3 This is a flow chart of a management cabinet control method combining a dual-drive algorithm and OCR recognition according to Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0057] The specific implementation methods of an intelligent management cabinet and a control method combining a dual-drive algorithm and OCR recognition of the present invention are described in detail below with reference to the accompanying drawings.

[0058] Example 1

[0059] See also Figure 1 , an embodiment provided by the present invention: an intelligent management cabinet combining dual-drive algorithm and OCR recognition, comprising: a management cabinet body 001, the management cabinet body comprising an electronic display screen 002, a screen cabinet and key cabinet 005, a tool storage cabinet 004 and an information management and control system; wherein 003 represents a text recognition device;

[0060] Electronic display screen for visual display of management cabinet workflow;

[0061] The tool storage cabinet is used for tool storage and monitoring; the management cabinet body includes 60 tool storage cabinets; each tool storage cabinet is marked with a unique identification number; tools include line patrol meters, screwdrivers, network cable pliers, lighting pens, emergency lights, cable ties and diagonal pliers, etc.

[0062] The screen cabinet key cabinet is used to store and monitor the screen cabinet keys; the screen cabinet key cabinet and tool storage cabinet are equipped with RFID devices to monitor and identify the removal and storage of screen cabinet keys and tools and to issue abnormal alarms;

[0063] The corresponding cabinet keys and tools stored in each cabinet key cabinet and tool storage cabinet are marked with high-density RFID tags;

[0064] The high-density RFID tag corresponding to the cabinet key includes the cabinet key number, the matching tool storage cabinet identification number and the RFID tag information of each tool stored in the tool storage cabinet;

[0065] The high-density RFID tags corresponding to tools include the identification number of the tool storage cabinet, the physical attribute information and functional attribute information of the tools; the physical attribute information of the tools includes size, weight, material, shape and manufacturing date / batch number; the functional attribute information includes maximum bearing capacity, voltage level, insulation performance, scope of use, instructions for use and maintenance requirements; the maximum bearing capacity indicates the maximum torque or maximum shear force of the tool; the voltage level indicates the safe voltage use range of the tool; the insulation performance indicates the insulation level of the tool; the scope of use indicates which type of wire or screw specification the corresponding tool is suitable for;

[0066] See also Figure 2 The information management and control system includes an information storage subsystem and a management cabinet control subsystem; the information storage subsystem includes an information repository and an information update module; the information repository is used to automatically file and save the label information of the cabinet keys and tools stored in the management cabinet; the information repository includes a cabinet key information sub-library and a tool information sub-library;

[0067] The cabinet key information sub-library is used to automatically file and save the label information of the cabinet key;

[0068] The tool information sub-library is used to automatically file and save the label information of tools through natural language algorithms and graph databases; each cabinet key information in the cabinet key information sub-library is connected with the tool information stored in the corresponding tool storage cabinet through the paired tool storage cabinet identification number; further, the construction process of the cabinet key information sub-library and the tool information sub-library in this embodiment is the same;

[0069] The information update module is used to update the label information corresponding to the tools newly stored in the tool storage cabinet to the tool information sub-library, and delete the label information of the obsolete tools from the tool information sub-library;

[0070] Furthermore, in this embodiment, the step of automatically filing and saving the label information of the tool in the tool information sub-library includes:

[0071] A1. Obtain label information corresponding to tools and instruments, annotate the obtained label information using the ternary method, and obtain a tool and instrument information data set;

[0072] A2. Use the pre-trained Bert model to build an information extraction model, input the acquired information into the information extraction model, and obtain the label information keyword triple set corresponding to each tool;

[0073] A3, extracting tool function attribute keywords from the tool corresponding label information keyword triple set, and using the function attribute keywords as classification labels to perform secondary labeling on each tool to obtain a tool classification data set; further, in this embodiment,

[0074] A4. Construct a tool classification model based on a support vector machine, input the tool classification data set into the tool classification model for training, obtain a trained tool classification model, and output the tool classification result;

[0075] A5. Use the tool classification result as the first-level node of the tool information sub-library, the tool name attribute as the second-level node, the tool storage cabinet identification number corresponding to the tool as the third-level node, the physical attribute information and functional attribute information as the fourth-level node, construct the tool storage status as the fifth-level node, and set the initial storage status to "unused";

[0076] A6. The hierarchical nodes constructed based on A5 save the tag information keyword triple set corresponding to each tool under the corresponding node, and construct a tool information sub-library with hierarchical nodes.

[0077] A7. The trained tool classification model is built into the tool information sub-library. When a new tool is stored in the tool storage cabinet, the tool classification model is used to update the label information corresponding to the new tool to the corresponding node.

[0078] A8. Obtain tool service life information and tool damage image information, annotate the tool damage image information with damage type and damage score, and obtain a tool damage detection data set;

[0079] A9. Use the random forest algorithm to build a damage detection model, input the tool damage detection data set into the damage detection model for training, and obtain a trained damage detection model;

[0080] A10. Set the elimination score threshold, embed the damage detection model into the tool information sub-library, and obtain the damage type and damage score of the corresponding tool through the built-in damage detection model based on the tool service life information and tool damage image information obtained in real time;

[0081] A11. If the damage score of the corresponding tool is greater than the elimination score threshold, an alarm is issued to the corresponding tool, and the corresponding tool label information is deleted from the tool information sub-library;

[0082] This process manages cabinet keys and tools by labeling them, and uses advanced natural language processing technology and graph database technology. The system can automatically extract key information and establish a structured database. In particular, for the management of tool information, by obtaining label information and using the pre-trained BERT model for information extraction, it can not only accurately identify the functional attributes of tools, but also build a detailed classification system. This process not only facilitates the rapid search and location of tools, but also makes the information organization more orderly and intuitive by setting up a multi-level node structure. In addition, the system also has the ability to dynamically update, which can promptly reflect the information of new tools and the elimination of old tools, ensuring the timeliness and accuracy of the data. Through the built-in damage detection model, the system can also monitor the status of tools to achieve timely detection and processing of damaged tools, thereby effectively avoiding potential safety hazards and ensuring operational safety.

[0083] The management cabinet control subsystem includes a work ticket generation module, an identification control module, an alarm module and a remote management module;

[0084] The work ticket generation module is used to generate and print the work ticket content; the work ticket generation module includes an intelligent matching unit, a work ticket generation unit and a work ticket printing unit;

[0085] The intelligent matching unit is used to match the tool information sub-library with the user's demands and maintenance parameters input in real time through the configured matching model to obtain the tool location information and label information corresponding to the user's demands;

[0086] Generate a printing unit, and generate a corresponding maintenance process and a list of safety measures for the corresponding maintenance process through a text generation algorithm according to the acquired tool attribute information, user demands and repair parameters; further, the text generation algorithm in this embodiment includes an n-gram model, T5 and BERT series;

[0087] A work ticket printing unit is used to generate a work ticket corresponding to the user's request and print out a work ticket entity according to the acquired tool location information, repair process and safety measures list of the corresponding repair process through a preset work ticket template;

[0088] Furthermore, in this embodiment, the steps of generating a maintenance process corresponding to the user's request and a safety measures list of the maintenance process include:

[0089] B1. Obtain user demand information, and extract corresponding maintenance parameter keyword information in the user demand information through the information extraction model built into the tool information sub-library; further, in this embodiment, the user demand information includes the corresponding equipment parameter information that needs to be repaired in the machine room and the corresponding physical and functional parameter information of the required tools;

[0090] B2. Use the SENet algorithm to build a matching model, embed the matching model into the tool information sub-library interface, match the maintenance parameter keyword information obtained with the tool label information stored in the tool information sub-library through the matching model embedded in the interface, and obtain the tool matching score corresponding to the user's request;

[0091] B3. Set a matching score threshold. If the tool matching score is greater than the matching score threshold, obtain the corresponding tool location information and label information. If the tool matching score is less than or equal to the matching score threshold, issue an alarm prompt that there is no corresponding tool.

[0092] B4. Input the acquired tool label information and the maintenance parameter keyword information corresponding to the user's request into the text generation algorithm to obtain the maintenance process corresponding to the user's request and the safety measures list of the maintenance process;

[0093] B5. Input the maintenance process corresponding to the acquired user demands and the corresponding safety measures list of the maintenance process into the preset maintenance process template to obtain the electronic version of the maintenance process and the corresponding safety measures list;

[0094] By accurately matching user demands with the data in the tool information sub-library, the system can quickly determine the appropriate tools and their locations, greatly shortening the preparation time. The matching model constructed using the SENet algorithm, combined with the set matching score threshold, ensures the accuracy and reliability of the matching results. In addition, with the help of advanced text generation algorithms such as the n-gram model, T5 and BERT, the system can automatically generate clear and concise maintenance processes and safety measures lists, which not only reduces the possibility of human error, but also ensures the safety of operators.

[0095] The recognition control module is used to identify and control the tool access process according to the acquired work ticket and the built-in deep learning algorithm; the recognition control module includes an information recognition unit and a control instruction unit; the information recognition unit is used to perform text recognition on the work ticket entity through a device with a built-in OCR text recognition model to obtain the work ticket text information; the control instruction unit is used to generate the user's demand tool storage cabinet opening and closing instructions through the PID algorithm according to the tool location information and tool attribute information in the acquired work ticket text information, and mark the tool storage status corresponding to the user's demand stored in the tool information sub-library as "in use";

[0096] Furthermore, in this embodiment, the step of generating a switch instruction for the tool storage cabinet requested by the user includes:

[0097] C1. Obtain historical work ticket image data and corresponding text data, and preprocess the acquired data;

[0098] C2. Construct a lightweight OCR text recognition model, input the preprocessed work ticket image data and the corresponding text data into the OCR text recognition model for training, and obtain a trained OCR text recognition model; further, the steps of constructing the lightweight OCR text recognition model in this embodiment include:

[0099] C21, use FPN structure and MobileNetV3 to build the backbone layer of the OCR text recognition model, use the pre-trained completion Bert model to build the text processing layer, use PACT to build the knowledge distillation layer of the OCR text recognition model, and use the fully connected network to build the output layer of the OCR text recognition model;

[0100] C22, input the acquired work ticket image data into the backbone layer to obtain the image segmentation probability containing text and the corresponding segmentation threshold;

[0101] C23, input the image segmentation probability and the corresponding segmentation threshold into the DB function to obtain the image binary vector containing the text;

[0102] C24, input the text data into the text processing layer to obtain the corresponding text vector;

[0103] C25, concatenate the obtained text vector and binary vector and input them into the knowledge distillation layer for lightweight processing, and input the lightweight processed vector into the output layer to obtain the text recognition result;

[0104] C26, using the loss corresponding to the image segmentation probability obtained in the C22 process and the loss corresponding to the image binary vector to construct the total loss function of the model, train the OCR text recognition model, and obtain the trained OCR text recognition model;

[0105] C3. The trained OCR text recognition model is built into the text recognition device configured in the management cabinet, and the text recognition device is used to extract text from the real-time printed work ticket entity to obtain the text information in the work ticket entity;

[0106] C4. Use the information extraction model to extract the identification number and location information of the corresponding tool storage cabinet from the text information in the work ticket entity, input the identification number and location information of the tool storage cabinet into the PID algorithm, and generate the corresponding tool storage cabinet switch control instruction;

[0107] C5. By generating the corresponding tool storage cabinet switch control instruction, the corresponding storage cabinet on the management cabinet is controlled to pop open. When the corresponding tool storage cabinet pops open, the storage status of the corresponding tool stored in the tool information sub-library is converted to "in use".

[0108] First, the lightweight OCR model's efficient text recognition capability can quickly and accurately extract key information from the work ticket, ensuring the accuracy of tool access instructions. At the same time, the lightweight OCR model further improves the smoothness of system operation. Secondly, the PID algorithm dynamically adjusts the switch instructions to ensure the precise control of the tool storage cabinet and effectively prevent the occurrence of misoperation. In addition, the automatic update of the tool status to "in use" helps to monitor the use of tools in real time and avoid waste of resources. The whole process not only greatly simplifies the operation process, but also improves work efficiency.

[0109] An alarm module is used to monitor and warn the storage process of the completed tools in real time; the alarm module includes an RFID identification unit and an alarm unit;

[0110] The RFID identification unit identifies and verifies the high-density RFID tags on the tools stored after use through the RFID equipment configured in the corresponding tool storage cabinet. If the corresponding stored tools belong to the tool storage cabinet stored at the current moment, the tool storage status corresponding to the user's request stored in the tool information sub-library is secondary converted. If not, the corresponding identification result is transmitted to the alarm unit; the alarm unit is used to give real-time alarm reminders on the tool storage process through the configured voice prompt algorithm according to the obtained identification results; the secondary conversion is specifically to convert the tool storage status from "in use" to "unused";

[0111] Furthermore, in this embodiment, the steps of real-time monitoring and alarming the storage process of the used tools include:

[0112] D1. After the corresponding tools are used, use the work ticket to open the corresponding tool storage cabinet through C3-C5, put the tools into the corresponding tool storage cabinet, and align the high-density RFID tag on the tools with the configured RFID device;

[0113] D2. Identify the high-density RFID tags placed on the tools through the configured RFID device, obtain the corresponding tag information, and compare the obtained tag information with the tool tag information stored in the tool information sub-library, and obtain the comparison result;

[0114] D3. According to the comparison result, if the tools put in belong to the tools storage cabinet at the current moment, the configured voice prompt algorithm is used to prompt that the tools are stored correctly, and the tool storage cabinet door can be closed normally;

[0115] D4. According to the comparison result, if the tools placed in the cabinet do not belong to the tools stored in the cabinet at the current moment, the configured voice prompt algorithm will prompt that the tools are stored incorrectly, and the corresponding tool storage cabinet door will be prohibited from closing. If the corresponding tool storage cabinet door is closed by mistake, the tool storage cabinet door where the tools are stored incorrectly will automatically pop open.

[0116] This process integrates RFID technology and voice prompt function to achieve effective monitoring and instant feedback on the tool return process; the RFID recognition unit can accurately identify and verify the returned tools, ensuring that they are correctly placed in the designated storage cabinet, thereby effectively avoiding safety hazards and management confusion caused by misplaced tools. Once improper placement of tools is detected, the alarm unit immediately activates the voice prompt to remind the user to correct the error in time, while preventing the abnormal closure of the storage cabinet door, ensuring the accuracy and reliability of tool management.

[0117] The remote management module uses the Internet of Things to query the device status of the management cabinet, configure the system settings, and diagnose faults remotely.

[0118] Example 2

[0119] See also Figure 3 Another embodiment provided by the present invention is a management cabinet control method combining dual-drive algorithm and OCR recognition, the steps comprising:

[0120] S1. Obtain tool attribute information and the location information of the corresponding storage cabinet, construct a tool information sub-library through natural language algorithm and graph database, and update the tool information sub-library in real time through the feedback model;

[0121] The real-time update process specifically includes: updating the label information of the newly stored tools to the tool information sub-library or deleting the label information of the obsolete tools from the tool information sub-library through the feedback model;

[0122] S2. Obtain user demands and maintenance parameters, and match them with the tool information sub-library through the configured matching model to obtain the tool location information and label information corresponding to the user demands;

[0123] S3. Generate the corresponding repair process and the safety measures list of the corresponding repair process through the text generation algorithm according to the tool attribute information, user demands and maintenance parameters, and use the preset work ticket template to generate the work ticket entity corresponding to the user demands;

[0124] S4. Based on the acquired work ticket entity, the text information of the work ticket is obtained by performing text recognition on the work ticket entity through the OCR text recognition model, and the tool position information and tool attribute information in the work ticket text information are used to generate the switch command of the tool storage cabinet requested by the user through the PID algorithm, and at the same time, the storage status of the corresponding tool is converted;

[0125] S5. When the tools are stored back in the corresponding tool storage cabinet after use, the high-density RFID tags on the tools stored after use are identified and verified by the built-in RFID device in the tool storage cabinet. If they are, the storage status of the tools is converted for the second time. If they are not, an alarm is given through the configured voice prompt algorithm. Furthermore, in this embodiment, the second conversion is specifically to convert the storage status of the tools from "in use" to "unused".

[0126] Example 3

[0127] A computer-readable storage medium stores computer instructions, which, when executed, execute a management cabinet control method combining a dual-drive algorithm and OCR recognition.

[0128] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, a management cabinet control method combining a dual-drive algorithm and OCR recognition is implemented.

[0129] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

[0130] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit the present invention. A person skilled in the art should understand that the specific implementation of the present invention can be modified or replaced by equivalents, but these modifications or changes are within the scope of protection of the pending claims.

Claims

1. An intelligent management cabinet combining dual-drive algorithm and OCR recognition, characterized in that: It includes: a management cabinet body, which includes a tool storage cabinet and an information management and control system; The tool storage cabinet is used for storing and monitoring tools; The information management and control system includes an information storage subsystem and a management cabinet control subsystem; the information storage subsystem includes an information repository and an information update module; the information repository includes a tool information sub-repository; the tool information sub-repository automatically files and saves the label information of tools through a natural language algorithm and a graph database, and updates the label information corresponding to the tools newly stored in the tool storage cabinet to the tool information sub-repository through the information update module, and deletes the label information corresponding to the obsolete tools from the information repository; The management cabinet control subsystem includes a work ticket generation module, an identification control module and an alarm module; the work ticket generation module includes an intelligent matching unit, a work ticket generation unit and a work ticket printing unit; The intelligent matching unit is used to match the tool information sub-library with the user's demands and maintenance parameters input in real time through the configured matching model to obtain the tool location information and label information corresponding to the user's demands; the generation and printing unit is used to generate the maintenance process corresponding to the user's demands and the safety measures list of the corresponding maintenance process through the text generation algorithm according to the acquired tool attribute information and the user's demands and repair parameters; The work ticket printing unit is used to generate a work ticket corresponding to the user's request and print out a work ticket entity according to the acquired tool location information, repair process and safety measures list of the corresponding repair process through a preset work ticket template; The identification and control module is used to identify and control the tool storage and access process based on the acquired work ticket and the built-in deep learning algorithm; the alarm module is used to monitor and alarm the storage process of the used tools in real time.

2. According to claim 1, the intelligent management cabinet combining dual-drive algorithm and OCR recognition is characterized in that: The identification control module includes an information identification unit and a control instruction unit; the information identification unit is used to perform text recognition on the work ticket entity through a device with a built-in OCR text recognition model to obtain the work ticket text information; the control instruction unit is used to generate a switch instruction for the user's tool storage cabinet through a PID algorithm based on the tool location information and tool attribute information in the obtained work ticket text information, and mark the tool storage status corresponding to the user's request stored in the tool information sub-library as "in use"; The alarm module includes an RFID identification unit and an alarm unit; the RFID identification unit identifies and verifies the high-density RFID tags on the stored tools after use through the RFID device configured in the tool storage cabinet; if the corresponding stored tools belong to the tool storage cabinet where they are currently stored, the tool storage status corresponding to the user's request stored in the tool information sub-library is converted for the second time; if it does not belong, the corresponding identification result is transmitted to the alarm unit; the alarm unit is used to give a real-time alarm reminder of the tool storage process through the configured voice prompt algorithm based on the obtained identification result.

3. According to claim 2, the intelligent management cabinet combining dual-drive algorithm and OCR recognition is characterized in that: The label information of the tool includes the identification number of the tool storage cabinet, the physical attribute information and the functional attribute information of the tool; the steps of automatically filing and saving the label information of the tool by the tool information sub-library include: A1. Obtain label information corresponding to tools and instruments, annotate the obtained label information using the ternary method, and obtain a tool and instrument information data set; A2. Use the pre-trained Bert model to build an information extraction model, input the acquired information into the information extraction model, and obtain the label information keyword triple set corresponding to each tool; A3. Extract tool function attribute keywords from the tool corresponding label information keyword triple set, and use the function attribute keywords as classification labels to perform secondary labeling on each tool to obtain a tool classification data set; A4. Construct a tool classification model based on a support vector machine, input the tool classification data set into the tool classification model for training, obtain a trained tool classification model, and output the tool classification result; A5. The tool classification result is used as the first-level node of the tool information sub-library, the tool name attribute is used as the second-level node, the tool storage cabinet identification number corresponding to the tool is used as the third-level node, the physical attribute information and the functional attribute information are listed in parallel as the fourth-level node, the tool storage status is constructed as the fifth-level node, and the initial storage status is set to "unused"; A6. Based on the hierarchical nodes constructed in A5, the tag information keyword triple set corresponding to each tool is saved under the corresponding node, and the tool information sub-library with hierarchical nodes is constructed.

4. According to claim 3, the intelligent management cabinet combining dual-drive algorithm and OCR recognition is characterized in that: The step of automatically filing and saving the label information of the tool and equipment information sub-library also includes: A7. The trained tool classification model is built into the tool information sub-library. When a new tool is stored in the tool storage cabinet, the label information corresponding to the new tool is updated to the corresponding node using the tool classification model. A8. Obtain tool service life information and tool damage image information, annotate the tool damage image information with damage type and damage score, and obtain a tool damage detection data set; A9. Use the random forest algorithm to build a damage detection model, input the tool damage detection data set into the damage detection model for training, and obtain a trained damage detection model; A10, setting an elimination score threshold, embedding a damage detection model into the tool information sub-library, and obtaining the damage type and damage score of the corresponding tool through the built-in damage detection model according to the tool service life information and tool damage image information obtained in real time; A11. If the damage score of the corresponding tool obtained is greater than the elimination score threshold, an alarm will be issued to the corresponding tool, and the corresponding tool label information will be deleted from the tool information sub-library.

5. According to claim 4, the intelligent management cabinet combining dual-drive algorithm and OCR recognition is characterized in that: The step of generating a maintenance process corresponding to the user's request and a list of safety measures for the maintenance process includes: B1. Obtain user demand information, and extract maintenance parameter keyword information corresponding to the user demand information through the information extraction model built into the tool information sub-library; B2. Use the SENet algorithm to build a matching model, embed the matching model into the interface of the tool information sub-library, match the obtained maintenance parameter keyword information with the tool label information stored in the tool information sub-library through the matching model embedded in the interface, and obtain the tool matching score corresponding to the user's request; B3. Set a matching score threshold. If the tool matching score is greater than the matching score threshold, obtain the corresponding tool location information and label information. If the tool matching score is less than or equal to the matching score threshold, issue an alarm prompt that there is no corresponding tool. B4. Input the acquired tool label information and the maintenance parameter keyword information corresponding to the user's request into the text generation algorithm to obtain the maintenance process corresponding to the user's request and the safety measures list of the maintenance process; B5. Input the maintenance process corresponding to the user's demands and the corresponding safety measures list of the maintenance process into the preset maintenance process template to obtain the electronic version of the maintenance process and the corresponding safety measures list.

6. The intelligent management cabinet combining dual-drive algorithm and OCR recognition according to claim 5, characterized in that: The step of generating a switch instruction for the tool storage cabinet requested by the user comprises: C1. Obtain historical work ticket image data and corresponding text data, and preprocess the acquired data; C2. Build a lightweight OCR text recognition model, input the preprocessed work ticket image data and the corresponding text data into the OCR text recognition model for training, and obtain a trained OCR text recognition model; C3. The trained OCR text recognition model is built into the text recognition device configured in the management cabinet, and the text recognition device is used to extract text from the real-time printed work ticket entity to obtain the text information in the work ticket entity; C4. Using the information extraction model to extract the identification number and location information of the tool storage cabinet from the text information in the work ticket entity, input the identification number and location information of the tool storage cabinet into the PID algorithm, and generate the corresponding tool storage cabinet switch control instruction; C5. By generating a corresponding tool storage cabinet switch control instruction, the corresponding storage cabinet on the management cabinet is controlled to pop open. When the corresponding tool storage cabinet pops open, the storage status of the corresponding tool stored in the tool information sub-library is converted to "in use".

7. The intelligent management cabinet combining dual-drive algorithm and OCR recognition according to claim 6, characterized in that: The steps of real-time monitoring and alarming the storage process of the used tools include: D1. After the corresponding tools are used, use the work ticket to open the corresponding tool storage cabinet through C3-C5, put the tools into the corresponding tool storage cabinet, and align the high-density RFID tag on the tools with the configured RFID device; D2. Identify the high-density RFID tags placed on the tools through the configured RFID device, obtain the corresponding tag information, and compare the obtained tag information with the tool tag information stored in the tool information sub-library, and obtain the comparison result; D3. According to the comparison result, if the tools put in belong to the tools storage cabinet at the current moment, the configured voice prompt algorithm is used to prompt that the tools are stored correctly, and the tool storage cabinet door can be closed normally; D4. According to the comparison result, if the tools placed in the cabinet do not belong to the tools stored in the cabinet at the current moment, the configured voice prompt algorithm will prompt that the tools are stored incorrectly, and the corresponding tool storage cabinet door will be prohibited from closing. If the corresponding tool storage cabinet door is closed by mistake, the tool storage cabinet door where the tools are stored incorrectly will automatically pop open.

8. A management cabinet control method combining dual-drive algorithm and OCR recognition according to any one of claims 1 to 7, characterized in that: include: S1. Obtain tool attribute information and location information of corresponding storage cabinets, construct the tool information sub-library through natural language algorithm and graph database, and update the tool information sub-library in real time through feedback model; S2. Obtain user demands and maintenance parameters, and match them with the tool information sub-library through the configured matching model to obtain the tool location information and label information corresponding to the user demands; S3. Generate the corresponding repair process and the safety measures list of the corresponding repair process according to the tool attribute information, user demands and maintenance parameters through the text generation algorithm, and generate the work ticket entity corresponding to the user demands using the preset work ticket template; S4. Based on the acquired work ticket entity, the OCR text recognition model is used to perform text recognition on the work ticket entity to obtain the work ticket text information, and the tool location information and tool attribute information in the work ticket text information are used to generate the user's demand for the tool storage cabinet through the PID algorithm, and at the same time, the storage state of the corresponding tool is converted; S5. When the tools are put back into the corresponding tool storage cabinet after use, the high-density RFID tags on the tools are identified and verified through the built-in RFID device in the tool storage cabinet. If they are, the storage status of the tools is converted again. If they are not, an alarm is given through the configured voice prompt algorithm.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, a management cabinet control method combining a dual-drive algorithm and OCR recognition as described in claim 8 is executed.

Images