A method and system for intelligent traceability management of surveying and mapping equipment

Abstract

An embodiment of the present invention relates to an intelligent traceability management method and system for surveying and mapping equipment, including: generating identity identification information of surveying and mapping equipment; collecting status information of the surveying and mapping equipment, wherein the status information includes usage status and user information of the surveying and mapping equipment; performing loan management on the surveying and mapping equipment, including updating the status information of the surveying and mapping equipment and recording the loan information; performing loan management on the surveying and mapping equipment, including updating the user information of the surveying and mapping equipment and recording the loan information; performing return management on the surveying and mapping equipment, including confirming the return status of the surveying and mapping equipment and recording the usage information.

Description

Technical Field

[0001] The present invention relates to the technical field of surveying and mapping equipment management, and in particular to an intelligent traceability management method and system for surveying and mapping equipment. Background Art

[0002] Surveying and mapping equipment plays a key role in engineering surveying, geographic information collection, environmental monitoring and other fields. With the development of surveying and mapping technology and the expansion of application scenarios, the number of various types of surveying and mapping equipment continues to increase, and the complexity of equipment management has also increased. At present, the main problems in surveying and mapping equipment management are as follows:

[0003] Equipment management is complex and inefficient. There are many types of surveying and mapping equipment, including total stations, GPS receivers, drones, etc. The traditional manual recording method is prone to errors and it is difficult to keep track of the equipment's usage status and whereabouts. The equipment borrowing and returning process lacks effective digital management methods, resulting in frequent equipment loss or untimely return.

[0004] The equipment is not used efficiently. Due to the lack of real-time monitoring and scientific analysis of equipment usage, idle equipment and equipment shortage coexist, resulting in a waste of resources. At the same time, it is impossible to accurately assess the use intensity and wear status of the equipment, making it difficult to formulate a reasonable maintenance plan.

[0005] Therefore, there is an urgent need for a technical solution that can realize intelligent management of the entire life cycle of surveying and mapping equipment. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for intelligent traceability management of surveying and mapping equipment, which can realize intelligent traceability management of equipment by digitally managing the entire process of equipment, thereby improving management efficiency and equipment utilization efficiency.

[0007] In order to achieve the above object, the present invention adopts the following technical scheme:

[0008] A method for intelligent traceability management of surveying and mapping equipment, comprising:

[0009] Generate identification information of surveying and mapping equipment;

[0010] Collecting status information of the surveying and mapping equipment, wherein the status information includes usage status and user information of the surveying and mapping equipment;

[0011] Performing loan management on the surveying and mapping equipment, including updating status information of the surveying and mapping equipment and recording loan information;

[0012] Performing loan management on the surveying and mapping equipment, including updating user information of the surveying and mapping equipment and recording loan information;

[0013] Performing return management on the surveying and mapping equipment, including confirming the return status of the surveying and mapping equipment and recording usage information;

[0014] After collecting the status information of the surveying and mapping equipment, the method further includes:

[0015] Performing finite state transduction modeling on the state of the surveying and mapping equipment and establishing state transfer rules;

[0016] Based on the state transition rule, marking the historical usage data of the surveying and mapping equipment as a state sequence;

[0017] Inputting the state sequence into a Bayesian inference model to calculate the damage probability of the surveying and mapping equipment;

[0018] According to the damage probability of the surveying and mapping equipment, a predictive maintenance report is generated, and the maintenance cycle of the surveying and mapping equipment is dynamically adjusted.

[0019] Optionally, the method further comprises:

[0020] storing usage data of the surveying and mapping equipment, wherein the usage data includes loan time, user, usage frequency, and return time;

[0021] Perform statistical analysis on the usage data and generate a usage report of the surveying and mapping equipment.

[0022] Optionally, calculating the damage probability of the surveying and mapping equipment includes:

[0023] The damage probability of the surveying and mapping equipment is calculated according to the following Bayesian formula:

[0024] P(damage|number of loans) = P(number of loans|damage) * P(damage) / P(number of loans)

[0025] in,

[0026] P(damage|number of loans) is the probability that the surveying and mapping equipment is damaged after being loaned a certain number of times;

[0027] P(number of loaned out | damage) is the probability of damage when the surveying and mapping equipment described in the historical data is loaned out a specific number of times;

[0028] P(damage) is the a priori probability of damage of the surveying and mapping equipment;

[0029] P (number of loaned times) is the total probability that the surveying and mapping equipment is loaned a certain number of times.

[0030] Optionally, performing finite state transduction modeling on the state of the surveying and mapping device and establishing state transition rules include:

[0031] Define the basic status of the surveying and mapping equipment, including storage status, loan status, loan status, return status and damage status;

[0032] Define the transition conditions between states, including loan triggering the transition from storage to loan, loan triggering the transition from loan to loan, and return triggering the transition from loan or loan to return;

[0033] Define the tokenization rules for state transfer to ensure that each state transfer process is fully recorded.

[0034] Optionally, the method further comprises:

[0035] Collecting real-time usage data of the surveying and mapping equipment, including usage frequency, environmental conditions, and operating information;

[0036] Inputting the usage data into a CSRNet model to generate a density map of the usage intensity of the surveying and mapping equipment;

[0037] Performing fuzzy reasoning on the state of the surveying and mapping equipment according to preset fuzzy rules;

[0038] Based on the density map and the fuzzy reasoning result, a status assessment result of the surveying and mapping equipment is output.

[0039] Optionally, the preset fuzzy rules include:

[0040] The fuzzy set of the use frequency of the surveying and mapping equipment includes low frequency: 0-30 times, medium frequency: 31-60 times, and high frequency: more than 61 times;

[0041] The fuzzy set of the temperature of the surveying and mapping equipment includes normal temperature: 0-50°C, high temperature: 51-70°C, and excessive temperature: above 71°C;

[0042] The fuzzy set of states of the surveying and mapping equipment includes light wear, medium wear, and heavy wear;

[0043] Performing fuzzy reasoning on the state of the surveying and mapping equipment includes:

[0044] When the surveying and mapping equipment is used at a high frequency and the temperature is too high, a severe wear state is inferred;

[0045] When the surveying and mapping equipment is used at a medium frequency and the temperature is relatively high, a medium wear state is inferred;

[0046] When the usage frequency of the surveying and mapping device is low and the temperature is normal, a slight wear state is inferred.

[0047] Optionally, after returning the surveying and mapping equipment, the method includes:

[0048] A non-iterative excitation coordination method is used to simultaneously coordinate multiple state excitation sources of the surveying and mapping device to obtain state change characteristics of the surveying and mapping device;

[0049] Based on the state change characteristics, a causal relationship chain between the state change of the surveying and mapping equipment and the damage event is established through a causal reasoning engine;

[0050] Generate a device traceability report based on the causal chain.

[0051] Optionally, the adopting of the non-iterative excitation coordination method to simultaneously coordinate multiple state excitation sources of the surveying and mapping device includes:

[0052] Defining the state stimulus source of the surveying and mapping equipment, including usage frequency, return status, and damage status;

[0053] Coordinate the multiple state excitation sources at one time to determine the state change nodes of the surveying and mapping equipment in the flow;

[0054] The state change node is associated with the abnormal state of the surveying and mapping equipment for analysis.

[0055] The present application also provides an intelligent traceability management system for surveying and mapping equipment, including:

[0056] An identity module, used to generate identity information for the surveying and mapping equipment and store the identity information in a database;

[0057] A status monitoring module, used to collect status information of the surveying and mapping equipment, and manage the lending, re-lending and returning of the surveying and mapping equipment;

[0058] A data processing module, used to receive the status information collected by the status monitoring module, store and analyze the usage data of the surveying and mapping equipment, generate a usage report, and send the usage report to the predictive maintenance module;

[0059] A predictive maintenance module, configured to receive the usage report, calculate the damage probability of the surveying and mapping equipment based on a Bayesian reasoning model, generate a predictive maintenance report, and send the predictive maintenance report to a condition assessment module;

[0060] A status assessment module, used for receiving the predictive maintenance report, monitoring the status of the surveying and mapping equipment based on the CSRNet model and fuzzy logic technology, generating a status assessment result, and sending the assessment result to the traceability tracking module;

[0061] A traceability module is used to receive the status assessment result, track the traceability target of the equipment based on the non-iterative incentive coordination method and the causal reasoning engine, and generate a device traceability report;

[0062] Among them, after collecting the status information of the surveying and mapping equipment, the intelligent traceability management system for surveying and mapping equipment is also used to: perform finite state transduction modeling on the status of the surveying and mapping equipment and establish state transition rules; based on the state transition rules, mark the historical usage data of the surveying and mapping equipment as a state sequence; input the state sequence into the Bayesian reasoning model to calculate the damage probability of the surveying and mapping equipment; generate a predictive maintenance report based on the damage probability of the surveying and mapping equipment, and dynamically adjust the maintenance cycle of the surveying and mapping equipment.

[0063] The beneficial effects of the present invention include:

[0064] 1. Improve management efficiency: Through digital management methods, the entire process of equipment borrowing, lending and returning can be automated, greatly improving management efficiency.

[0065] 2. Enhanced traceability: Based on state transfer rules and causal chain, the usage history of the device can be accurately tracked and responsibility can be clearly identified.

[0066] 3. Realize predictive maintenance: Use the Bayesian reasoning model to predict the probability of equipment damage, combine the CSRNet model and fuzzy logic technology to evaluate the equipment status, realize predictive maintenance, and extend the service life of the equipment.

[0067] 4. Improve resource utilization: Optimize equipment scheduling and improve resource utilization efficiency through real-time monitoring and analysis of equipment usage status. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is the overall architecture diagram of the intelligent traceability management system for surveying and mapping equipment of the present invention;

[0069] Figure 2 This is a flow chart of the intelligent traceability management method for surveying and mapping equipment of the present invention;

[0070] Figure 3 The workflow diagram of the Bayesian reasoning model and finite state transduction tokenization of the present invention;

[0071] Figure 4 It is a workflow diagram of the CSRNet model and fuzzy logic technology of the present invention;

[0072] Figure 5 It is a workflow diagram of the non-iterative incentive coordination method and causal reasoning of the present invention;

[0073] Figure 6 This is a schematic diagram of the state transition of the surveying and mapping equipment of the present invention;

[0074] Figure 7 An example of using an intensity density map for the surveying and mapping device of the present invention;

[0075] Figure 8 This is an example diagram of the state assessment result of the surveying and mapping equipment of the present invention;

[0076] Figure 9a System home page interface diagram of the intelligent traceability management system for surveying and mapping equipment of the present invention;

[0077] Figure 9b Device classification interface diagram of the intelligent traceability management system for surveying and mapping equipment of the present invention;

[0078] Fig.9c Equipment lending management interface diagram of the intelligent traceability management system for surveying and mapping equipment of the present invention;

[0079] Figure 9d Diagram of the interface of the use of traceability equipment of the intelligent traceability management system for surveying and mapping equipment of the present invention. DETAILED DESCRIPTION

[0080] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0081] like Figure 1 As shown, the intelligent traceability management system for surveying and mapping equipment provided by the present invention includes an identification module 11, a status monitoring module 12, a data processing module 13, a predictive maintenance module 14, a status assessment module 15 and a traceability module 16. The system also includes a database 18 for storing equipment information, status records and analysis results.

[0082] Specifically, the identity identification module is used to generate identity identification information for the surveying and mapping equipment and store the identity identification information in a database;

[0083] A status monitoring module, used to collect status information of the surveying and mapping equipment, and to manage the lending, re-lending and returning of the surveying and mapping equipment;

[0084] A data processing module, used to receive the status information collected by the status monitoring module, store and analyze the usage data of the surveying and mapping equipment, generate a usage report, and send the usage report to the predictive maintenance module;

[0085] A predictive maintenance module, configured to receive the usage report, calculate the damage probability of the surveying and mapping equipment based on a Bayesian reasoning model, generate a predictive maintenance report, and send the predictive maintenance report to a condition assessment module;

[0086] A status assessment module, used for receiving the predictive maintenance report, monitoring the status of the surveying and mapping equipment based on the CSRNet model and fuzzy logic technology, generating a status assessment result, and sending the assessment result to the traceability tracking module;

[0087] The traceability module is used to receive the status assessment result, track the traceability target of the equipment based on the non-iterative incentive coordination method and the causal reasoning engine, and generate an equipment traceability report.

[0088] Figure 9a-9b This is an example diagram of an APP front-end interface in an embodiment of the present application. Figures 9a-9d , the intelligent traceability management system includes:

[0089] Equipment classification management: This function is mainly to classify and manage all equipment, such as total stations, levels, depth sounders, GPS, drones, leak detectors and other instruments and equipment, and manage them by adding, deleting, modifying, checking, setting QR code labels, etc. according to basic attributes such as unified numbers and names.

[0090] Equipment inventory management: Add, delete, modify, check, and manage all instruments and equipment according to basic attributes such as number, name, quantity, type, and status.

[0091] Equipment collection: Users scan the QR code through WeChat to automatically identify the device and proceed with the collection process, so that the project where the equipment is located and the corresponding user can be traced.

[0092] Equipment return: The user can directly click the device return button on the mobile terminal to complete the equipment return management process and record a log for it.

[0093] Asset inventory: All equipment will be designed with inventory statistics functions, such as equipment inventory quantity, equipment in use quantity, equipment idle quantity, etc., to intuitively understand the use of the company's related equipment and instruments.

[0094] My Instrument: This function can intuitively display basic information such as the project the technician is in, the instruments and equipment used, and the length of use.

[0095] like Figure 2 As shown, the basic process of the intelligent traceability management method for surveying and mapping equipment of the present invention includes:

[0096] The first step is to generate the identity information of the surveying and mapping equipment.

[0097] The identification module 11 first collects the basic information of the surveying and mapping equipment, including the equipment type, model, factory number, etc. Based on this information, the system generates a unique equipment number and establishes an equipment information file in the database 18. The equipment information file contains the basic attributes of the equipment, purchase information, maintenance records, etc. At the same time, the system also generates a corresponding QR code identification for each device, which contains key information such as the equipment number, so as to facilitate the subsequent rapid identification and management of the equipment by scanning the QR code.

[0098] The second step is to collect the status information of the surveying and mapping equipment.

[0099] The status monitoring module 12 collects the device status in real time in the following ways: on the one hand, the device identity information is obtained by scanning the QR code of the device; on the other hand, the working parameters of the device are collected in real time through the sensor network, such as device temperature, usage time, working environment and other data. The system divides the collected status information into two categories: usage status information and user information. Usage status information includes whether the device is in the library, whether it is borrowed, whether it is in use, whether there is an abnormality, etc.; user information includes the identity information, contact information, department of the current borrower, etc. This information is updated in real time and stored in the database to form a dynamic record of the device status.

[0100] The third step is to manage lending.

[0101] When receiving an application for borrowing equipment, the system first verifies the identity and authority of the borrower. After verification, the status monitoring module 12 performs the following operations: first, check the current status of the equipment to confirm that the equipment can be borrowed; then, update the equipment status information to the "borrowed" status; then, record detailed borrowing information, including borrower information (name, department, contact information), borrowing time, expected return time, borrowing purpose, etc.; finally, generate a borrowing record and store it in the database. During this period, the system will automatically send a borrowing confirmation notice to the borrower and send a return reminder before the expected return date.

[0102] The fourth step is to manage the loan.

[0103] When the equipment needs to be lent between users, the status monitoring module 12 executes the following process: first, the original user submits a lent application in the system and specifies the new user; after the system verifies the authority of the new user, it updates the user information of the equipment; at the same time, it records the lent time, lent reason and other information; finally, it generates a lent record, which is associated with the original borrowing record to form a complete chain of responsibility. The system automatically notifies the new user to receive the equipment and requires him to confirm the equipment status to ensure the standardization and traceability of the equipment handover.

[0104] The fifth step is to carry out return management.

[0105] When the device is returned, the status monitoring module 12 performs the following operations: first, verify the identity of the returned device; then, the administrator checks the device status and records whether the device is intact, whether it is damaged, etc.; then, the system updates the device status to the "returned" status; finally, a return record is generated, which contains information such as usage time, return time, and device status. If the device is abnormal, the system will start the abnormality handling process, record the abnormal situation, and push the relevant information to the maintenance personnel. At the same time, the system will automatically count the usage data of the device, including usage frequency, usage time, user distribution, etc., which will be used for subsequent equipment usage analysis and predictive maintenance.

[0106] For example, Example 1: Intelligent management of total station

[0107] In surveying and mapping projects, the total station is one of the key equipment. By numbering the total station and generating a QR code, engineers can scan the QR code to log in to the system and record the borrowing information when using it. The administrator can check the borrowing status of the total station and the current user information at any time through the system, and record the equipment status when the equipment is returned.

[0108] Example 2: Drone borrowing and lending management

[0109] Drones are used for large-scale surveying and mapping. The user information is quickly recorded by scanning the QR code during the borrowing and lending process. If a drone needs to be lent to other engineers, they only need to submit a loan application through the system, and the recipient can scan and confirm to complete the loan operation.

[0110] Example 3: Surveying and mapping data recording and analysis

[0111] The system automatically records the usage data of each device, including usage time, number of borrowings, etc. Administrators can understand the usage of the device through the reports generated by the system and make decisions on device maintenance and updates.

[0112] The following is a specific example:

[0113] Example: Total Station (Model: Leica TS60) Management Process of a Surveying and Mapping Unit

[0114] Step 1: Generate identity information

[0115] When the administrator enters the equipment into the warehouse, he / she enters the following equipment information:

[0116] Equipment type: Total station;

[0117] Equipment model: Leica TS60;

[0118] Factory number: TS202403001;

[0119] Purchase date: March 1, 2024;

[0120] Equipment value: 350,000 yuan;

[0121] Based on the above information, the system generates a unique equipment number: TSJ-TS60-2024001, and generates a corresponding QR code. The QR code contains information such as the equipment number, model, and factory number, and the QR code is printed and attached to the equipment box. The system also creates an equipment file for the total station in the database, recording the basic information, technical parameters, and calibration records of the equipment.

[0122] Step 2: Collect status information

[0123] The system collects device status in real time through built-in sensors:

[0124] Equipment temperature: 23℃;

[0125] Usage time: 87 hours in total;

[0126] Battery level: 85%

[0127] Current status: In stock;

[0128] This status information is updated to the database in real time, and managers can view the real-time status of the equipment at any time through the management interface.

[0129] Step 3: Loan Management

[0130] At 9:00 am on March 15, 2024, surveying engineer Zhang Gong submitted a borrowing application:

[0131] 1. The system verifies Mr. Zhang’s borrowing authority (confirming that he is an engineer in the surveying and mapping department and has the authority to borrow equipment);

[0132] 2. Record borrowing information:

[0133] Borrower: Zhang Gong

[0134] Department: Surveying and Mapping Department;

[0135] Contact number: 138xxxxxxxx;

[0136] Borrowing time: 2024-03-15 09:00;

[0137] Expected return: 2024-03-17 18:00;

[0138] Application projects: urban topographic mapping projects;

[0139] Use location: New urban construction site;

[0140] 3. The system generates a borrowing number: JY202403150001;

[0141] 4. Update the device status to "borrowed";

[0142] 5. The system automatically sets a return reminder: a return reminder will be sent to Mr. Zhang at 10:00 am on March 17, 2024.

[0143] Step 4: Loan Management

[0144] At 15:00 on March 16, 2024, due to the adjustment of the project schedule, Engineer Zhang needs to loan the equipment to Engineer Li in the same department:

[0145] 1. Mr. Zhang submits a loan application in the system:

[0146] Original user: Zhang Gong;

[0147] New user: Mr. Li;

[0148] Reason for transfer: The project schedule has been adjusted and Engineer Li is required to continue to complete the measurement work;

[0149] Transfer time: 2024-03-16 15:00;

[0150] 2. The system verifies Mr. Li’s authority;

[0151] 3. Generate a loan record, number: ZJ202403160001;

[0152] 4. The system notifies Mr. Li to receive the equipment;

[0153] 5. Mr. Li scans the QR code of the device to confirm receipt;

[0154] 6. The system updates the user information to Mr. Li, but retains the association with Mr. Zhang's borrowing record;

[0155] Step 5: Return Management

[0156] At 17:30 on March 17, 2024, Mr. Li returned the equipment:

[0157] 1. The administrator scans the device QR code to confirm the device identity;

[0158] 2. Check the device status and record:

[0159] Equipment appearance: intact;

[0160] Function test: normal;

[0161] Cumulative usage time: increased by 26.5 hours;

[0162] Battery level: 25%

[0163] 3. The system generates a return record:

[0164] Return order number: GH202403170001;

[0165] Return time: 2024-03-17 17:30;

[0166] Duration: 26.5 hours

[0167] Total number of users: 2 (Engineer Zhang and Engineer Li);

[0168] Device status: normal;

[0169] 4. The system updates the device status to "Returned";

[0170] 5. Automatically generate usage statistics reports:

[0171] The total duration of this loan: 56.5 hours; users involved: 2 people; completed measurement task: urban topographic mapping; equipment status: normal, no maintenance required;

[0172] These data records provide a basis for subsequent predictive maintenance. After systematic analysis of usage data, it was found that the total station has been used for a total of 87 hours since it was put into use. The damage probability calculated by the Bayesian reasoning model is low, and the next calibration maintenance time is recommended to be performed when it reaches 200 hours of use.

[0173] Through the above five steps, the system realizes the full process management of surveying and mapping equipment from borrowing to returning, and each link has detailed data records to ensure the standardization and traceability of equipment management. These basic processes provide reliable data support for subsequent predictive maintenance, status assessment and traceability analysis.

[0174] The data processing module 13 receives the status information collected by the status monitoring module 12 and stores and analyzes the usage data. Figure 2 As shown in process 230, the usage data includes information such as loan time, user, usage frequency, and return time, and the system generates a device usage report based on this data.

[0175] like Figure 3 As shown, the predictive maintenance module 14 uses the Bayesian scaling law and finite state transduction tokenization technology to perform predictive maintenance.

[0176] The specific explanation is as follows:

[0177] Bayesian scaling law: Bayesian scaling law can be used to adjust the prior probability of the model in contextual learning, and use historical data to infer the probability of future events. For example, the historical frequency of equipment use can be used to predict future equipment damage or maintenance needs, thus establishing a dynamic prediction model.

[0178] Finite State Transducer (FST) Tokenization: Finite State Transducer (FST) is an automaton that converts an input sequence into an output sequence. Combined with tokenization technology, it can track complex event sequences, such as the loan, loan, and return status of equipment. Through tokenization, the transfer conditions and operations in each state can be automatically recorded and responded to.

[0179] This step specifically includes:

[0180] Performing finite state transduction modeling on the state of the surveying and mapping equipment and establishing state transfer rules;

[0181] Based on the state transition rule, marking the historical usage data of the surveying and mapping equipment as a state sequence;

[0182] Inputting the state sequence into a Bayesian inference model to calculate the damage probability of the surveying and mapping equipment;

[0183] According to the damage probability of the surveying and mapping equipment, a predictive maintenance report is generated, and the maintenance cycle of the surveying and mapping equipment is dynamically adjusted.

[0184] Wherein, calculating the damage probability of the surveying and mapping equipment includes:

[0185] The damage probability of the surveying and mapping equipment is calculated according to the following Bayesian formula:

[0186] P(damage|number of loans) = P(number of loans|damage) * P(damage) / P(number of loans)

[0187] in,

[0188] P(damage|number of loans) is the probability that the surveying and mapping equipment is damaged after being loaned a certain number of times;

[0189] P(number of loaned out | damage) is the probability of damage when the surveying and mapping equipment described in the historical data is loaned out a specific number of times;

[0190] P(damage) is the a priori probability of damage of the surveying and mapping equipment;

[0191] P (number of loaned times) is the total probability that the surveying and mapping equipment is loaned a certain number of times.

[0192] In addition, the state of the surveying and mapping equipment is modeled by finite state transduction and state transition rules are established, specifically including:

[0193] Define the basic status of the surveying and mapping equipment, including storage status, loan status, loan status, return status and damage status;

[0194] Define the transition conditions between states, including loan triggering the transition from storage to loan, loan triggering the transition from loan to loan, and return triggering the transition from loan or loan to return;

[0195] Define the tokenization rules for state transfer to ensure that each state transfer process is fully recorded.

[0196] Firstly, the state of the surveying and mapping equipment is modeled by finite state transduction.

[0197] like Figure 6 As shown in the figure, the system defines five basic states of equipment: storage state S1, loan state S2, loan state S3, return state S4 and damage state S5. The transition between states is triggered by specific conditions. For example, the loan operation triggers the transition from S1 to S2, and the loan operation triggers the transition from S2 to S3.

[0198] The system marks the historical usage data of the device according to the state transition rules to form a state sequence.

[0199] Specifically, in the process of surveying and mapping equipment management, in order to accurately track the flow of equipment, the system defines five basic states and corresponding state transition rules. Figure 6 As shown in the figure, each state represents a specific stage in the life cycle of the device. The in-stock state S1 means that the device is in inventory and can be borrowed; the loaned state S2 means that the device has been borrowed by a user; the loaned state S3 means that the device is transferred between different users; the returned state S4 means that the device has been used and returned to the inventory; the damaged state S5 means that the device is faulty and needs repair.

[0200] The transitions between these states follow strict rules and trigger conditions. When a user applies to borrow a device and is approved, the transition from the storage state S1 to the loan state S2 is triggered. If the current user needs to transfer the device to another user, the system will trigger the transition from the loan state S2 to the loan state S3, and record the new user information. When the user completes the work and returns the device, the system transfers the state from the loan state S2 or the loan state S3 to the return state S4. If a fault is found during the use of the device, the system will transfer the state to the damaged state S5 and start the repair process.

[0201] The system records the transfer process of the equipment status in real time to form a state sequence. Taking a total station as an example, its state sequence may be as follows: equipment storage (S1) → first loan (S2) → loan to others (S3) → returned to storage (S4) → re-storage (S1) → loaned again (S2) → equipment failure (S5) → maintenance completed (S1). This state sequence clearly records the use history of the equipment, and each state transfer contains detailed information such as timestamp, trigger condition, operator, etc.

[0202] To facilitate analysis and prediction, the system tokenizes the state sequence. For example, "1" can be used to represent a normal transfer (such as a loan from S1 to S2), and "0" can be used to represent an abnormal transfer (such as a fault from S2 directly to S5). Through this tokenization process, the system can identify regularities and abnormal patterns in the use of equipment, providing data support for predictive maintenance.

[0203] The analysis of state sequences can also reveal some important characteristics of equipment use. For example, frequent S2→S3 transitions (borrowed to loaned) may indicate that equipment use scheduling is not reasonable; frequent S2→S5 transitions (borrowed to damaged) may indicate that the equipment is overused or improperly used. These characteristics provide important references for the formulation of equipment management strategies.

[0204] Through this management method based on state transfer, the system can not only grasp the current state of the equipment in real time, but also predict possible problems of the equipment by analyzing the historical state sequence, thereby achieving preventive maintenance and scientific management. At the same time, the complete state transfer record also provides a reliable basis for tracing the responsibility of equipment use.

[0205] The predictive maintenance module 14 inputs the state sequence into the Bayesian reasoning model and calculates the damage probability of the equipment according to the above Bayesian formula. Based on the calculation results, the system dynamically adjusts the maintenance cycle of the equipment.

[0206] The following is a specific example:

[0207] Assume that a surveying and mapping device (such as a total station) is frequently loaned out over a period of time. We hope to use a Bayesian inference model to predict whether the device has a high risk of damage, so as to decide whether maintenance is needed in advance.

[0208] 1. Problem description:

[0209] Using the Bayesian inference model, the probability of damage to the total station is predicted based on the historical usage data of the equipment (such as the number of times it has been loaned out) and the current usage status, and maintenance decisions are made based on this prediction.

[0210] 2. Known conditions:

[0211] Prior probability of the device (the probability of the device being damaged when not borrowed): Assuming that the probability of the device being damaged when not borrowed is 10%, that is:

[0212] P(damage) = 0.1

[0213] Historical data of equipment when it was loaned out (likelihood): Based on historical records, we found that the probability of equipment damage increases each time a device is loaned out. By analyzing historical data, we obtained the following data:

[0214] The probability of a device breaking down after one use is 15%; the probability of a device breaking down after five uses is 30%; the probability of a device breaking down after ten uses is 50%. These data represent the conditional probability (P(number of borrowings|breakage)).

[0215] Assume that the current device has been borrowed 8 times.

[0216] 3. Bayesian formula:

[0217] The core formula of the Bayesian inference model is:

[0218]

[0219] in:

[0220] P(damage|number of loans) is what we want to calculate, which represents the probability of the device being damaged after being loaned out 8 times;

[0221] P(number of loans | damage) is the likelihood given in the historical data, i.e., the probability of damage when the device is loaned out 8 times;

[0222] P(damage) is the a priori probability of damage to the equipment, assumed to be 10%;

[0223] P(number of loans) is the total probability that the device is loaned out 8 times, which can be estimated from all possible situations.

[0224] 4. Step calculation:

[0225] Step 1: Determine Historical Likelihood

[0226] According to historical data, the probability of damage when the equipment is loaned out 8 times is close to 40%, that is:

[0227] P(number of loans = 8 | damage) = 0.4

[0228] Step 2: Calculate the terms in the Bayesian formula

[0229] Prior probability:

[0230] The prior probability of equipment damage is P(damage) = 0.1.

[0231] The total probability P of a device being loaned out 8 times (number of loans = 8):

[0232] Assume that there is a 30% chance that the device will be loaned out 8 times during its entire life cycle, so P(number of loans = 8) = 0.3.

[0233] Step 3: Apply the Bayesian formula to calculate the posterior probability

[0234] Substitute into the Bayesian formula:

[0235]

[0236] The results show that the probability of damage to a device is 13.3% after it has been loaned out eight times.

[0237] Finally, a predictive maintenance report can be generated based on this probability: the system generates regular maintenance and predictive reports based on the equipment's usage data and prediction results, listing a list of equipment that requires priority maintenance.

[0238] Report content: equipment number and name; current status (borrowed, returned, damaged, etc.); predicted maintenance needs and damage probability; last maintenance time and recommended next maintenance time. The frequency of use, damage probability prediction and maintenance timeline of the equipment are displayed in charts, providing equipment managers with a clear overview of the equipment status.

[0239] For example, predictive maintenance with drones:

[0240] Finite state transduction tokenization technology is used to record each borrowing and lending of drones. Combined with the Bayesian inference model, the system analyzes the historical borrowing frequency and damage data of drones to derive the damage probability of drones. After a drone is lent out for more than a certain number of times, the system automatically generates maintenance reminders, and the administrator performs equipment inspection and maintenance based on the reminders.

[0241] Another example: Dynamic maintenance cycle of total station:

[0242] Each time a total station is borrowed, the system recalculates its maintenance cycle based on the Bayesian model. If the frequency of use of a total station suddenly increases, the system will dynamically shorten its maintenance cycle to avoid equipment damage due to frequent use.

[0243] like Figure 4 As shown, the state assessment module 15 uses the CSRNet model and fuzzy logic technology to perform refined monitoring of the device state.

[0244] Technical term explanation:

[0245] CSRNet (Crowd Counting via Scale-Adaptive Convolutional Neural Network): CSRNet is a convolutional neural network model that was originally used to handle crowd counting tasks and performs well in image density estimation. In device status monitoring, the density estimation function of CSRNet can be used to evaluate the state change density of the device and identify the usage intensity and potential abnormalities of the device.

[0246] Fuzzy logic technology: Fuzzy logic is a method for dealing with uncertainty and fuzzy information. By defining fuzzy sets and fuzzy rules, the state of the equipment can be described in detail. For example, in equipment monitoring, fuzzy logic can handle state changes such as "light wear", "medium wear", and "heavy wear", thereby providing more flexible state assessment.

[0247] The specific steps include:

[0248] Collect real-time usage data of the surveying and mapping equipment, including usage frequency, environmental conditions, and operation information; input the usage data into the CSRNet model to generate a density map of the usage intensity of the surveying and mapping equipment; perform fuzzy reasoning on the state of the surveying and mapping equipment according to preset fuzzy rules; and output the state assessment result of the surveying and mapping equipment based on the density map and the fuzzy reasoning result.

[0249] The preset fuzzy rules include:

[0250] The fuzzy set of the use frequency of the surveying and mapping equipment includes low frequency: 0-30 times, medium frequency: 31-60 times, and high frequency: more than 61 times;

[0251] The fuzzy set of the temperature of the surveying and mapping equipment includes normal temperature: 0-50°C, high temperature: 51-70°C, and excessive temperature: above 71°C;

[0252] The fuzzy set of states of the surveying and mapping equipment includes light wear, medium wear, and heavy wear;

[0253] Perform fuzzy reasoning on the state of the surveying and mapping equipment, specifically:

[0254] When the surveying and mapping equipment is used at a high frequency and the temperature is too high, a severe wear state is inferred;

[0255] When the surveying and mapping equipment is used at a medium frequency and the temperature is relatively high, a medium wear state is inferred;

[0256] When the usage frequency of the surveying and mapping device is low and the temperature is normal, a slight wear state is inferred.

[0257] In an embodiment of the present application, the CSRNet model receives real-time usage data of a device, including usage frequency, environmental conditions, operating information, etc., and generates a density map of the device usage intensity.

[0258] In the status assessment of surveying and mapping equipment, the application of the CSRNet model is an innovation. CSRNet (Congested Scene Recognition Network) was originally a deep learning model for crowd density estimation in crowded scenes. This invention creatively applies it to the assessment of the usage intensity of surveying and mapping equipment. The model uses the front-end VGG16 as a feature extractor and the back-end uses a dilated convolutional network for density map estimation.

[0259] In this invention, the CSRNet model is improved to process the multi-dimensional usage data of surveying and mapping equipment. The model input includes: the frequency of equipment use (number of times per day), environmental conditions (temperature, humidity, vibration, etc.), and operation information (operation duration, operation type, etc.). These data are first converted into feature maps, and then feature extraction and fusion are performed through a multi-layer convolutional network.

[0260] Specifically, first of all, the input data preprocessing of the CSRNet model is the key first step. The original data contains three categories of information: usage frequency data (number of uses per day, duration of a single use, distribution of usage periods, etc.), environmental condition data (temperature, humidity, vibration intensity, dust concentration, etc.), and operation information data (operation duration, operation type, operation intensity, etc.). These heterogeneous data need to be standardized to map data of different scales to a unified numerical range, usually the [0,1] interval. For example, temperature data is mapped from (-20℃, 80℃) to [0,1], and usage frequency is mapped from (0, 100 times / day) to [0,1].

[0261] Secondly, the process of converting the preprocessed data into feature maps is a key step. The system uses a time window sliding method to divide the continuous time series data into fixed-size time segments, each of which contains data for all monitoring parameters. These data are reorganized into a three-dimensional tensor with the dimensions [number of channels, time step, number of features]. Among them, the number of channels corresponds to different monitoring parameters (such as frequency of use, temperature, etc.), the time step represents the size of the observation window, and the number of features represents the number of derived features of each parameter.

[0262] Again, the feature extraction network adopts the improved VGG16 structure. Compared with the original VGG16, the main improvements are as follows:

[0263] 1. Input layer adaptation: transform the original 3-channel RGB image input into multi-channel time series data input, where each channel corresponds to a device status parameter;

[0264] 2. Convolution kernel design: According to the characteristics of surveying and mapping equipment data, specific convolution kernel parameters are designed to better capture the time correlation of data and the correlation between parameters;

[0265] 3. Pooling strategy: Different pooling strategies are used in the time dimension to retain important temporal features;

[0266] 4. Activation function: Use LeakyReLU instead of traditional ReLU to handle negative features.

[0267] Then, the feature fusion network uses a dilated convolution structure, which has the following advantages:

[0268] 1. Expanding the receptive field: By setting different dilation rates, the receptive field can be expanded without increasing the number of parameters;

[0269] 2. Multi-scale features: capturing changes in device status at different time scales;

[0270] 3. Parameter association: effectively model the association between different monitoring parameters;

[0271] 4. Computational efficiency: Compared with traditional convolution, dilated convolution reduces the amount of computation while maintaining performance.

[0272] Next, after feature fusion, the model also introduces an attention mechanism. This mechanism can:

[0273] 1. Adaptive weighting: Dynamically weight the importance of different time periods and different parameters; 2. Highlight key information: Emphasize abnormal patterns and key state transition points; 3. Suppress noise: Reduce the impact of irrelevant features; 4. Improve interpretability: Explain the decision basis of the model through the distribution of attention weights.

[0274] Finally, the output layer design uses a density estimation method. The advantages of this design are:

[0275] 1. Continuous representation: Using a continuous density map to represent the intensity of equipment use is more accurate than discrete classification; 2. Spatial correlation: The spatial correlation information between parameters is retained; 3. Visualization-friendly: The output in the form of a density map is easy to intuitively understand and analyze; 4. Flexible thresholds: Different alarm thresholds can be flexibly set according to actual needs.

[0276] Through this multi-level feature extraction and fusion architecture, the improved CSRNet model is able to:

[0277] 1. Accurately capture the spatiotemporal characteristics of equipment usage patterns; 2. Effectively model the complex relationships between multiple monitoring parameters; 3. Accurately predict the distribution of equipment usage intensity; 4. Timely detect potential abnormal conditions; 5. Provide reliable data support for equipment management decisions;

[0278] This deep learning-based method has stronger adaptability and generalization ability than the traditional rule-based method, and can better serve the intelligent management needs of surveying and mapping equipment. At the same time, the design of the model fully considers the needs of actual application scenarios and achieves a good balance between accuracy, real-time and interpretability.

[0279] In summary, the key improvements of the model are:

[0280] 1. Input layer adaptation: The original image input is transformed into multi-channel time series data input, and each channel corresponds to a device status parameter; 2. Feature extraction: The improved VGG16 network is used, and the first 13 convolutional layers are retained, but the convolution kernel parameters are modified to adapt to the device data characteristics; 3. Back-end network: 4 dilated convolutional layers are used, and the dilation rates are 2, 2, 2, 2, respectively, to effectively extract the spatiotemporal characteristics of device usage patterns; 4. Density estimation: The output layer generates a device usage intensity density map, and the color depth indicates the distribution of usage intensity;

[0281] This improvement enables the model to:

[0282] Accurately capture the spatiotemporal distribution characteristics of equipment usage; identify peak intervals of usage intensity; predict possible high-load periods; and provide data support for predictive maintenance.

[0283] Example of report content:

[0284] Total station number: 001; Usage density: high; Current state: heavy wear; Recommendation: schedule maintenance immediately;

[0285] Case 1: Drone equipment status monitoring

[0286] Data collection: UAV usage data, including flight time, ambient temperature, operating load, etc., are collected in real time.

[0287] CSRNet application: The CSRNet model is used to generate a drone flight density map to identify high-intensity flight areas. The density map shows that the drone frequently performs high-load tasks within a certain period of time, resulting in a high flight density.

[0288] Fuzzy logic reasoning: Combining the frequency of use of the drone and the temperature of the flight environment, the fuzzy reasoning engine determines that the drone is currently in a "medium wear" state.

[0289] Output: The system generates a status report and recommends that the administrator maintain the drone to avoid excessive wear and tear due to continued use.

[0290] Case 2: Total Station Status Monitoring and Maintenance

[0291] Data collection: The usage data of the total station (such as number of measurements, ambient humidity, power usage, etc.) is collected in real time.

[0292] CSRNet Application: The CSRNet model generates a usage density estimate for the total station, showing that the device has been used frequently and at a high density recently, indicating high-load usage.

[0293] Fuzzy logic reasoning: Based on input features such as frequency of use and temperature, fuzzy reasoning determines that the total station has reached a “severe wear” state.

[0294] Output: The system automatically generates maintenance reports, prompting administrators to arrange equipment maintenance to prevent damage or measurement errors.

[0295] Combining the density estimation function of CSRNet and the state reasoning of fuzzy logic technology, this solution can achieve refined monitoring of surveying and mapping equipment and timely detect high-intensity use and potential damage risks. Through real-time monitoring and prediction, equipment maintenance plans are more accurate and timely, effectively extending equipment life and reducing failures and unnecessary downtime.

[0296] The sample code is as follows:

[0297] import torch

[0298] import torch.nn as nn

[0299] import torch.nn.functional as F

[0300] class CSRNetForEquipment(nn.Module):

[0301] def __init__(self, input_channels=8):

[0302] super(CSRNetForEquipment, self).__init__()

[0303] # The front end uses a modified VGG16 network

[0304] self.frontend_feat = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M',512, 512, 512]

[0305] self.frontend = self._make_layers(self.frontend_feat, in_channels=input_channels)

[0306] # Use dilated convolutional network for the backend

[0307] self.backend_feat = [512, 512, 512, 256, 128, 64]

[0308] self.backend = self._make_backend_layers(self.backend_feat, in_channels=512)

[0309] # Output layer

[0310] self.output_layer = nn.Conv2d(64, 1, kernel_size=1)

[0311] def forward(self, x):

[0312] # x shape: (batch_size, channels, time_steps, features)

[0313] x = self.frontend(x)

[0314] x = self.backend(x)

[0315] x = self.output_layer(x)

[0316] return x

[0317] def _make_layers(self, cfg, in_channels):

[0318] layers = []

[0319] for v in cfg:

[0320] if v == 'M':

[0321] layers += [nn.MaxPool2d(kernel_size=2, stride=2)]

[0322] else:

[0323] conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=1)

[0324] layers += [conv2d, nn.ReLU(inplace=True)]

[0325] in_channels = v

[0326] return nn.Sequential(*layers)

[0327] def _make_backend_layers(self, cfg, in_channels):

[0328] layers = []

[0329] for v in cfg:

[0330] conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=2, dilation=2)

[0331] layers += [conv2d, nn.ReLU(inplace=True)]

[0332] in_channels = v

[0333] return nn.Sequential(*layers)

[0334] # Data preprocessing class

[0335] class EquipmentDataProcessor:

[0336] def __init__(self):

[0337] self.feature_names =

[0338] 'usage_frequency',# Usage frequency

[0339] 'temperature',# Temperature

[0340] 'humidity',# humidity

[0341] 'vibration',# vibration

[0342] 'operation_time',# operation time

[0343] 'operation_type',# Operation type

[0344] 'environment_dust',# Environment dust

[0345] 'power_consumption'# Power consumption ]

[0347] def preprocess_data(self, data_dict):

[0348] # Convert raw data to model input format

[0349] # data_dict: dictionary containing each feature

[0350] processed_data = []

[0351] for feature in self.feature_names:

[0352] if feature in data_dict:

[0353] # Normalization

[0354] normalized_value = self._normalize_feature(

[0355] data_dict[feature],

[0356] feature )

[0358] processed_data.append(normalized_value)

[0359] else:

[0360] processed_data.append(0.0)

[0361] return torch.tensor(processed_data).float()

[0362] def _normalize_feature(self, value, feature_name):

[0363] # Feature normalization processing

[0364] normalization_params = {

[0365] 'usage_frequency': {'min': 0, 'max': 100},

[0366] 'temperature': {'min': -20, 'max': 80},

[0367] 'humidity': {'min': 0, 'max': 100},

[0368] 'vibration': {'min': 0, 'max': 10},

[0369] 'operation_time': {'min': 0, 'max': 24},

[0370] 'operation_type': {'min': 0, 'max': 5},

[0371] 'environment_dust': {'min': 0, 'max': 1000},

[0372] 'power_consumption': {'min': 0, 'max': 1000}

[0373] }

[0374] params = normalization_params.get(feature_name)

[0375] if params:

[0376] normalized = (value - params['min']) / (params['max'] - params['min'])

[0377] return max(0.0, min(1.0, normalized))

[0378] return value

[0379] # Example

[0380] def generate_density_map(device_data, model_path='equipment_csrnet.pth'):

[0381] # Initialize the model and data processor

[0382] model = CSRNetForEquipment(input_channels=8)

[0383] model.load_state_dict(torch.load(model_path))

[0384] model.eval()

[0385] processor = EquipmentDataProcessor()

[0386] # Preprocess the data

[0387] processed_data = processor.preprocess_data(device_data)

[0388] # Add batch and timestep dimensions

[0389] processed_data = processed_data.unsqueeze(0).unsqueeze(2)

[0390] # Generate density map

[0391] with torch.no_grad():

[0392] density_map = model(processed_data)

[0393] return density_map.squeeze().numpy()

[0394] # Example Usage

[0395] if __name__ == '__main__':

[0396] # Test data

[0397] test_data = {

[0398] 'usage_frequency': 75, # 75 times / day

[0399] 'temperature': 45,# 45℃

[0400] 'humidity': 65, # 65%

[0401] 'vibration': 3.5,# 3.5g

[0402] 'operation_time': 8.5, # 8.5 hours

[0403] 'operation_type': 2, # Type 2 operation

[0404] 'environment_dust': 150,# 150μg / m³

[0405] 'power_consumption': 450# 450W

[0406] }

[0407] # Generate density map

[0408] density_map = generate_density_map(test_data)

[0409] print("Equipment usage intensity density map shape:", density_map.shape)

[0410] print("Maximum usage intensity:", density_map.max())

[0411] print("Average usage intensity:", density_map.mean())

[0412] In this code implementation, the CSRNetForEquipment class defines the improved CSRNet model structure, which mainly includes the following features:

[0413] 1. Multi-channel input: The model receives 8 different device status parameters as input channels, including usage frequency, temperature, humidity, etc.

[0414] 2. Front-end network: Use the modified VGG16 structure to extract features, remove the fully connected layer, and retain the convolutional layer part

[0415] 3. Backend network: Use dilated convolutional network for feature fusion, and the dilation rate is 2 to increase the receptive field

[0416] 4. Data preprocessing: The EquipmentDataProcessor class is responsible for converting raw equipment data into a format that can be processed by the model.

[0417] The following is a detailed description of the above code:

[0418] 1. Model architecture design

[0419] The overall architecture of the CSRNetForEquipment model is divided into three main parts: the front-end feature extraction network, the back-end dilated convolutional network, and the output layer. This design draws on the successful experience of the original CSRNet in crowd density estimation and is specially optimized for the characteristics of surveying and mapping equipment.

[0420] The front-end network uses a modified VGG16 structure, which contains a combination of multiple convolutional layers and pooling layers. The specific configuration is:

[0421] [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512]

[0422] The numbers represent the number of output channels of the convolutional layer, and 'M' represents the maximum pooling layer. This configuration can extract deep features of the device status layer by layer.

[0423] The backend network adopts dilated convolution design:

[0424] [512, 512, 512, 256, 128, 64]

[0425] By using dilated convolution with a dilation rate of 2, the receptive field can be expanded without increasing the number of parameters, and the spatial dependencies of device states can be better captured.

[0426] 2. Data preprocessing mechanism

[0427] The EquipmentDataProcessor class implements data preprocessing functions, mainly processing 8 types of equipment status parameters:

[0428] Usage frequency (usage_frequency): record the number of times the device is used every day; temperature (temperature): device operating temperature; humidity (humidity): environmental humidity level; vibration (vibration): vibration intensity when the device is running; operation time (operation_time): duration of a single use; operation type (operation_type): different types of operation modes; environment dust (environment_dust): dust concentration in the environment; power consumption (power_consumption): energy consumption of the device;

[0429] For each parameter, the preprocessor defines a specific normalization range, ensuring that all input data is mapped to the [0,1] interval:

[0430] ```Python

[0431] normalization_params = {

[0432] 'usage_frequency': {'min': 0, 'max': 100},

[0433] 'temperature': {'min': -20, 'max': 80},

[0434] 'humidity': {'min': 0, 'max': 100}, ...

[0435] }

[0436] 3. Density map generation process

[0437] The generate_density_map function shows the complete density map generation process:

[0438] 1) First load the pre-trained model parameters;

[0439] 2) Create a data processor instance;

[0440] 3) Convert raw device data into a standard format;

[0441] 4) Generate a density map through model forward propagation;

[0442] The test data used in the example reflects the typical usage scenario of a surveying and mapping device:

[0443] The daily usage frequency is 75 times, indicating that the usage intensity is high; the operating temperature is 45℃, which is within the normal working temperature range; the ambient humidity is 65%, which is within the normal range; the vibration intensity is 3.5g, indicating that the equipment is running stably; the daily operation time is 8.5 hours, which is the typical working time on a working day; the ambient dust concentration is 150μg / m³, which is within the acceptable range; the power consumption is 450W, indicating that the equipment is in normal working condition;

[0444] 4. Practical application value

[0445] The practical application value of this model is mainly reflected in:

[0446] Real-time monitoring: Ability to generate density maps of equipment usage intensity in real time, intuitively displaying the equipment usage status;

[0447] Anomaly detection: By analyzing the high-value areas of the density map, potential overuse problems can be discovered in a timely manner;

[0448] Predictive maintenance: Combine historical data to predict possible equipment failures and guide the formulation of maintenance plans;

[0449] Usage optimization: Based on density map analysis, optimize the use and scheduling of equipment to improve resource utilization efficiency;

[0450] 5. Usage suggestions

[0451] In actual use, it is recommended that:

[0452] 1) Regularly update model parameters to adapt to changes in device usage patterns;

[0453] 2) Adjust the normalization parameter range based on actual usage scenarios;

[0454] 3) Establish a density map threshold warning mechanism to issue a timely warning when the usage intensity exceeds the threshold;

[0455] 4) Combine the density map analysis results with other monitoring indicators to achieve multi-dimensional status assessment;

[0456] Through this deep learning-based method, the system can more accurately evaluate the usage status of surveying and mapping equipment and provide strong support for the intelligent management of equipment.

[0457] Through this design, the model can generate a density map that reflects the intensity of equipment use, where high-value areas represent periods or state combinations with high equipment use intensity. This information is of great guiding significance for equipment maintenance planning and use optimization. For example, when the density map shows that the use intensity is abnormally high in certain periods or state combinations, the system will issue a warning in a timely manner and recommend maintenance inspections or adjustments to the use plan.

[0458] This deep learning-based equipment status assessment method has stronger adaptability and predictive capabilities than traditional rule-based methods, and can better serve the intelligent management of surveying and mapping equipment.

[0459] like Figure 7 As shown, the density map shows the distribution of device usage intensity with different colors.

[0460] The state evaluation module 15 also infers the device state according to the preset fuzzy rules. Figure 8 As shown in the figure, the system establishes a fuzzy set of three dimensions: frequency of use, temperature, and device status. When the device is used at a high frequency (more than 61 times) and the temperature is too high (more than 71°C), the system infers that it is in a state of severe wear; when the device is used at a medium frequency (31-60 times) and the temperature is high (51-70°C), the system infers that it is in a state of medium wear; when the device is used at a low frequency (0-30 times) and the temperature is normal (0-50°C), the system infers that it is in a state of slight wear.

[0461] like Figure 5 As shown, the traceability module 16 uses a non-iterative incentive coordination method and a causal reasoning engine to perform traceability analysis.

[0462] Among them, the non-iterative incentive coordination method is an efficient multi-objective optimization method that is suitable for quickly converging to the optimal solution without multiple iterations. It coordinates multiple incentive sources at one time, allowing the system to quickly adjust various goals in a complex environment and reduce computing resource consumption. When used in a traceability system, the incentive source can represent multiple state changes of the device (such as loan, transfer, and return), thereby helping to quickly locate key nodes in the flow of equipment.

[0463] Causal reasoning engine: Causal reasoning is a reasoning method based on causal relationship analysis. It analyzes the causal chain between events to find the root cause of an event. In device traceability, the causal reasoning engine is used to identify the causal relationship from one state (such as loaned) to another state (such as damaged), thereby helping to locate the root cause of device damage or abnormality.

[0464] The specific steps include:

[0465] 1. Using a non-iterative excitation coordination method to simultaneously coordinate multiple state excitation sources of the surveying and mapping device to obtain state change characteristics of the surveying and mapping device;

[0466] Definition of multi-state excitation source: In equipment traceability, the excitation source can be multiple dimensions of equipment status (such as equipment usage frequency, equipment return status, whether it is damaged, etc.). Through the non-iterative excitation coordination method, the system coordinates multiple state excitation sources in the equipment flow process at one time to quickly determine the state change node of the equipment in the flow.

[0467] Example: The equipment is borrowed, loaned, returned, etc. many times, and each state can be regarded as a source of motivation.

[0468] Fast state convergence: The non-iterative incentive coordination method can quickly aggregate the various states of the equipment in the complex flow process through one coordination, and determine which states or events are related to equipment damage and abnormalities. For example, whether the high frequency of equipment borrowing is directly related to abnormal state.

[0469] Example: The system can identify which borrowing or lending of a device during its circulation process may be the key node that may cause it to be damaged.

[0470] 2. Based on the state change characteristics, a causal relationship chain between the state change of the surveying and mapping equipment and the damage event is established through a causal reasoning engine;

[0471] Causal relationship definition: The system establishes a causal relationship model for the equipment circulation process, defining the causal chain between different equipment states and damage or abnormality. For example, the possible causal relationship between multiple borrowing and frequent transfer of equipment and equipment damage.

[0472] Model input: The system inputs the equipment status data and the results of the non-iterative incentive coordination output into the causal reasoning engine to derive the possible cause chain of equipment damage.

[0473] Cause and effect example:

[0474] Reason 1: The device is damaged after being loaned out more than 10 times;

[0475] Reason 2: The device was not returned on time after a loan, resulting in an abnormal status;

[0476] Reason 3: The equipment was damaged due to improper use (such as operation in a high temperature environment) during the borrowing process.

[0477] Reasoning example: The equipment may be damaged due to a loan that was not returned on time, which may have caused the equipment to be exposed to harsh environments for a long time without supervision. The causal reasoning engine will link this event to the ultimate cause of the equipment damage.

[0478] 3. Generate a device traceability report based on the causal chain.

[0479] Contents of the traceability report:

[0480] Device No.: 001

[0481] Status records: loaned (5 times), lent (3 times), returned (1 time delayed), damaged;

[0482] Key reason: Because the equipment was not returned on time after being loaned out, it was exposed to an inappropriate environment, increasing the probability of damage.

[0483] The subsequent system can also use the causal reasoning results to predict the future usage status of the device and determine whether it is likely to have abnormal status. For example, if a device is frequently borrowed and has a history of delayed return, the system will infer that the device is likely to be damaged or abnormal in the future.

[0484] For example, if a device has been loaned out more than 10 times and its temperature was too high during the most recent use, the system predicts that the probability of damage is 30%, and recommends that the administrator check or perform maintenance in advance.

[0485] Finally, the system combines the results of causal reasoning analysis and abnormal state prediction to automatically generate equipment maintenance suggestions and present them to the administrator through the interface.

[0486] For example, if equipment number 001 has been frequently borrowed recently and was not returned in time the last time it was used, it is recommended to arrange maintenance immediately to avoid further damage to the equipment.

[0487] Case 1: Total Station Traceability and Maintenance

[0488] Data collection: The status data of the total station such as borrowing, returning and transferring are collected in real time through sensors and management systems.

[0489] Application of the non-iterative incentive coordination method: The system coordinates the borrowing and lending status of the total station at one time, and identifies that a certain lending operation may be the key node that causes damage to the total station.

[0490] Analysis by the causal reasoning engine: The causal reasoning engine identified that a delay in a loan caused the equipment to be exposed to a high temperature environment, and inferred that the loan was the root cause of the damage to the total station.

[0491] Output result: The system generates a traceability report, indicating that the damage to the equipment was caused by environmental factors after the transfer, and provides maintenance recommendations.

[0492] Case 2: UAV equipment status tracking and prediction

[0493] Data collection: The drone’s flight time, environmental conditions, borrowing records and other data are collected and recorded in real time.

[0494] Application of the non-iterative incentive coordination method: The system quickly analyzes the state changes of the drone and identifies that its multiple high-frequency borrowing is the source of the incentive that causes the abnormal state.

[0495] Application of causal reasoning engine: The reasoning engine determined that the frequent borrowing of the drone caused increased internal battery loss and that failures might occur in the future.

[0496] Output: The system predicts that the drone will fail 30% of the time in the next 10 uses and recommends an immediate battery check.

[0497] Wherein, the non-iterative excitation coordination method is used to simultaneously coordinate multiple state excitation sources of the surveying and mapping equipment, including:

[0498] Defining the state stimulus source of the surveying and mapping equipment, including usage frequency, return status, and damage status;

[0499] Coordinate the multiple state excitation sources at one time to determine the state change nodes of the surveying and mapping equipment in the flow;

[0500] The state change node is associated with the abnormal state of the surveying and mapping equipment for analysis.

[0501] The system first defines the state excitation sources of the equipment, including usage frequency, return status, damage status, etc. By coordinating multiple state excitation sources at one time, the system obtains the state change characteristics of the equipment and determines the key state change nodes. Based on the state change characteristics, the causal reasoning engine establishes a causal relationship chain between the equipment state change and the damage event, and finally generates a device traceability report.

[0502] In the intelligent traceability management of surveying and mapping equipment, the coordination of state excitation sources and the construction of causal chains are key links. This process involves the one-time coordination of multiple state excitation sources, the extraction of state change features, and the construction of causal chains. The following is a detailed explanation of this process:

[0503] 1. Definition and characteristics of state excitation sources

[0504] The system first defines multiple state stimulus sources, each of which represents a dimension of device state change:

[0505] Frequency of use incentive: reflects the intensity of device usage, including the number of times a day it is used, the duration of a single use, etc.;

[0506] Return status incentives: including on-time return, delayed return, and completeness of return;

[0507] Damage state incentives: including functional damage, appearance damage, performance degradation, etc.;

[0508] Environmental condition incentives: including environmental factors such as temperature, humidity, and vibration;

[0509] Operation standard incentive: reflects whether the user's operation complies with the standard;

[0510] 2. One-time coordination mechanism

[0511] The traditional iterative coordination method requires multiple iterations to reach a stable state, while the present invention adopts an innovative one-time coordination method, which specifically includes:

[0512] a) State vector construction:

[0513] Organize the state data of each stimulus source into a state vector:

[0514] S = [Sf, Sr, Sd, Se, So]

[0515] Among them, Sf is the usage frequency state, Sr is the return state, Sd is the damage state, Se is the environment state, and So is the operation specification state.

[0516] b) Weight matrix design:

[0517] Based on the historical data of the equipment, the weight matrix W between the excitation sources is constructed to reflect the influence relationship between different excitation sources:

[0518] W = [w11 w12 w13 w14 w15]

[0519] [w21 w22 w23 w24 w25]

[0520] [w31 w32 w33 w34 w35]

[0521] [w41 w42 w43 w44 w45]

[0522] [w51 w52 w53 w54 w55]

[0523] Wherein, wij represents the influence weight of the i-th excitation source on the j-th excitation source.

[0524] 3. State change feature extraction

[0525] Based on the state data after one-time coordination, the system extracts state change features through the following steps:

[0526] a) Time series feature extraction:

[0527] Use sliding window method to capture the temporal pattern of state changes;

[0528] Calculate the rate of change and acceleration of state variables;

[0529] Identify state mutation points and gradual trends;

[0530] b) Correlation feature extraction:

[0531] Analyze the correlation between different motivational sources;

[0532] Identify patterns of synchronous changes between excitation sources;

[0533] Extract multi-dimensional state combination features;

[0534] 4. Determination of key state change nodes

[0535] The system uses a multi-criteria decision-making method to determine key state change nodes:

[0536] a) Node importance score:

[0537] Score = α * ΔS + β * R + γ * I

[0538] in:

[0539] ΔS is the amplitude of state change;

[0540] R is the duration of the change;

[0541] I is the degree of influence on other states;

[0542] α, β, and γ are weight coefficients;

[0543] b) Node screening criteria:

[0544] The state change amplitude exceeds the preset threshold;

[0545] The duration of the change meets the minimum duration requirement;

[0546] The impact on other states is significant;

[0547] 5. Causal chain construction

[0548] Based on the identified key nodes, the system builds a causal chain:

[0549] a) Direct causal relationship:

[0550] Analyze the timing relationship between adjacent state change nodes;

[0551] Calculate state transition probability;

[0552] Establish a state transfer rule base;

[0553] b) Indirect causation:

[0554] Use graph structures to represent complex causal relationships;

[0555] Identify potential causal pathways;

[0556] Calculate causal strength;

[0557] 6. Generate traceability report

[0558] The system generates a detailed traceability report based on the causal chain, including:

[0559] Timeline of key status change nodes;

[0560] The triggers and effects of state changes;

[0561] Responsibility attribution analysis;

[0562] preventive advice;

[0563] This state analysis method based on one-time coordination has the following advantages over traditional methods:

[0564] 1. High computational efficiency: multiple iterations are avoided, greatly improving processing speed;

[0565] 2. Strong real-time performance: able to quickly respond to status changes and detect anomalies in a timely manner;

[0566] 3. High accuracy: The accuracy of status assessment is improved through comprehensive analysis of multi-dimensional data;

[0567] 4. Good explainability: can clearly show the cause and effect relationship of state changes, which is easy to understand and make decisions;

[0568] This method is particularly suitable for the intelligent management of surveying and mapping equipment, and can effectively support predictive maintenance and risk prevention of equipment. By accurately identifying and analyzing key status change nodes, the system can detect potential problems in advance, so as to take corresponding preventive measures and improve the efficiency and reliability of equipment management.

[0569] In another specific embodiment, the one-time coordination mechanism realizes the synchronous coordination of multiple state excitation sources by constructing a global state mapping matrix (GSMM). This mechanism is different from the traditional iterative coordination method. It can complete the coordination of all state excitation sources in a single calculation, greatly improving the calculation efficiency and real-time performance.

[0570] Specifically, the one-time coordination mechanism works as follows:

[0571] First, after constructing the state vector S and weight matrix W, the system introduces the coordination matrix C:

[0572] C = W * S + λ * D

[0573] in:

[0574] W is the weight matrix, reflecting the influence relationship between various excitation sources

[0575] S is the state vector, which contains the current states of all excitation sources

[0576] λ is a balancing factor used to adjust the coordination strength

[0577] D is the bias matrix, used to introduce prior knowledge

[0578] Then, the system processes the coordination matrix using a nonlinear activation function f:

[0579] S' = f(C)

[0580] Where f can be selected as Sigmoid or Tanh function to map the coordination result to a suitable value range. This step can effectively handle state extreme values ​​and outliers.

[0581] Finally, the system retains the original state information through the residual connection mechanism:

[0582] S_final = α * S' + (1-α) * S

[0583] Where α is the residual coefficient, which is used to balance the coordinated state and the original state.

[0584] Through this innovative coordination mechanism, the system can more accurately and efficiently capture the changing characteristics of equipment status, providing a reliable data basis for subsequent causal analysis. This method is particularly suitable for the management and monitoring of complex systems such as surveying and mapping equipment with multi-dimensional status characteristics.

[0585] Through the implementation of the above technical solution, the present invention realizes the intelligent traceability management of surveying and mapping equipment, improves the equipment management efficiency, enhances the traceability capability, realizes predictive maintenance, and improves the resource utilization rate. The specific implementation of the present invention is not limited to the above embodiment. Those skilled in the art can make various equivalent substitutions and modifications based on the technical solution of the present invention, and these equivalent substitutions and modifications also belong to the scope of protection claimed by the present invention.

Claims

1. A method for intelligent traceability management of surveying and mapping equipment, characterized in that: include: Generate identification information of surveying and mapping equipment; Collecting status information of the surveying and mapping equipment, wherein the status information includes usage status and user information of the surveying and mapping equipment; Performing loan management on the surveying and mapping equipment, including updating status information of the surveying and mapping equipment and recording loan information; Performing loan management on the surveying and mapping equipment, including updating user information of the surveying and mapping equipment and recording loan information; Performing return management on the surveying and mapping equipment, including confirming the return status of the surveying and mapping equipment and recording usage information; After collecting the status information of the surveying and mapping equipment, the method further includes: Performing finite state transduction modeling on the state of the surveying and mapping equipment and establishing state transfer rules; Based on the state transition rule, marking the historical usage data of the surveying and mapping equipment as a state sequence; Inputting the state sequence into a Bayesian inference model to calculate the damage probability of the surveying and mapping equipment; Generate a predictive maintenance report based on the damage probability of the surveying and mapping equipment, and dynamically adjust the maintenance cycle of the surveying and mapping equipment; The method further comprises: Collecting real-time usage data of the surveying and mapping equipment, including usage frequency, environmental conditions, and operating information; Inputting the usage data into a CSRNet model to generate a density map of the usage intensity of the surveying and mapping equipment; Performing fuzzy reasoning on the state of the surveying and mapping equipment according to preset fuzzy rules; Based on the density map and the fuzzy reasoning result, outputting a status assessment result of the surveying and mapping equipment; Wherein, after the surveying and mapping equipment is returned and managed, the method includes: A non-iterative excitation coordination method is used to simultaneously coordinate multiple state excitation sources of the surveying and mapping device to obtain state change characteristics of the surveying and mapping device; Based on the state change characteristics, a causal relationship chain between the state change of the surveying and mapping equipment and the damage event is established through a causal reasoning engine; Generate a device traceability report based on the causal chain.

2. The method according to claim 1, characterized in that: The method further comprises: storing usage data of the surveying and mapping equipment, wherein the usage data includes loan time, user, usage frequency, and return time; Perform statistical analysis on the usage data and generate a usage report of the surveying and mapping equipment.

3. The method according to claim 1, characterized in that The calculating the damage probability of the surveying and mapping equipment comprises: The damage probability of the surveying and mapping equipment is calculated according to the following Bayesian formula: P(damage|number of loans) = P(number of loans|damage) * P(damage) / P(number of loans) in, P(damage|number of loans) is the probability that the surveying and mapping equipment is damaged after being loaned a certain number of times; P(number of loaned out | damage) is the probability of damage when the surveying and mapping equipment described in the historical data is loaned out a specific number of times; P(damage) is the a priori probability of damage of the surveying and mapping equipment; P (number of loaned times) is the total probability that the surveying and mapping equipment is loaned a certain number of times.

4. The method according to claim 1, characterized in that The finite state transduction modeling of the state of the surveying and mapping equipment and the establishment of state transfer rules include: Define the basic status of the surveying and mapping equipment, including storage status, loan status, loan status, return status and damage status; Define the transition conditions between states, including loan triggering the transition from storage to loan, loan triggering the transition from loan to loan, and return triggering the transition from loan or loan to return; Define the tokenization rules for state transfer to ensure that each state transfer process is fully recorded.

5. The method according to claim 1, characterized in that The preset fuzzy rules include: The fuzzy set of the use frequency of the surveying and mapping equipment includes low frequency: 0-30 times, medium frequency: 31-60 times, and high frequency: more than 61 times; The fuzzy set of the temperature of the surveying and mapping equipment includes normal temperature: 0-50°C, high temperature: 51-70°C, and excessive temperature: above 71°C; The fuzzy set of states of the surveying and mapping equipment includes light wear, medium wear, and heavy wear; Performing fuzzy reasoning on the state of the surveying and mapping equipment includes: When the surveying and mapping equipment is used at a high frequency and the temperature is too high, a severe wear state is inferred; When the surveying and mapping equipment is used at a medium frequency and the temperature is relatively high, a medium wear state is inferred; When the usage frequency of the surveying and mapping device is low and the temperature is normal, a slight wear state is inferred.

6. The method according to claim 1, characterized in that The method of using a non-iterative excitation coordination method to simultaneously coordinate multiple state excitation sources of the surveying and mapping equipment includes: Defining the state stimulus source of the surveying and mapping equipment, including usage frequency, return status, and damage status; Coordinate the multiple state excitation sources at one time to determine the state change nodes of the surveying and mapping equipment in the flow; The state change node is associated with the abnormal state of the surveying and mapping equipment for analysis.

7. An intelligent traceability management system for surveying and mapping equipment, characterized in that: include: An identity module, used to generate identity information for the surveying and mapping equipment and store the identity information in a database; A status monitoring module, used to collect status information of the surveying and mapping equipment, and manage the lending, re-lending and returning of the surveying and mapping equipment; A data processing module, used to receive the status information collected by the status monitoring module, store and analyze the usage data of the surveying and mapping equipment, generate a usage report, and send the usage report to the predictive maintenance module; A predictive maintenance module, configured to receive the usage report, calculate the damage probability of the surveying and mapping equipment based on a Bayesian reasoning model, generate a predictive maintenance report, and send the predictive maintenance report to a condition assessment module; A status assessment module, used for receiving the predictive maintenance report, monitoring the status of the surveying and mapping equipment based on the CSRNet model and fuzzy logic technology, generating a status assessment result, and sending the assessment result to the traceability tracking module; A traceability module is used to receive the status assessment result, track the traceability target of the equipment based on the non-iterative incentive coordination method and the causal reasoning engine, and generate a device traceability report; Among them, after collecting the status information of the surveying and mapping equipment, the intelligent traceability management system for surveying and mapping equipment is also used to: perform finite state transduction modeling on the status of the surveying and mapping equipment and establish state transition rules; based on the state transition rules, mark the historical usage data of the surveying and mapping equipment as a state sequence; input the state sequence into the Bayesian reasoning model to calculate the damage probability of the surveying and mapping equipment; generate a predictive maintenance report based on the damage probability of the surveying and mapping equipment, and dynamically adjust the maintenance cycle of the surveying and mapping equipment; collect real-time usage data of the surveying and mapping equipment, including usage frequency, environmental conditions, and operation information; input the usage data into the CSRNet model to generate a density map of the usage intensity of the surveying and mapping equipment; perform fuzzy reasoning on the status of the surveying and mapping equipment according to preset fuzzy rules; and output the status evaluation result of the surveying and mapping equipment based on the density map and the fuzzy reasoning result; After the surveying and mapping equipment is returned, the intelligent traceability management system for surveying and mapping equipment is also used to: A non-iterative excitation coordination method is used to simultaneously coordinate multiple state excitation sources of the surveying and mapping device to obtain state change characteristics of the surveying and mapping device; Based on the state change characteristics, a causal relationship chain between the state change of the surveying and mapping equipment and the damage event is established through a causal reasoning engine; Generate a device traceability report based on the causal chain.

Images