Intelligent counting and locking method and system for limited space operation

Abstract

The invention discloses a checking and locking method and system for limited space operation. The method comprises the following steps: acquiring a reference image of a space cleaning state before operation; starting an ending process after operation, and sequentially executing personnel counting, tool counting, field cleaning verification based on comparison with a reference image and entrance locking verification; comprehensive judgment is carried out according to the verification result, only when all steps are passed, an authorization token containing a digital signature is generated and sent to the electronic work ticket system, and closing limitation of the electronic work ticket system is relieved; if not, a locking signal is sent to prevent the operation ticket from being closed, and correction guidance is generated. According to the method, the physical security verification result is forcibly associated with the electronic work ticket process through a digital signature mechanism, personnel or tools are effectively prevented from being left behind, and a traceable digital record is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial safety production intelligent management and control, in particular to a checking and locking method and system for verifying the end process of limited space operation by technical means and performing safety linkage with the operation permission state. BACKGROUND

[0002] Limited space operation, such as operation inside tanks, towers, kettles, tanks, pipelines, boilers, manholes, etc., is one of the operation types with higher risk level in industrial production. The internal environment is complex, and in addition to the risks during the operation process, there are also major safety hazards in the end stage of the operation. The following challenges often exist in the existing safety management practices:

[0003] Personnel left behind: After operation in a large or complex structure limited space, relying on manual roll call or oral confirmation, there is a major risk that the operation personnel are forgotten in the space and accidentally closed.

[0004] Tools and equipment left behind: tools, equipment, materials carried by the operation personnel are left behind in the limited space, which may cause mechanical damage when the equipment is restarted, block the pipeline, or even trigger a chemical reaction or become a source of ignition, posing a serious production safety hazard.

[0005] Ineffective closure of the entrance: after the operation, the manhole door, cover plate and other physical entrances are not closed in a timely and effective manner as required, which may lead to the entry of unrelated personnel or the entry of external harmful substances, forming a new source of danger.

[0006] Incomplete site cleaning: waste, packaging materials, temporary facilities generated during the operation process are not thoroughly cleaned, which may become a source of hidden dangers such as fire, blockage or tripping. If the site cleaning verification is simply dependent on a target detection model, false positives or false negatives may occur due to the complex structure of the limited space, such as inherent pipelines and supports, and the reliability is insufficient.

[0007] The traditional safety management mode mainly relies on paper or electronic operation tickets and manual safety checklists. After the operation, the guardian or operation responsible person manually checks and confirms the corresponding items. The effectiveness of this method highly depends on the responsibility and memory of the execution personnel, and lacks objective technical verification means and mandatory constraint mechanisms.

[0008] More importantly, even if the enterprise introduces an independent digital personnel positioning system, a tool management system or an AI video monitoring system, these systems usually only provide data query and alarm functions, and the verification results lack automatic and mandatory technical association with the electronic work ticket system as the key control point of the work process. This means that the guardian can still perform the "close" operation in violation of the electronic work ticket system based on subjective judgment or under pressure, causing the last line of defense of safety management to be bypassed. The separation of this management process and the physical safety state in technology means that the digital achievements of safety verification cannot form mandatory constraints on the closing process of work permission, and may cause safety risks due to human negligence or illegal operations. SUMMARY

[0009] Technical problem to be solved: The present application provides a counting and locking method and system for limited space work, aiming to solve the safety hazards caused by relying on manual confirmation in the prior art, and the technical problem that the safety verification system and the work permission management system are separated from each other and cannot form a technical closed loop.

[0010] Technical solution:

[0011] To solve the above problems, the present application provides a computer-implemented method for counting and locking for limited space work. The method establishes an intelligent closed-loop management mechanism that technically associates the verification results of multiple key physical safety verification steps in the closing stage of limited space work with the life cycle state of the electronic work ticket through a digital signature mechanism.

[0012] The method starts a closing counting process after receiving a work completion instruction. Before the work starts, the system collects and stores baseline image data representing the "clean and initial" state of the limited space after confirming that the space has been cleaned. The closing process includes personnel counting, tool counting, site cleaning verification and entrance locking verification. Among them, the site cleaning verification identifies new objects generated by work activities as leftovers by comparing and analyzing the post-work site image with the baseline image data before the work, which helps to improve the accuracy of detection.

[0013] The system comprehensively judges the results of all the above steps, and the judgment result is passed only when all the steps are verified. If the judgment is passed, the system generates a digital signature-based authorization token containing the timestamp and the work order identifier using its own private key, and sends it to the electronic work order management system. The electronic work order management system can only remove the closing operation restriction on the work order after successfully verifying the token using the pre-set public key. If any step fails to verify, the system sends a locking signal to the electronic work order management system to forcibly prevent it from closing. At the same time, the system analyzes the specific reasons for the failure and sends clear correction guidance (e.g., "Person 'Zhang San' has not been counted" or "A suspected wrench is found at image coordinates (x, y)") to the mobile terminal of the on-site personnel, guiding them to complete the rectification and re-verify, forming a closed-loop management process of "finding problems - intelligent guidance - correction - re-verification".

[0014] At the same time, the method also associates and stores all the verification data, image and video evidence, correction records, and authorization tokens generated in each verification step with the unique identifier of the electronic work order, forming a complete and traceable digital record.

[0015] Advantages

[0016] Compared with the prior art, the technical scheme of the present application has the following advantages:

[0017] 1. Establishes a technical closed loop and security association: The present application introduces a digital signature authorization token mechanism to convert the passing result of physical safety verification into a technical prerequisite for closing the electronic work order based on encryption technology. This helps to avoid the possibility of human bypassing safety checks technically, tightly binds the physical safety state and the management process, and solves the problem of the separation of the management process and the physical safety state in the prior art.

[0018] 2. Improves the accuracy and objectivity of verification: By using computer vision analysis methods based on reference image comparison, the false positive and false negative problems that may occur in complex environments using traditional target detection can be effectively addressed, improving the accuracy and reliability of on-site cleaning verification. Combined with face recognition, identification scanning and other technologies, subjective manual confirmation is converted into objective machine recognition and judgment, and the result is more reliable.

[0019] 3. Realizes intelligent guidance type closed loop error correction: When verification fails, the present application not only alarms and locks, but also provides correction guidance accurate to the specific personnel, tools or positions of the remaining objects. This intelligent closed-loop management mode of "finding - guiding - correcting - re-verifying" improves the efficiency of on-site problem disposal and the compliance of work completion, enabling the system to actively guide on-site rectification.

[0020] 4. Form a detailed, traceable record: Store the data generated by all verification steps, such as images, videos, identification results, correction records, and authorization tokens containing digital signatures, in association with electronic work tickets and with time stamps, forming a complete, reliable, and auditable digital record that provides objective evidence for safety audits or accident investigations. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced as follows.

[0022] Figure 1 The flowchart of the inventory and locking method for limited space operation in an embodiment of the present application.

[0023] Figure 2 The structural block diagram of the inventory and locking system for limited space operation in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments.

[0025] In the description of the present application, some terms are defined as follows:

[0026] Limited space: refers to a space that is closed or partially closed, relatively isolated from the outside world, has a narrow entrance and exit, and workers cannot work in it for a long time.

[0027] Mobile terminal: can be a handheld or wearable device including a processor, an image acquisition unit, and a communication module, such as an industrial-grade explosion-proof smart phone, a tablet computer, or AI glasses.

[0028] Designated personnel: refers to personnel who are assigned the responsibility of final verification according to safety regulations, such as an operation guardian or an operation supervisor.

[0029] Electronic work ticket management system: refers to a software system for electronic whole-life cycle management of industrial site operation permits.

[0030] Embodiment one

[0031] This embodiment details a kind of inventory and locking method for limited space operation as shown in Figure 1 The method is executed in a computer system (backend server), which interacts with field personnel through a mobile terminal and communicates with an external electronic work ticket management system through a standard API interface.

[0032] Preparation phase: Collect baseline images

[0033] Before the official start of the limited space operation, designated personnel first confirm that the inside of the limited space has been cleaned and there is no any operation-related tools, materials or waste. After confirmation, use the mobile terminal to take multi-angle photos or record a panoramic video of the inside of the limited space that has been cleaned and is ready for operation. These image or video data are uploaded to the backend server as baseline image data representing the "clean and initial state" of the space, and are stored in association with the electronic work order to be generated.

[0034] S1: Receive operation completion instruction and start end-checking process.

[0035] The site operation leader clicks "Apply for completion" through the application on his mobile terminal. The instruction is sent to the electronic work order management system, which immediately sends a work completion instruction containing the unique identifier of the work order to the backend server of the present scheme. After the backend server listens to the instruction, it officially starts the end-checking process and pushes the task instruction to the mobile terminal of the designated personnel.

[0036] S2: Perform personnel checking step.

[0037] The designated personnel enter the "personnel checking" interface on the mobile terminal. The system first obtains the "preset entering personnel list" associated with the present operation from the electronic work order management system. The designated personnel uses the mobile phone camera to perform face scanning on each of the operation personnel who has left. The face recognition engine in the application matches the collected face with the personnel information in the list in real time to form a list of personnel who have left. When all personnel in the preset entering personnel list are identified and confirmed, the system determines that the verification result of this step is passed.

[0038] S3: Perform tool checking step.

[0039] The system obtains the "preset tool list" from the electronic work order management system. The designated personnel use the camera or NFC sensing area of the mobile terminal to scan the two-dimensional code or RFID / NFC tag on each of the tools taken out one by one to form a list of tools taken out. When all tools in the preset tool list are scanned and confirmed, the system determines that the verification result of this step is passed.

[0040] As an alternative or supplementary scheme, the tool checking step (S3) can be combined with the personnel checking step (S2) for execution. When the designated personnel uses the mobile terminal to identify the face of a person to confirm that the person has left, the application interface can simultaneously display the tool list under the name of the person or preset for the present operation, and the designated personnel can check and confirm which tools the person has actually taken out. The system background aggregates all the tools confirmed to be taken out by the personnel to form a list of tools taken out.

[0041] The skilled person can understand that if some tools are damaged only partially during the operation, the designated personnel can make special marks on the mobile terminal and take photos for evidence. The system will record this mark and consider it as a special item that has been processed during the tool inventory step, without triggering the inventory failure alarm.

[0042] S4: Perform the site cleaning verification step.

[0043] Enter the "site verification" interface in the application. The designated personnel use the mobile terminal to take photos or record videos inside the limited space, forming current state images or video data and uploading them. The AI analysis engine of the backend server receives the data and compares it with the baseline image data stored before the operation. The comparison analysis aims to identify new objects that have appeared relative to the baseline image and do not belong to the inherent structure of the space (such as fixed pipes, supports), and determine them as leftovers. If the AI analysis engine determines that no such newly appearing objects are found in the current state image, the verification result of this step is passed.

[0044] To improve the robustness of the computer vision analysis model, the comparison analysis can further include image preprocessing steps such as illumination normalization, perspective transformation correction, etc. to eliminate interference caused by different shooting conditions. In addition, the comparison analysis can not be limited to pixel-level differences, and more preferably uses deep learning-based feature vector comparison to compare deep semantic features of images, thereby more accurately identifying objects while ignoring non-critical background changes.

[0045] S5: Perform the entrance locking verification step.

[0046] After the designated personnel physically close and lock the entrance (e.g., tighten all manhole cover bolts), they use the mobile terminal to take a photo of the locked entrance and upload it. The AI analysis engine of the backend server analyzes the photo using an image recognition model to determine whether the entrance is in the preset locked state (e.g., all bolts are in place and tightened). If it is determined to be in compliance, the verification result of this step is passed. Alternatively, the system can also complete the verification by receiving a "locked" state signal from the smart lock installed at the entrance.

[0047] S6: Comprehensive judgment.

[0048] The comprehensive judgment and decision module of the backend server performs logical judgment on the verification results of S2, S3, S4, and S5. Only when the verification results of all four steps are "passed", the result of the comprehensive judgment is "passed".

[0049] S7: If all are passed, send a digital signature authorization token.

[0050] If the comprehensive judgment result is "pass", the comprehensive judgment and decision module will call the encryption module inside the server, use the pre-stored server private key to digitally sign a data packet containing the unique identifier of the electronic work order, the current timestamp and the "authorized closing" instruction, and generate a time-limited authorization token (such as JWT format). The authorization token is sent to the electronic work order management system through the API. The electronic work order management system uses the pre-set server public key to decrypt and verify the token. After successful verification, the system automatically removes the closing operation restriction on the work order, and the work order manager can perform the final closing operation on the interface.

[0051] S8: If any step fails, send a lock signal and perform intelligent guidance.

[0052] If the comprehensive judgment result is "fail", for example, in S4, the AI comparison finds a new object. The backend server immediately sends a lock signal to the electronic work order management system, making it remain or enter a locked state, and the "close" button of the work order will remain unavailable.

[0053] At the same time, the intelligent guidance and alarm module will generate specific correction guidelines and push them to designated personnel through mobile terminals. For example:

[0054] Personnel counting fails: "Alarm: Personnel counting is not completed. According to the records, work personnel 'Li Si' (employee ID: 12345) has not been identified. Please confirm his position and safety status."

[0055] On-site cleaning verification fails: "Alarm: A suspected left-over object is found on site. AI analysis found a new object near the coordinate area (342, 512) in image file IMG_5678.jpg, suspected to be a'red glove'. Please send personnel to clean up and re-execute the on-site verification."

[0056] After the on-site personnel complete the rectification according to the guidelines, they need to re-execute the verification steps that failed until all steps are verified.

[0057] S9: Associate storage, form records.

[0058] All verification data generated by the steps, including personnel identification records, tool and equipment inventory lists, on-site verification baseline images and current state images, AI analysis results, entrance lock photos, and each failure alarm and correction record, the final generated authorization token, etc., are strongly associated with the unique identifier of the electronic work order and stored in the database in encrypted form, forming a complete and traceable digital record.

[0059] It is understood by those skilled in the art that in the case of poor network signal, the mobile terminal can be configured to complete all the inventory and verification steps locally in advance, cache all the data and preliminary results. After the network is restored, the data is synchronized with the processing and decision unit, and the latter completes the final comprehensive judgment and sends the authorization token / lock signal.

[0060] Embodiment Two

[0061] The present embodiment provides an inventory and locking system for implementing the above method, which has the structure as shown in Figure 2 The system mainly includes a data acquisition and interaction unit 100 (such as a mobile terminal) and a processing and decision unit 200 (such as a backend server cluster).

[0062] The data acquisition and interaction unit 100 is installed with a special application program (App) and integrates identity information acquisition, tool information acquisition, on-site image acquisition, and man-machine interaction modules.

[0063] The processing and decision unit 200 is deployed in the cloud or enterprise data center and is responsible for core logic processing, intelligent analysis, and decision-making, including:

[0064] The data interface and state management module 201 is responsible for secure API communication with the external electronic work order management system, receives instructions and initial data, manages the state machine of the end process, and is responsible for sending the locking signal and the authorization token after digital signature. This module preferably communicates with the external electronic work order management system through a secure HTTPS-based RESTful API to exchange JSON format data packets. This module contains a cryptography component for generating and verifying digital signatures.

[0065] The comprehensive judgment and decision module 202 is responsible for summarizing the verification results of each verification step, making the final logical judgment, and triggering the generation of the authorization token or the sending of the locking signal according to the judgment result.

[0066] The AI analysis engine 203 is embedded with one or more pre-trained computer vision models. Its core function is to perform comparison analysis based on benchmark image data for on-site residual object detection. In addition, it can also perform image recognition tasks for entrance locking status. This engine can be deployed on a server equipped with a GPU to ensure operation efficiency.

[0067] The data storage and association module 204 is responsible for associating all process data, result data, and evidence files with the electronic work order ID, encrypting and archiving, and building a secure and traceable database.

[0068] The intelligent guidance and warning module 205: when the comprehensive judgment and decision module 202 determines that there is a step that does not pass, this module not only generates a warning, but also calls a preset rule library to generate an operable correction guide for a specific failure cause, and pushes it to the data acquisition and interaction unit 100.

[0069] Through the cooperative work of the above-mentioned units and modules, the embodiment fully realizes the safety management method with the technical closed loop and the intelligent guidance and correction function proposed by the application.

Claims

1. A computer-based method for inventory and locking in confined space operations, characterized in that, The method is run in a computer system, the computer system is connected with an electronic work order management system, and the method comprises: Before the start of the limited space work, after confirming that there is no work-related residual in the limited space, one or more baseline image data representing the cleaning and initial state of the limited space are collected and stored; S1: after receiving a work completion instruction associated with an electronic work order, a closing and counting process is started; S2: a personnel counting step is performed, a preset entering personnel list associated with the electronic work order is obtained from the electronic work order management system, and a left personnel list is formed by collecting the biological characteristics of the personnel leaving the limited space or the identity credentials held by them as identity identification information, it is determined whether each personnel in the preset entering personnel list is included in the left personnel list, and a personnel counting verification result is generated based on the determination result; S3: a tool counting step is performed, a preset tool list associated with the electronic work order is obtained from the electronic work order management system, and a tool list taken out of the limited space is obtained to form a tool list taken out of the limited space, it is determined whether each tool in the preset tool list is included in the tool list taken out of the limited space, and a tool counting verification result is generated based on the determination result; S4: a site cleaning verification step is performed, current state image or video data of the limited space work site is collected, and a computer vision analysis model is used to compare and analyze the current state image or video data with the baseline image data to identify objects that newly appear relative to the baseline image data and do not belong to the inherent structure of the limited space, and the objects are determined as residual, the residual includes one or more of tools, materials, waste or temporary facilities, and a verification result of the site cleaning verification step is generated based on the determination result of whether the residual exists; S5: an entrance locking verification step is performed, current state information of the physical entrance of the limited space provided by an image collection device or a state sensor is obtained, and it is determined whether the physical entrance has been in a preset locked state according to the state information, and a verification result about whether the physical entrance has been locked is generated; S6: according to the verification results generated by the personnel counting, tool counting, site cleaning verification and entrance locking verification steps, a comprehensive judgment is made, wherein the result of the comprehensive judgment is passed only when the verification results of the personnel counting, tool counting, site cleaning verification and entrance locking verification are all passed; S7: if the result of the comprehensive judgment is passed, an authorized token containing a digital signature is generated and sent to the electronic work order management system, and the authorized token is used to release the closing operation restriction of the electronic work order after being verified by the electronic work order management system. S8: If the result of the comprehensive judgment is not passed, a lock signal is sent to the electronic work order management system to prevent the electronic work order from being closed, and instructions containing correction guidelines are generated to guide the site to complete rectification and re-verification according to the steps and specific reasons for not passing; S9: The verification results generated in the personnel counting, tool counting, site cleaning verification and entrance locking verification steps, and the sending records of the authorization token or lock signal are associated with and stored with the unique identifier of the electronic work order to form a traceable digital record.

2. The method of claim 1, wherein, The correction guidelines generated in the S8 step include at least one of the following: When the personnel counting fails, specify the specific personnel name or identity information that is not counted; When the tool counting fails, specify the specific tool name or number that is not counted; When the site cleaning verification fails, specify the suspected location and category of the remaining items in the current state image or video data; When the entrance locking verification fails, specify the specific situation that the physical entrance does not meet the preset locking state.

3. The method of claim 1, wherein, The authorization token generated in the S7 step is generated by signing the data containing the unique identifier of the electronic work order, the timestamp and the authorization instruction with the private key of the computer system; the electronic work order management system verifies the authorization token by using the public key of the computer system.

4. The method of claim 1, wherein, The personnel counting step specifically includes: A designated person uses a mobile terminal to identify all personnel who have left through the face recognition function, forming the list of personnel who have left.

5. The method of claim 1, wherein, The way to obtain the information of the tools taken out of the limited space includes: A designated person uses a mobile terminal to collect the two-dimensional code, bar code or RFID / NFC tag attached to all taken-out tools as an identity, forming the list of taken-out tools.

6. The method of claim 1, wherein, The tool counting step (S3) is combined with the personnel counting step (S2) to execute, and the way to obtain the information of the tools taken out of the limited space includes: When the personnel counting step is executed, in response to an identified left personnel, an operation of checking and confirming on the human-computer interaction interface from the preset list of taken-in tools, and associating the operation record with the identity of the identified left personnel, the checked tools are recorded as taken out; The system background aggregates all recorded tools to form the list of taken-out tools.

7. The method of claim 1, wherein, The computer vision analysis model is an image comparison model that identifies the remaining items by calculating the pixel difference, structural similarity difference or feature vector difference between the current state image data and the reference image data.

8. The method of claim 1, wherein, The way to obtain the current state information of the physical entrance includes: a designated person uses a mobile terminal to collect the current state image of the physical entrance as the state information, and uses an image recognition model to analyze the image to determine whether the locking device of the physical entrance is in the preset locking state.

9. An inventory and lockout system for use in confined space operations, comprising: The system includes: a data collection and interaction unit configured to perform data collection and human-machine interaction at the work site; and a processing and decision unit in communication connection with the data collection and interaction unit and an external electronic work order management system; the processing and decision unit is configured to perform the method of any one of claims 1 to 8; in particular, the processing and decision unit comprises: a data interface and state management module for communicating with the electronic work order management system, receiving work completion instructions and initial inventory associated with the electronic work order, and managing the execution state of the final inventory process; a comprehensive judgment and decision module configured to receive verification results generated by each verification step and make comprehensive judgments; when all verification results are passed, an authorized token containing a digital signature is generated; when at least one verification result is not passed, a lock signal is generated, and the authorized token or lock signal is sent to the electronic work order management system through the data interface and state management module; an AI analysis engine configured to run at least one computer vision analysis model to analyze image or video data obtained from the data collection and interaction unit, thereby performing on-site cleaning verification and entrance locking verification; an intelligent guidance and warning module configured to generate instructions containing correction guidelines according to the specific reason for the failure when the verification result of any step is not passed, and present them to the user through the data collection and interaction unit; a data storage and association module for associating and storing the verification data generated by each verification step, the correction and re-verification records, and the sent authorized token or lock signal with the unique identifier of the electronic work order to form traceable digital records.

Images