Method and system for recognizing wearing state of protective articles of maintenance personnel based on video monitoring

Abstract

This application relates to the field of industrial equipment maintenance technology, and provides a method and system for identifying the protective equipment (PE) wearing status of maintenance personnel based on video surveillance. The method includes a central control module responding to the activation of a maintenance work order, parsing the work order to obtain the corresponding work area, work type identifier, and work time window, and sending the work type identifier and work time window to a target edge analysis module corresponding to the work area. The target edge analysis module activates the video surveillance equipment in the work area at the beginning of the work time window to acquire surveillance images, and identifies the PE wearing status of each worker within the work time window based on the work type identifier and the surveillance images. This method can concentrate limited computing power on work processes where real safety risks exist.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial equipment maintenance, and particularly relates to a method and system for identifying wearing state of protective articles of maintenance personnel based on video monitoring. BACKGROUND

[0002] In the field of gas power generation, the maintenance operation of unit equipment is a key link to ensure safe operation, and whether personnel correctly wear protective articles such as safety helmets and safety belts is directly related to the safety of the operation. In order to strengthen on-site supervision, automatic identification technology based on video monitoring has become an important auxiliary means. However, the existing video identification system usually continuously and indiscriminately analyzes all target objects in the monitoring screen. This mode fails to distinguish between high-risk maintenance operation periods and areas and regular low-risk states, resulting in that a large amount of computing resources is consumed in the analysis of irrelevant scenes, and the limited analysis capability cannot be concentrated on specific operation activities that truly have safety hazards. SUMMARY

[0003] The main purpose of the present application is to provide a method and system for identifying wearing state of protective articles of maintenance personnel based on video monitoring, which aims to solve the problems proposed in the background.

[0004] In a first aspect, the present application provides a method for identifying wearing state of protective articles of maintenance personnel based on video monitoring, which is used in a system for identifying wearing state of protective articles of maintenance personnel based on video monitoring, the system comprising a central control module and a plurality of edge analysis modules, and the method comprising: The central control module responds to the validation of a maintenance operation ticket, analyzes the maintenance operation ticket to obtain an operation area, an operation type identifier and an operation time window corresponding to the maintenance operation ticket, and sends the operation type identifier and the operation time window to a target edge analysis module corresponding to the operation area; The target edge analysis module activates a video monitoring device of the operation area at a starting moment of the operation time window to obtain a monitoring image, and identifies wearing state of protective articles of each operation personnel based on the operation type identifier and the monitoring image within the operation time window.

[0005] In a second aspect, the present application further provides a system for identifying wearing state of protective articles of maintenance personnel based on video monitoring, which comprises a central control module and a plurality of edge analysis modules; The central control module responds to the validation of a maintenance operation ticket, analyzes the maintenance operation ticket to obtain an operation area, an operation type identifier and an operation time window corresponding to the maintenance operation ticket, and sends the operation type identifier and the operation time window to a target edge analysis module corresponding to the operation area; The target edge analysis module activates the video monitoring device of the work area at the starting moment of the work time window to obtain a monitoring image, and identifies the protective article wearing state of each worker based on the work type identifier and the monitoring image within the work time window.

[0006] The embodiment provides a video monitoring-based maintenance personnel protective article wearing state identification method and system, which comprises the following steps: in response to the validity of a maintenance work ticket, a central control module analyzes the maintenance work ticket to obtain a work area corresponding to the maintenance work ticket, a work type identifier, a work time window, and sends the work type identifier and the work time window to a target edge analysis module corresponding to the work area, so as to convert an abstract production work plan into accurate space-time and task instructions that can drive a video analysis system, and help to realize the synchronous start of an analysis task and a production business; and the target edge analysis module activates a video monitoring device of the work area at the starting moment of the work time window to obtain a monitoring image, and identifies the protective article wearing state of each worker based on the work type identifier and the monitoring image within the work time window, so that the input of image analysis and calculation resources is limited to the time and space range in which actual maintenance work activities exist, the huge calculation load caused by indiscriminate analysis of all-weather video streams in the whole factory area in the traditional method is avoided, and limited computing power can be concentrated on work links that are truly at risk. BRIEF DESCRIPTION OF DRAWINGS

[0007] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0008] Figure 1 A flowchart of a video monitoring-based maintenance personnel protective article wearing state identification method provided by an embodiment of the present application is shown in the figure. Figure 2 A structural schematic block diagram of a video monitoring-based maintenance personnel protective article wearing state identification system provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0009] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0010] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0011] This application provides a method and system for identifying the wearing status of protective equipment by maintenance personnel based on video surveillance.

[0012] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for recognizing the wearing status of protective equipment (PE) for maintenance personnel based on video surveillance, provided in an embodiment of this application. The method is used in a video surveillance-based PE wearing status recognition system for maintenance personnel. The system includes a central control module and multiple edge analysis modules, such as... Figure 1 As shown in this embodiment, the method for recognizing the wearing status of protective equipment for maintenance personnel based on video surveillance includes: Step S1: The central control module responds to the activation of the maintenance work ticket, parses the maintenance work ticket, obtains the work area, work type identifier, and work time window corresponding to the maintenance work ticket, and sends the work type identifier and work time window to the target edge analysis module corresponding to the work area.

[0014] The maintenance work ticket refers to an electronic work voucher that has been marked as "permitted" after a complete approval process in the production management system. It exists in a structured data format and includes information such as work location, work content, and planned time. Specifically, the central control module continuously monitors the status change events of maintenance work tickets in the production management system through a predefined application programming interface. When a ticket status changes to active, the central control module receives the complete data packet associated with that ticket and parses it. The parsing process includes verifying the integrity of the data packet format, then extracting the values ​​of specific fields from the data packet according to predefined field mapping rules; extracting the original location description from the field describing the work location and converting it into a work area identifier uniformly used within the system according to a predefined encoding mapping table; extracting the original type description from the field describing the nature or risk of the work and converting it into a work type identifier uniformly used within the system according to a predefined type mapping table; extracting time information from the planned start time and planned end time fields and converting it into a unified timestamp format used within the system to form the work time window; after completing the parsing, the central control module queries a preconfigured region-edge analysis module mapping table based on the obtained work area identifier to determine the network address of the target edge analysis module responsible for the work area; finally, the central control module encapsulates the work type identifier and the work time window into a control command message and sends the message to the command receiving interface of the target edge analysis module through a secure enterprise intranet.

[0015] Step S2: The target edge analysis module activates the video surveillance equipment in the work area at the beginning of the work time window to obtain surveillance images, and identifies the protective equipment wearing status of each worker based on the work type identifier and the surveillance images within the work time window.

[0016] Specifically, after receiving a control command message from the central control module, the target edge analysis module parses out the operation time window. Internally, the module initiates a timed task that, at the start of the operation time window, sends wake-up and control commands to network cameras within its jurisdiction that are related to the current operation area. These commands cause the cameras to adjust to a preset viewing angle and begin outputting video streams. As a video stream client, the target edge analysis module receives real-time video streams from these cameras and decodes them into a continuous sequence of monitoring image frames. Throughout the entire operation time window, the target edge analysis module uses the operation type identifier as input to trigger and execute an integrated identification process. This process uses the operation type identifier to determine the category of protective equipment requiring attention and analyzes the continuously input monitoring images to determine whether each worker in the image is correctly wearing the prescribed protective equipment.

[0017] The method provided in this embodiment, on the one hand, involves the central control module responding to the activation of the maintenance work order, parsing the maintenance work order to obtain the corresponding work area, work type identifier, and work time window, and sending the work type identifier and work time window to the target edge analysis module corresponding to the work area. This transforms the abstract production work plan into precise spatiotemporal and task instructions that can drive the video analysis system, facilitating the synchronous start of analysis tasks and production operations. On the other hand, the target edge analysis module activates the video monitoring equipment in the work area at the beginning of the work time window to acquire monitoring images. Within the work time window, it identifies the protective equipment wearing status of each worker based on the work type identifier and the monitoring images. This limits the investment of image analysis computing resources to the time and space range where actual maintenance work activities occur, avoiding the huge computational load caused by indiscriminate analysis of the entire plant's video stream around the clock in traditional methods. It can concentrate limited computing power on work links where there are real safety risks.

[0018] In some embodiments, the central control module parses the maintenance work ticket to obtain the work area, work type identifier, and work time window corresponding to the maintenance work ticket, including: Step S111: Read the field representing spatial location from the maintenance work ticket, and convert the field into a preset work area identifier according to the preset encoding mapping rule.

[0019] Specifically, in the received maintenance work order, the central control module locates and reads the fields describing the work location. These fields may be named "Equipment Code", "Plant Location", or "Work Location Description". The central control module obtains the raw string values ​​of these fields. Then, it queries a pre-set coding mapping rule table, which stores the correspondence between the raw location description and the unified and simplified work area identifier within the system. Through a lookup and matching, the raw string is mapped to a standardized work area identifier, such as "AREA_TURBINE_01".

[0020] Step S112: Read the fields representing the characteristics of the operation from the maintenance work ticket, and convert the fields into preset operation type identifiers according to preset type mapping rules. Specifically, the central control module locates and reads the fields describing the operation content in the maintenance work ticket, such as "operation item", "risk type" or "work ticket type"; the central control module obtains the original string values ​​of these fields; then, it queries a preset type mapping rule table, which stores the correspondence between the original operation description and the operation type identifiers defined internally by the system that represent different risk categories; through a search and match, the original string is mapped to a standardized operation type identifier, such as "JOB_HIGH_ALTITUDE".

[0021] Step S113: Read the planned start time field and planned end time field from the maintenance work ticket, and construct the work time window based on the planned start time field and the planned end time field.

[0022] Specifically, the central control module locates and reads the fields named "Planned Start Time" and "Planned End Time" from the maintenance work ticket; the values ​​of these fields are usually date and time strings that conform to a certain standard format (such as ISO 8601); the module parses these strings into a unified timestamp format (such as Unix timestamp) used by the system's internal processing; the parsed start timestamp and end timestamp together define the work time window for this work activity.

[0023] The method provided in this embodiment, on the one hand, achieves a standardized understanding of production management system data from different sources or formats by reading specific fields from maintenance work tickets and converting them according to mapping rules; on the other hand, by converting unstructured business descriptions (such as location and work content) into standardized identifiers predefined within the system, it provides an effective data foundation for all subsequent automated processing based on this information (such as area matching).

[0024] In some embodiments, the central control module sends the job type identifier and job time window to the target edge analysis module corresponding to the job area, including: Step S121: Based on the work area identifier, query the preset mapping table between work areas and target edge analysis modules to determine the target edge analysis modules responsible for the work area and their network addresses. Specifically, the central control module maintains a static configuration table named the mapping table between work areas and target edge analysis modules; this table records the responsibility relationship between each work area identifier and an edge analysis module instance, and stores the network communication addresses (such as IP addresses and port numbers) of these edge analysis modules; after obtaining the work area identifier converted in step S111, the central control module performs a query operation in this mapping table; by accurately matching the work area identifier, it retrieves all target edge analysis module records responsible for the area and extracts their network addresses from them.

[0025] Step S122: Encapsulate the job type identifier and the job time window into a control command message. Specifically, the central control module creates a new data message object; fills the message with the job type identifier obtained in step S112, the job time window constructed in step S113, and the unique maintenance job ticket number used for global tracking as key data payload; and adds necessary header information, such as command type, version number, and timestamp, to the message according to a predefined communication protocol format (such as JSON over HTTP), thus completing the assembly of the control command message.

[0026] Step S123: Send the control command message to the target edge analysis module based on the network address.

[0027] Specifically, the central control module establishes a network connection (such as a TCP connection) pointing to the network address of the target edge analysis module obtained in step S121. Through this connection, the control command message assembled in step S122 is sent out as the payload. The message is routed through the enterprise's internal network and finally arrives at the service process listening to the corresponding port and protocol on the target edge analysis module, thereby completing the accurate delivery of the command. The method provided in this embodiment, on the one hand, realizes automatic discovery and addressing from the logical operation area to the physical execution unit (edge ​​analysis module) by querying a preset mapping relationship table; on the other hand, by encapsulating key parameters into standardized control command messages and sending them over the network, a reliable and clear command and communication mechanism from the central control node to the edge computing node is established, ensuring that the analysis task can be accurately and timely triggered and executed.

[0028] In some embodiments, the target edge analysis module identifies the protective equipment wearing status of each worker based on the job type identifier and the monitoring image, including: Step S221: Extract the corresponding protective equipment list from the preset protective equipment list database based on the operation type identifier; the protective equipment list lists multiple protective equipment.

[0029] The protective equipment list database is a rule-based database stored locally on the edge analysis module. Its structure is a relational table, recording the correspondence between different job type identifiers and the set of required protective equipment names. Specifically, when the target edge analysis module initiates the identification process, it uses the received job type identifier as the query key; the module accesses the local protective equipment list database, performs a precise query, and retrieves a record associated with that job type identifier; it then reads the protective equipment list from a specified field of that record. This list is a string array or list that explicitly lists all the protective equipment names that are mandatory for this type of work, such as ["safety helmet", "safety belt", "protective gloves"].

[0030] Step S222: Analyze the monitoring image to identify the wearing status of each worker in the monitoring image wearing each protective equipment in the protective equipment list, and generate wearing status information corresponding to each worker.

[0031] Specifically, this step involves complex computer vision processing. The target edge analysis module processes the continuously input monitoring images; first, it uses a pre-loaded human detection model to locate the regions of all personnel in each frame of the image; then, combined with the real-time location data provided by the personnel positioning system, it binds the personnel regions in the image to specific worker identifications through coordinate transformation; next, for each identified worker, based on the protective equipment list obtained in step S221, it calls the corresponding pre-trained target detection model one by one to analyze its image region; these target detection models will output the status judgment of each type of protective equipment for the corresponding person, such as "helmet - worn" and "safety belt - not worn"; finally, the personnel identification and the status judgment of all their protective equipment are summarized to form a structured record of wearing status information.

[0032] Step S223: For each of the aforementioned workers, determine whether the worker has committed any violation based on the corresponding wearing status information.

[0033] Specifically, the target edge analysis module obtains the wearing status information of all workers generated in step S222; for each wearing status information, the module iterates through each item in the protective equipment list and checks the judgment result corresponding to the status information; according to the predefined rules, if the status of any protective equipment in the list is judged as "not worn" or "inappropriate wearing", the module logically judges that the worker has violated regulations; the module generates a Boolean violation mark for each worker and records which protective equipment(s) are non-compliant.

[0034] Step S224: If a violation occurs, generate a violation report and alarm information based on the wearing status information, and determine the target recipient of the alarm information.

[0035] Specifically, when step S224 determines that a worker has committed a violation, the module immediately initiates the event generation logic. The module generates a structured violation report based on the specific violation item (i.e., non-compliant protective equipment item). The report includes fields such as timestamp, personnel identification, work ticket number, and violation details. At the same time, the module determines the alarm information based on the violation type (such as "not wearing a safety helmet" or "not wearing a safety belt") according to the preset alarm rules, and queries the personnel organizational relationship table to determine the on-site personnel (target recipients) who should receive this alarm, such as the team leader, safety officer, or worker.

[0036] Step S225: Send the violation report to the central control module so that the central control module stores the violation report in the operator's safety file data unit and sends the alarm information to the target receiving object.

[0037] Specifically, the target edge analysis module transmits the violation report generated in step S224 back to the central control module via the network. After receiving the report, the central control module locates the corresponding personnel's record in the centrally stored safety file database based on the personnel identifier in the report, and inserts the violation report as a historical data entry into the record, completing the archiving process. Simultaneously, the target edge analysis module or the central control module, based on the target receiving object information determined in step S224, sends the generated alarm information to the mobile terminal application of the relevant responsible personnel in real time via a message push service. The method provided in this embodiment, on the one hand, strictly binds the identified target with the specific operational risk by dynamically extracting the protective equipment list based on the work type identifier, avoiding the waste of identification resources; on the other hand, through the complete process from image analysis and status judgment to event generation and reporting, it realizes the automated discovery, recording, and immediate alarm of violations, forming a management closed loop from perception to handling; furthermore, by archiving the violation report to the personnel safety file, it provides a data foundation for personnel safety performance evaluation and traceability.

[0038] In some embodiments, analyzing the monitoring image to identify the wearing status of each worker in the monitoring image wearing each piece of protective equipment in the protective equipment list, and generating wearing status information corresponding to each worker, includes: Step S2221: Determine the image bounding boxes and pixel coordinates of each worker in the monitoring image using a preset human detection model. The human detection model is a pre-trained computer vision model used to locate human regions in an image, such as a model optimized based on YOLO or SSD algorithms. The pixel coordinates of the image bounding boxes represent the position of the bounding box in the two-dimensional coordinate system of the image, typically composed of the pixel coordinates of the top-left and bottom-right corners of the box. Specifically, the target edge analysis module inputs each frame of the acquired monitoring image into the loaded human detection model; the model processes the image and outputs one or more detection results; each result includes an image bounding box and the confidence score that the box belongs to the category "human"; the pixel coordinates of the image bounding boxes are extracted and used to mark the area where each worker is located in the image.

[0039] Step S2222: Obtain the spatial location coordinates with unique identifiers broadcast by the positioning devices worn by each worker. Specifically, at the work site, each worker wears a positioning device with a unique identifier; these positioning devices broadcast their acquired spatial location coordinates with unique identifiers, and the target edge analysis module obtains these spatial location coordinates with unique identifiers through the data interface of the integrated positioning system.

[0040] Step S2223: For each of the image bounding boxes, determine the target spatial coordinates corresponding to the pixel coordinates of the image bounding box in each of the spatial coordinates based on the preset camera calibration parameters, and assign the identifier corresponding to the target spatial coordinates to the image bounding box.

[0041] The preset camera calibration parameters refer to a set of parameter matrices obtained in advance through camera calibration technology, describing the camera's internal optical characteristics (such as focal length and principal point) and external position and pose (relative to the world coordinate system), used to establish the geometric mapping relationship between image pixel coordinates and world spatial coordinates. Specifically, the target edge analysis module uses the preset camera calibration parameters (including intrinsic and extrinsic parameter matrices) to construct a projection model from world three-dimensional coordinates to image two-dimensional coordinates, or its inverse transformation model; for each image bounding box obtained in step S2221, the module calculates the pixel coordinates of its center point; using the projection relationship, the image center point is projected backward into the world coordinate system, forming a ray originating from the camera's optical center; the module finds the coordinate point spatially closest to this ray among all real-time acquired spatial position coordinates; the coordinates of this closest point are determined as the target spatial position coordinates corresponding to the image bounding box; subsequently, the module assigns the unique person identifier carried by the target spatial position coordinates to the image region that generated the bounding box, thereby completing the identity binding from "person in the image" to "actual person".

[0042] Step S2224: Extract the target detection model corresponding to each protective item in the protective equipment list from the preset target detection model database.

[0043] The target detection model database is a model repository stored locally on the edge analysis module. For each type of protective equipment (such as helmets, seat belts, and reflective vests) that needs to be identified, a pre-trained deep learning model is pre-stored to detect and classify the state of that equipment in an image. Specifically, the target edge analysis module iterates through each item in the protective equipment list extracted in step S21. For each protective equipment name in the list, the module searches the target detection model database to retrieve the pre-trained target detection model file associated with that name. These model files are then loaded into the memory of the computing device, ready for subsequent image analysis.

[0044] Step S2225: For each of the image bounding boxes, target detection is performed within the image bounding box based on each of the target detection models to obtain the wearing status of each protective equipment in the protective equipment list worn by the workers within the image bounding box. The wearing status of each protective equipment in the protective equipment list worn by the workers within the image bounding box is matched with the identifier corresponding to the image bounding box to obtain the wearing status information of the workers within the image bounding box.

[0045] Specifically, for each image bounding box assigned a personnel identifier, the module takes the image region enclosed by the bounding box as input and sequentially feeds it into each target detection model loaded in step S2224 for processing. Each model detects and classifies the protective equipment it is responsible for, and outputs a judgment result. For example, for the safety helmet model, it may output "exists and is worn correctly", "exists but is not worn correctly", or "not detected". After processing all the protective equipment models, the module obtains a series of status judgment results for the personnel within the image bounding box regarding all the protective equipment in the list. Finally, the module packages and associates this series of status judgment results with the personnel identifier assigned to the bounding box in step S223 to form a complete wearing status information containing "personnel identifier" and "status of each item". The method provided in this embodiment, on the one hand, improves processing efficiency and accuracy by using human detection models and target detection models for hierarchical identification, locating personnel first and then analyzing details; on the other hand, by combining positioning device data with preset camera calibration parameters for coordinate space alignment, it achieves precise association between the anonymous personnel area detected in the image and the identity of the actual workers, solving the pain point of difficult identity recognition in video surveillance; furthermore, by extracting models from the target detection model database on demand, the system can flexibly adapt to different protective equipment identification tasks with different operational requirements.

[0046] In some embodiments, the wearing status is any one of wearing qualified, wearing unqualified, or not wearing, and the step of determining whether the worker has violated regulations based on the wearing status information of the worker includes: Step S2231: Iterate through the wearing status of each protective item in the protective equipment list corresponding to the wearing status information.

[0047] Specifically, when processing each piece of wearing status information generated from step S225, the target edge analysis module uses the protective equipment list obtained in step S221 as a standard checklist. The target edge analysis module reads each protective equipment name in the list in sequence, then searches for the status field corresponding to the name in the currently processed wearing status information and reads its value. The status value is a predefined discrete enumeration value, specifically one of "wearing qualified", "wearing unqualified" or "not worn". For example, for the protective equipment "safety helmet", its corresponding status value may be "wearing qualified".

[0048] Step S2232: If any of the protective equipment is in an unqualified wearing status or not worn in the wearing status information, it is determined that the worker has committed a violation. If all protective equipment in the wearing status information is in a qualified wearing status, it is determined that the worker has not violated any regulations. The method provided in this embodiment, on the one hand, explicitly divides the wearing status into three categories: "qualified," "unqualified," and "not worn," providing a clear and unambiguous input standard for the judgment logic; on the other hand, through traversal comparison and conditional judgment, it realizes a deterministic and repeatable automatic violation judgment rule, replacing manual visual inspection and eliminating subjective judgment differences; furthermore, through the rule of "triggering any unqualified item," it ensures the strict implementation of safety supervision standards, as any failure of a protective measure will be detected by the system.

[0049] In some embodiments, generating violation reports and alarm information based on the wearing status information, and determining the target recipients of the alarm information, includes: Step S2241: Determine from the wearing status information that the wearing status is not a qualified wearing of the target protective equipment.

[0050] Specifically, based on the judgment result of step S2232, the target edge analysis module filters out all protective equipment items whose status value is not "fitted properly" from the wearing status information that triggers the violation judgment; these filtered items constitute the "target protective equipment" set for this violation event; for example, if the wearing status information shows "helmet: fit properly, seat belt: not worn, protective gloves: unfit properly", then "seat belt" and "protective gloves" are identified as target protective equipment.

[0051] Step S2242: Generate the violation report based on the wearing status of the protective equipment of each target.

[0052] Specifically, the target edge analysis module creates a structured data object as a violation report; the report contains at least the following fields: a unique violation event number, a timestamp of occurrence, an associated worker identifier, and a violation details list; in the violation details list, a record is created for each target protective equipment identified in step S2241, the record containing the name of the equipment and its specific wearing status ("inadequate wearing" or "not worn").

[0053] Step S2243: For each of the target protective equipment, if the wearing status of the target protective equipment is unqualified, generate alarm information to correctly wear the target protective equipment, and determine the operator corresponding to the identifier of the wearing status information as the target receiving object.

[0054] Specifically, for items in the target protective equipment set that are in a "misfit" state (such as "protective gloves: misfit"), the target edge analysis module generates a guidance alarm message. The text content of this message instructs the operator to "Please wear [protective equipment name] correctly". At the same time, the module sets the target recipient of the alarm message to the operator corresponding to the personnel identifier in the wearing status information. This means that the alarm will directly remind the operator to correct the non-standard behavior through the smart device or team communication device they are wearing.

[0055] Step S2244: For each of the target protective equipment, if the target protective equipment is not worn, generate alarm information for the target protective equipment to the operator corresponding to the identifier corresponding to the wearing status information, and determine the safety management personnel of the work area as the target receiving object.

[0056] Specifically, for items in the target protective equipment set whose status is "not worn" (such as "safety belt: not worn"), the module generates a demand alarm message. The text content of this message indicates "The worker [personnel identifier] is not wearing [protective equipment name], which needs to be provided." At the same time, the target edge analysis module sets the target recipient of the alarm message to the safety management personnel responsible for the work area (whose identity information is obtained from the pre-configured personnel organization table). This means that the alarm will notify the on-site supervisor to intervene, which may involve stopping the work, providing equipment, or conducting on-site training. The method provided in this embodiment, on the one hand, makes on-site intervention measures more targeted by distinguishing between the two different states of "inadequate wearing" and "not wearing" and generating alarm messages with differentiated content; on the other hand, by setting different target recipients (the worker or the safety management personnel) for different types of violations, it realizes the hierarchical push and responsibility division of alarm information, directing the responsibility for correcting general non-standard behaviors to the worker himself, while directing the responsibility for handling serious deficiencies (not wearing) to the on-site manager, thus optimizing the scheduling of safety management resources.

[0057] Please see Figure 2 , Figure 2 This is a schematic diagram of a video surveillance-based system for recognizing the wearing status of protective equipment for maintenance personnel, provided as an embodiment of this application. Figure 2 As shown, the video surveillance-based maintenance personnel protective equipment wearing status recognition system 100 includes a central control module and multiple edge analysis modules; The central control module responds to the activation of the maintenance work ticket by parsing the maintenance work ticket to obtain the work area, work type identifier, and work time window corresponding to the maintenance work ticket, and sends the work type identifier and work time window to the target edge analysis module corresponding to the work area. The target edge analysis module activates the video surveillance equipment in the work area at the beginning of the work time window to acquire surveillance images, and identifies the protective equipment wearing status of each worker based on the work type identifier and the surveillance images within the work time window.

[0058] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and each module described above can be referred to the corresponding process in the aforementioned embodiment of the method for recognizing the wearing status of protective equipment for maintenance personnel based on video surveillance, and will not be repeated here.

[0059] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0060] It should also be understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. It should be noted that, herein, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0061] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. The above descriptions are merely specific implementations of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for identifying the wearing status of protective equipment by maintenance personnel based on video surveillance, characterized in that, The method is used in a video surveillance-based system for recognizing the wearing status of protective equipment by maintenance personnel. The system includes a central control module and multiple edge analysis modules. The method includes: The central control module responds to the activation of the maintenance work ticket by parsing the maintenance work ticket to obtain the work area, work type identifier, and work time window corresponding to the maintenance work ticket, and sends the work type identifier and work time window to the target edge analysis module corresponding to the work area. The target edge analysis module activates the video surveillance equipment in the work area at the beginning of the work time window to acquire surveillance images, and identifies the protective equipment wearing status of each worker based on the work type identifier and the surveillance images within the work time window.

2. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 1, characterized in that, The target edge analysis module identifies the protective equipment wearing status of each worker based on the job type identifier and the monitoring image, including: Based on the work type identifier, the corresponding protective equipment list is retrieved from the preset protective equipment list database; the protective equipment list lists multiple protective equipment items. The monitoring images are analyzed to identify the wearing status of each worker in the monitoring images wearing each piece of protective equipment in the protective equipment list, and to generate wearing status information for each worker. For each of the aforementioned workers, it is determined whether the worker has committed any violation based on the corresponding wearing status information; If a violation occurs, a violation report and alarm information are generated based on the wearing status information, and the target recipient of the alarm information is determined. The violation report is sent to the central control module, so that the central control module stores the violation report in the operator's safety file data unit and sends the alarm information to the target receiving object.

3. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 2, characterized in that, The analysis of the monitoring images to identify the wearing status of each worker in the monitoring images wearing each piece of protective equipment from the protective equipment list, and to generate wearing status information corresponding to each worker, includes: The image bounding box and pixel coordinates of each worker in the monitoring image are determined by a preset human detection model; the spatial position coordinates with unique identifiers broadcast by the positioning device worn by each worker are obtained; for each image bounding box, the target spatial position coordinates corresponding to the pixel coordinates of the image bounding box are determined in the spatial position coordinates based on preset camera calibration parameters, and the identifier corresponding to the target spatial position coordinates is assigned to the image bounding box. Extract the target detection model corresponding to each protective item in the list of protective equipment from the preset target detection model database; For each of the image bounding boxes, target detection is performed within the image bounding box based on each of the target detection models to obtain the wearing status of each protective equipment in the protective equipment list of the workers within the image bounding box. The wearing status of each protective equipment in the protective equipment list of the workers within the image bounding box is then matched with the identifier corresponding to the image bounding box to obtain the wearing status information of the workers within the image bounding box.

4. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 3, characterized in that, The wearing status is any one of wearing qualified, wearing unqualified, or not wearing. The step of determining whether the worker has violated regulations based on the worker's wearing status information includes: Iterate through the wearing status of each piece of protective equipment in the protective equipment list, as shown in the wearing status information. If any of the protective equipment is found to be improperly worn or not worn in the wearing status information, the worker is deemed to have committed a violation. If all protective equipment in the wearing status information is in a qualified wearing status, it is determined that the worker has not violated any regulations.

5. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 4, characterized in that, The process of generating violation reports and alarm information based on the wearing status information, and determining the target recipients of the alarm information, includes: The wearing status information indicates that the target protective equipment is not worn properly. The violation report is generated based on the wearing status of the protective equipment of each target. For each of the target protective equipment, if the wearing status of the target protective equipment is unqualified, an alarm message is generated to correctly wear the target protective equipment, and the operator corresponding to the identifier corresponding to the wearing status information is identified as the target receiving object; For each of the target protective equipment, if the target protective equipment is not worn, an alarm message for the target protective equipment is generated and provided to the operator corresponding to the identifier corresponding to the wearing status information, and the safety management personnel of the work area is identified as the target receiving object.

6. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 1, characterized in that, The central control module parses the maintenance work ticket to obtain the corresponding work area, work type identifier, and work time window, including: Read the field representing spatial location from the maintenance work ticket, and convert the field into a preset work area identifier according to the preset encoding mapping rule; The field representing the characteristics of the operation is read from the maintenance work ticket, and the field is converted into a preset operation type identifier according to the preset type mapping rule; the planned start time field and the planned end time field are read from the maintenance work ticket, and the operation time window is constructed based on the planned start time field and the planned end time field.

7. The method for identifying the wearing status of protective equipment for maintenance personnel based on video surveillance according to claim 6, characterized in that, The central control module sends the job type identifier and job time window to the target edge analysis module corresponding to the job area, including: Based on the work area identifier, a preset mapping table between work areas and target edge analysis modules is queried to determine the target edge analysis module responsible for the work area and its network address; the work type identifier and the work time window are encapsulated into a control command message; the control command message is sent to the target edge analysis module based on the network address.

8. A video surveillance-based system for recognizing the wearing status of protective equipment by maintenance personnel, the system comprising a central control module and multiple edge analysis modules, characterized in that, The central control module responds to the activation of the maintenance work ticket by parsing the maintenance work ticket to obtain the work area, work type identifier, and work time window corresponding to the maintenance work ticket, and sends the work type identifier and work time window to the target edge analysis module corresponding to the work area. The target edge analysis module activates the video surveillance equipment in the work area at the beginning of the work time window to acquire surveillance images, and identifies the protective equipment wearing status of each worker based on the work type identifier and the surveillance images within the work time window.

Images