An online intelligent management method, system and device for hot work supervision

Abstract

The present invention discloses an online intelligent management method, system and device for hot work supervision, which relates to the field of intelligent supervision technology for construction projects and includes the following steps: S1, creating a hot work management system database; S2, configuring a monitoring box and a distribution box with a QR code after the operation application and approval are passed, and binding the monitoring box to the corresponding work order; S3, after the installation is completed at the operation site, scanning the QR code of the distribution box, binding the distribution box and the operator's authority to the corresponding work order, and opening the distribution box; S4, transmitting the monitoring video data of the monitoring box, the temperature data of the distribution box, and the smoke concentration data of the smoke sensor to the database, and performing algorithm analysis; S5, generating an alarm message based on the analysis results. The present invention realizes online supervision, efficient monitoring, and saves manpower and time costs; improves safety management capabilities and reduces the occurrence of safety accidents; helps analyze the path of fire spread, the location of the fire source and related risks, and provides evidence reference for fire investigation and analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent supervision of construction projects, and in particular to an online intelligent management method, system and device for supervision of hot work operations. Background Art

[0002] Currently, hot work and hot work management are primarily conducted offline, making it difficult for construction unit managers to obtain real-time information on on-site hot work. Furthermore, the number of safety management personnel within construction companies is limited, making comprehensive oversight of project-level hot work locations difficult. Hot work operations are complex and located at numerous locations, making comprehensive oversight difficult. Video surveillance is difficult to fully cover hot work areas, making it difficult to capture complete process footage with handheld cameras and storage challenges. Furthermore, smoke detectors are not connected to the internet, making it difficult to receive immediate remote fire information.

[0003] The existing technical solutions include:

[0004] 1. During the project management process, the operator submits a paper version of the hot work application to the management personnel offline. After approval, the operator can start the hot work;

[0005] 2. During hot work, management personnel can only go to the specific hot work site to check the operation status;

[0006] 3. Whether the hot work is completed depends on the quality of the operators and the on-site inspection of the management personnel.

[0007] The following defects and deficiencies exist in existing hot work and hot work management:

[0008] 1. The number of safety management personnel in each enterprise is limited, making it difficult to achieve comprehensive supervision of project-level hot spots;

[0009] 2. Hot work operations are complex and involve numerous hot work locations, making comprehensive supervision difficult;

[0010] 3. It is difficult to fully cover the hot work area with video surveillance. Handheld cameras cannot generate complete process image data, and storage is difficult.

[0011] 4. The smoke detectors are not connected to the Internet, making it difficult to receive fire information remotely and immediately. Summary of the Invention

[0012] To overcome the problems existing in the related art, the present invention provides an online intelligent management method for hot work supervision, comprising the following steps:

[0013] S1. Create a hot work management system database;

[0014] S2. The operator initiates a job application through the client, which is then approved by the management staff. After approval, a monitoring box and distribution box with a QR code are configured for this job, and a corresponding work order is generated. The monitoring box and the corresponding work order are also bound.

[0015] S3: The operator picks up the monitoring box and distribution box offline, arrives at the designated location to complete the installation, scans the QR code on the distribution box, sends a command to the server, binds the distribution box and operator permissions to the corresponding work order, and then turns on the distribution box.

[0016] S4, storing the monitoring video data of the monitoring box, the distribution box temperature data of the distribution box, and the smoke concentration data of the smoke detector in the database, and performing algorithm analysis on the monitoring video data, the distribution box temperature data, and the smoke concentration data of the smoke detector;

[0017] S5. Generate alarm information during the hot work process based on the results of the algorithm analysis.

[0018] Furthermore, the algorithm analysis process of the monitoring video data is as follows:

[0019] (1) Obtain surveillance video images;

[0020] (2) Capture images from surveillance video at a rate of one image per second;

[0021] (3) Sharpen the image;

[0022] (4) Convert the image into a tensor format image;

[0023] (5) Use 6 categories of flame recognition models to identify the image and output the recognition results;

[0024] (6) Judge the recognition results and output the results.

[0025] Furthermore, the method further includes training the six categories of flame recognition models, training the flame recognition models according to six training strategies, and obtaining six categories of flame recognition models.

[0026] Furthermore, the six training strategies are:

[0027] A1, the flame direction is upward, and the picture is marked as flame a1 of abnormal operation;

[0028] B1, only flames but no workers, the picture is marked as abnormal operation flame b1;

[0029] C1, when the flame height reaches half of the human body height, the image is marked as abnormal operation flame c1;

[0030] A2, a person faces the flame, and the picture is marked as flame a2 of normal operation;

[0031] B2, the human hand is parallel to the height of the flame, and the picture is marked as normal operation flame b2;

[0032] C2, flame connected to the device, the picture is marked as normal operation flame c2.

[0033] Furthermore, the specific process of generating the alarm information is as follows: based on the recognition result, it is determined whether there is a person and a fire in the image at the same time. If there is only a flame but no person, it is identified as a fire hazard and an alarm message is generated; if there is a person and a fire at the same time, it is further determined whether the height of the flame reaches half of the person's height. If the height of the flame reaches half of the person's height, it is identified as a fire hazard and an alarm message is generated; if the height of the flame does not reach half of the person's height, it is further determined whether the flame is connected to the device or equipment. If the flame is connected to the device or equipment, it is identified as normal operation; if the flame is not connected to the device or equipment, it is further determined whether the person is facing the flame. If the person is facing the flame, it is identified as normal operation; if the person is not facing the flame, it is further determined whether the person's hand is parallel to the height of the flame. If the person's hand is parallel to the height of the flame, it is identified as normal operation; if the person's hand is not parallel to the height of the flame, it is further determined whether the direction of the flame is upward. If the direction of the flame is not upward, it is identified as normal operation; if the direction of the flame is upward, it is identified as a fire and an alarm message is generated.

[0034] Furthermore, the hot work management system database includes user information, hot work operation information, equipment information and hot work operation statistical reports.

[0035] Furthermore, a high-definition camera and an intelligent smoke sensor are provided in the monitoring box, and a distribution box controller and a multifunctional sensor are provided in the distribution box.

[0036] Furthermore, it also includes dividing the risks of hot work into three levels, namely: level one risk, level two risk and level three risk.

[0037] Furthermore, it includes a login module, an approval module, a supervision module, an analysis module and an early warning module. The login module is used by the operator to initiate a job application through the client, the approval module is used by the management personnel to approve the job, the monitoring module is used to obtain monitoring video data, distribution box temperature data and smoke sensor smoke concentration data, the analysis module is used to perform algorithmic analysis on the monitoring video data, distribution box temperature data and smoke sensor smoke concentration data, and the early warning module is used to generate alarm information according to the algorithmic analysis results.

[0038] Furthermore, it includes a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any of the above methods is executed.

[0039] Beneficial effects of the technical solution of the present invention:

[0040] 1. Realize full supervision of on-site hot work, improve the safety and fire safety level of the construction site, and reduce the occurrence of safety accidents;

[0041] 2. Realize online supervision, efficient monitoring, and save manpower and time costs;

[0042] 3. Historical video playback can be used to trace the cause and development of the fire, allowing us to better understand the cause and evolution of the accident and take appropriate measures to prevent similar incidents from happening again;

[0043] 4. By replaying the video, you can analyze the path of fire spread, the location of the fire source, and the potential risks caused by possible human behavior. This not only helps protect the safety of personnel, but also provides strong evidence and reference for fire investigation and accident analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 Flowchart of online intelligent management of hot work supervision;

[0045] Figure 2 Fire hazard logic judgment flow chart;

[0046] Figure 3 Structure diagram of the online intelligent management system for hot work supervision;

[0047] Figure 4 Structural diagram of the online intelligent management device for hot work supervision. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Examples of the embodiments are shown in the accompanying drawings, in which the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be understood as limiting the present invention.

[0049] The present invention provides an online intelligent management method for hot work supervision, such as Figure 1 As shown, the online intelligent management method for hot work supervision includes the following steps.

[0050] S1. Create a hot work management system database. The specific data in the hot work management system database includes: user information, hot work operation information, equipment information, and hot work operation statistical reports. The following specifically defines user information, hot work operation information, equipment information, and hot work operation statistical reports.

[0051] S11. User information includes: user ID, password, role (hot work operator, hot work administrator), operator permissions, and administrator permissions. Roles are categorized as hot work operator and hot work administrator. Hot work operator permissions include the ability to submit work requests, scan the QR code to open and close the distribution box, request power reconnection, and obtain and process alarm data. Hot work administrator permissions include the ability to review requests, delete work orders, obtain real-time smoke detection, monitoring, distribution box switches, obtain multi-function sensor status, and obtain alarm status data.

[0052] S12. Hot work information includes: work type, work content, hot work level, work location, application permit period, fire watcher, operating unit, photo certifying working conditions, and approving personnel. Fire watcher: A fire watcher is someone who monitors fires during hot work, specifically to prevent fires. Photo certifying working conditions: A detailed photo of the hot work environment can be taken using a camera-enabled device.

[0053] S13. Device information includes: hot work monitoring MAC address, smart smoke sensor MAC address, distribution box controller MAC address, distribution box controller Bluetooth name, multi-function sensor MAC address, distribution box switch status, distribution box QR code information, and monitoring box QR code information. Hot work monitoring MAC address: The monitoring box contains a high-definition camera; the HD camera's MAC address is the hot work monitoring MAC address. Smart smoke sensor MAC address: During hot work operations, the smart smoke sensor should be placed above the hot work point. The smart smoke sensor's MAC address is the smart smoke sensor's MAC address. The distribution box is equipped with a distribution box controller and multi-function sensor. The distribution box controller has a MAC address and Bluetooth name, and the multi-function sensor also has a MAC address.

[0054] S14. Hot work statistics report includes working status statistics report and alarm statistics report, as shown in Table 1 and Table 2:

[0055] Table 1 Working status statistics report

[0056] Serial number Job Type Job Description Hot Work Level Work Location License Period Work Unit sponsor Firewatch Process Status Distribution box status Distribution box name Toolbox Number 1 Waterproofing work Garage roof waterproofing Level 3 hot work 1# Building garage roof Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Zhang San Li Xin in progress Powered DHPD-001 DHJK-002 2 Welding work Rebar welding Level 3 hot work 5th floor, Building 2 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Li Si Chen Yu in progress Powered DHPD-002 DHJK-008 3 Cutting operation Rebar cutting Secondary hot work 6th Floor, Building 3 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Wang Wu Ding Wen Completed Power off DHPD-003 DHJK-009

[0057] Table 2 Alarm statistics report

[0058] Serial number Alarm time Alarm Type Job Type Job Description Hot Work Level Work Location License Period Work Unit sponsor Firewatch Whether to process 1 2023-12-8 15：00 Smoke alarm Waterproofing work Garage roof waterproofing Level 3 hot work 1# Building garage roof Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Zhang San Li Xin yes 2 2023-12-8 16：00 Timeout alarm Welding work Rebar welding Level 3 hot work 5th Floor, Building 2 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Li Si Chen Yu no 3 2023-12-8 18：00 SOS alarm Cutting operations Rebar cutting Secondary hot work 6th Floor, Building 3 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Wang Wu Ding Wen yes

[0059] S2. The operator initiates a job application through the client, which is then approved by the management staff. After approval, a monitoring box and distribution box with a QR code are configured for this job, and a corresponding work order is generated. The monitoring box is bound to the work order and can only be used for this job and cannot be used for other job locations.

[0060] When the operator initiates a work application through the client, he / she fills in the following information: work type, work content, fire level, work location, application permit period, fire watcher, work unit, photo of the work conditions certification, and approval personnel, to ensure that the corresponding location can be reached immediately for rescue in the event of danger, as shown in Table 3:

[0061] Table 3 Hot Work Application Form

[0062] Serial number Job Type Job Description Hot Work Level Work Location Application period for permission Work Unit Approver Firewatch Photos proving working conditions 1 Waterproofing work Garage roof waterproofing Level 3 hot work 1# Building garage roof Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Horseman Li Xin 2 Welding work Rebar welding Level 3 hot work 5th Floor, Building 2 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Horseman Chen Yu 3 Cutting operations Rebar cutting Secondary hot work 6th Floor, Building 3 Start: 2023-12-8 14:00 End: 2023-12-8 19:00 Construction Company A Horseman Ding Wen

[0063] The monitoring box is bound to the corresponding work order. Specifically, the hot work work order is bound to the monitoring box QR code number, monitoring box name, hot work monitoring camera device number, and smart smoke sensor MAC address, as shown in Table 4:

[0064] Table 4 Binding table of hot work work order and monitoring box

[0065] Hot Work Order Monitoring box QR code name Monitoring box name Fire monitoring MAC address Smart smoke sensor MAC address C1373F2D C856960E-4C2D-9DCB-63E4-F8417D3AED94 DHJK-001 C5-54-B2-C9 A1-54-B2-C5 C1373F5F DHJK-002 DHJK-002 DF-59-2B-B7 A1-54-2B-C8 C1373F7K DHJK-003 DHJK-003 A5-D5-B2-F9 A1-96-B2-F5

[0066] S3: The operator picks up the monitoring box and distribution box offline, arrives at the designated location to complete the installation, scans the QR code on the distribution box, sends a command to the server, binds the distribution box and operator permissions to the corresponding work order, and then turns on the distribution box.

[0067] The distribution box is equipped with a multi-function sensor. Through the MAC address of the multi-function sensor and NB-IoT data transmission, the server can obtain the distribution box temperature, humidity, atmospheric pressure, altitude, and light intensity data of the multi-function sensor.

[0068] Distribution box switch status: When the distribution box controller is powered on, the AC receiver defaults to the disconnected state. When the software sends a closing signal to the controller, the controller closes the AC receiver, and the switch is now closed. When the software sends a disconnecting signal to the controller, the controller controls the AC receiver to disconnect, and the switch is now disconnected.

[0069] The QR code number of the distribution box corresponds to the distribution box number. One QR code number corresponds to one distribution box controller MAC address, distribution box multi-function sensor MAC address, distribution box switch status and distribution box Bluetooth name;

[0070] The hot work personnel receive the monitoring box and distribution box from the management staff, open the monitoring equipment in the monitoring box, and scan the code to start the power supply of the distribution box. The specific steps are as follows:

[0071] (1) By judging the status of the camera in the monitoring box in the database, it is determined whether the camera monitoring image data has been received. When the camera in the monitoring box is turned on, the distribution box can be opened by scanning the code; if the camera in the monitoring box is not turned on, the distribution box cannot be opened;

[0072] (2) Scan the code to obtain the status of the distribution box. If the distribution box is already in use, it cannot be opened;

[0073] (3) Scan the QR code in an internet-connected state to obtain the controller number, Bluetooth name, multi-function sensor number, and smoke sensor number of the distribution box bound to the QR code, and bind all these devices to this work order through a QR code;

[0074] (4) After the code is scanned, the system determines that the binding with the device is complete, and the distribution box can be turned on for power supply;

[0075] (5) After the distribution box scan is turned on, the work order is in progress. The distribution box and monitoring box are bound to this work order and can no longer be used for other operations;

[0076] The specific binding form between the hot work work order and the distribution box is shown in Table 5:

[0077] Table 5 Binding table of hot work work order and distribution box

[0078] QR code number QR code name Controller Number Bluetooth Name Switch status Distribution box temperature Distribution box humidity Light intensity of distribution box Distribution box atmospheric pressure Distribution box altitude C1373F2D DHPD-001 A1-B2 A1-B2 open 23℃ 23%rh 300Lux 101.32kPa 14.5m C1373F5F DHPD-002 A1-B2 A1-B2 open 23℃ 23%rh 323Lux 101.32kPa 14.5m C1373F7K DHPD-003 A1-B2 A1-B2 open 23℃ 23%rh 343Lux 101.32kPa 14.5m

[0079] S4. The monitoring video data from the monitoring box, the distribution box temperature data, and the smoke detector smoke concentration data are stored in the database, and algorithmic analysis is performed on the monitoring video data, distribution box temperature data, and smoke detector smoke concentration data. First, the hot work risk is divided into three levels, with the highest risk being level 1 risk and the lowest risk being level 3 risk. The specific definitions of the three levels of hot work risk are shown in Table 6:

[0080] Table 6 Risk Levels of Hot Work

[0081] Risk Management Judgment indicator A: working environment temperature Judgment indicator B: Monitoring AI to identify flames Judgment indicator C: smoke concentration Judgment index D: distribution box temperature Level judgment method Level 1 risk Actual temperature ≥50℃ Identify fire hazards caused by abnormal operation Actual concentration ≥1dB / m Actual temperature ≥65℃ When any of the ABCD parameters is reached, it is considered a level 1 risk. Secondary risk 50℃>actual temperature≥45℃ Identify fire hazards caused by abnormal operation 1dB / m>actual concentration≥0.5dB / m 65℃>actual temperature≥57℃ If two of ABCD are met at the same time, it is judged as a secondary risk. Level 3 risk 57℃>actual temperature≥50℃ Identify fire hazards caused by abnormal operation Actual concentration <0.5dB / m 57℃>actual temperature≥50℃ If five of ABCD are met at the same time, it is judged as level 3 risk.

[0082] The working environment temperature and smoke concentration are collected by an intelligent smoke sensor, while the distribution box temperature is collected by a multi-function sensor. Specifically, the intelligent smoke sensor can be a photoelectric smoke detector, which uses the photoelectric effect to detect smoke. It primarily consists of a light source, a photosensor, and a signal processing circuit. The light source emits a beam of infrared light. The photosensitive element receives the light signal, which is then analyzed by the signal processing circuit. If smoke is detected, the alarm triggers an audible or visual signal. The operating principle of a photoelectric smoke detector is based on the photoelectric effect, which describes the phenomenon in which light strikes certain substances, generating electrons or electrical charges. In a photoelectric smoke detector, infrared light from the light source strikes smoke particles in the air. The smoke particles absorb the light energy and scatter it. The scattered light is received by the photosensor, generating electrons or electrical charges, which are converted into electrical signals and output to the signal processing circuit. The signal processing circuit analyzes and processes the received electrical signals. First, it amplifies the electrical signal output by the photosensor to enhance the signal strength. Then, it compares the input signal with a set threshold to determine whether smoke has been detected. When the electrical signal exceeds a threshold, the signal processing circuit triggers an alarm to emit an audible or visual signal, alerting people to the presence of smoke. Monitoring AI for flame recognition involves machine learning and intelligent program logic. The specific model training methods and logic algorithms involved are explained in detail later.

[0083] S5. Generate an alarm message during the hot work process based on the results of the algorithm analysis. If the algorithm analysis results in a level 1 risk, a level 2 risk, or a level 3 risk, the platform generates an alarm message based on the results of the algorithm analysis, specifically:

[0084] Level 1 risk: SMS reminder + app and PC platform alarm reminder + automatic power off of the distribution box + sound and light alarm;

[0085] Secondary risk: App and PC platform early warning reminder + sound and light alarm;

[0086] Level 3 risk: App and PC platform early warning reminder + sound and light alarm;

[0087] For level 2 and level 3 risks, management personnel can judge them through real-time monitoring data and real-time monitoring videos and decide whether to conduct remote intervention. They can also remotely control the on and off of the current in the distribution box through the communication module to control the progress of hot work.

[0088] In addition to the monitoring data in Table 6, operation timeout is also one of the system alarm contents. Management personnel can determine whether to intervene based on the actual situation monitoring to prevent hot workers from working with hot work for too long. The early warning mechanism here is relatively flexible and not specifically limited.

[0089] At the same time, when operating personnel encounter emergencies, they can issue an SOS alarm. By triggering the SOS distress button on the monitoring device, the backend management personnel can receive the alarm information. The alarm mechanism here is also relatively flexible and not specifically limited.

[0090] The following is an explanation of the specific algorithm for monitoring the flame recognition by the AI ​​in step S4 and the specific method for making logical judgments based on the recognition results:

[0091] First, model training is performed using existing image format data to train the model to accurately identify flames. The present invention uses six training strategies to train six models, ultimately enabling the six models to identify different types of flames and perform different types of result labeling.

[0092] The six training strategies are as follows:

[0093] 1. Identify flames that are not operating normally:

[0094] A1, identify the direction of the flame, the flame direction is upward, and it is identified as the flame a1 of abnormal operation;

[0095] B1, only flames but no operator determines it as abnormal operation flame b1;

[0096] C1, with a person as the reference object, when the flame height reaches half of the human body height, it is identified as an abnormal operation flame c1;

[0097] 2. Identify the flame as working normally:

[0098] A2, identifying the relationship between people and fire, a person facing a flame is judged as a normal flame a2;

[0099] B2, the height of the human hand is parallel to the flame, which is recognized as the flame of normal operation b2;

[0100] C2, the flame is connected to the device and is identified as the flame c2 of normal operation;

[0101] The specific training process is:

[0102] (1) Mark the flames, workers, and equipment in the picture;

[0103] (2) Input images of types A1, B1, C1, A2, B2, and C2 to the six models respectively, and output the corresponding results a1, b1, c1, a2, b2, and c2;

[0104] (3) Stop training until the model training reaches a certain recognition accuracy;

[0105] Next, the monitoring AI flame recognition process: Capture video images through surveillance cameras → Capture images at a rate of one frame per second → Transmit images to the AI ​​box via wireless network → Sharpen the images → Convert the images to tensor format → Process and analyze the images using six different strategies based on deep learning algorithms → Identify the images using six flame recognition models and output the results. After recognition, a single image may be labeled with six or fewer labels.

[0106] Finally, the identification results are logically judged according to the priority of the identification results, such as Figure 2 As shown, the specific logical judgment method is as follows: based on the recognition result, it is judged whether there are people and fire in the image at the same time. If there is only flame but no people, it is identified as a fire hazard and an alarm message is generated; if there are people and fire at the same time, it is further judged whether the height of the flame reaches half of the person's height. If the height of the flame reaches half of the person's height, it is identified as a fire hazard and an alarm message is generated; if the height of the flame does not reach half of the person's height, it is further judged whether the flame is connected to the device. If the flame is connected to the device, it is identified as normal operation; if the flame is not connected to the device, it is further judged whether the person is facing the flame. If the person is facing the flame, it is identified as normal operation; if the person is not facing the flame, it is further judged whether the person's hand is parallel to the height of the flame. If the person's hand is parallel to the height of the flame, it is identified as normal operation; if the person's hand is not parallel to the height of the flame, it is further judged whether the direction of the flame is upward. If the direction of the flame is not upward, it is identified as normal operation; if the direction of the flame is upward, it is identified as a fire, and an alarm message is generated.

[0107] In addition, the present invention provides an online intelligent management system for hot work supervision, such as Figure 3 As shown, it includes a login module, an approval module, a supervision module, an analysis module and an early warning module. The login module is used by operators to initiate job applications through the client, the approval module is used by managers to approve jobs, the monitoring module is used to obtain monitoring video data, distribution box temperature data and smoke sensor smoke concentration data, the analysis module is used to perform algorithmic analysis on the monitoring video data, distribution box temperature data and smoke sensor smoke concentration data, and the early warning module is used to generate alarm information according to the algorithmic analysis results.

[0108] At the same time, it provides an online intelligent management device for hot work supervision, such as Figure 4 As shown, it includes a processor and a memory, the processor and the memory are connected via a bus, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any of the above methods is executed.

[0109] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. An online intelligent management method for hot work supervision, characterized in that: The following steps are included: S1. Create a hot work management system database; the specific data in the hot work management system database includes: user information, hot work operation information, equipment information and hot work operation statistical reports; S2. The operator initiates a job application through the client, which is then approved by the management staff. After approval, a monitoring box and distribution box with a QR code are configured for this job, and a corresponding work order is generated. The monitoring box and the corresponding work order are also bound. S3. The operator receives the monitoring box and distribution box offline, arrives at the designated location to complete the installation, scans the QR code on the distribution box, sends a command to the server, binds the distribution box and operator permissions to the corresponding work order, and turns on the distribution box. The hot work operator receives the monitoring box and distribution box from the management staff, turns on the monitoring equipment in the monitoring box, and scans the code to turn on the power supply of the distribution box. The specific steps are as follows: (1) By judging the status of the camera in the monitoring box in the database, it is determined whether the camera monitoring image data has been received. When the camera in the monitoring box is turned on, the distribution box can be opened by scanning the code; if the camera in the monitoring box is not turned on, the distribution box cannot be opened; (2) Scan the code to obtain the status of the distribution box. If the distribution box is already in use, it cannot be opened; (3) Scan the QR code when there is an Internet connection to obtain the controller number, Bluetooth name, multi-function sensor number, and smoke sensor number of the distribution box bound to the QR code, and bind all these devices to this work order through a QR code; (4) After the code is scanned, the system determines that the binding with the device is complete, and the distribution box can be turned on for power supply; (5) After the distribution box scan is turned on, the work order is in progress. The distribution box and monitoring box are bound to this work order and cannot be used for other operations. S4, storing the monitoring video data of the monitoring box, the distribution box temperature data of the distribution box, and the smoke concentration data of the smoke detector in the database, and performing algorithm analysis on the monitoring video data, the distribution box temperature data, and the smoke concentration data of the smoke detector; S5. Generate an alarm message during the hot work process based on the results of the algorithm analysis; The algorithm analysis process of the monitoring video data is as follows: (1) Obtain surveillance video images; (2) Capture images of surveillance video at a frequency of one frame per second; (3) sharpen the image; (4) Convert the image into a tensor format image; (5) Use 6 types of flame recognition models to identify the image and output the recognition results; The six types of flame recognition model training strategies are: A1, the flame direction is upward, and the picture is marked as flame a1 of abnormal operation; B1, only flames but no workers, the picture is marked as abnormal operation flame b1; C1, when the flame height reaches half of the human body height, the image is marked as abnormal operation flame c1; A2, a person faces the flame, and the picture is marked as flame a2 of normal operation; B2, the human hand is parallel to the height of the flame, and the picture is marked as normal operation flame b2; C2, the flame is connected to the device, and the picture is marked as the normal operation flame c2; (6) Judge the recognition results and output the results; The specific process of generating the alarm information is as follows: judging whether there is a person and a fire in the image at the same time based on the recognition result, if there is only a flame but no person, it is recognized as a fire hazard and an alarm information is generated; if there is a person and a fire at the same time, further judging whether the height of the flame reaches half of the person's height, if the height of the flame reaches half of the person's height, it is recognized as a fire hazard and an alarm information is generated; if the height of the flame does not reach half of the person's height, further judging whether the flame is connected to a device or equipment, if the flame is connected to a device or equipment, it is recognized as normal operation; if the flame is not connected to a device or equipment, further judging whether a person is facing the flame, if the person is facing the flame, it is recognized as normal operation; if the person is not facing the flame, further judging whether the person's hand is parallel to the height of the flame, if the person's hand is parallel to the height of the flame, it is recognized as normal operation; if the person's hand is not parallel to the height of the flame, further judging whether the direction of the flame is upward, if the direction of the flame is not upward, it is recognized as normal operation; If the flame is directed upward, it is identified as a fire and an alarm message is generated.

2. The online intelligent management method for hot work supervision according to claim 1 is characterized in that: The hot work management system database contains user information, hot work operation information, equipment information and hot work operation statistical reports.

3. The online intelligent management method for hot work supervision according to claim 2 is characterized in that: The monitoring box is provided with a high-definition camera and an intelligent smoke sensor, and the distribution box is provided with a distribution box controller and a multifunctional sensor.

4. The online intelligent management method for hot work supervision according to claim 3 is characterized in that: It also includes dividing the risks of hot work into three levels, specifically: level one risk, level two risk and level three risk.

5. An online intelligent management system capable of running the online intelligent management method for hot work supervision according to any one of claims 1 to 4, characterized in that: It includes a login module, an approval module, a monitoring module, an analysis module and an early warning module. The login module is used by operators to initiate job applications through the client; the approval module is used by managers to approve jobs; the monitoring module is used to obtain monitoring video data, distribution box temperature data and smoke sensor smoke concentration data; the analysis module is used to perform algorithmic analysis on monitoring video data, distribution box temperature data and smoke sensor smoke concentration data; and the early warning module is used to generate alarm information based on the algorithmic analysis results.

6. An online intelligent management device for hot work supervision, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 4 is executed.

Images