Carbon plant fire-fighting and security integrated control system based on AI video

Abstract

A carbon plant fire-fighting and security integrated control system based on AI video relates to the technical field of fire-fighting and security control, and comprises an Internet of Things sensing layer, a network service layer, an alarm service layer and an application layer, the Internet of Things sensing layer is used for sensing enterprise traditional fire-fighting facility signals and monitoring field water pressure, liquid level, temperature and flame video data; the network service layer analyzes the collected data and images in a comparison and recognition mode, external data access special port associated equipment is created through network resources, the alarm service layer comprises a retrieval module, a processing module and an alarm module, and the application layer comprises a service management module, a mobile office module and a visualization module. Potential safety hazards can be accurately recognized in real time, measures are taken in time for handling, and therefore accidents are effectively prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of fire protection and security control, and in particular to an integrated fire protection and security control system for a carbon plant based on AI video. Background Art

[0002] The fire safety management of the carbon plant has the characteristics of easy occurrence of fire safety accidents, single means of fire IoT perception, and low efficiency in handling fire safety incidents. The materials involved in the production and storage of the carbon plant are flammable, explosive, toxic and harmful, with many hidden risks and risks, prone to fire safety accidents, and incomplete identification of fire safety risks. If the hidden dangers of accidents are not dealt with in time, it is easy to cause frequent fire safety accidents, causing huge losses or impacts. The traditional fire safety monitoring system cannot meet the needs of users for comprehensive perception of fire safety. It is urgent to add information hosts or communication modules to realize data interconnection, enrich data diversity, enrich fire IoT perception means, and evaluate the status of fire safety in multiple dimensions. When fire safety hazards appear in the traditional fire safety monitoring system, it is basically necessary for manpower to go to the scene to confirm, which often leads to time lags, missed events, and even non-confirmation and inaction. A lot of manpower is consumed but still cannot be effectively guaranteed. It is urgent to connect the fire protection system and the security system to realize the business linkage between fire protection and security. Summary of the invention

[0003] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a carbon plant fire protection and security integrated control system based on AI video, which can identify potential safety hazards in real time and accurately, and take timely measures to deal with them, thereby effectively preventing the occurrence of accidents.

[0004] In order to achieve the above object, the main technical solutions adopted by the present invention include: A carbon plant fire protection and security integrated control system based on AI video, including an Internet of Things perception layer, a network service layer, an alarm service layer, and an application layer. The Internet of Things perception layer includes image, video data acquisition equipment, on-site water pressure, liquid level, gas, temperature, flame detection equipment, and traditional fire protection facility signal acquisition equipment. The Internet of Things perception layer is used to perceive the signals of traditional fire protection facilities of the enterprise, monitor on-site water pressure, liquid level, temperature, and flame video data; the network service layer includes a data comparison module, an image recognition module, a network resource module, an analysis resource module, a backup resource module, and a security disaster recovery resource pool. The network service layer analyzes the collected data and images by comparison and recognition, uses network resources to create an external data access dedicated port association device, and the analysis resource module identifies the data of the data comparison module, the image recognition module, and the network resource module, and transmits the data that needs to be alarmed to the alarm service layer, and transmits the data that needs to be backed up to the backup resource module. The data used to establish the resource pool is transported to the security disaster recovery resource pool, and the security disaster recovery resource pool is used to establish the resource pool. The pool is used to store actual alarm accident data and form reports; the alarm service layer includes a retrieval module, a disposal module, and an alarm module. The retrieval module retrieves and marks the set hidden dangers and performs graded priority warning processing. The disposal module coordinates with the fire-fighting equipment to deal with the identified fire safety hazards or accidents, and uploads them to the application layer and centralized management layer through the alarm module. The application layer includes a business management module, a mobile office module, and a visualization module. The business management module is used to coordinate and process business processes related to fire protection. The mobile office module is connected to the business management module through the network, receives tasks and notifications issued by the business management module in real time, checks relevant fire equipment information and emergency plans, and reports the on-site situation and processing progress in a timely manner. The visualization module is connected to the business management module to obtain fire-fighting facilities and emergency plan data, and obtains real-time image information from AI video surveillance. After processing and integration, the distribution of fire-fighting facilities, real-time monitoring screens, and fire risk area information of the carbon plant are displayed in the form of graphics, images, and videos.

[0005] Furthermore, the IoT sensing layer includes network intelligent cameras, smoke and temperature detectors, gas detectors, intelligent liquid level detectors, intelligent water pressure detectors, and traditional fire-fighting facilities. The traditional fire-fighting facilities include smoke alarms, temperature alarms, manual alarms, and alarm bells.

[0006] Furthermore, the data of the IoT perception layer targets six typical safety hazards in carbon plants, including: safety hazards caused by personnel violations, safety hazards caused by door and wall perimeters, safety hazards caused by on-site operations, identification of safety hazards caused by construction in production areas, identification of safety hazards caused by equipment operation, and fire conditions and fire hazards.

[0007] Furthermore, the data comparison module establishes a database, sets a threshold, compares the collected data with the threshold of the database, and extracts problematic data. The database can update data through learning and training, delete false alarm data, and retain valid data.

[0008] Furthermore, the image recognition module process includes creating a text hidden danger library - creating a reference image hidden danger library - obtaining the image to be identified - image registration - image hidden danger library update. Through on-site sample collection, the collected original data and safety data are used as safety data samples, which are compared with the collected data and then sent to the analysis resource module. The image database uses the loss function to repeatedly iterate the training of the detection model.

[0009] Furthermore, the network service layer associates the device by creating a dedicated port for external data access, adopts wired and NB-IoT, Loa wireless methods, uses COAP and MQTT communication protocols, and connects its signals to the platform PC and mobile terminals; by creating an event subscription forwarding interface, the warning events issued by the IoT device are pushed to related services or applications in real time.

[0010] Furthermore, the business management module coordinates and processes business processes related to firefighting, including the formulation of maintenance plans for firefighting equipment, arrangements for personnel training, and the formulation and updating of emergency plans.

[0011] Furthermore, the system also includes a centralized management layer, which includes interconnected PC operation station modules and mobile APP modules. The PC operation station module centrally manages and monitors the entire fire protection system, including viewing AI video surveillance images, obtaining the operating status of fire protection equipment, and analyzing historical data; the mobile APP module includes recording equipment inspection conditions, uploading real-time data and pictures, understanding fire conditions, receiving emergency task instructions, and feedback on processing progress.

[0012] Furthermore, the centralized management level identifies whether there is a fire based on the on-site monitoring images and image analysis of the front-end video surveillance equipment; obtains fire alarm information based on the status of the front-end IoT sensing equipment and the alarm data. When a fire is confirmed, event information and action instructions are pushed to relevant personnel, displayed in GIS maps and two-dimensional models, and linked to video cameras to display rescue information on the map. The beneficial effect of the present invention is that the AI ​​video-based carbon plant fire protection and security integrated control system provided by the present invention can identify potential safety hazards in real time and accurately, and take timely measures to deal with them, thereby effectively preventing the occurrence of accidents, and realizing the access of various fire protection equipment and facilities, security equipment and facilities, monitoring equipment, detection equipment and other equipment, meeting the needs of carbon production enterprises to obtain real-time data on fire protection, security, hidden dangers, environment, occupational health and other data at the production site, realizing full-time and full-area monitoring of the production site, and extending the company's fire protection and security management tentacles. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a structural schematic diagram of the integrated fire protection and security control system for a carbon plant based on AI video of the present invention.

[0014] In the figure: 1 is the IoT perception layer, 2 is the network service layer, 3 is the alarm service layer, 4 is the application layer, and 5 is the centralized management layer. DETAILED DESCRIPTION

[0015] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below through specific implementation modes in conjunction with the accompanying drawings.

[0016] The present invention provides an integrated fire protection and security control system for a carbon plant based on AI video, such as Figure 1 As shown, it includes an IoT perception layer 1, a network service layer 2, an alarm service layer 3, an application layer 4, and a centralized management layer 5. The present invention uses AI video, IoT, big data, cloud computing and other technical means to break through the information islands of traditional security video surveillance and fire alarms, and realizes the interconnection between the fire protection system and the security system through system development, transformation, and docking. It relies on the Internet to remotely transmit information to the administrator terminal, builds a "three-dimensional intelligent" fire prevention and control system, and comprehensively improves the level of fire prevention and control.

[0017] The IoT perception layer 1 includes image and video data acquisition equipment, on-site water pressure, liquid level, gas, temperature, flame detection equipment, and traditional fire-fighting facility signal acquisition equipment. The IoT perception layer is used to perceive the signals of traditional fire-fighting facilities in enterprises and monitor on-site water pressure, liquid level, temperature, and flame video data. Specifically, the IoT perception layer 1 includes network intelligent cameras, smoke and temperature detectors, gas detectors, intelligent liquid level detectors, intelligent water pressure detectors, and traditional fire-fighting facilities. The traditional fire-fighting facilities include smoke alarms, temperature alarms, manual alarms, and alarm bells. Collect suspicious early warning information of fires and alarm signals issued by equipment for summary collection. For the six typical safety hazards of carbon plants, they mainly include: safety hazards of personnel violation status, safety hazard status of door and wall perimeters, safety hazard status of on-site operation operations, safety hazard identification of production area construction, safety hazard status identification of equipment operation, fire status and fire hazards. IoT perception layer 1 is to collect and send to network service layer 2 through traditional early warning signals, fire-fighting facility signals, and data of monitoring on-site water pressure, liquid level, temperature, and video data from the basic equipment end.

[0018] The network service layer 2 is a service layer that operates in an analysis and pooling manner, and provides a service layer for resources such as data comparison, image recognition, network resources, analysis resources, backup resources, security and disaster recovery, and divides resource pools according to business, and provides corresponding services and analysis. Data comparison and image recognition are two core functions of the network service layer 2. Specifically, the network service layer includes a data comparison module, an image recognition module, a network resource module, an analysis resource module, a backup resource module, and a security disaster recovery resource pool. The network service layer analyzes the collected data and images in a comparison and recognition manner, uses network resources to create an external data access dedicated port association device, and the analysis resource module identifies the data of the data comparison module, the image recognition module, and the network resource module, and transmits the data that needs to be alarmed to the alarm service layer, and transmits the data that needs to be backed up to the backup resource module. The data used to establish the resource pool is transmitted to the security disaster recovery resource pool, and the security disaster recovery resource pool is used to store the real alarm accident data and form a report. Ultimately, these data are used to achieve early warning and alarm functions with the support of the network resource layer 2 analysis resource module. The alarm data after analysis needs to be sent to the alarm service layer 3 for further alarm and warning.

[0019] The data comparison module establishes a database, sets a threshold, compares the collected data with the threshold of the database, and extracts problematic data. The database can update data through learning, delete false alarm data, and retain valid data.

[0020] The data of the alarm service layer 3 mentioned above can be fed back to the security disaster recovery resource pool of the network service layer 2, and this data is also the core function of this invention. The image recognition module process includes creating a text hidden danger library-creating a benchmark image hidden danger library-obtaining the image to be identified-image registration-image hidden danger library update. Through on-site sample collection, the collected original data and security data are used as security data samples, and are sent to the analysis resource module after being compared with the collected data. The image database uses the loss function to repeatedly iterate the training detection model. If the data is found to have security risks, it is sent to the alarm module. The original benchmark image hidden danger library established by the image recognition module uses digital image processing methods to perform different transformations on the training samples, increase the diversity of the data, enrich the benchmark image hidden danger library, and continuously upgrade and learn the data. On this basis, the available loss function is used to repeatedly iterate the training detection model to improve the image registration level, and automatically improve the benchmark hidden danger image library to improve the accuracy of hidden danger identification. The perfect hidden danger type eventually forms a set of completely reliable resource pool whitelists for enterprises. The training detection model is established and optimized, and the existing Faster-RCNN network is used to detect specific objects in the image. The key task of image detection is the generation of candidate regions, which can be generated using an exhaustive method. A is a pixel point on the image, and B is another pixel point to the lower right of A. Points A and B can determine a rectangular frame, denoted as AB. A is at the upper left corner of the image, and B traverses all positions except A to generate rectangular frames A1B1, A1B2, ..., A1B n ; A is somewhere in the middle of the image, B traverses all positions to the lower right of A, generating a rectangular box A k B1, A k B2,…,A k B n .

[0021] In the detection process, the number of detected targets is controlled by continuously adjusting the threshold, thereby changing the precision P and recall R. In target detection of multiple object categories, mAP (mean average precision) is generally used as the accuracy indicator, that is, a PR curve is obtained by adjusting the threshold for each category to calculate the average precision AP, and finally the average value of all APs is mAP: F1 is the harmonic mean of precision P and recall R, F1= For the detection of multiple object categories, the average of different categories is taken as the F1 value. According to the evaluation results, in actual use, the threshold is set to 0.9.

[0022] After the above steps, a detection model with high accuracy is obtained, and the model can be deployed on a public cloud server or a private server and called through an independent REST API. The network smart camera installed on site takes pictures of the scene, and the background automatically analyzes and intelligently gives possible hidden dangers. The above data is finally returned to the security disaster recovery resource pool of the network service layer 2 for backup.

[0023] In the network service layer 2, the data processed by the network resource module and the analysis resource module enter the alarm service layer 3. The alarm service layer 3 includes a retrieval module, a disposal module and an alarm module. The retrieval module can search and mark the hidden dangers of key concern to achieve hierarchical priority warning processing. The disposal module links the fire-fighting equipment with the identified fire safety hidden dangers or accidents to deal with the hidden dangers or accidents, and uploads them to the application layer and the centralized management layer through the alarm module. The alarm module can intelligently report and notify the corresponding management personnel to facilitate the timely handling of safety hazards. In this way, the intelligent tracking and management of the identified safety hazards can be realized.

[0024] The data comparison module and image recognition module of the network service layer 2 back up and store the alarm and warning signals in each data. Through continuous learning, a complete set of safety and disaster recovery resource pool modules are finally formed, forming a complete set of carbon plant data, which is the core data foundation of the carbon plant disaster recovery system. The system can also call relevant information on demand with the fire safety integration platform, emergency dispatch and other safety and environmental protection systems of carbon enterprises, support the comprehensive decision-making and rapid disposal capabilities of emergency accidents, and curb the occurrence and spread of major accidents. Help carbon enterprises to open up the process channels of fire safety risk warning (before the event), assessment and judgment decision-making (after the event), hidden danger disposal and event accident linkage response control (during the event), accident backtracking and summary (after the event), realize the collection of all elements at the source, the monitoring of the whole scene of the process, the analysis of the whole model of the results, the prediction, warning and pre-control of the whole process of risks, comprehensively improve the comprehensive management and control capabilities of fire safety risks, lay the foundation for carbon enterprises to achieve green production and inherent safety, and help carbon enterprises realize intelligent fire safety management.

[0025] Data from network service layer 2 and alarm service layer 3 enter application layer 4 for management. Application layer 4 can be regarded as office OA, including business management module, mobile office module and visualization module. The business management module is mainly responsible for coordinating and processing various business processes related to fire protection, such as the formulation of maintenance plans for fire protection equipment, personnel training arrangements, formulation and updating of emergency plans, etc. It is the core management part of the entire system, providing basic business rules and data support for other modules. The mobile office module provides a convenient mobile work platform for fire protection staff. Through mobile phones or other mobile devices, staff can receive tasks and notifications issued by the business management module in real time, can quickly check relevant fire protection equipment information, emergency plans, etc. on site, and can also report the situation and processing progress on site in a timely manner. The mobile office module is connected to the business management module in real time through the network to ensure the timely transmission and update of information. The visualization module displays the distribution of fire protection facilities, real-time monitoring images, fire risk areas and other information of the carbon plant through intuitive graphics, images and videos. It obtains data such as fire protection facilities and emergency plans from the business management module, and obtains real-time image information from AI video surveillance. After processing and integration, it is presented to users in a clear and easy-to-understand form. The staff can quickly understand the fire protection status of the entire carbon plant through the visualization module, providing an intuitive basis for decision-making and action.

[0026] The business management module, mobile office module and visualization module use big data cloud platform technology to achieve data sharing of the entire plant, push switch signals of different units in different areas through WEB and APP, and promptly push alarm information (time, type, location, etc.) to designated fire safety responsible persons. At the same time, the fire control room or fire department of the entire plant can timely grasp the fire safety status of the networked units through the client, which is convenient for centralized supervision, management, statistics, analysis and display. It supports five-level alarm push, that is, according to the importance of fire incidents, the response speed of fire incidents, the efficiency of fire incident handling, and fire supervision and management requirements, the objects of five-level alarm push can be customized.

[0027] The data collected by the application layer 4 is sent to the centralized management layer 5, which includes interconnected PC operation station modules and mobile APP modules. The PC operation station module centrally manages and monitors the entire fire protection system, including viewing AI video surveillance images, obtaining the operating status of fire protection equipment, and analyzing historical data; the mobile APP module includes recording the inspection status of the equipment, uploading real-time data and pictures, understanding the fire situation, receiving emergency task instructions, and feedback on the processing progress. The PC operation station module is usually deployed in the control center or related office space. It has powerful processing power and display functions, and can fully and detailedly display various data and information of the fire protection system. The staff can use the PC operation station module to centrally manage and monitor the entire fire protection system, including viewing AI video surveillance images, obtaining the operating status of fire protection equipment, analyzing historical data, etc. It also provides a rich operation interface and function options to facilitate the staff to perform complex system configuration, task scheduling and emergency command operations. The mobile APP module provides great flexibility and convenience for the staff. By installing the corresponding APP on the mobile device, the staff can obtain key information of the fire protection system anytime and anywhere. For example, during daily inspections, inspectors can use mobile apps to record equipment inspections and upload real-time data and images. When an emergency occurs, even if the relevant personnel are not in the control center, they can quickly understand the general situation of the fire, receive emergency task instructions, and promptly feedback the progress of the processing through the mobile app. The two modules synchronize and interact with each other in real time through the network. Important instructions and configuration information generated by the PC operation station module can be pushed to the mobile app module in a timely manner, and the field data collected by the mobile app module can also be quickly transmitted back to the PC operation station module, realizing seamless information sharing and collaborative work.

[0028] Based on the on-site monitoring images and image analysis technology of the front-end video surveillance equipment, it can identify whether there is a fire; based on the status of the front-end IoT sensing equipment such as smoke sensors, temperature sensors, flame detectors, gas monitors, and fire alarm controllers, and the acquisition of alarm data, fire alarm information can be obtained in a timely manner. When a fire is confirmed, the event information and action instructions are pushed to the relevant personnel as soon as possible, and displayed in GIS maps and two-dimensional models, automatically linked to the surrounding video cameras, and the internal personnel location information, emergency resource distribution information, and surrounding water sources, police force, and rescue force distribution information are displayed on the map to provide support for emergency command; emergency command personnel can check event information and resource distribution in the first time, emergency handling personnel can handle on-site according to the system action instructions, and on-site personnel can escape in an emergency according to the system alarm information and escape route information.

[0029] The centralized management layer 5 realizes the interconnection and data fusion of the fire protection system and the security system through system development, transformation and docking. At the same time, the development and docking of security do not affect any fire protection function and fully meet the requirements of fire protection technical standards. The platform can display information such as the location of the fire control room, affiliated unit, alarm source, alarm time, alarm level, alarm type, and alarm number in a pop-up window. Fire management personnel can view the real-time video and video playback of the points associated with the alarm equipment through video; view the actual location of the alarm point plan through the plan, which is convenient for quick review and confirmation. Fire management personnel can also make handling opinions on the alarm according to the actual situation to confirm whether it is an alarm or a false alarm.

[0030] The centralized management layer 5 predicts fire safety trends under existing conditions through multi-dimensional analysis of fire data, optimizes the allocation of human and material resources, effectively links networked devices, links access control systems to quickly open doors, speeds up personnel evacuation, and ensures the safety of life and property; links vehicle barriers to quickly open, facilitates the entry of fire rescue forces, and improves rescue efficiency.

[0031] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. Alterations, modifications, substitutions and variations of the above embodiments by a person skilled in the art are all within the scope of the present invention.

Claims

1. An integrated fire protection and security control system for a carbon plant based on AI video, characterized in that: It includes an IoT perception layer, a network service layer, an alarm service layer, and an application layer. The IoT perception layer includes image and video data acquisition equipment, on-site water pressure, liquid level, gas, temperature, flame detection equipment, and traditional fire-fighting facility signal acquisition equipment. The IoT perception layer is used to perceive the signals of traditional fire-fighting facilities in enterprises and monitor on-site water pressure, liquid level, temperature, and flame video data; the network service layer includes a data comparison module, an image recognition module, a network resource module, an analysis resource module, a backup resource module, and a security disaster recovery resource pool. The network service layer analyzes the collected data and images by comparison and recognition, uses network resources to create external data access dedicated port association equipment, and the analysis resource module identifies the data of the data comparison module, image recognition module, and network resource module, and transmits the data that needs to be alarmed to the alarm service layer, and transmits the data that needs to be backed up to the backup resource module. The data of the source pool is transmitted to the security disaster recovery resource pool, which is used to store the actual alarm accident data and form reports; the alarm service layer includes a retrieval module, a disposal module, and an alarm module. The retrieval module retrieves and marks the set hidden dangers and performs hierarchical priority warning processing. The disposal module links the identified fire safety hidden dangers or accidents with the fire-fighting equipment to deal with the hidden dangers or accidents, and uploads them to the application layer and centralized management layer through the alarm module. The application layer includes a business management module, a mobile office module, and a visualization module. The business management module is used to coordinate and process business processes related to fire protection. The mobile office module is connected to the business management module through the network, receives tasks and notifications issued by the business management module in real time, checks relevant fire equipment information and emergency plans, and reports the on-site situation and processing progress in a timely manner. The visualization module is connected to the business management module to obtain fire-fighting facilities and emergency plan data, obtains real-time image information from AI video surveillance, and after processing and integration, displays the distribution of fire-fighting facilities, real-time monitoring screens, and fire risk area information of the carbon plant in the form of graphics, images, and videos.

2. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The IoT sensing layer includes network intelligent cameras, smoke and temperature detectors, gas detectors, intelligent liquid level detectors, intelligent water pressure detectors, and traditional fire-fighting facilities. The traditional fire-fighting facilities include smoke alarms, temperature alarms, manual alarms, and alarm bells.

3. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The data of the IoT perception layer targets six typical safety hazards in carbon plants, including: safety hazards caused by personnel violations, safety hazards caused by door and wall perimeters, safety hazards caused by on-site operations, identification of safety hazards caused by construction in production areas, identification of safety hazards caused by equipment operation, and fire conditions and fire hazards.

4. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The data comparison module establishes a database, sets a threshold, compares the collected data with the threshold of the database, and extracts problematic data. The database can be updated through learning and training, and false alarm data can be deleted to retain valid data.

5. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The image recognition module process includes creating a text hidden danger library - creating a reference image hidden danger library - obtaining the image to be identified - image registration - image hidden danger library update. Through on-site sample collection, the collected original data and safety data are used as safety data samples, which are compared with the collected data and then sent to the analysis resource module. The image database uses the loss function to repeatedly iterate the training of the detection model.

6. According to claim 1, the integrated fire protection and security control system for carbon plants based on AI video is characterized in that: The network service layer creates a dedicated port for external data access and associates the device, adopts wired mode and NB-IoT, Loa wireless mode, uses COAP and MQTT communication protocols, and connects its signal to the platform PC and mobile terminal; By creating an event subscription forwarding interface, the warning events issued by IoT devices can be pushed to related services or applications in real time.

7. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The business management module coordinates and processes business processes related to fire protection, including the formulation of maintenance plans for fire protection equipment, the arrangement of personnel training, and the formulation and updating of emergency plans.

8. According to claim 1, the integrated fire protection and security control system for a carbon plant based on AI video is characterized in that: The system also includes a centralized management layer, which includes interconnected PC operation station modules and mobile APP modules. The PC operation station module centrally manages and monitors the entire fire protection system, including viewing AI video surveillance images, obtaining the operating status of fire protection equipment, and analyzing historical data; the mobile APP module includes recording equipment inspection conditions, uploading real-time data and pictures, understanding fire conditions, receiving emergency task instructions, and feedback on processing progress.

9. The AI ​​video-based integrated fire and security control system for a carbon plant according to claim 8 is characterized in that: The centralized management level identifies whether there is a fire based on the on-site monitoring images and image analysis of the front-end video surveillance equipment. It obtains fire alarm information based on the status of the front-end IoT sensing equipment and the alarm data. When a fire is confirmed, it pushes event information and action instructions to relevant personnel, displays them in GIS maps and two-dimensional models, and links video cameras to display rescue information on the map.

Images