Fire remote alarm monitoring and processing system based on video linkage

Abstract

This invention discloses a video-linked remote fire alarm monitoring and processing system, belonging to the field of fire monitoring and alarm technology. The proposed solution includes a terminal acquisition and linkage module, a data transmission module, a remote monitoring and processing center module, a database module, an image recognition and analysis module, and a smoke analysis module. These modules are interconnected to form a closed-loop architecture. This invention fundamentally solves many pain points of traditional fire alarm systems. Through the integration of multiple technologies, it achieves accurate fire identification; visual and intelligent linkage significantly improves emergency response speed; and the collaborative construction of multiple modules establishes a comprehensive fire control system. Simultaneously, robust data management ensures data security and provides support for the review and optimization of fire-fighting work. Overall, it achieves precise, rapid, and intelligent remote fire alarm monitoring, significantly improving modern fire safety prevention and control capabilities and adapting to the development needs of building intelligence and urban fire safety construction.

Description

Technical Field

[0001] This invention relates to the field of fire monitoring and alarm technology, and in particular to a fire remote alarm monitoring and processing system based on video linkage. Background Technology

[0002] With the continuous advancement of building intelligence and urban fire safety construction, traditional fire alarm systems mostly rely on single-point alarms using sensors such as smoke detectors and heat detectors. This results in problems such as limited alarm information, lack of on-site visual verification, high false alarm rates, and slow remote response. When a sensor triggers an alarm, the fire monitoring center can only obtain the alarm location information and cannot grasp the actual fire situation on-site in a timely manner. False alarms can easily lead to ineffective dispatch of manpower and resources, and the information lag in the actual fire situation can also cause delays in response. At the same time, the video monitoring and fire alarm modules of traditional systems are independent of each other. Video analysis can only achieve basic image acquisition and lacks professional fire feature recognition and analysis capabilities. It cannot achieve in-depth linkage verification between alarm signals and on-site video features, requiring manual operation for viewing and judgment, which further reduces the efficiency of emergency response and makes it difficult to meet the intelligent, rapid, and precise requirements of modern fire safety prevention and control. Summary of the Invention

[0003] The present invention proposes a video-linked remote alarm monitoring and processing system for fire protection, which solves the above-mentioned shortcomings of the prior art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A video-linked fire alarm monitoring and processing system includes a terminal acquisition and linkage module, a data transmission module, a remote monitoring and processing center module, a database module, an image recognition and analysis module, and a smoke analysis module. The modules are interconnected to form a closed-loop architecture.

[0006] Furthermore, the terminal acquisition and linkage module is deployed at the monitoring site and adopts a modular design. The core integrates the flue gas analysis module and is divided into a sensor acquisition unit, a video acquisition unit, a field linkage execution unit, and a terminal control unit. Each unit and the flue gas analysis module operate independently and achieve coordinated linkage through the terminal control unit. The terminal control unit is electrically connected to each unit and the flue gas analysis module to receive feedback signals and issue control commands, thereby realizing integrated control of acquisition, analysis, and linkage.

[0007] The sensor acquisition unit includes a smoke sensor, a temperature sensor, a combustible gas sensor, a flame sensor, and a smoke concentration sensor. Each sensor is distributed and deployed in key areas of the monitoring site to collect environmental parameters in real time. When a parameter exceeds a preset threshold, an alarm trigger signal is sent to the terminal control unit to provide basic data support for the image recognition and analysis module and the smoke analysis module.

[0008] The video acquisition unit includes a high-definition network camera, an infrared thermal imaging camera, and a panoramic camera. The high-definition network camera is used to acquire clear images of the alarm area, the infrared thermal imaging camera is used to capture the ambient temperature distribution and hidden fires, and the panoramic camera is used to acquire the overall picture of the monitoring area. The video acquisition unit supports automatic zoom, 360° rotation, and night vision functions. It can quickly switch the acquisition angle and adjust the focus according to the instructions of the terminal control unit to accurately capture detailed information of the alarm area and simultaneously transmit the acquired video data to the image recognition and analysis module of the remote monitoring and processing center module.

[0009] The on-site linkage execution unit includes on-site audible and visual alarm devices, emergency lighting devices, fire door control devices, smoke exhaust fans, sprinkler systems, and fire broadcasts. Each execution device is deployed according to functional zones, receives instructions from the terminal control unit, and combines the judgment results of the image recognition analysis module and the smoke analysis module to realize functions such as on-site warning, personnel evacuation, and initial fire control.

[0010] The terminal control unit adopts a PLC controller as the core control unit of the terminal acquisition and linkage module. It presets the linkage strategy, receives feedback signals from the sensor acquisition unit and the flue gas analysis module, controls the video acquisition unit to acquire and transmit the data to the image recognition and analysis module, drives the on-site linkage execution unit to perform actions, and realizes communication docking with the data transmission module to complete data uploading and command reception.

[0011] Furthermore, the data transmission module adopts a wired and wireless redundant communication method, including a gateway communication unit, an encrypted transmission unit, and a protocol conversion unit. Each unit works in concert to ensure the stability, security, and compatibility of data transmission between modules.

[0012] The gateway communication unit adopts an industrial-grade gateway to realize communication interfaces between various units of the terminal acquisition and linkage module, the flue gas analysis module and various units of the remote monitoring and processing center module, and the image recognition and analysis module, establishing a two-way data transmission channel and supporting synchronous transmission of multiple terminals and multiple data types.

[0013] The encrypted transmission unit uses the AES-256 encryption algorithm to encrypt the alarm data, video data, smoke analysis module detection data, image recognition analysis module analysis data, and equipment operation data collected by the terminal, to prevent data leakage or tampering and ensure data transmission security.

[0014] The protocol conversion unit converts the different communication protocols of the terminal acquisition and linkage module, the flue gas analysis module, the remote monitoring and processing center module, and the image recognition and analysis module into the TCP / IP standardized protocol, so as to realize the interconnection and interoperability between the units and modules and ensure efficient data transmission.

[0015] Furthermore, the remote monitoring and processing center module, as the core control unit of the system, integrates an image recognition and analysis module, including a data receiving and parsing unit, an intelligent judgment unit, an alarm handling unit, a remote linkage unit, and a visualization control unit. Each unit and the image recognition and analysis module has a clear division of labor and works collaboratively. Combined with data from the smoke analysis module, it achieves full-process control of data parsing, fire situation assessment, specialized analysis, alarm push notification, remote dispatch, and real-time monitoring. The fire level is determined by a weighted comprehensive score, and a false alarm threshold is set. When the feature matching degree is lower than the threshold, it is judged as a false alarm. The core formula for the integrated fire situation assessment and false alarm identification is as follows:

[0016] (1) Formula for total score of comprehensive fire situation assessment:

[0017] ;

[0018] (2) Formula for identifying false alarms:

[0019] when ;

[0020] when Enter the level classification;

[0021] (3) Formula for classifying fire severity levels:

[0022] ;

[0023] in, (Significant fire threshold). (Threshold for major fires); The overall score is based on the sensor data. Weights for sensor data; The overall score is calculated based on the flue gas analysis data. Weights for flue gas analysis data; The overall score for image recognition data; Weights for image recognition data; Score the device status data; Weights for device status data; The threshold for determining false alarms; The score is based on the comprehensive assessment of the fire situation;

[0024] The data receiving and parsing unit receives encrypted data transmitted by the data transmission module, decrypts and parses it, and extracts alarm location, sensor parameters, video data, smoke analysis module detection data, image recognition analysis module analysis data and equipment operating status data, and transmits them synchronously to the intelligent judgment unit, image recognition analysis module and database module.

[0025] The intelligent judgment unit incorporates a fire detection algorithm and a false alarm identification algorithm. Combining sensor-collected data, smoke analysis module detection data, and image recognition analysis module analysis results, it accurately determines whether a fire is real and outputs the fire level. The false alarm identification algorithm identifies false alarms by comparing historical alarm data, environmental parameters, smoke data, image analysis data, and video footage, and automatically deactivates the alarm. It also records the cause of the false alarm for system parameter optimization. The fire detection algorithm is as follows:

[0026] ;

[0027] The formula for effectively identifying and determining a fire is as follows:

[0028] ;

[0029] ;

[0030] in, The original confidence level for image recognition; For the verification coefficient of flue gas data linkage; The final confidence level for fire identification; Set a confidence threshold for fire detection;

[0031] The algorithm for calculating the fire spread rate is as follows:

[0032] Formula for inter-frame area change rate:

[0033] ;

[0034] γ is the rate of change of the flame / smoke area on a per-second scale, reflecting the spread trend of the fire. γ is positive when the area increases and 0 when the area remains unchanged.

[0035] The final formula for calculating the fire spread rate is as follows:

[0036] ;

[0037] in, The area of ​​the flame / smoke region in the nth frame of the image; For the first Area of ​​the flame / smoke region in the frame image; This refers to the inter-frame time interval. This is a correction factor for the type of fire. To the speed at which the fire spreads; Pixel-to-actual distance conversion factor;

[0038] If determined to be smoldering Taking 0.1, the calculation result represents the slow spread rate of smoldering;

[0039] If determined to be an open flame Taking 0.6 accurately reflects the rapid diffusion characteristics of open flame;

[0040] The image recognition and analysis module, as a core specialized module, works in conjunction with the video acquisition unit and the smoke analysis module. It includes an image preprocessing subunit, a feature extraction subunit, a fire identification subunit, and a combustible material prediction subunit to achieve refined analysis of video footage.

[0041] The alarm handling unit includes a hierarchical push subunit, a work order generation subunit, and a handling tracking subunit. The hierarchical push subunit pushes alarm information to the terminals of staff at the corresponding level based on the fire level and the analysis results of the image recognition analysis module and the smoke analysis module. The work order generation subunit automatically generates an emergency handling work order, which clarifies the handling task, responsible personnel, and time nodes. The handling tracking subunit tracks the work order execution progress in real time, records the handling process, and forms a closed loop for handling.

[0042] The remote linkage unit supports linkage with terminal equipment of fire and rescue agencies, property management personnel, and maintenance units. It can share on-site video, fire data, smoke analysis module detection data, image recognition analysis module analysis results and handling progress in real time, receive remote command instructions, adjust the actions of on-site linkage execution units, and remotely control the acquisition angle and focal length of the video acquisition unit and the sampling frequency of the smoke analysis module to achieve remote visual command and precise analysis.

[0043] The visualization control unit adopts a large screen display interface to display the distribution of the monitoring area, the operating status of each unit's equipment, alarm information, video footage, detection data from the flue gas analysis module, analysis results from the image recognition analysis module, and the progress of the handling in real time. It supports multi-screen switching, area zooming, historical video playback, and flue gas data trend analysis, which makes it easier for staff to intuitively grasp the on-site situation and make quick handling decisions.

[0044] Furthermore, the image preprocessing subunit performs noise reduction, enhancement, and frame extraction on the images transmitted by the video acquisition unit to improve image clarity. The feature extraction subunit extracts core features such as flame outline, smoke morphology, and color changes from the image to distinguish flames from other high-temperature light sources, smoke from other sources, and dust from other sources. The fire identification subunit combines feature data and smoke analysis module detection data to accurately identify the type of fire and its spread rate. The combustible material prediction subunit predicts the type of combustible material based on flame characteristics, smoke morphology, and smoke composition data transmitted by the smoke analysis module, providing accurate support for the formulation of response strategies. The analysis results are simultaneously transmitted to the intelligent judgment unit and the alarm response unit.

[0045] Furthermore, the database module adopts a distributed storage architecture, including a real-time storage unit, a historical storage unit, a data retrieval unit, and a backup management unit. It focuses on storing relevant data from the image recognition and analysis module and the flue gas analysis module, while also storing terminal-collected data, alarm records, analysis results, handling procedures, video recordings, and equipment operation data in categories, supporting data classification retrieval, source tracing analysis, and backup management.

[0046] The real-time storage unit uses a Redis database to store current alarm data, video stream data, smoke analysis module detection data, image recognition analysis module analysis data, and equipment operating status data, ensuring real-time data retrieval with a response time of ≤1 second.

[0047] The historical storage unit uses a MySQL database to store historical alarm records, handling procedures, video recordings, historical analysis results from the image recognition and analysis module, historical detection data from the smoke analysis module, and system logs. The storage period is no less than one year, which meets the relevant requirements of fire protection codes for data storage.

[0048] The data retrieval unit supports data retrieval based on time, region, equipment type, fire level, smoke composition, and other conditions. It can quickly retrieve relevant data from the image recognition and analysis module and the smoke analysis module, providing support for fire safety assessment and accident tracing.

[0049] The backup management unit adopts a dual mode of local backup and cloud backup, and regularly backs up the database data to prevent data loss and ensure data security and integrity.

[0050] Furthermore, the flue gas analysis module, as a core special module, includes a flue gas sampler, a flue gas component detector, and a data preprocessing unit. The flue gas sampler is deployed in areas of the monitoring site that are prone to generating flue gas, and collects flue gas samples from the site in real time.

[0051] The smoke component detector adopts electrochemical detection technology and a combination of timed sampling and triggered sampling. The detection accuracy is ≤1ppm and the response time is ≤3 seconds. It can simultaneously detect more than 8 common fire smoke components, supports automatic calibration to ensure the accuracy of detection data, and is used to detect the core components and concentrations in smoke, distinguishing fire smoke from normal environmental smoke. Under normal conditions, it samples once every 5 minutes. After receiving an alarm signal, the sampling frequency is increased to once every 10 seconds, accurately capturing the trend of smoke changes and providing accurate data support for the image recognition and analysis module.

[0052] The data preprocessing unit performs noise reduction and calibration on the detected flue gas data, removes interfering data, and transmits the standardized flue gas data to the terminal control unit. Simultaneously, it uploads the data to the image recognition and analysis module of the remote monitoring and processing center module to collaboratively complete the fire situation determination and response decision.

[0053] Furthermore, the fire identification subunit of the image recognition and analysis module adopts a deep learning model, which has been trained with more than 100,000 fire scene samples. Combined with the detection data of the smoke analysis module, the fire identification accuracy is ≥98%, which can effectively distinguish flames, smoke and other interference factors.

[0054] The combustion material prediction subunit has a built-in combustion material feature database. Combined with flame color, smoke concentration, and smoke composition data transmitted from the smoke analysis module, it achieves accurate prediction of combustion material types with an accuracy rate of ≥90%. The combustion material similarity matching algorithm is based on the combustion material feature database, extracts image feature factors and smoke feature factors, calculates the matching degree between real-time features and standard features in the database using cosine similarity, and obtains a weighted sum to obtain a comprehensive similarity. The combustion material with the highest matching degree and above the threshold is the prediction result. The cosine similarity calculation formula is as follows:

[0055] ;

[0056] Image features: ,calculate The closer the value is to 1, the higher the image feature matching degree.

[0057] Smoke characteristics: ,calculate The closer the value is to 1, the higher the matching degree of the flue gas characteristics;

[0058] Cosine similarity can effectively quantify the similarity between two feature vectors, avoiding the influence of differences in units on the matching results;

[0059] The formula for the overall similarity of combustible materials is as follows:

[0060] ;

[0061] The formula for predicting combustible materials is as follows:

[0062] ;

[0063] ;

[0064] Note: , , ;

[0065] in, For real-time image feature vectors; For the first in the database Standard image feature vectors of combustible materials; For real-time and the Cosine similarity of image features of combustible materials; Image feature weights; This represents the real-time flue gas feature vector. For the first in the database Standard flue gas feature vectors for combustible materials; For real-time and the Cosine similarity of flue gas characteristics of similar combustibles; Weights for flue gas characteristics; For real-time and the Overall similarity of combustibles; A similarity threshold is set for combustible materials.

[0066] Compared with existing technologies, the beneficial effects of this invention are:

[0067] 1. This invention integrates sensor, flue gas analysis and image recognition technologies. The image recognition module has been trained with 100,000+ fire samples, achieving an accuracy rate of ≥98% and a combustion prediction accuracy rate of ≥90%. Furthermore, through multi-dimensional data weighted analysis and false alarm identification algorithms, it can accurately distinguish between fire and interference factors such as dust and light, solving the problems of single-point alarm information and high false alarm rate in traditional systems, and avoiding ineffective scheduling of manpower and resources.

[0068] 2. This invention uses a video acquisition unit paired with a large visual control screen to achieve real-time transmission of on-site images, multi-screen switching, and detail magnification. The remote center can intuitively grasp the fire situation. Combined with the calculation of fire spread speed and the prediction of combustibles, it can quickly formulate a response strategy. At the same time, alarms are pushed in a graded manner and work orders are automatically generated, forming a closed loop for response and solving the problems of information lag and low efficiency of manual operation in traditional systems.

[0069] 3. This invention integrates multiple units and a flue gas analysis module through a terminal acquisition and linkage module, and integrates an image recognition module in a remote center. Each module achieves data interoperability through redundant communication. From on-site data acquisition, analysis and judgment to on-site linkage execution and remote dispatch, it forms an integrated control of the entire process, replacing the traditional system's independent alarm and video monitoring mode, and realizing intelligent linkage of fire prevention and control.

[0070] 4. In this invention, the database adopts a distributed architecture, with separate real-time and historical storage units, which meets the needs of fast data retrieval and long-term storage. It also supports multi-condition retrieval and is equipped with local + cloud dual backup to ensure data security and integrity. Historical data can be used for fire safety assessment and accident tracing, and can also provide a basis for system algorithm optimization, which meets the requirements of fire protection regulations for data management.

[0071] In summary, this invention fundamentally solves many pain points of traditional fire alarm systems. It achieves accurate fire identification through the integration of multiple technologies, significantly improves emergency response speed through visualization and intelligent linkage, and constructs a full-process fire control system through multi-module collaboration. At the same time, the comprehensive data management not only ensures data security but also provides support for the review and optimization of fire protection work. Overall, it realizes the precision, speed, and intelligence of remote fire alarm monitoring, greatly improves modern fire safety prevention and control capabilities, and adapts to the development needs of building intelligence and urban fire safety construction. Attached Figure Description

[0072] Figure 1 This is a flowchart of the overall system of a video-linked remote alarm monitoring and processing system for fire protection proposed in this invention.

[0073] Figure 2 This is a flowchart of the unit modules of a video-linked remote alarm monitoring and processing system for fire protection proposed in this invention. Detailed Implementation

[0074] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0075] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0076] Example

[0077] Reference Figure 1-2 A video-linked fire alarm monitoring and processing system includes a terminal acquisition and linkage module, a data transmission module, a remote monitoring and processing center module, a database module, an image recognition and analysis module, and a smoke analysis module. The modules are interconnected to form a closed-loop architecture.

[0078] In this invention, the terminal acquisition and linkage module is deployed at the monitoring site and adopts a modular design. The core integrates the flue gas analysis module and is divided into a sensor acquisition unit, a video acquisition unit, a field linkage execution unit, and a terminal control unit. Each unit and the flue gas analysis module operate independently and achieve coordinated linkage through the terminal control unit. The terminal control unit is electrically connected to each unit and the flue gas analysis module to receive feedback signals and issue control commands, thereby realizing integrated control of acquisition, analysis, and linkage.

[0079] The sensor acquisition unit includes smoke sensors, temperature sensors, combustible gas sensors, flame sensors, and smoke concentration sensors. Each sensor is distributed and deployed in key areas of the monitoring site to collect environmental parameters in real time. When a parameter exceeds a preset threshold, an alarm trigger signal is sent to the terminal control unit, providing basic data support for the image recognition and analysis module and the smoke analysis module.

[0080] The video acquisition unit includes a high-definition network camera, an infrared thermal imaging camera, and a panoramic camera. The high-definition network camera is used to acquire clear images of the alarm area, the infrared thermal imaging camera is used to capture the ambient temperature distribution and hidden fires (such as smoldering), and the panoramic camera is used to acquire the overall picture of the monitored area. The video acquisition unit supports automatic zoom, 360° rotation, and night vision functions. It can quickly switch the acquisition angle and adjust the focus according to the instructions of the terminal control unit to accurately capture detailed information of the alarm area and simultaneously transmit the acquired video data to the image recognition and analysis module of the remote monitoring and processing center module.

[0081] The on-site linkage execution unit includes on-site audible and visual alarm devices, emergency lighting devices, fire door control devices, smoke exhaust fans, sprinkler systems, and fire broadcasts. Each execution device is deployed according to functional zones, receives instructions from the terminal control unit, and combines the judgment results of the image recognition analysis module and the smoke analysis module to realize functions such as on-site warning, personnel evacuation, and initial fire control.

[0082] The terminal control unit adopts a PLC controller as the core control unit of the terminal acquisition and linkage module. It presets the linkage strategy, receives feedback signals from the sensor acquisition unit and the flue gas analysis module, controls the video acquisition unit to acquire and transmit the data to the image recognition and analysis module, drives the on-site linkage execution unit to perform actions, and realizes communication interface with the data transmission module to complete data uploading and command reception.

[0083] In this invention, the data transmission module adopts a wired and wireless redundant communication method, including a gateway communication unit, an encrypted transmission unit, and a protocol conversion unit. Each unit works in concert to ensure the stability, security, and compatibility of data transmission between modules.

[0084] The gateway communication unit adopts an industrial-grade gateway to realize communication interfaces between various units of the terminal acquisition and linkage module, the flue gas analysis module and various units of the remote monitoring and processing center module, and the image recognition and analysis module. It establishes a two-way data transmission channel and supports the synchronous transmission of multiple terminals and multiple data types (including analysis data from the two main modules).

[0085] The encrypted transmission unit uses the AES-256 encryption algorithm to encrypt alarm data, video data, smoke analysis module detection data, image recognition analysis module analysis data, and equipment operation data collected by the terminal, to prevent data leakage or tampering and ensure data transmission security.

[0086] The protocol conversion unit converts the different communication protocols (such as the Modbus protocol of the sensor, the ONVIF protocol of the video acquisition unit, and the RS485 protocol of the flue gas analysis module) of each unit of the terminal acquisition and linkage module, each unit of the flue gas analysis module and the remote monitoring and processing center module, and the image recognition and analysis module into the TCP / IP standardized protocol, so as to realize the interconnection and interoperability between each unit and module and ensure efficient data transmission.

[0087] In this invention, the remote monitoring and processing center module serves as the core control unit of the system. It integrates an image recognition and analysis module, including a data receiving and parsing unit, an intelligent judgment unit, an alarm handling unit, a remote linkage unit, and a visualization control unit. Each unit and the image recognition and analysis module has a clear division of labor and works collaboratively. Combined with data from the smoke analysis module, it achieves full-process control of data parsing, fire situation assessment, specialized analysis, alarm push notifications, remote dispatch, and real-time monitoring. The fire level is determined by a weighted comprehensive score, and a false alarm threshold is set. When the feature matching degree is below the threshold, it is judged as a false alarm. The core formula for the integrated fire situation assessment and false alarm identification is as follows:

[0088] (1) Formula for total score of comprehensive fire situation assessment:

[0089] ;

[0090] (2) Formula for identifying false alarms:

[0091] when ;

[0092] when Enter the level classification;

[0093] (3) Formula for classifying fire severity levels:

[0094] ;

[0095] in, (Significant fire threshold). (Threshold for major fires); The overall score is based on the sensor data. Weights for sensor data; The overall score is calculated based on the flue gas analysis data. Weights for flue gas analysis data; The overall score for image recognition data; Weights for image recognition data; Score the device status data; Weights for device status data; The threshold for determining false alarms; The score is based on the comprehensive assessment of the fire situation;

[0096] The data receiving and parsing unit receives encrypted data transmitted by the data transmission module, decrypts and parses it, and extracts alarm location, sensor parameters, video data, smoke analysis module detection data, image recognition analysis module analysis data and equipment operating status data, and transmits them synchronously to the intelligent judgment unit, image recognition analysis module and database module.

[0097] The intelligent assessment unit incorporates a fire detection algorithm and a false alarm identification algorithm. Combining sensor data, smoke analysis module data, and image recognition analysis module results, it accurately determines whether a fire is real and outputs the fire level (general fire, large fire, major fire). The false alarm identification algorithm compares historical alarm data, environmental parameters, smoke data, image analysis data, and video footage to identify false alarms (such as dust interference, equipment malfunction, or normal smoke fluctuations) and automatically deactivates the alarm. It also records the cause of the false alarm for system parameter optimization. The fire detection algorithm is as follows:

[0098] ;

[0099] The formula for effectively identifying and determining a fire is as follows:

[0100] ;

[0101] ;

[0102] in, The original confidence level for image recognition; For the verification coefficient of flue gas data linkage; The final confidence level for fire identification; Set a confidence threshold for fire detection;

[0103] The algorithm for calculating the fire spread rate is as follows:

[0104] Formula for inter-frame area change rate:

[0105] ;

[0106] γ is the rate of change of the flame / smoke area on a per-second scale, reflecting the spread trend of the fire. γ is positive when the area increases and 0 when the area remains unchanged.

[0107] The final formula for calculating the fire spread rate is as follows:

[0108] ;

[0109] in, The area of ​​the flame / smoke region in the nth frame of the image; For the first Area of ​​the flame / smoke region in the frame image; This refers to the inter-frame time interval. This is a correction factor for the type of fire. To the speed at which the fire spreads; Pixel-to-actual distance conversion factor;

[0110] If determined to be smoldering Taking 0.1, the calculation result represents the slow spread rate of smoldering;

[0111] If determined to be an open flame Taking 0.6 accurately reflects the rapid diffusion characteristics of open flame;

[0112] The image recognition and analysis module, as a core specialized module, works in conjunction with the video acquisition unit and the smoke analysis module. It includes an image preprocessing subunit, a feature extraction subunit, a fire identification subunit, and a combustible material prediction subunit to achieve refined analysis of video footage.

[0113] The alarm handling unit includes a hierarchical push subunit, a work order generation subunit, and a handling tracking subunit. The hierarchical push subunit pushes alarm information (including on-site video, alarm location, fire level, smoke composition, and predicted combustibles) to the corresponding level of personnel terminals (such as on-site personnel, fire control room, and emergency management department) based on the fire level and the analysis results of the image recognition and smoke analysis modules. The work order generation subunit automatically generates emergency handling work orders, clearly defining the handling tasks, responsible personnel, and time nodes. The handling tracking subunit tracks the work order execution progress in real time, records the handling process, and forms a closed loop for handling.

[0114] The remote linkage unit supports linkage with terminal equipment of fire and rescue agencies, property management personnel, and maintenance units. It can share on-site video, fire data, smoke analysis module detection data, image recognition analysis module analysis results and handling progress in real time, receive remote command instructions, adjust the actions of on-site linkage execution units, and remotely control the acquisition angle and focal length of the video acquisition unit and the sampling frequency of the smoke analysis module to achieve remote visual command and precise analysis.

[0115] The visual control unit uses a large screen display interface to show the distribution of the monitoring area, the operating status of each unit's equipment, alarm information, video footage, detection data from the flue gas analysis module, analysis results from the image recognition analysis module, and the progress of the handling in real time. It supports multi-screen switching, area zooming, historical video playback, and flue gas data trend analysis, making it easy for staff to intuitively grasp the on-site situation and make quick handling decisions.

[0116] In this invention, the image preprocessing subunit performs noise reduction, enhancement, and frame extraction on the images transmitted by the video acquisition unit to improve image clarity. The feature extraction subunit extracts core features such as flame outline, smoke morphology, and color changes from the images to distinguish flames from other high-temperature light sources, smoke from other sources, and dust from other sources. The fire identification subunit combines feature data and data detected by the smoke analysis module to accurately identify the type of fire (such as open flame or smoldering) and its spread rate. The combustible material prediction subunit predicts the type of combustible material (such as wood, plastic, or oil and gas) based on flame characteristics, smoke morphology, and smoke composition data transmitted by the smoke analysis module, providing accurate support for the formulation of disposal strategies. The analysis results are simultaneously transmitted to the intelligent judgment unit and the alarm disposal unit.

[0117] In this invention, the database module adopts a distributed storage architecture, including a real-time storage unit, a historical storage unit, a data retrieval unit, and a backup management unit. It focuses on storing relevant data from the image recognition and analysis module and the flue gas analysis module, while also storing terminal-collected data, alarm records, analysis results, handling procedures, video recordings, and equipment operation data in categories, supporting data classification retrieval, source tracing analysis, and backup management.

[0118] The real-time storage unit uses a Redis database to store current alarm data, video stream data, smoke analysis module detection data, image recognition analysis module analysis data, and equipment operating status data, ensuring real-time data retrieval with a response time of ≤1 second.

[0119] The historical storage unit uses a MySQL database to store historical alarm records, handling procedures, video recordings, historical analysis results from the image recognition and analysis module, historical detection data from the smoke analysis module, and system logs. The storage period is no less than one year, meeting the relevant requirements of fire protection codes for data storage.

[0120] The data retrieval unit supports data retrieval based on time, region, equipment type, fire level, smoke composition, and other conditions. It can quickly retrieve relevant data from the image recognition and analysis module and the smoke analysis module, providing support for fire safety assessment and accident tracing.

[0121] The backup management unit adopts a dual mode of local backup and cloud backup, regularly backing up database data to prevent data loss and ensure data security and integrity.

[0122] In this invention, the flue gas analysis module is the core module, which includes a flue gas sampler, a flue gas component detector, and a data preprocessing unit. The flue gas sampler is deployed in areas of the monitoring site that are prone to generating flue gas (such as warehouses, power distribution rooms, and kitchens) to collect flue gas samples in real time.

[0123] The smoke composition detector employs electrochemical detection technology combined with timed sampling and triggered sampling, achieving a detection accuracy of ≤1ppm and a response time of ≤3 seconds. It can simultaneously detect more than 8 common fire smoke components, supports automatic calibration to ensure the accuracy of detection data, and is used to detect the core components (such as CO, CO2, formaldehyde, benzene series compounds, etc.) and concentrations in smoke, distinguishing between fire smoke and normal environmental smoke. Under normal conditions, it samples once every 5 minutes, and after receiving an alarm signal, the sampling frequency increases to once every 10 seconds, accurately capturing the trend of smoke changes and providing accurate data support for the image recognition and analysis module.

[0124] The data preprocessing unit reduces noise and calibrates the detected flue gas data, removes interfering data, and transmits the standardized flue gas data to the terminal control unit. Simultaneously, it uploads the data to the image recognition and analysis module of the remote monitoring and processing center module to collaboratively complete fire situation determination and response decisions.

[0125] In this invention, the fire identification subunit of the image recognition and analysis module adopts a deep learning model (such as YOLOv8), which has been trained with 100,000+ fire scene samples. Combined with the detection data of the smoke analysis module, the fire identification accuracy is ≥98%, which can effectively distinguish flames, smoke and other interference factors (such as lights, dust and steam).

[0126] The combustible material prediction subunit has a built-in combustible material feature database. Combined with flame color, smoke concentration, and smoke composition data transmitted from the smoke analysis module, it achieves accurate prediction of combustible material types with an accuracy rate of ≥90%. The combustible material similarity matching algorithm is based on the combustible material feature database, extracts image feature factors and smoke feature factors, calculates the matching degree between real-time features and standard features in the database using cosine similarity, and obtains a weighted sum to get a comprehensive similarity. The combustible material with the highest matching degree and above the threshold is the prediction result. The cosine similarity calculation formula is as follows:

[0127] ;

[0128] Image features: ,calculate The closer the value is to 1, the higher the image feature matching degree.

[0129] Smoke characteristics: ,calculate The closer the value is to 1, the higher the matching degree of the flue gas characteristics;

[0130] Cosine similarity can effectively quantify the similarity between two feature vectors, avoiding the influence of differences in units on the matching results;

[0131] The formula for the overall similarity of combustible materials is as follows:

[0132] ;

[0133] The formula for predicting combustible materials is as follows:

[0134] ;

[0135] ;

[0136] Note: , , ;

[0137] in, For real-time image feature vectors; For the first in the database Standard image feature vectors of combustible materials; For real-time and the Cosine similarity of image features of combustible materials; Image feature weights; This represents the real-time flue gas feature vector. For the first in the database Standard flue gas feature vectors for combustible materials; For real-time and the Cosine similarity of flue gas characteristics of similar combustibles; Weights for flue gas characteristics; For real-time and the Overall similarity of combustibles; A similarity threshold is set for combustible materials.

[0138] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A fire alarm remote monitoring and processing system based on video linkage, characterized in that, It includes a terminal acquisition and linkage module, a data transmission module, a remote monitoring and processing center module, a database module, an image recognition and analysis module, and a flue gas analysis module; The terminal acquisition and linkage module is deployed at the monitoring site and adopts a modular design, with the core integrated flue gas analysis module. The data transmission module adopts a redundant wired and wireless communication method to realize the interconnection and interoperability between various units and modules, and ensure efficient data transmission. The remote monitoring and processing center module serves as the core control unit of the system. It integrates an image recognition and analysis module to achieve full-process control of data parsing, fire situation assessment, special analysis, alarm push, remote dispatch, and real-time monitoring. The database module adopts a distributed storage architecture to store relevant data from the image recognition and analysis module and the flue gas analysis module. It also stores terminal-collected data, alarm records, analysis results, handling procedures, video recordings, and equipment operation data in categories, supporting data classification retrieval, source tracing analysis, and backup management. The flue gas analysis module is integrated into the terminal acquisition and linkage module and is used to collect flue gas samples and detect component concentrations. The image recognition and analysis module is integrated into the remote monitoring and processing center module, and works together to achieve integrated operation of data collection, transmission, analysis, handling, and storage, thereby improving the accuracy of fire identification and the efficiency of handling.

2. The fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The terminal acquisition and linkage module includes: The sensor acquisition unit is used to collect environmental parameters and send alarm signals. The video acquisition unit is used to capture on-site images and transmit them to the image recognition and analysis module. The on-site coordination and execution unit is used to execute on-site warning, evacuation and fire control actions; The terminal control unit is used to manage the coordinated operation of each unit and the flue gas analysis module.

3. The fire alarm remote monitoring and processing system based on video linkage according to claim 2, characterized in that, The sensor acquisition unit includes smoke, temperature, combustible gas, flame, and smoke concentration sensors, which are distributed and deployed in key monitoring areas. When parameters exceed thresholds, an alarm trigger signal is sent, providing basic data for the image recognition and analysis module and the smoke analysis module.

4. A fire alarm remote monitoring and processing system based on video linkage according to claim 2, characterized in that, The video acquisition unit includes a high-definition network, infrared thermal imaging, and a panoramic camera. It supports automatic zoom, 360° rotation, and night vision. It accurately acquires images in response to commands from the terminal control unit and transmits them synchronously to the image recognition and analysis module for detailed analysis.

5. A fire alarm remote monitoring and processing system based on video linkage according to claim 2, characterized in that, The on-site linkage execution unit includes audible and visual alarms, emergency lighting, fire door control, smoke exhaust fans, sprinkler systems, and fire broadcasts. It responds to instructions from the terminal control unit and completes on-site emergency actions based on the analysis results of the two main modules.

6. A fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The data transmission module includes: The gateway communication unit establishes a bidirectional data transmission channel; The encrypted transmission unit uses the AES-256 algorithm to encrypt data. The protocol conversion unit converts various protocols into the TCP / IP standardized protocol, ensuring interoperability between modules.

7. A fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The remote monitoring and processing center module includes: The data receiving and parsing unit is used to decrypt and parse various types of data; The intelligent assessment unit combines the analysis results from the image recognition and analysis module and the smoke analysis module to assess the fire level and identify false alarms. Alarm handling unit, hierarchical push and tracking of handling work orders; Remote linkage unit, which links external terminals and controls field equipment; The visual control unit displays various monitoring data and the analysis results from the image recognition and analysis module and the flue gas analysis module.

8. A fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The image recognition and analysis module includes image preprocessing, feature extraction, fire identification, and combustible material prediction subunits. Combined with data from the smoke analysis module, it enables accurate identification of fire type, combustible material, and calculation of spread rate.

9. A fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The database module includes: The real-time storage unit stores various real-time data and data information analyzed by the image recognition and analysis module and the flue gas analysis module. The historical storage unit stores historical data, logs, and historical analysis results from the image recognition and analysis module and the flue gas analysis module. The data retrieval unit supports multi-condition data retrieval; The backup management unit adopts a dual backup mode, including local and cloud backups.

10. A fire alarm remote monitoring and processing system based on video linkage according to claim 1, characterized in that, The flue gas analysis module includes a flue gas sampler, a flue gas component detector, and a data preprocessing unit. After collecting flue gas samples, detecting component concentrations, and preprocessing the data, it transmits the data to the terminal control unit and the image recognition and analysis module of the remote monitoring terminal to collaboratively complete the fire situation assessment.

Images