Centralized spraying system and method based on AI video monitoring

Abstract

The invention belongs to the technical field of emergency rescue, and relates to a centralized spraying system and method based on AI video monitoring. Comprising an AI video monitoring module, a water supply guarantee module, a spraying eye washer, an alarm device and a control module, the control module is respectively connected with the AI video monitoring module, the water supply guarantee module, the spraying eye washer and the alarm device; and the spraying eye washer is connected with the water supply guarantee module. According to the invention, full-process automatic emergency response from accident sensing, automatic judgment to spray starting is realized, so that the timeliness and accuracy of emergency processing and the overall safety protection level are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of emergency rescue technology and relates to a centralized sprinkler system and method based on AI video monitoring. Background Technology

[0002] In high-risk workplaces such as thermal power plants and chemical plants, sudden injuries (such as chemical burns, fainting, and falls) or hazardous chemical leaks are characterized by their suddenness, rapid hazard, and extremely short response window. Traditional sprinkler and eyewash systems generally use manual triggering, relying on on-site personnel to actively pull levers or press buttons to activate the device after discovering a hazard. While some upgraded systems have introduced infrared sensors or button alarms, human intervention is still required, resulting in significant response delays and an inability to identify the type and severity of the accident. Furthermore, existing systems are mostly deployed independently at single points, lacking centralized water supply assurance, water quality monitoring, and environmental adaptability control capabilities, making it difficult to meet the comprehensive requirements of modern intelligent safety management systems for "early identification, rapid response, stable supply, and traceability."

[0003] However, the aforementioned traditional solutions rely entirely on manual triggering, resulting in inherent defects such as delayed response, susceptibility to subjective judgment, and inoperability at night or during periods of unattended operation. In particular, when personnel have lost their ability to move (e.g., unconscious, curled up in severe pain) or when there are no obvious odors or visible signs in the early stages of a leak, they cannot provide automatic early warning and immediate response, leading to missed golden rescue time and exacerbating the consequences of harm. Summary of the Invention

[0004] To address the problems in existing technologies, this invention provides a centralized sprinkler system and method based on AI video surveillance, which realizes fully automated emergency response from accident detection and automatic judgment to sprinkler activation, thereby significantly improving the timeliness, accuracy, and overall safety protection level of emergency response.

[0005] To achieve the above objectives, the present invention employs the following technical solution: In a first aspect, the present invention provides a centralized sprinkler system based on AI video surveillance, including an AI video surveillance module, a water supply guarantee module, a sprinkler eyewash station, an alarm device, and a control module. The control module is connected to the AI ​​video monitoring module, the water supply guarantee module, the eyewash station, and the alarm device, respectively; the eyewash station is connected to the water supply guarantee module.

[0006] Preferably, the water supply guarantee module includes a water supply pipeline and a water pressure detection unit and a control valve unit installed on the water supply pipeline; one end of the water supply pipeline is connected to the spray eyewash station, and the other end is connected to the water supply source; the control valve unit is connected to the water pressure detection unit and the control module respectively.

[0007] Preferably, the water supply pipeline is also equipped with an electric heat tracing device and a temperature sensor; both the electric heat tracing device and the temperature sensor are connected to the control module.

[0008] Preferably, a turbidity sensor is also installed on the water supply pipeline; the turbidity sensor is connected to the control module.

[0009] Preferably, the inner and outer walls of the water supply pipeline are provided with a hot-dip galvanized layer.

[0010] Preferably, it also includes a communication module and a central control room monitoring terminal; the control module is connected to the central control room monitoring terminal through the communication module.

[0011] Secondly, the present invention provides a centralized sprinkler system based on AI video surveillance, comprising the following steps: The AI ​​video surveillance module collects video data within the monitored area, analyzes the video data to identify personnel injury status or hazardous chemical spill accident status, and issues identification signals. The control module receives recognition signals from the AI ​​video monitoring module; When the identification signal indicates a person is injured or a hazardous chemical spill occurs, the control module performs the following operations: Send a first control command to the eyewash station to control the eyewash station to start spraying; Send a second control command to the alarm device to control the alarm device to trigger an audible and visual alarm; And send a third control command to the water supply guarantee module to control the water supply guarantee module to provide flushing water.

[0012] Preferably, the method for analyzing the video data to identify the state of personnel injury or hazardous chemical spill includes: The video data is input into the trained recognition model; The recognition model analyzes the images in the video data. If a preset abnormal behavior pattern or abnormal state of an object is detected, it is determined to be a state of personal injury or a state of hazardous chemical leakage accident.

[0013] Preferably, the abnormal behavior pattern includes at least one of the following: falling down, abnormal limb curling, abnormal running, and remaining still for a long time near the monitored area; The abnormal condition of the item includes at least one of the following: abnormal dumping, damage, leakage of hazardous chemical containers, and diffusion of unidentified smoke or liquid.

[0014] Preferably, the training method for the recognition model includes: Collect historical video data, which includes normal work behavior of personnel, injury behavior of personnel, and normal and abnormal states of hazardous chemical containers; The categories of human behavior, the status of hazardous chemical containers, and leakage characteristics in the historical video data are labeled to form a training dataset and a validation dataset for training. The training dataset is input into the initial model for multiple rounds of iterative training to optimize the model parameters and obtain the trained model. The trained model is tested using the validation dataset. When the accuracy of the test results reaches a preset threshold, the trained recognition model is obtained.

[0015] Compared with the prior art, the present invention has the following beneficial effects: By setting up an AI video monitoring module to monitor the work area in real time, the system can proactively identify accident states such as personnel injuries or hazardous chemical leaks, achieving non-contact automatic accident judgment and response. A water supply guarantee module ensures a stable and reliable water supply. Sprinkler eyewash stations serve as terminal execution devices, providing timely and effective rinsing and eyewash functions after accident confirmation, enabling rapid physical emergency intervention. An alarm device provides audible and visual alerts when an accident is triggered, guiding injured personnel to quickly locate the sprinkler position, achieving on-site warning and location guidance. A control module centrally coordinates and links all functional modules, automatically triggering emergency response procedures based on monitoring signals, achieving intelligent system integration and unified control. This invention solves the technical problems of traditional systems that rely on manual triggering, cannot proactively identify personnel injuries or chemical leaks, lack environmental adaptability, and lack remote collaborative command capabilities. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0020] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0021] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper," "lower," "horizontal," or "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use, they are only for the convenience of describing the present 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, and therefore should not be construed as a limitation of the present invention. Furthermore, terms such as "first" and "second" are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] Furthermore, the use of the term "horizontal" does not imply that the component must be absolutely horizontal, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0023] In the description of the embodiments of the present invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0024] The present invention will now be described in further detail with reference to the accompanying drawings: The primary objective of this invention is to provide a centralized sprinkler system based on AI video surveillance, comprising an AI video surveillance module, a water supply guarantee module, a sprinkler eyewash station, an alarm device, and a control module. The control module is connected to the AI ​​video monitoring module, the water supply guarantee module, the eyewash station, and the alarm device, respectively; the eyewash station is connected to the water supply guarantee module.

[0025] The AI ​​video surveillance module continuously collects dynamic video streams within the monitored area. It employs at least one high-resolution camera capable of supporting wide dynamic range and low-light imaging, deployed 3-5 meters above the monitored area, with a field of view covering all potentially hazardous work surfaces. The core function of the AI ​​video surveillance module is not simple image acquisition, but rather the real-time interpretation of semantic information such as human posture, movement trajectory, and object status within the image. Its output is a structured recognition signal (including event type, timestamp, and spatial coordinates), rather than raw pixel data.

[0026] The water supply module is used to stably deliver rinsing water to the eyewash station. It is rigidly connected to the eyewash station via corrosion-resistant pipes (such as 304 stainless steel or UPVC) to ensure that the water flow path is free of stagnation and secondary pollution. The eyewash station integrates a top wide-area spray head (coverage diameter ≥30cm) and double-sided eyewash nozzles (spacing conforms to the average interpupillary distance of the human eye), producing a soft mist or laminar flow of water to avoid impact damage. Its installation height meets ergonomic requirements, and the entire unit has an IP66 protection rating and antistatic surface treatment, making it suitable for flammable and explosive environments.

[0027] Taking thermal power plants as an example, the common installation locations of sprinkler eyewash stations are shown in Table 1: Table 1 Common Installation Locations of Spray Eyewash Stations in Thermal Power Plants

[0028] The alarm device provides audible and visual warnings, including a red LED warning light and an explosion-proof buzzer, both of which activate and deactivate synchronously. It is installed above the normal working line of sight (≥2.2m) to ensure strong penetration even in noisy, smoky, or complex lighting environments. The control module uses an industrial-grade programmable logic controller (PLC) or an embedded Linux industrial computer, equipped with multiple communication interfaces such as RS485 / Modbus, TCP / IP, and DI / DO. It features a pre-built multi-threaded task scheduling mechanism, capable of simultaneously handling concurrent tasks such as video signal parsing commands, valve opening and closing control, alarm activation and deactivation, and water supply parameter feedback.

[0029] The water supply guarantee module includes a water supply pipeline and a water pressure detection unit and a control valve unit installed on the water supply pipeline; one end of the water supply pipeline is connected to the spray eyewash station, and the other end is connected to the water supply source; the control valve unit is connected to the water pressure detection unit and the control module respectively.

[0030] The water supply pipeline is a fluid channel used to transport flushing water. The water pressure detection unit is an integrated electronic pressure sensing device, including a pressure sensor core, signal conditioning circuit, and digital communication interface (such as RS485 or CAN bus). Its range covers 0 to 2.5 MPa, with an accuracy class of not less than ±0.5%FS (full scale) and a sampling frequency of not less than 10 Hz. This unit is fixed to the key measuring point of the water supply pipeline by threaded connection or clamp flange, preferably arranged on the water supply pipeline 300 mm to 500 mm upstream of the inlet of the eyewash station, so as to truly reflect the available pressure at the end. Its output pressure analog or digital signal is uploaded to the control module in real time to determine whether the pipeline is in a startable condition (for example, a pressure value of ≥0.3 MPa for 1 second is considered an effective water supply pressure threshold).

[0031] The control valve unit is an electrically actuated fluid on / off / regulation device, comprising a valve body, a drive mechanism (such as a stepper motor or brushless DC motor), a position feedback component, and a drive control circuit. The drive mechanism receives switching commands (such as DO signals) or analog / Modbus commands from the control module to remotely start / stop or adjust the opening. When the water pressure detection unit detects abnormal pressure in the water supply pipeline, it can select to directly drive the control valve unit for a rapid response according to a preset strategy to ensure basic safety. Furthermore, external operators can also drive the control valve unit via the control module for a rapid response, enabling remote intervention.

[0032] The water supply pipeline is also equipped with an electric heat tracing device and a temperature sensor; both the electric heat tracing device and the temperature sensor are connected to the control module. Since the temperature sensor can accurately reflect the local thermal state of the pipeline and feed this state back to the control module in real time, the control module issues precise and timely control commands to drive the electric heat tracing device to heat only when necessary and at the necessary intensity. Furthermore, the electric heat tracing device itself has a temperature self-limiting characteristic, which can avoid the risk of overheating.

[0033] A turbidity sensor is also installed on the water supply pipeline; the turbidity sensor is connected to the control module. After receiving real-time monitoring data from the turbidity sensor, the control module compares and analyzes the data based on preset safety thresholds. When the detected value continuously exceeds the set limit, the control module can automatically start the sewage discharge procedure, that is, send a command to the control valve unit to open the drain outlet to discharge the water stagnant in the pipeline; or drive the water pump to run briefly to achieve pipeline flushing and circulation.

[0034] Both the inner and outer walls of the water supply pipeline are equipped with hot-dip galvanized layers. The hot-dip galvanized layer on the inner wall effectively isolates the direct contact between the water flow and the metal substrate, inhibiting the internal electrochemical corrosion process caused by dissolved oxygen, chloride ions, etc. in the water; the hot-dip galvanized layer on the outer wall is mainly used to resist external humid air, condensate, salt spray, and corrosive liquids that may be splashed, preventing the spread of rust caused by the external environment.

[0035] The system of the present invention also includes a communication module and a central control room monitoring terminal; the control module is connected to the central control room monitoring terminal through the communication module.

[0036] The control module continuously acquires recognition signals from the AI ​​video monitoring module, sensor data such as water pressure / temperature / turbidity from the water supply module, start / stop status of the sprinkler and eyewash stations, and trigger records from alarm devices. This data is then encapsulated into structured data packets (including timestamps, event types, locations, and images) at preset intervals (e.g., 1 second / time) or in an event-driven mode (e.g., at the moment of alarm triggering). These packets are encoded by the communication module and pushed to the central control room monitoring terminal. Simultaneously, the terminal can send configuration commands (e.g., adjusting AI model sensitivity thresholds, setting electric heat tracing start / stop temperature ranges) or manual intervention commands (e.g., forcibly shutting down the electric heat tracing device) to the control module via a reverse channel. This invention enables the dynamic mapping of the operating status, real-time alarm events, and environmental parameters of dispersed sprinkler systems to a centralized management interface without requiring on-site human supervision. This allows safety management personnel to immediately grasp the location, type, and development trend of accidents, and simultaneously retrieve AI video streams from the corresponding areas for manual verification, thereby quickly activating emergency plans, allocating rescue resources, and recording the handling process.

[0037] The communication module can flexibly employ wireless or wired network technologies to construct a reliable information transmission channel, depending on the actual deployment environment and transmission requirements. In areas where cabling is convenient and stability requirements are extremely high, wired network technologies such as industrial Ethernet can be used to provide high-bandwidth, low-latency, and highly interference-resistant deterministic data transmission guarantees. In vast areas where equipment is dispersed, layout is complex, or cabling is difficult, wireless communication technologies such as Wi-Fi, 4G / 5G, or LoRa can be selected to achieve flexible and economical network coverage and device access. This module supports heterogeneous network convergence, ensuring seamless exchange and stable aggregation of data between different types of networks through built-in protocol conversion and data adaptation functions. Ultimately, it efficiently and reliably transmits on-site control commands, status information, and video data to the central control room monitoring terminal, providing a smooth information link for centralized supervision and remote decision-making.

[0038] The second objective of this invention is to provide a centralized sprinkler system based on AI video surveillance, such as... Figure 1 As shown, it includes the following steps: The AI ​​video surveillance module collects video data within the monitored area, analyzes the video data to identify personnel injury status or hazardous chemical spill accident status, and issues identification signals. The control module receives recognition signals from the AI ​​video monitoring module; When the identification signal indicates a person is injured or a hazardous chemical spill occurs, the control module performs the following operations: Send a first control command to the eyewash station to control the eyewash station to start spraying; Send a second control command to the alarm device to control the alarm device to trigger an audible and visual alarm; And send a third control command to the water supply guarantee module to control the water supply guarantee module to provide flushing water.

[0039] This method utilizes an AI video monitoring module to proactively identify abnormal conditions such as personnel injuries or chemical leaks. The control module then simultaneously triggers a coordinated response from the sprinkler, alarm, and water supply systems, achieving a fundamental shift from passive response to proactive protection. This fully automated process not only significantly shortens the response time from accident discovery to emergency response, reducing delays and uncertainties caused by human intervention, but also ensures the accuracy and coordination of emergency actions through integrated system control. This provides more timely and reliable safety guarantees for on-site personnel, comprehensively strengthening the factory's emergency management capabilities.

[0040] The method for analyzing the video data to identify the state of personal injury or hazardous chemical spill includes: The video data is input into the trained recognition model; The recognition model analyzes the images in the video data. If a preset abnormal behavior pattern or abnormal state of an object is detected, it is determined to be a state of personal injury or a state of hazardous chemical leakage accident.

[0041] The abnormal behavior patterns include at least one of the following: falling down, abnormal limb curling, abnormal running, and remaining still for a long time near the monitored area. The abnormal condition of the item includes at least one of the following: abnormal dumping, damage, leakage of hazardous chemical containers, and diffusion of unidentified smoke or liquid.

[0042] This method can accurately capture behavioral patterns indicating injury or discomfort, such as falls, abnormal postures, abnormal movement, or prolonged stillness. It also effectively identifies the states of objects and the environment that indicate an accident, such as spilled or damaged chemical containers, leaks, and even abnormal fumes or liquid diffusion. This recognition mechanism, based on multi-dimensional visual features and behavioral patterns, significantly improves the accuracy and comprehensiveness of accident status assessment, effectively reduces the risk of false alarms and missed alarms, and enables the system to more intelligently and reliably identify real hazards, thus providing a solid technical basis for the subsequent automatic triggering of precise emergency responses.

[0043] The training method for the recognition model includes: Collect historical video data, which includes normal work behavior of personnel, injury behavior of personnel, and normal and abnormal states of hazardous chemical containers; The categories of human behavior, the status of hazardous chemical containers, and leakage characteristics in the historical video data are labeled to form a training dataset and a validation dataset for training. The training dataset is input into the initial model for multiple rounds of iterative training to optimize the model parameters and obtain the trained model. The trained model is tested using the validation dataset. When the accuracy of the test results reaches a preset threshold, the trained recognition model is obtained.

[0044] The training method for this recognition model involves systematically collecting and labeling historical video data covering normal operations, typical accidents, and various abnormal states. This constructs a comprehensive and high-quality training and validation dataset, ensuring that the model can fully cover the diverse scenarios and potential risk characteristics of the complex operating environment of thermal power plants during the learning process. Through multiple rounds of iterative training and rigorous validation testing, the model continuously optimizes its recognition capabilities, ultimately achieving the preset accuracy requirements. This training mode, based on real-world data, effectively improves the model's generalization ability and robustness in recognizing various personnel behaviors and object states in practical applications, laying a solid and reliable core algorithm foundation for the system to achieve high-precision, low-false-alarm intelligent monitoring in actual operation.

[0045] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A centralized sprinkler system based on AI video surveillance, characterized in that, Includes an AI video surveillance module, a water supply guarantee module, a sprinkler eyewash station, an alarm device, and a control module; The control module is connected to the AI ​​video monitoring module, the water supply guarantee module, the eyewash station, and the alarm device, respectively; the eyewash station is connected to the water supply guarantee module.

2. The centralized sprinkler system based on AI video surveillance according to claim 1, characterized in that, The water supply guarantee module includes a water supply pipeline and a water pressure detection unit and a control valve unit installed on the water supply pipeline; one end of the water supply pipeline is connected to the spray eyewash station, and the other end is connected to the water supply source; the control valve unit is connected to the water pressure detection unit and the control module respectively.

3. A centralized sprinkler system based on AI video surveillance according to claim 2, characterized in that, The water supply pipeline is also equipped with an electric heat tracing device and a temperature sensor; both the electric heat tracing device and the temperature sensor are connected to the control module.

4. A centralized sprinkler system based on AI video surveillance according to claim 2, characterized in that, A turbidity sensor is also installed on the water supply pipeline; the turbidity sensor is connected to the control module.

5. A centralized sprinkler system based on AI video surveillance according to claim 2, characterized in that, The inner and outer walls of the water supply pipeline are both covered with hot-dip galvanized layers.

6. A centralized sprinkler system based on AI video surveillance according to claim 1, characterized in that, It also includes a communication module and a central control room monitoring terminal; the control module is connected to the central control room monitoring terminal through the communication module.

7. A centralized sprinkler system based on AI video surveillance, characterized in that, The system based on any one of claims 1 to 6 includes the following steps: The AI ​​video surveillance module collects video data within the monitored area, analyzes the video data to identify personnel injury status or hazardous chemical spill accident status, and issues identification signals. The control module receives recognition signals from the AI ​​video monitoring module; When the identification signal indicates a person is injured or a hazardous chemical spill occurs, the control module performs the following operations: Send a first control command to the eyewash station to control the eyewash station to start spraying; Send a second control command to the alarm device to control the alarm device to trigger an audible and visual alarm; And send a third control command to the water supply guarantee module to control the water supply guarantee module to provide flushing water.

8. A centralized sprinkler system based on AI video surveillance according to claim 7, characterized in that, The method for analyzing the video data to identify the state of personal injury or hazardous chemical spill includes: The video data is input into the trained recognition model; The recognition model analyzes the images in the video data. If a preset abnormal behavior pattern or abnormal state of an object is detected, it is determined to be a state of personal injury or a state of hazardous chemical leakage accident.

9. A centralized sprinkler system based on AI video surveillance according to claim 8, characterized in that, The abnormal behavior pattern includes at least one of the following: falling down, abnormal limb curling, abnormal running, and remaining still for a long time near the monitored area. The abnormal condition of the item includes at least one of the following: abnormal dumping, damage, leakage of hazardous chemical containers, and diffusion of unidentified smoke or liquid.

10. A centralized sprinkler system based on AI video surveillance according to claim 8, characterized in that, The training method for the recognition model includes: Collect historical video data, which includes normal work behavior of personnel, injury behavior of personnel, and normal and abnormal states of hazardous chemical containers; The categories of human behavior, the status of hazardous chemical containers, and leakage characteristics in the historical video data are labeled to form a training dataset and a validation dataset for training. The training dataset is input into the initial model for multiple rounds of iterative training to optimize the model parameters and obtain the trained model. The trained model is tested using the validation dataset. When the accuracy of the test results reaches a preset threshold, the trained recognition model is obtained.

Images