Fire water monitor intelligent control system and method based on AI

Abstract

The invention provides an AI-based fire water monitor intelligent control system and method, and the system comprises a fire source detection module which is used for recognizing and positioning a main fire source and outputting a recognition result, and the fire source detection module comprises a laser range finder, a dual-light camera and a computer; the control processing module is used for obtaining the recognition result and generating a control instruction in combination with a first algorithm, a second algorithm and a third algorithm, and the control processing module comprises a CPU, an input unit, an output unit, a communication unit and a power module; and the execution module is used for executing angle rotation of the motor according to the control instruction, and the execution module comprises a pitching motor control unit and a rotating motor control unit. According to the AI-based intelligent control system and method for the fire water monitor, the control instruction is generated by obtaining the recognition result and combining the first algorithm, the second algorithm and the third algorithm, rapid and accurate fire source recognition and efficient fire extinguishing control are achieved, and the intelligent level and the fire extinguishing efficiency of the system are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of intelligent fire fighting and fire extinguishing technology, and in particular to an AI-based fire water cannon intelligent control system and method. Background Art

[0002] In recent years, with the progress of society and the continuous development of science and technology, my country has made great progress in dealing with fires. However, with the continuous acceleration of urbanization and the continuous growth of population, the number of fires in my country and the losses caused by them are still on the rise. Fire extinguishing is still one of the problems in today's world. At present, the operation process of the traditional fire water cannon fire extinguishing system is usually as follows: When a fire occurs, the on-site personnel will sound the fire alarm after discovering the fire. After receiving the fire alarm signal, the fire control center will dispatch firefighters to the scene. After the firefighters arrive at the scene, they will manually operate the fire water cannon to extinguish the fire according to the fire situation.

[0003] This traditional fire water cannon fire extinguishing system has the following disadvantages:

[0004] 1. Slow response: After a fire occurs, it takes time for humans to discover the fire, sound the fire alarm, and have firefighters rush to the scene, which results in a slow response of the fire extinguishing system.

[0005] 2. Low fire-fighting accuracy: Due to reliance on manual operation, the positioning and spraying accuracy of fire-fighting water monitors are difficult to guarantee, which can easily lead to waste of water resources and may delay the fire-fighting opportunity.

[0006] 3. Affected by environmental factors: In conditions of smoke, dim light, etc., manual operation is prone to misjudgment, affecting the fire extinguishing effect.

[0007] In order to overcome the shortcomings of traditional fire water cannon fire extinguishing systems, in recent years, with the rapid development of artificial intelligence technology, some researchers have begun to apply artificial intelligence technology to fire water cannon control systems and have made certain progress.

[0008] For example, a fire monitor control system based on machine vision uses a camera to collect images of the fire scene, identifies the location of the fire source through an image recognition algorithm, and controls the fire monitor to automatically point to the fire source to extinguish the fire. This method can improve the response speed and fire extinguishing accuracy of the fire monitor, but it requires a high accuracy of the image recognition algorithm and may have the risk of misjudgment in complex fire situations.

[0009] Another example is a sensor-based fire monitor control system, which installs various sensors on the fire monitor, such as temperature sensors, smoke sensors, etc., to determine the location and size of the fire source by collecting sensor data, and control the fire monitor to extinguish the fire. This method can overcome the influence of environmental factors such as smoke and dim light, but the installation and maintenance costs of the sensor are high, and the sensitivity and reliability of the sensor need to be further improved.

[0010] Although the existing fire water monitor control system based on artificial intelligence technology and sensor technology has made some progress, it still has the following shortcomings:

[0011] 1. The accuracy of fire source identification is not high enough: Existing image recognition algorithms and sensors have certain errors in identifying fire sources, especially in complex fire situations, which are prone to missed detection or misjudgment.

[0012] 2. Imperfect fire extinguishing control strategy: The existing fire extinguishing control strategy is usually relatively simple and cannot be flexibly adjusted according to changes in the fire situation, resulting in low fire extinguishing efficiency.

[0013] 3. The system is not highly intelligent: The existing fire water cannon control system lacks effective fire analysis and decision-making mechanisms, making it difficult to achieve autonomous control and coordinated operations.

[0014] Therefore, it is necessary to provide an AI-based fire water cannon intelligent control system and method that can solve the above problems. Summary of the invention

[0015] The present invention provides an AI-based fire water monitor intelligent control system and method, which achieves fast and accurate fire source identification and efficient fire extinguishing control by acquiring the recognition results and combining the first algorithm, the second algorithm and the third algorithm to generate control instructions, thereby improving the intelligence level and fire extinguishing efficiency of the system.

[0016] An embodiment of the present invention provides an AI-based fire water monitor intelligent control system, the system comprising:

[0017] A fire source detection module, which is used to identify and locate the main fire source and output the identification result, and the fire source detection module includes a laser rangefinder, a dual-light camera and a computer;

[0018] A control processing module, which is used to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, and the control processing module includes a CPU, an input unit, an output unit, a communication unit and a power module;

[0019] An execution module, which is used to execute the angular rotation of the motor according to the control instruction, and the execution module includes a pitch motor control unit and a rotation motor control unit;

[0020] The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information;

[0021] The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame;

[0022] The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source;

[0023] Wherein, the first algorithm comprises:

[0024] Using the dual-light camera to obtain images of the fire scene;

[0025] Using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, wherein the target is the fire source;

[0026] Using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera;

[0027] The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

[0028] Preferably, the target detection model is an improved YOLOV8n target detection model. In the FSC-Net architecture based on YOLOV8n, a core component BasicStage is introduced. The core component BasicStage includes a PConv layer.

[0029] The number of memory accesses of the PConv layer is:

[0030]

[0031] Among them, h and w represent the width and height of the feature map respectively, k represents the size of the convolution kernel, and c p Indicates the number of channels for PConv convolution;

[0032] The Focaler-IoU loss function is introduced to replace the CIoU loss function. The specific formula is as follows:

[0033]

[0034] Among them, IoU focaler represents the Focaler-IoU loss function, [d,u]∈[0,1].

[0035] Preferably, the second algorithm comprises:

[0036] When there are multiple fire sources in the image of the fire scene, the fire source in the largest calibration frame is determined as the main fire source according to the size of the fire source calibration frame. When the main fire source is extinguished, a new main fire source is re-determined according to the size of the fire source calibration frame.

[0037] Preferably, the third algorithm comprises:

[0038] The coordinate system is determined for the center pixel point position of the calibration frame of the main fire source and the center pixel point position of the main shooting subject image. After determining the quadrant position of the main fire source at the center point of the screen, the control processing module sends an instruction to the dual-light camera to move the dual-light camera to the quadrant position until the pixel point of the calibration frame of the main fire source coincides with the center pixel point of the main shooting subject image. The control processing module sends an instruction to the execution module to enable the fire water monitor to spray water.

[0039] Preferably, the third algorithm further includes:

[0040] The jet landing point is used as a calibration frame, and the center position of the jet calibration frame and the center pixel point position of the subject image are determined by the coordinate system. After determining the quadrant position of the jet at the center point of the screen, the control processing module sends an instruction to the execution module to move the fire water monitor to the quadrant position until the pixel point of the jet calibration frame coincides with the center pixel point of the subject image, so that the main fire source and the water source are in the same frame.

[0041] Preferably, the third algorithm further includes:

[0042] When no fire source is identified, the control processing module sends an instruction to the dual-light camera to rotate the dual-light camera to photograph potential fire sources.

[0043] Preferably, the dual-light camera is a natural light and infrared dual-light camera.

[0044] Preferably, the pitch motor control unit is used to receive instructions from the control processing module and control the pitch motor to run forward and reverse to adjust the fire water monitor to rotate up and down to the jet angle.

[0045] Preferably, the rotating motor control unit is used to receive instructions from the control processing module and control the rotating motor to run forward and reverse to adjust the fire water monitor to rotate left and right to a swing angle.

[0046] The embodiment of the present invention also provides an AI-based fire water monitor intelligent control method, the method comprising:

[0047] Using a fire source detection module to identify and locate the main fire source and output the identification result, the fire source detection module includes a laser rangefinder, a dual-light camera and a computer;

[0048] Using a control processing module to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, the control processing module including a CPU, an input unit, an output unit, a communication unit and a power module;

[0049] Using an execution module to execute the angle rotation of the motor according to the control instruction, the execution module includes a pitch motor control unit and a rotation motor control unit;

[0050] The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information;

[0051] The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame;

[0052] The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source;

[0053] Wherein, the first algorithm comprises:

[0054] Using the dual-light camera to obtain images of the fire scene;

[0055] Using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, wherein the target is the fire source;

[0056] Using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera;

[0057] The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

[0058] Compared with the prior art, the technical solution of the embodiment of the present invention has the following beneficial effects:

[0059] The AI-based fire monitor intelligent control system provided by the embodiment of the present invention comprises a fire source detection module, which is used to identify and locate the main fire source and output the identification result, and the fire source detection module comprises a laser rangefinder, a dual-light camera and a computer; a control processing module, which is used to obtain the identification result and generate a control instruction in combination with a first algorithm, a second algorithm and a third algorithm, and the control processing module comprises a CPU, an input unit, an output unit, a communication unit and a power supply module; an execution module, which is used to execute the angle rotation of the motor according to the control instruction, and the execution module comprises a pitch motor control unit and a rotation motor control unit. By obtaining the identification result and generating a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, fast and accurate fire source identification and efficient fire extinguishing control are achieved, thereby improving the intelligence level and fire extinguishing efficiency of the system, and being particularly suitable for complex fire situations;

[0060] Furthermore, when there are multiple fire sources in the image of the fire scene, the fire source with the largest calibration frame is determined as the main fire source according to the size of the fire source calibration frame. When the main fire source is extinguished, a new main fire source is re-determined according to the size of the fire source calibration frame, thereby giving priority to the main fire source and improving the fire extinguishing efficiency.

[0061] Furthermore, the coordinate system is determined for the center pixel point position of the calibration frame of the main fire source and the center pixel point position of the main subject image. After determining the quadrant position of the main fire source at the center point of the screen, the control processing module sends an instruction to the dual-light camera to move the dual-light camera to the quadrant position until the pixel point of the calibration frame of the main fire source coincides with the center pixel point of the main subject image. The control processing module sends an instruction to the execution module to cause the fire monitor to spray water, thereby improving the aiming rate of the fire monitor for the main fire source;

[0062] Furthermore, the jet landing point is used as a calibration frame, and the center position of the jet calibration frame and the center pixel point position of the subject image are determined by the coordinate system. After determining the quadrant position of the jet at the center point of the screen, the control processing module sends an instruction to the execution module to move the fire water monitor to the quadrant position until the pixel point of the jet calibration frame coincides with the center pixel point of the subject image, so that the main fire source and the water source are in the same frame, thereby further improving the aiming rate of the fire water monitor for the main fire source. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention, not all embodiments. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0064] Figure 1 A schematic diagram of a module of an AI-based fire-fighting water monitor intelligent control system provided by one embodiment of the present invention;

[0065] Figure 2 A schematic flow chart of a first algorithm in an AI-based fire water monitor intelligent control system provided by one embodiment of the present invention;

[0066] Figure 3 A schematic diagram of a flow chart of a second algorithm in an AI-based fire water monitor intelligent control system provided by an embodiment of the present invention;

[0067] Figure 4 A schematic flow chart of a third algorithm in an AI-based fire water cannon intelligent control system provided for one embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0070] Based on the problems existing in the prior art, an embodiment of the present invention provides an AI-based fire water cannon intelligent control system, which achieves fast and accurate fire source identification and efficient fire extinguishing control by acquiring the recognition results and combining the first algorithm, the second algorithm and the third algorithm to generate control instructions, thereby improving the intelligence level and fire extinguishing efficiency of the system.

[0071] Figure 1 A schematic diagram of a module of an AI-based fire water monitor intelligent control system provided by an embodiment of the present invention, Figure 2A schematic diagram of a flow chart of a first algorithm in an AI-based fire water cannon intelligent control system provided by an embodiment of the present invention. Figure 1 and Figure 2 , an embodiment of the present invention provides a module schematic diagram of an AI-based fire water monitor intelligent control system, the system comprising:

[0072] A fire source detection module 11, which is used to identify and locate the main fire source and output the identification result. The fire source detection module 11 includes a laser rangefinder 111, a dual-light camera 112 and a computer 113;

[0073] A control processing module 12, which is used to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, and the control processing module includes a CPU 121, an input unit 122, an output unit 123, a communication unit 124 and a power module 125;

[0074] The execution module 13 is used to execute the angular rotation of the motor according to the control instruction. The execution module includes a pitch motor control unit 131 and a rotation motor control unit 132 .

[0075] The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information;

[0076] The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame;

[0077] The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source;

[0078] Wherein, the first algorithm comprises:

[0079] Step S201: using the dual-light camera to obtain an image of the fire scene;

[0080] Step S202: using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, where the target is the fire source;

[0081] Step S203: using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera;

[0082] Step S204: The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

[0083] Specifically, the control processing module 12 may be a PLC (Programmable Logic Controller) controller.

[0084] In the specific implementation, the target detection model is an improved YOLOV8n target detection model. In the FSC-Net architecture based on YOLOV8n, a core component BasicStage is introduced. The core component BasicStage includes a PConv layer.

[0085] The number of memory accesses of the PConv layer is:

[0086]

[0087] Among them, h and w represent the width and height of the feature map respectively, k represents the size of the convolution kernel, and c p Indicates the number of channels for PConv convolution;

[0088] The Focaler-IoU loss function is introduced to replace the CIoU loss function. The specific formula is as follows:

[0089]

[0090] Among them, IoU focaler represents the Focaler-IoU loss function, [d,u]∈[0,1].

[0091] In the specific implementation, PConv convolution channel number c p The number of memory accesses of PConv is set to 1 / 4 of that of regular convolution. The remaining channels do not participate in the calculation, so there is no need to access memory, so PConv can better utilize the computing power on the device.

[0092] By using the Focaler-IoU loss function, edge regression can be improved to solve the problem of sample imbalance in target detection. The Focaler-IoU function dynamically adjusts the focus of the loss function so that the model can adaptively learn according to the difficulty of the sample to suit a variety of detection tasks.

[0093] Figure 3 A schematic diagram of a flow chart of a second algorithm in an AI-based fire water cannon intelligent control system provided by an embodiment of the present invention. Figure 3 , the second algorithm comprises:

[0094] Step S301: when there are multiple fire sources in the image of the fire scene, determine the fire source with the largest calibration frame as the main fire source according to the size of the fire source calibration frame;

[0095] Step S302: After the main fire source is extinguished, a new main fire source is re-determined according to the size of the fire source calibration frame.

[0096] Figure 4 A schematic diagram of a flow chart of a third algorithm in an AI-based fire water monitor intelligent control system provided by an embodiment of the present invention. Figure 4 , the third algorithm comprises:

[0097] Step S401: Determine the coordinate system of the center pixel point position of the calibration frame of the main fire source and the center pixel point position of the main shooting subject image. After determining the quadrant position of the main fire source at the center point of the screen, the control processing module sends an instruction to the dual-light camera to move the dual-light camera to the quadrant position until the pixel point of the calibration frame of the main fire source coincides with the center pixel point of the main shooting subject image. The control processing module sends an instruction to the execution module to enable the fire water monitor to spray water.

[0098] Step S402: Use the jet landing point as a calibration frame, and determine the center position of the jet calibration frame and the center pixel point position of the subject image by coordinate system. After determining the quadrant position of the jet at the center point of the screen, the control processing module sends an instruction to the execution module to move the fire water monitor to the quadrant position until the pixel point of the jet calibration frame coincides with the center pixel point of the subject image, so that the main fire source and the water source are in the same frame.

[0099] Step S403: When no fire source is identified, the control processing module sends an instruction to the dual-light camera to rotate the dual-light camera to photograph a potential fire source.

[0100] In a specific implementation, the dual-light camera is a natural light and infrared dual-light camera, thereby ensuring that the system can work at any time period, and at the same time, the natural light image and the infrared image are compared with each other to improve the recognition accuracy.

[0101] In a specific implementation, the pitch motor control unit is used to receive instructions from the control processing module and control the pitch motor to run forward and reverse to adjust the fire water monitor to rotate up and down to the jet angle.

[0102] In a specific implementation, the rotating motor control unit is used to receive instructions from the control processing module and control the rotating motor to run forward and reverse to adjust the fire water monitor to rotate left and right to a swing angle.

[0103] The embodiment of the present invention also provides an AI-based fire water monitor intelligent control method, the method comprising:

[0104] Using a fire source detection module to identify and locate the main fire source and output the identification result, the fire source detection module includes a laser rangefinder, a dual-light camera and a computer;

[0105] Using a control processing module to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, the control processing module including a CPU, an input unit, an output unit, a communication unit and a power module;

[0106] Using an execution module to execute the angle rotation of the motor according to the control instruction, the execution module includes a pitch motor control unit and a rotation motor control unit;

[0107] The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information;

[0108] The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame;

[0109] The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source;

[0110] Wherein, the first algorithm comprises:

[0111] Using the dual-light camera to obtain images of the fire scene;

[0112] Using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, wherein the target is the fire source;

[0113] Using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera;

[0114] The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

[0115] In summary, the AI-based fire-fighting water cannon intelligent control system provided by the embodiment of the present invention comprises a fire source detection module, which is used to identify and locate the main fire source and output the identification result, and the fire source detection module comprises a laser rangefinder, a dual-light camera and a computer; a control processing module, which is used to obtain the identification result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, and the control processing module comprises a CPU, an input unit, an output unit, a communication unit and a power supply module; an execution module, which is used to execute the angle rotation of the motor according to the control instruction, and the execution module comprises a pitch motor control unit and a rotation motor control unit. By obtaining the identification result and generating a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, fast and accurate fire source identification and efficient fire extinguishing control are achieved, thereby improving the intelligence level and fire extinguishing efficiency of the system, and being particularly suitable for complex fire situations;

[0116] Furthermore, when there are multiple fire sources in the image of the fire scene, the fire source with the largest calibration frame is determined as the main fire source according to the size of the fire source calibration frame. When the main fire source is extinguished, a new main fire source is re-determined according to the size of the fire source calibration frame, thereby giving priority to the main fire source and improving the fire extinguishing efficiency.

[0117] Furthermore, the coordinate system is determined for the center pixel point position of the calibration frame of the main fire source and the center pixel point position of the main subject image. After determining the quadrant position of the main fire source at the center point of the screen, the control processing module sends an instruction to the dual-light camera to move the dual-light camera to the quadrant position until the pixel point of the calibration frame of the main fire source coincides with the center pixel point of the main subject image. The control processing module sends an instruction to the execution module to cause the fire monitor to spray water, thereby improving the aiming rate of the fire monitor for the main fire source;

[0118] Furthermore, the jet landing point is used as a calibration frame, and the center position of the jet calibration frame and the center pixel point position of the subject image are determined by the coordinate system. After determining the quadrant position of the jet at the center point of the screen, the control processing module sends an instruction to the execution module to move the fire water monitor to the quadrant position until the pixel point of the jet calibration frame coincides with the center pixel point of the subject image, so that the main fire source and the water source are in the same frame, thereby further improving the aiming rate of the fire water monitor for the main fire source.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-based fire water monitor intelligent control system, characterized in that: The system comprises: A fire source detection module, which is used to identify and locate the main fire source and output the identification result, and the fire source detection module includes a laser rangefinder, a dual-light camera and a computer; A control processing module, which is used to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, and the control processing module includes a CPU, an input unit, an output unit, a communication unit and a power module; An execution module, which is used to execute the angular rotation of the motor according to the control instruction, and the execution module includes a pitch motor control unit and a rotation motor control unit; The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information; The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame; The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source; Wherein, the first algorithm includes: Using the dual-light camera to obtain images of the fire scene; Using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, wherein the target is the fire source; Using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera; The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

2. The AI-based fire water monitor intelligent control system is characterized by: The target detection model is an improved YOLOV8n target detection model. In the FSC-Net architecture based on YOLOV8n, a core component BasicStage is introduced. The core component BasicStage includes a PConv layer. The number of memory accesses of the PConv layer is: Among them, h and w represent the width and height of the feature map respectively, k represents the size of the convolution kernel, and c p Indicates the number of channels for PConv convolution; The Focaler-IoU loss function is introduced to replace the CIoU loss function. The specific formula is as follows: Among them, IoU focaler Denotes the Focaler-IoU loss function, [d,u]∈[0,1].

3. The AI-based fire water monitor intelligent control system according to claim 1 is characterized in that: The second algorithm comprises: When there are multiple fire sources in the image of the fire scene, the fire source in the largest calibration frame is determined as the main fire source according to the size of the fire source calibration frame. When the main fire source is extinguished, a new main fire source is re-determined according to the size of the fire source calibration frame.

4. The AI-based fire water monitor intelligent control system according to claim 3 is characterized in that: The third algorithm comprises: The coordinate system is determined for the center pixel point position of the calibration frame of the main fire source and the center pixel point position of the main shooting subject image. After determining the quadrant position of the main fire source at the center point of the screen, the control processing module sends an instruction to the dual-light camera to move the dual-light camera to the quadrant position until the pixel point of the calibration frame of the main fire source coincides with the center pixel point of the main shooting subject image. The control processing module sends an instruction to the execution module to enable the fire water monitor to spray water.

5. The AI-based fire water monitor intelligent control system according to claim 4 is characterized in that: The third algorithm also includes: The jet landing point is used as a calibration frame, and the center position of the jet calibration frame and the center pixel point position of the subject image are determined by the coordinate system. After determining the quadrant position of the jet at the center point of the screen, the control processing module sends an instruction to the execution module to move the fire water monitor to the quadrant position until the pixel point of the jet calibration frame coincides with the center pixel point of the subject image, so that the main fire source and the water source are in the same frame.

6. The AI-based fire water monitor intelligent control system according to claim 4 is characterized in that: The third algorithm also includes: When no fire source is identified, the control processing module sends an instruction to the dual-light camera to rotate the dual-light camera to photograph potential fire sources.

7. The AI-based fire water monitor intelligent control system according to claim 1 is characterized in that: The dual-light camera is a natural light and infrared dual-light camera.

8. The AI-based fire water monitor intelligent control system according to claim 1 is characterized in that: The pitch motor control unit is used to receive instructions from the control processing module and control the pitch motor to run forward and reverse to adjust the fire water monitor to rotate up and down to the jet angle.

9. The AI-based fire water monitor intelligent control system according to claim 1 is characterized in that: The rotating motor control unit is used to receive instructions from the control processing module and control the rotating motor to run forward and reverse to adjust the fire water monitor to rotate left and right to a swing angle.

10. An AI-based fire water monitor intelligent control method, characterized in that: The method comprises: Using a fire source detection module to identify and locate the main fire source and output the identification result, the fire source detection module includes a laser rangefinder, a dual-light camera and a computer; Using a control processing module to obtain the recognition result and generate a control instruction in combination with the first algorithm, the second algorithm and the third algorithm, the control processing module including a CPU, an input unit, an output unit, a communication unit and a power module; Using an execution module to execute the angle rotation of the motor according to the control instruction, the execution module includes a pitch motor control unit and a rotation motor control unit; The first algorithm is a fire source identification and location algorithm, which is used to obtain fire source information; The second algorithm is a main fire source determination algorithm, which is used to determine the main fire source according to the fire source calibration frame; The third algorithm is a fire extinguishing control algorithm, which is used to achieve accurate fire extinguishing according to the fire source information and the determined main fire source; Wherein, the first algorithm includes: Using the dual-light camera to obtain images of the fire scene; Using a target detection model to compare and analyze the image of the fire scene, identify the target in the image, and output the position and confidence of the target, wherein the target is the fire source; Using the laser rangefinder to obtain distance information and angle information of the target relative to the dual-light camera; The position and confidence of the target as well as the distance information and angle information are integrated to obtain complete fire source information.

Images