Emergency evacuation guiding system and method based on node measurement and intelligent identification

Abstract

The invention discloses an emergency evacuation guiding system and method based on node measurement and intelligent identification. The method comprises the following steps: firstly, acquiring and sending indoor node temperature; whether a temperature abnormal area exists or not is judged according to the temperature data, if yes, camera rotation control and fire source intelligent identification are carried out, once a fire disaster is confirmed, fire disaster warning is carried out, an evacuation indication board rotates to guide an exit, and if not, indoor node temperature collection and sending are carried out again; when the system finds that a fire source exists in a room according to a fire source recognition result of the camera, the system gives out a fire alarm, and meanwhile, the motor driving circuit controls the indicator motor to rotate to drive the evacuation indicator to rotate to point to the exit direction; meanwhile, the camera intelligently recognizes the crowding state of the evacuation exit, when the number of evacuated people is large, the camera feeds back the recognition result to the microprocessor, and the microprocessor controls the indication board to rotate to other exits with small people flow according to the feedback result; and finally, the whole evacuation process is completed.

Description

Technical Field

[0001] This invention belongs to the field of emergency evacuation technology, specifically relating to an emergency evacuation guidance system and method based on node measurement and intelligent identification. Background Technology

[0002] With rapid economic development, high-rise buildings are springing up everywhere. While meeting people's daily needs, safety issues have also arisen. Escape is difficult in the event of a fire in an indoor building. Currently, traditional evacuation guidance systems are ill-suited to this development, exhibiting the following problems: 1. Because escape routes are pre-designed based on the building's internal structure, traditional evacuation systems provide fixed directions and cannot be used flexibly; 2. Escape routes and fire detection alarm devices are not interconnected and operate independently, potentially leading people to the fire source. Therefore, it is crucial to design an evacuation system that can correlate fire source identification with evacuation guidance and can change directions in real time. Summary of the Invention

[0003] To address the aforementioned problems, this invention proposes an emergency evacuation guidance system and method based on node measurement and intelligent identification. First, this invention designs a node temperature measurement module capable of real-time monitoring of temperature information at various indoor locations. Based on this, it proposes an improved SUSAN algorithm based on Michelson contrast to achieve intelligent flame identification, improving the accuracy and reliability of fire detection. Second, this invention designs a data receiving and control module, which serves as the core of the entire system. This module links fire detection and identification with evacuation signs, making guidance directions more accurate and avoiding errors caused by fixed directions. Finally, this invention introduces congestion recognition at evacuation exits, using the YOLOv8 algorithm to identify the congestion status of evacuees at the current exit. This invention can provide more reasonable evacuation guidance directions in multi-exit situations, improving indoor evacuation efficiency.

[0004] The above objectives are achieved through the following technical solutions:

[0005] This invention first provides an emergency evacuation guidance system based on node measurement and intelligent identification. The system includes a node temperature measurement module, a data receiving and control module, an early warning indication module, a rotatable evacuation sign module, and a rotatable camera module. The node temperature measurement module is arranged at intervals indoors to collect temperature data from multiple nodes and send it to the data receiving and control module. The data receiving and control module processes the collected temperature data from each node to determine whether an abnormal temperature area has occurred. The data receiving and control module is used to control the camera to rotate via a motor, so that the abnormal temperature area is within the camera's observation area. The data receiving and control module is also used to issue fire warnings and simultaneously control the evacuation sign to rotate and point towards the exit.

[0006] Furthermore, the node temperature measurement module includes: a DS18B20 temperature sensor, an STC89C52 microcontroller, and a LORA wireless module; wherein, both the temperature sensor and the wireless module are wired to the microcontroller, the temperature sensor is responsible for collecting indoor node temperature data, the microcontroller is responsible for transmitting the temperature data to the wireless module, and the wireless module is responsible for wirelessly transmitting the data to the data receiving and control module.

[0007] Furthermore, the data receiving and control module includes a microcontroller and a motor drive circuit, with the microcontroller being an STM32F103; the rotatable camera module includes a camera motor and a camera, the camera motor being a 42-stepper motor and the camera being an OV5640, the camera motor and the camera being fixedly connected, and the rotation speed of the camera motor being 5° / s; the warning indication module consists of a red LED indicator and a loudspeaker, wherein the red LED indicator and the loudspeaker are both connected to the microprocessor via cables; the rotatable evacuation sign module includes a sign motor and a sign with indication markings, the sign motor being a 42-stepper motor, the sign motor and the sign being fixedly connected, and the rotation speed being 5° / s.

[0008] The present invention also provides a method for emergency evacuation using the above-mentioned emergency evacuation guidance system based on node measurement and intelligent identification, the method comprising the following steps:

[0009] S1. Indoor node temperature acquisition and transmission: The temperature sensor in the node temperature measurement module is responsible for acquiring indoor node temperature data, the microcontroller is responsible for transmitting the temperature data to the wireless module, and the wireless module is responsible for wirelessly transmitting it to the data receiving and control module.

[0010] S2. Node temperature data processing: The data receiving and control module receives the temperature data sent in step S1 and determines whether an abnormal temperature area has occurred. If an abnormal area is found, proceed to step S3.

[0011] S3. Camera rotation control and intelligent fire source identification: The data receiving control module controls the camera to rotate through the camera motor, so that the temperature abnormal area is within the camera's observation area. At this time, the camera identifies whether there is a fire source in the temperature abnormal area. If there is, proceed to step S4; otherwise, return to step S1. After the camera acquires an image with a fire source, it transmits it to the microcontroller in the data receiving control module via wired connection. After obtaining the image with a fire source, the microprocessor uses the improved SUSAN algorithm based on Michelson contrast to achieve intelligent fire source identification. If the identification result shows that there is indeed a fire source, proceed to step S4; otherwise, proceed to step S1.

[0012] S4. Fire warning and evacuation sign rotation guides the exit: Based on the results of the camera's fire source identification, if a fire source is found indoors, the system will issue a fire warning. At the same time, the motor drive circuit controls the sign motor to rotate, causing the evacuation sign to rotate and point towards the exit.

[0013] S5. Intelligent recognition of congestion at evacuation exits by camera: After being rotated, both the evacuation sign and the camera point towards the exit. At this time, the evacuation sign guides people to the exit, and the camera identifies the congestion at the exit. Based on the camera image information, the YOLOv8 algorithm is used to identify the congestion at the exit. If the number of evacuees is less than the threshold of the evacuation exit, the sign will always point towards the exit until the evacuation is completed. If the number of evacuees is greater than the threshold of the evacuation exit, proceed to step S6.

[0014] S6. When the number of evacuees exceeds the threshold of the evacuation exit, the camera will feed back the recognition result to the microprocessor, and the microprocessor will control the sign to rotate to another exit with less traffic based on the feedback result.

[0015] S7: Follow the directions on the rotating signs to complete the evacuation process.

[0016] Furthermore, step S2 specifically includes the following process:

[0017] S21. The data receiving and control module calculates the indoor average temperature based on the temperature data received in step S1: , in This represents the average indoor temperature at time t; Indicates the number of indoor installed node temperature measurement modules; Indicates the Temperature values ​​at each node;

[0018] S22. Record the temperature value of each node. and average indoor temperature contrast;

[0019] S23. When the difference between the temperature value of each node and the average indoor temperature value is within 10 degrees Celsius, it is considered that there is no temperature anomaly area, and the temperature value of the next moment is processed.

[0020] S24. When the difference between the node temperature and the indoor average temperature is greater than 10 degrees Celsius, proceed to step S3 for further processing.

[0021] Furthermore, the intelligent identification of fire sources based on the improved SUSAN algorithm using Michelson contrast, as described in step S3, specifically includes the following process:

[0022] S31. Acquire a photo of the fire source with RGB pixel values ​​using a camera, and perform grayscale processing on the photo using a weighted method, as calculated below: , in, This indicates that the coordinates in the image are grayscale value; This indicates that the coordinates in the image are The R pixel value at the location; This indicates that the coordinates in the image are The G-pixel value at the location; This indicates that the coordinates in the image are The B pixel value at the position;

[0023] S32. Fire source area identification based on Michelson contrast:

[0024] Perform a Michelson contrast transform on the grayscale value of each location in the photo. The transformation formula is as follows: , in, This represents the grayscale value after Michelson contrast transformation. This represents the average grayscale value of the entire photo;

[0025] Based on this, find the maximum gray value after transformation. and minimum value ;

[0026] Map the MICSON contrast transform value corresponding to each coordinate position in the photo to a new grayscale value: , This represents the grayscale value of any pixel in the photo. The above process enhances the grayscale value of the image and inverts the grayscale value. The lower the grayscale value, the greater the possibility that it is a flame area.

[0027] Next, a grayscale threshold needs to be determined to determine whether the flame area or the background area in the photo is in the image;

[0028] S33. Grayscale threshold Adaptive selection:

[0029] Design a grayscale threshold Adaptive selection method: , in, This represents the adjustment coefficient, selected from... ; This represents the grayscale threshold used to determine whether a flame area or a background area is in a photo. This represents the maximum grayscale value of the i-th image. This represents the maximum grayscale value of the i-th pixel in the photo. Since a photo consists of many pixels, there will be many points with the same grayscale value. This represents 1 / 200 of the number of pixels in the photo.

[0030] S34. SUSAN Region Determination and Area Calculation: This involves determining the grayscale value and threshold of each pixel in the image. The comparison is then used to determine whether it is a USAN region. The specific formula is as follows: , in, Represents the discriminant function; This represents the grayscale value of any pixel in the photo. This represents the grayscale value of the center pixel in the photo; Indicates the coordinates of the center position of the photo; Indicates the grayscale threshold;

[0031] Based on the above, the area of ​​the USAN region at each point is calculated using the following formula: , in, Represented by coordinates The central pixel area; Indicates the area of ​​the USAN region;

[0032] S35. Flame Edge Detection: Calculate the area of ​​the USAn region for each pixel. The point is compared with a set geometric threshold j to determine whether it is the edge of a flame. If so, If it is, then it is a flame area; otherwise, it is a background area. The setting is ,in This represents the maximum grayscale value of all pixels in the entire photo.

[0033] Furthermore, the specific process for identifying the crowd congestion status at the exit based on the YOLOv8 algorithm described in step S5 is as follows:

[0034] S51. Load Model: Construct and load the YOLOv8 network model for identifying people;

[0035] S52. Input Data: Read the image information of each frame in the camera video and input the images into the YOLOv8 network model;

[0036] S53. Perform detection: Identify people in the image and mark their locations;

[0037] S54. Calculation Results: Based on the personnel labeling information, the number of people is counted and the personnel density is calculated.

[0038] The advantages of this invention compared to the prior art are:

[0039] (1) This invention links fire detection and identification with evacuation guidance system, making the guidance direction more accurate and avoiding the problem of traditional fixed guidance signs leading people to the fire source area;

[0040] (2) This invention combines physical node measurement method and camera intelligent flame recognition method to improve the accuracy and reliability of fire recognition, thereby avoiding problems such as false alarms;

[0041] (3) The present invention introduces the recognition of congestion at evacuation exits, which can provide more reasonable evacuation guidance in the case of multiple exits and improve evacuation efficiency. Attached Figure Description

[0042] Figure 1 This is a system module block diagram of the present invention;

[0043] Figure 2 This is a schematic diagram of the installation of the present invention;

[0044] Figure 3 This is a flowchart of the method of the present invention;

[0045] Figure 4 This is a flowchart of the improved SUSAN algorithm based on Michelson contrast proposed in this invention. Detailed Implementation

[0046] The present invention will be further described below with reference to the accompanying drawings and specific examples.

[0047] like Figure 1As shown, this invention first provides an emergency evacuation guidance system based on node measurement and intelligent identification. The system includes a node temperature measurement module, a data receiving and control module, an early warning indication module, a rotatable evacuation sign module, and a rotatable camera module. The node temperature measurement module is spaced apart indoors to collect temperature data from multiple nodes and send it to the data receiving and control module. The data receiving and control module processes the collected temperature data from each node to determine if an abnormal temperature area has occurred. The data receiving and control module also controls the camera to rotate via a motor, ensuring that the abnormal temperature area is within the camera's observation range. Furthermore, the data receiving and control module issues fire warnings and simultaneously controls the evacuation sign to rotate and point towards the exit. Figure 2 This is a schematic diagram of the installation of the present invention; Figure 2 The attached diagram shows the following reference numerals: 1. Data receiving and control module; 2. Rotatable evacuation sign module; 3. Rotatable camera module; 4. Node temperature measurement module; 5. Early warning indicator module. Wherein:

[0048] The node temperature measurement module includes: a DS18B20 temperature sensor, an STC89C52 microcontroller, and a LORA wireless module. Both the temperature sensor and the wireless module are wired to the microcontroller. The temperature sensor collects indoor node temperature data, the microcontroller transmits the temperature data to the wireless module, and the wireless module wirelessly transmits the data to the data receiving and control module. The specific installation location of the node temperature measurement module is as follows: Figure 2 As shown.

[0049] The data receiving and control module mainly includes: microcontroller model: STM32F103, and motor drive circuit.

[0050] The rotatable camera includes: a 42-stepper motor and an OV5640 camera. The motor and camera are fixedly connected, and the rotation speed is 5° / s. The motor and camera are connected via cables and a microcontroller in the data receiving control module.

[0051] Early warning indicator module: This module consists of a red LED indicator and a loudspeaker. Both the red LED indicator and the loudspeaker are connected to a microprocessor via cables. Upon the occurrence of a fire, the microprocessor sends a signal to the red LED indicator and the loudspeaker, at which point the red LED indicator illuminates and the loudspeaker sounds an alarm.

[0052] The rotatable evacuation sign module includes: a 42-stepper motor and a sign with indicator markings. The motor and sign are fixedly connected, and the rotation speed is 5° / s. The motor is connected to a microcontroller via a cable.

[0053] The system of the present invention, in the case of a known indoor layout and safe evacuation exits, can promptly and accurately identify the location of the fire source in the event of a fire, and at the same time, can guide people to evacuate towards the safe exits in real time.

[0054] Based on the aforementioned emergency evacuation guidance system based on node measurement and intelligent identification, this invention provides an emergency evacuation guidance method based on node measurement and intelligent identification, such as... Figure 3 As shown, this method combines fire detection and evacuation guidance, providing reasonable evacuation directions in real-world environments and improving evacuation efficiency. The specific implementation steps are as follows:

[0055] S1. Indoor Node Temperature Acquisition and Transmission: This step mainly completes the temperature acquisition and transmission processing for multiple indoor nodes. This step is primarily implemented in the node temperature measurement module. The temperature acquisition data for each indoor node is processed separately using... , , express.

[0056] S2. Node Temperature Data Processing: This step mainly processes the temperature data collected from each node to determine if any abnormal temperature areas have occurred. Node temperature data processing is primarily implemented in the data receiving and control module, such as... Figure 1 As shown. The node temperature data processing is completed in the microcontroller, and the specific process is as follows:

[0057] S21. Calculate the average indoor temperature based on step 1: ,

[0058] in Indicates the average indoor temperature at time t; indicates the number of indoor temperature measurement modules installed at nodes. Indicates the Temperature values ​​at each node.

[0059] S22. Record the temperature value of each node. and average indoor temperature contrast.

[0060] S23. When the difference between the temperature value at each node and the average indoor temperature is within 10 degrees Celsius, it is considered that there is no temperature anomaly area. At this time, continue processing the temperature value at the next time step.

[0061] S24. When the difference between the node temperature and the indoor average temperature is greater than 10 degrees Celsius, proceed to step 3 below for processing.

[0062] S3. Camera rotation control and intelligent fire source recognition:

[0063] Based on the above steps, the data receiving and control module controls the camera to rotate via a motor, bringing the temperature anomaly area within the camera's observation range. At this point, the camera identifies whether a fire source exists within this temperature anomaly area. If so, it proceeds to step S4; otherwise, it returns to step S1.

[0064] The camera rotation control is mainly implemented in the rotatable camera module.

[0065] Fire source identification processing: After the camera acquires an image containing a fire source, it transmits it via wired connection to the microcontroller in the data receiving and control module. Upon receiving the image containing the fire source, the microprocessor uses the improved SUSAN algorithm based on Michelson contrast, as proposed in this invention, to intelligently identify the fire source. If the identification result confirms the presence of a fire source, proceed to step S4; otherwise, proceed to step S1.

[0066] The improved SUSAN algorithm based on Michelson contrast proposed in this invention is implemented as follows:

[0067] S31. Obtain a photo containing a fire source:

[0068] A photo of the fire source with RGB pixel values ​​is acquired using a camera. The photo is then processed into grayscale using a weighted method, as detailed below: , in, This indicates that the coordinates in the image are The grayscale value; RGB represents the coordinates in the image. The RGB pixel value at the position. This indicates that the coordinates in the image are The R pixel value at the location; This indicates that the coordinates in the image are The G-pixel value at the location; This indicates that the coordinates in the image are The B-pixel value at position .

[0069] S32. Fire source area identification based on Michelson contrast:

[0070] Perform a Michelson contrast transform on the grayscale value of each location in the photo. The transformation formula is as follows: , in, This represents the grayscale value after Michelson contrast transformation. This represents the average grayscale value of the entire image. Based on this, find the maximum grayscale value after transformation. and minimum value .

[0071] Map the MICSON contrast transform value corresponding to each coordinate position in the photo to a new grayscale value: , This represents the grayscale value of any pixel in the photo. The above process enhances the grayscale values ​​of the image and inverts them; the lower the grayscale value, the greater the likelihood that it represents a flame area.

[0072] Next, a grayscale threshold needs to be determined to distinguish whether the flame area or the background area in the photo is indicative of the fire area.

[0073] S33. Threshold Adaptive selection:

[0074] Considering the contrast between the target and the background in flame photographs, as well as the complexity of the texture and structure of the background area, the grayscale threshold... The selection of the threshold value has a significant impact on the edge detection results. Therefore, this invention designs a threshold value... Adaptive selection method: , in, This represents the adjustment coefficient, selected from... ; This represents the grayscale threshold used to determine whether the flame area or the background area in the photo is indicative of a grayscale value. This represents the maximum grayscale value of the i-th image. This represents the maximum grayscale value of the i-th pixel in the photo. A photo is composed of many pixels, so there will be many points with the same grayscale value. This represents 1 / 200 of the number of pixels in the photo.

[0075] S34. SUSAN Area Determination and Area Calculation:

[0076] The grayscale value and threshold of each pixel in the photo The comparison is then used to determine whether it is a USAN region. The specific formula is as follows: , in, Represents the discriminant function; This represents the grayscale value of any pixel in the photo. This represents the grayscale value of the center pixel in the photo; Indicates the coordinates of the center position of the photo; This represents the threshold.

[0077] Based on the above, the area of ​​the USAN region at each point is calculated using the following formula: , in, Represented by coordinates The central pixel area; This indicates the area of ​​the USAN region.

[0078] S35. Flame Edge Detection:

[0079] Finally, the area of ​​the USAn region for each pixel is obtained. The point is compared with a set geometric threshold j to determine whether it is on the edge of a flame. If... If the area is positive, it is the flame area; otherwise, it is the background area. The setting is , where represents the maximum grayscale value of all pixels in the entire photo.

[0080] The specific implementation process is as follows: Figure 4 As shown.

[0081] S4. Rotating fire warning and evacuation signs guide exits:

[0082] Based on the camera's fire source detection, if a fire source is detected indoors, the system will issue a fire warning, and simultaneously, the evacuation sign will rotate to point towards the exit. The fire warning is issued by the early warning module, and the rotation of the evacuation sign to point towards the exit is controlled by the rotatable evacuation sign module. Figure 3 As shown. Specifically:

[0083] After a fire occurs, the microprocessor of the early warning indicator module sends a signal to the red LED indicator and the loudspeaker. At this time, the red LED indicator lights up and the loudspeaker sounds an alarm.

[0084] The rotatable evacuation sign module features signs with indicator marks used to indicate exit locations.

[0085] S5. Cameras intelligently identify congestion at evacuation exits:

[0086] After being rotated, both the sign and the camera point towards the exit. At this point, the sign guides people to the exit, and the camera identifies the congestion at the exit. Based on the camera's image information, this invention uses YOLOv8 to identify the congestion at the exit. If the number of people being evacuated is small, the sign remains pointing towards the exit until the evacuation is complete. If the number of people being evacuated is large, proceed to step S6.

[0087] The specific process for recognizing crowd congestion at exits based on YOLOv8 is as follows:

[0088] S51. Load Model: Construct and load the YOLOv8 network model for identifying people;

[0089] S52. Input Data: Read the image information of each frame in the camera video and input the images into the YOLOv8 network model;

[0090] S53. Perform detection: Identify people in the image and mark their locations;

[0091] S54. Calculation Results: Based on the personnel labeling information, the number of people is counted and the personnel density is calculated.

[0092] S6. Evacuation signs rotate to guide other exits: When there are a large number of evacuees, the camera will feed back the recognition results to the microprocessor, which will then control the signs to rotate to other exits with less traffic based on the feedback results.

[0093] S7. Evacuation Complete: Follow the directions on the rotating signs to complete the evacuation process.

Claims

1. An emergency evacuation guidance system based on node measurement and intelligent identification, characterized in that, The system includes a node temperature measurement module, a data receiving and control module, an early warning and indication module, a rotatable evacuation sign module, and a rotatable camera module. The node temperature measurement modules are arranged at intervals indoors to collect temperature data from multiple nodes and send it to the data receiving and control module. The data receiving and control module processes the collected temperature data from each node to determine whether an abnormal temperature area has occurred. The data receiving and control module is used to control the rotation of the camera via a motor to ensure that the abnormal temperature area is within the camera's observation range. The data receiving and control module is also used to issue fire warnings and simultaneously control the evacuation signs to rotate and point towards the exit.

2. The emergency evacuation guidance system based on node measurement and intelligent identification according to claim 1, characterized in that, The node temperature measurement module includes: a DS18B20 temperature sensor, an STC89C52 microcontroller, and a LORA wireless module. The temperature sensor and the wireless module are both wired to the microcontroller. The temperature sensor is responsible for collecting indoor node temperature data, the microcontroller is responsible for transmitting the temperature data to the wireless module, and the wireless module is responsible for wirelessly transmitting the data to the data receiving and control module.

3. The emergency evacuation guidance system based on node measurement and intelligent identification according to claim 1, characterized in that, The data receiving and control module includes a microcontroller and a motor drive circuit. The microcontroller is an STM32F103. The rotatable camera module includes a camera motor and a camera. The camera motor is a 42-stepper motor, and the camera is an OV5640. The camera motor and the camera are fixedly connected. The rotation speed of the camera motor is 5° / s. The warning indication module consists of a red LED indicator and a loudspeaker, both of which are connected to a microprocessor via cables. The rotatable evacuation sign module includes a sign motor and a sign with indicator markings. The sign motor is a 42-step motor, and the sign motor and the sign are fixedly connected, with a rotation speed of 5° / s.

4. A method for emergency evacuation using the emergency evacuation guidance system based on node measurement and intelligent identification as described in any one of claims 1-3, characterized in that, The method includes the following steps: S1. Indoor node temperature acquisition and transmission: The temperature sensor in the node temperature measurement module is responsible for acquiring indoor node temperature data, the microcontroller is responsible for transmitting the temperature data to the wireless module, and the wireless module is responsible for wirelessly transmitting it to the data receiving and control module. S2. Node temperature data processing: The data receiving and control module receives the temperature data sent in step S1 and determines whether an abnormal temperature area has occurred. If an abnormal area is found, proceed to step S3. S3. Camera rotation control and intelligent fire source identification: The data receiving control module controls the camera to rotate through the camera motor, so that the temperature abnormal area is within the camera's observation area. At this time, the camera identifies whether there is a fire source in the temperature abnormal area. If there is, proceed to step S4; otherwise, return to step S1. After the camera acquires an image with a fire source, it transmits it to the microcontroller in the data receiving control module via wired connection. After obtaining the image with a fire source, the microprocessor uses the improved SUSAN algorithm based on Michelson contrast to achieve intelligent fire source identification. If the identification result shows that there is indeed a fire source, proceed to step S4; otherwise, proceed to step S1. S4. Fire warning and evacuation sign rotation guides the exit: Based on the results of the camera's fire source identification, if a fire source is found indoors, the system will issue a fire warning. At the same time, the motor drive circuit controls the sign motor to rotate, causing the evacuation sign to rotate and point towards the exit. S5. Intelligent recognition of congestion at evacuation exits by camera: After being rotated, both the evacuation sign and the camera point towards the exit. At this time, the evacuation sign guides people to the exit, and the camera identifies the congestion at the exit. Based on the camera image information, the YOLOv8 algorithm is used to identify the congestion at the exit. If the number of evacuees is less than the threshold of the evacuation exit, the sign will always point towards the exit until the evacuation is completed. If the number of evacuees is greater than the threshold of the evacuation exit, proceed to step S6. S6. When the number of evacuees exceeds the threshold of the evacuation exit, the camera will feed back the recognition result to the microprocessor, and the microprocessor will control the sign to rotate to another exit with less traffic based on the feedback result. S7: Follow the directions on the rotating signs to complete the evacuation process.

5. The method for emergency evacuation according to claim 4, characterized in that, Step S2 specifically includes the following process: S21. The data receiving and control module calculates the indoor average temperature based on the temperature data received in step S1: ， in This represents the average indoor temperature at time t; Indicates the number of indoor installed node temperature measurement modules; Indicates the first Temperature values ​​at each node; S22. Record the temperature value of each node. and average indoor temperature contrast; S23. When the difference between the temperature value of each node and the average indoor temperature value is within 10 degrees Celsius, it is considered that there is no temperature anomaly area, and the temperature value of the next moment is processed. S24. When the difference between the node temperature and the indoor average temperature is greater than 10 degrees Celsius, proceed to step S3 for further processing.

6. The method for emergency evacuation according to claim 4, characterized in that, The improved SUSAN algorithm based on Michelson contrast for intelligent fire source identification described in step S3 specifically includes the following process: S31. Acquire a photo of the fire source with RGB pixel values ​​using a camera, and perform grayscale processing on the photo using a weighted method, as calculated below: ， in, This indicates that the coordinates in the image are grayscale value; This indicates that the coordinates in the image are The R pixel value at the location; This indicates that the coordinates in the image are The G-pixel value at the location; This indicates that the coordinates in the image are The B pixel value at the position; S32. Fire source area identification based on Michelson contrast: Perform a Michelson contrast transform on the grayscale value of each location in the photo. The transformation formula is as follows: ， in, This represents the grayscale value after Michelson contrast transformation. This represents the average grayscale value of the entire photo; Based on this, find the maximum gray value after transformation. and minimum value ; Map the MICSON contrast transform value corresponding to each coordinate position in the photo to a new grayscale value: ， This represents the grayscale value of any pixel in the photo. The above process enhances the grayscale value of the image and inverts the grayscale value. The lower the grayscale value, the greater the possibility that it is a flame area. Next, a grayscale threshold needs to be determined to determine whether the flame area or the background area in the photo is in the image; S33. Grayscale threshold Adaptive selection: Design a grayscale threshold Adaptive selection method: ， in, This represents the adjustment coefficient, selected from... ; This represents the grayscale threshold used to determine whether a flame area or a background area is in a photo. This represents the maximum grayscale value of the i-th image. This represents the maximum grayscale value of the i-th pixel in the photo. Since a photo consists of many pixels, there will be many points with the same grayscale value. This represents 1 / 200 of the number of pixels in the photo. S34. SUSAN Region Determination and Area Calculation: This involves determining the grayscale value and threshold of each pixel in the image. The comparison is then used to determine whether it is a USAN region. The specific formula is as follows: ， in, Represents the discriminant function; This represents the grayscale value of any pixel in the photo. This represents the grayscale value of the center pixel in the photo; Indicates the coordinates of the center position of the photo; Indicates the grayscale threshold; Based on the above, the area of ​​the USAN region at each point is calculated using the following formula: ， in, Represented by coordinates The central pixel area; Indicates the area of ​​the USAN region; S35. Flame Edge Detection: Calculate the area of ​​the USAn region for each pixel. The point is compared with a set geometric threshold j to determine whether it is the edge of a flame. If so, If it is, then it is a flame area; otherwise, it is a background area. The setting is ,in This represents the maximum grayscale value of all pixels in the entire photo.

7. The method for emergency evacuation according to claim 4, characterized in that, The specific process for identifying the crowding status at the exit based on the YOLOv8 algorithm in step S5 is as follows: S51. Load Model: Construct and load the YOLOv8 network model for identifying people; S52. Input Data: Read the image information of each frame in the camera video and input the images into the YOLOv8 network model; S53. Perform detection: Identify people in the image and mark their locations; S54. Calculation Results: Based on the personnel labeling information, the number of people is counted and the personnel density is calculated.

Images