A tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system

Abstract

The present application discloses a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system, which includes: an image acquisition unit for acquiring first vehicle picture information of a vehicle entering a target tunnel, and the image acquisition unit is also used for acquiring second vehicle picture information of a vehicle exiting the target tunnel; an electric vehicle identification unit for determining the distribution information of electric vehicles and the distribution information of other vehicles in the target tunnel based on a vehicle identification model according to the first vehicle picture information and the second vehicle picture information; a tunnel environment collection unit for acquiring the tunnel environment information of the target tunnel; a risk prediction unit for determining a risk prediction level based on a risk prediction model according to the distribution information of electric vehicles, the distribution information of other vehicles and the tunnel environment information; and a control unit for controlling the waiting response level of the linkage fire extinguishing system based on the risk prediction level.

Description

Technical Field

[0001] This specification relates to the field of emergency handling. More specifically, the present application relates to a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system. Background Art

[0002] With the wide application of electric vehicles (such as electric cars and electric trucks) in traffic, the operation safety issues of electric vehicles in tunnels have gradually attracted wide attention. As the core component of electric vehicles, lithium batteries may cause fires under conditions such as high temperature, impact or abnormal charge and discharge. Especially in a closed tunnel environment, once a fire occurs, it is extremely likely to cause serious consequences. Most traditional tunnel fire monitoring and extinguishing systems rely on single environmental monitoring means, such as smoke detectors and temperature sensors, and do not conduct risk assessment and treatment for the particularity of electric vehicles. In addition, the presence of special vehicles (such as dangerous goods transport vehicles) may also exacerbate the fire risk, but the existing systems lack refined management and monitoring methods for different vehicle types.

[0003] Existing tunnel fire monitoring systems usually rely on fire detectors at fixed positions to trigger alarms by detecting parameters such as temperature and smoke concentration in the tunnel. However, this method often has problems such as delayed response and inaccurate positioning. Especially for the fire hazards of electric vehicle batteries, traditional fire detectors are difficult to perceive risks in advance. The existing systems lack real-time monitoring of electric vehicles and special vehicles and cannot effectively respond to the fire risks caused by different types of vehicles. At the same time, the fire extinguishing system usually only responds based on simple environmental data, lacks a dynamic and comprehensive risk assessment mechanism, cannot adjust the fire extinguishing strategy in a timely manner according to different fire risk levels, and is difficult to provide a precise fire extinguishing plan.

[0004] Therefore, it is necessary to propose a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system to solve at least some of the above problems. Summary of the Invention

[0005] A series of simplified concepts are introduced in the Summary of the Invention section, which will be further described in detail in the Detailed Description section. The Summary of the Invention section of the present application does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the protection scope of the claimed technical solution.

[0006] In a first aspect, the present application proposes a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system, including:

[0007] An image acquisition unit, which is used to acquire the first vehicle picture information of a vehicle driving into a target tunnel, and the image acquisition unit is also used to acquire the second vehicle picture information of a vehicle driving out of the target tunnel;

[0008] An electric vehicle identification unit, configured to determine the distribution information of electric vehicles and the distribution information of other vehicles in the target tunnel based on the vehicle identification model according to the above-mentioned first vehicle picture information and the above-mentioned second vehicle picture information. The above-mentioned distribution information of electric vehicles includes the number information of electric vehicles, the model information of electric vehicles, and the position information of electric vehicles in the tunnel. The above-mentioned distribution information of other vehicles includes the number information of special vehicles, the type information of special vehicles, and the position information of special vehicles in the tunnel;

[0009] A tunnel environment collection unit, configured to obtain the tunnel environment information of the above-mentioned target tunnel;

[0010] A risk prediction unit, configured to determine the risk prediction level based on the risk prediction model according to the above-mentioned distribution information of electric vehicles, the above-mentioned distribution information of other vehicles, and the above-mentioned tunnel environment information;

[0011] A control unit, configured to control the waiting response level of the linkage fire extinguishing system based on the above-mentioned risk prediction level.

[0012] In a feasible implementation manner, the specific steps for the above-mentioned electric vehicle identification unit to determine the distribution information of electric vehicles in the target tunnel include:

[0013] Based on the above-mentioned first vehicle picture information, the above-mentioned second vehicle picture information, and the picture shooting time information, determine the existing vehicle information in the target tunnel;

[0014] Based on the above-mentioned existing vehicle information, determine the above-mentioned number information of electric vehicles, the above-mentioned number information of special vehicles, the above-mentioned model information of electric vehicles, and the above-mentioned type information of special vehicles;

[0015] Based on the above-mentioned existing vehicle information, the speed information of the existing vehicle entering the tunnel, and the above-mentioned picture shooting time information, determine the position information of electric vehicles in the tunnel and the position information of special vehicles in the tunnel.

[0016] In a feasible implementation manner, it is characterized in that the above-mentioned determining the above-mentioned number information of electric vehicles, the above-mentioned number information of special vehicles, the above-mentioned model information of electric vehicles, and the above-mentioned type information of special vehicles based on the above-mentioned existing vehicle information includes:

[0017] Perform a preprocessing operation on the picture information corresponding to the existing vehicle information to obtain preprocessed picture information;

[0018] Input the above-mentioned preprocessed picture information into the first recognition module to obtain the above-mentioned number information of electric vehicles and the above-mentioned number information of special vehicles, where the above-mentioned first recognition module is established based on the Faster R-CNN model;

[0019] Input the above preprocessed picture information into the second recognition module to obtain the above electric vehicle model information and the above special vehicle type information, where the above second recognition module is established based on the SVM model.

[0020] In a feasible implementation manner, determining the position information of the electric vehicle in the tunnel and the position information of the special vehicle in the tunnel based on the above existing vehicle information, the speed information of the existing vehicle entering the tunnel, and the above picture shooting time information includes:

[0021] Send the first existing vehicle information with the speed information less than the preset speed to the first prediction module to obtain the first predicted position corresponding to the first existing vehicle information, where the above first prediction module is established based on the Kalman filter algorithm;

[0022] Send the second existing vehicle information with the speed information greater than or equal to the preset speed to the second prediction module to obtain the second predicted position corresponding to the second existing vehicle information, where the above second prediction module is established based on the long short-term memory network;

[0023] Establish integrated position information according to the first predicted position and the second predicted position;

[0024] Determine the position information of the electric vehicle in the tunnel and the position information of the special vehicle in the tunnel according to the integrated position information and the vehicle type lane and sequence information of the existing vehicle entering the tunnel.

[0025] In a feasible implementation manner, the above risk prediction model includes a density risk prediction module, an environmental risk prediction module, and a risk fusion module;

[0026] The specific steps for the above risk prediction unit to determine the risk prediction level based on the risk prediction model according to the above electric vehicle distribution information, the above other vehicle distribution information, and the tunnel environment information include:

[0027] Input the above electric vehicle distribution information and the above other vehicle distribution information into the above density risk prediction module to obtain a density risk factor;

[0028] Input the above tunnel environment information into the above environmental risk prediction module to obtain an environmental risk factor;

[0029] Input the above density risk factor and the above environmental risk factor into the above risk fusion module to obtain the above risk prediction level.

[0030] In a feasible implementation manner, the above inputting the above electric vehicle distribution information and the above other vehicle distribution information into the above density risk prediction module to obtain a density risk factor includes:

[0031] Grid the target tunnel to obtain the gridded information of the target tunnel;

[0032] Based on the above gridded information of the target tunnel, the above electric vehicle distribution information, and the above other vehicle distribution information, establish the gridded information of vehicle distribution;

[0033] Determine the density weight information of different vehicle types;

[0034] Determine the above density risk factor according to the above gridded information of vehicle distribution and the above density weight information.

[0035] In a feasible implementation manner, the density weight information of electric vehicles is obtained by weighted fusion based on the battery type information and collision test data information of electric vehicles.

[0036] In a feasible implementation manner, the above tunnel environment information includes temperature information, humidity information, ventilation condition information, and air quality information;

[0037] The above inputting the above tunnel environment information into the above environmental risk prediction module to obtain the environmental risk factor includes:

[0038] Perform weighted fusion based on the above temperature information, the above humidity information, the above ventilation condition information, and the above air quality information to obtain the above environmental risk factor.

[0039] In a feasible implementation manner, the above inputting the above density risk factor and the above environmental risk factor into the above risk fusion module to obtain the above risk prediction level includes:

[0040] Determine the first weight information corresponding to the above density risk factor and the second weight information corresponding to the above environmental risk factor based on historical fire risk information;

[0041] Calculate a risk score based on the above density risk factor, the above environmental risk factor, the above first weight information, and the above second weight information;

[0042] Determine the above risk level based on the above risk score.

[0043] In a feasible implementation manner, the specific steps for the above control unit to control the waiting response level of the linkage fire extinguishing system based on the above risk prediction level include:

[0044] Control the wind speed information, temperature monitoring frequency information, fire detection detector sensitivity information, and fire extinguishing equipment status information of the ventilation system of the linkage fire extinguishing system based on the above risk prediction level, where the above fire extinguishing equipment status information includes the operation status information of the automatic fire extinguishing device, the fire extinguishing area division status information, and the fire extinguishing agent injection volume status information.

[0045] In summary, through the image acquisition unit and the electric vehicle recognition unit, this application realizes the real-time recognition and monitoring of electric vehicles and special vehicles in the target tunnel. Compared with traditional fire detection systems that only rely on fixed detectors, this application can accurately obtain specific information about vehicles in the tunnel, including the number of electric vehicles, electric vehicle models, location information, and relevant information about special vehicles. This enables the system to identify the distribution of electric vehicles and special vehicles in advance, thereby more precisely evaluating potential fire risks. This application particularly takes into account the fire risk of electric vehicle batteries. Combining vehicle distribution and environmental conditions, it can locate and monitor the positions of high-risk vehicles. This targeted monitoring greatly improves the accuracy of fire risk assessment. The tunnel environment collection unit can obtain information such as temperature, humidity, ventilation conditions, and air quality in the tunnel in real time. Compared with traditional technologies that only rely on a single smoke detector or temperature sensor, this system integrates multiple environmental factors and integrates environmental information and vehicle distribution information through the risk prediction unit to conduct a more comprehensive fire risk assessment. By integrating environmental and vehicle information, the system can dynamically evaluate the probability of a fire occurring in the tunnel, making risk prediction more accurate. The combination of environmental data and vehicle data enables the system not only to detect potential fire hazards but also to adjust the monitoring strategy according to the specific tunnel conditions. For example, when the temperature in the tunnel rises abnormally but the ventilation conditions are good, the system can reduce the sensitivity of the fire risk warning to avoid unnecessary false alarms. This application calculates the vehicle density and environmental risk factors through the density risk prediction module and the environmental risk prediction module respectively, and conducts a comprehensive risk assessment in the risk fusion module. Different from traditional fire detection systems, this system can dynamically adjust the accuracy of fire prediction based on density information such as vehicle type, quantity, and location, combined with the environmental conditions of the tunnel. In addition, different weight information is set for different types of vehicles (such as electric vehicles, special vehicles) in the application, making risk prediction more accurate. In particular, the battery type and collision test data of electric vehicles are also introduced into the setting of risk weights to better address the battery fire risk. This method effectively avoids the disadvantages of traditional fire systems that cannot distinguish vehicle types and cannot adjust the risk level according to actual situations. The control unit in this system can accurately control the parameters of the interlock fire extinguishing system based on the real-time risk prediction level, including the wind speed of the ventilation system, the frequency of temperature monitoring, the sensitivity of fire detectors, and the status of fire extinguishing equipment. Compared with traditional systems that passively wait for a fire to occur before triggering fire extinguishing, this application can actively adjust the response level of the fire extinguishing system according to the predicted fire risk to achieve precise fire extinguishing. When the fire risk level rises, the system can not only activate the fire extinguishing device but also select the appropriate type and spraying amount of fire extinguishing agent according to the specific vehicle distribution and environmental conditions to achieve zoned fire extinguishing and intelligent scheduling, thereby minimizing the damage caused by the fire to the tunnel and vehicles to the greatest extent.Compared with the prior art, the tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system of the present application can not only accurately monitor the vehicle distribution and environmental information in the tunnel, conduct a comprehensive fire risk assessment, but also intelligently control the response measures of the fire extinguishing equipment according to the predicted fire risk level. Its advantages are reflected in dynamic real-time fire monitoring, intelligent risk assessment and proactive fire extinguishing response, effectively improving the safety and efficiency of fire prevention and control of electric vehicles and special vehicles in the tunnel.

[0046] The network information intelligent analysis and regulation system and method based on artificial intelligence proposed in the present application. Other advantages, objectives and features of the present application will be partially reflected by the following description, and partially will be understood by those skilled in the art through the research and practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of this specification. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0048] Figure 1 It is a structural schematic diagram of a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system provided by an embodiment of the present application;

[0049] Figure 2 It is a flow schematic diagram of the specific steps for an electric vehicle identification unit to determine the distribution information of electric vehicles in a target tunnel provided by an embodiment of the present application;

[0050] Figure 3 It is a flow schematic diagram of the specific steps for determining the number information of electric vehicles, the number information of special vehicles, the model information of electric vehicles and the type information of special vehicles based on the existing vehicle information provided by an embodiment of the present application;

[0051] Figure 4 It is a flow schematic diagram for determining the position information of electric vehicles in the tunnel and the position information of special vehicles in the tunnel provided by an embodiment of the present application;

[0052] Figure 5 It is a flow schematic diagram of a risk prediction level provided by an embodiment of the present application;

[0053] Figure 6 It is a flow schematic diagram for obtaining the density risk factor provided by an embodiment of the present application;

[0054] Figure 1 The corresponding relationship between the reference numerals and the drawing names in the figure is:

[0055] 10 Tunnel Vehicle-mounted Lithium Battery Fire Monitoring, Early Warning and Linkage Fire Extinguishing System; 101 Image Acquisition Unit; 102 Electric Vehicle Identification Unit; 103 Tunnel Environment Collection Unit; 104 Risk Prediction Unit; 105 Control Unit. Detailed Implementation Manner

[0056] In the description of the present application, the claims and the above-mentioned drawings, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices. The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0057] Please refer to Figure 1 , which is a structural schematic diagram of a tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system 10 provided by an embodiment of the present application. Specifically, it may include:

[0058] An image acquisition unit 101, which is used to acquire the first vehicle picture information of the vehicle driving into the target tunnel, and the image acquisition unit is also used to acquire the second vehicle picture information of the vehicle driving out of the target tunnel;

[0059] An electric vehicle identification unit 102, which is used to determine the distribution information of electric vehicles and the distribution information of other vehicles in the target tunnel based on a vehicle identification model according to the first vehicle picture information and the second vehicle picture information. The distribution information of electric vehicles includes the number information of electric vehicles, the model information of electric vehicles, and the position information of electric vehicles in the tunnel. The distribution information of other vehicles includes the number information of special vehicles, the type information of special vehicles, and the position information of special vehicles in the tunnel;

[0060] A tunnel environment collection unit 103, which is used to acquire the tunnel environment information of the target tunnel;

[0061] A risk prediction unit 104, which is used to determine the risk prediction level based on a risk prediction model according to the distribution information of electric vehicles, the distribution information of other vehicles, and the tunnel environment information;

[0062] The control unit 105 is configured to control the waiting response level of the interlock fire extinguishing system based on the above-mentioned risk prediction level.

[0063] Exemplarily, the image acquisition unit 101 is configured to acquire picture information of vehicles in the tunnel. Specifically, it can acquire the first vehicle picture information of the vehicle entering the target tunnel and the second vehicle picture information of the vehicle exiting the target tunnel. These picture information provide data support for subsequent vehicle identification and distribution information analysis. Through these image data, the vehicle conditions entering and exiting the tunnel can be monitored in real time.

[0064] The electric vehicle identification unit 102 processes the first vehicle picture information and the second vehicle picture information provided by the image acquisition unit 101 based on a preset vehicle identification model to determine the distribution information of electric vehicles and the distribution information of other vehicles in the target tunnel.

[0065] Electric vehicle distribution information: includes the number, model information of electric vehicles in the target tunnel, and the specific location information of electric vehicles in the tunnel. Other vehicle distribution information: mainly includes the number, type information of special vehicles, and the specific location information of these vehicles in the tunnel. Special vehicles may include dangerous goods transport vehicles, rescue vehicles, etc., which have an important impact on the fire risk.

[0066] The tunnel environment collection unit 103 is used to acquire the tunnel environment information of the target tunnel. These environmental information usually includes key information such as temperature, humidity, ventilation condition, and air quality. The environmental factors in the tunnel may directly affect the fire risk, so the collection of environmental data is crucial for accurately assessing the risk.

[0067] The risk prediction unit 104 conducts a comprehensive assessment of the fire risk based on a risk prediction model, in combination with the electric vehicle distribution information, other vehicle distribution information, and tunnel environment information, and determines the risk prediction level. The risk prediction level can reflect the magnitude of the current fire risk in the tunnel, and is divided into multiple levels such as low, medium, and high. This unit provides real-time risk pre-judgment by integrating multiple data sources such as vehicle density and environmental conditions, providing a basis for the automatic control of subsequent fire extinguishing measures.

[0068] The control unit 105 controls the response behavior of the entire system according to the risk prediction level output by the risk prediction unit 104, specifically controlling the waiting response level of the interlock fire extinguishing system. The control parameters may include:

[0069] Wind speed information of the ventilation system: adjusts the ventilation wind speed in the tunnel according to the risk level to ensure the rapid discharge of smoke and harmful gases.

[0070] Temperature monitoring frequency information: increases the temperature monitoring frequency when the risk level rises to detect abnormal temperature changes in a timely manner.

[0071] Fire detector sensitivity information: As the risk level increases, the sensitivity of the fire detector will be appropriately increased to respond more quickly to potential signs of a fire.

[0072] Fire extinguishing equipment status information: When the risk reaches a certain level, the fire extinguishing system will enter different standby states, and the automated operations include preloading of fire extinguishing agents, division of the fire extinguishing area, and adjustment of the fire extinguishing agent injection volume.

[0073] In summary, through the image acquisition unit and the electric vehicle identification unit, this application realizes the real-time identification and monitoring of electric vehicles and special vehicles in the target tunnel. Compared with traditional fire detection systems that rely solely on fixed detectors, this application can accurately obtain specific information about the vehicles in the tunnel, including the number of electric vehicles, electric vehicle models, location information, and relevant information about special vehicles. This enables the system to identify the distribution of electric vehicles and special vehicles in advance, thereby more precisely assessing potential fire risks. This application particularly takes into account the fire risk of electric vehicle batteries. Combining vehicle distribution and environmental conditions, it can locate and monitor the positions of high-risk vehicles. This targeted monitoring greatly improves the accuracy of fire risk assessment. The tunnel environment collection unit can obtain real-time information such as temperature, humidity, ventilation status, and air quality in the tunnel. Compared with traditional technologies that rely solely on single smoke detection or temperature sensors, this system integrates multiple environmental factors and, through the risk prediction unit, integrates environmental information and vehicle distribution information for a more comprehensive fire risk assessment. By integrating environmental and vehicle information, the system can dynamically evaluate the probability of a fire occurring in the tunnel, making risk prediction more accurate. The combination of environmental data and vehicle data enables the system not only to detect potential fire hazards but also to adjust the monitoring strategy according to the specific tunnel conditions. For example, when the temperature in the tunnel rises abnormally but the ventilation status is good, the system can reduce the sensitivity of the fire risk warning to avoid unnecessary false alarms. This application calculates the vehicle density and environmental risk factors through the density risk prediction module and the environmental risk prediction module respectively, and conducts a comprehensive risk assessment in the risk fusion module. Different from traditional fire detection systems, this system can dynamically adjust the accuracy of fire prediction based on density information such as vehicle type, quantity, and location, combined with the environmental conditions of the tunnel. In addition, different weight information is set for different types of vehicles (such as electric vehicles, special vehicles) in the application, making risk prediction more accurate. In particular, the battery type and collision test data of electric vehicles are also introduced into the setting of risk weights to better address the battery fire risk. This method effectively avoids the disadvantages of traditional fire systems that cannot distinguish vehicle types and cannot adjust the risk level according to actual situations. The control unit in this system can accurately control the parameters of the linkage fire extinguishing system based on the real-time risk prediction level, including the wind speed of the ventilation system, the frequency of temperature monitoring, the sensitivity of fire detectors, and the status of fire extinguishing equipment. Compared with traditional systems that passively wait for a fire to occur and then trigger fire extinguishing, this application can actively adjust the response level of the fire extinguishing system according to the predicted fire risk to achieve precise fire extinguishing. When the fire risk level rises, the system can not only activate the fire extinguishing device but also select the appropriate type and injection volume of the fire extinguishing agent according to the specific vehicle distribution and environmental conditions to achieve zoned fire extinguishing and intelligent scheduling, thereby minimizing the damage caused by the fire to the tunnel and vehicles to the greatest extent.Compared with the prior art, the tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system of the present application can not only accurately monitor the vehicle distribution and environmental information in the tunnel, conduct a comprehensive fire risk assessment, but also intelligently control the response measures of the fire extinguishing equipment according to the predicted fire risk level. Its advantages are reflected in dynamic real-time fire monitoring, intelligent risk assessment and proactive fire extinguishing response, effectively improving the safety and efficiency of fire prevention and control for electric vehicles and special vehicles in the tunnel.

[0074] In some examples, such as Figure 2 shown, the specific steps for the above-mentioned electric vehicle identification unit to determine the distribution information of electric vehicles in the target tunnel include:

[0075] S110. Based on the above-mentioned first vehicle picture information, the above-mentioned second vehicle picture information and the picture shooting time information, determine the existing vehicle information in the target tunnel;

[0076] S120. Based on the above-mentioned existing vehicle information, determine the above-mentioned electric vehicle quantity information, the above-mentioned special vehicle quantity information, the above-mentioned electric vehicle model information and the above-mentioned special vehicle type information;

[0077] S130. Based on the above-mentioned existing vehicle information, the speed information of the existing vehicle entering the tunnel and the above-mentioned picture shooting time information, determine the position information of the electric vehicle in the tunnel and the position information of the special vehicle in the tunnel.

[0078] Exemplarily, first, the electric vehicle identification unit analyzes the first vehicle picture information and the second vehicle picture information provided by the image acquisition unit, and combines the picture shooting time information to determine the existing vehicle information in the target tunnel. These existing vehicle information includes basic data such as picture data and shooting time of all vehicles entering and leaving the tunnel. Through these data, the system can identify all vehicles in the current tunnel as the basis for subsequent vehicle analysis.

[0079] Next, based on the above-mentioned existing vehicle information that has been obtained, the electric vehicle identification unit further analyzes the attributes of each vehicle and identifies electric vehicles and special vehicles therefrom. The system determines the electric vehicle quantity information in the tunnel, that is, the number of electric vehicles existing in the current tunnel, according to the vehicle appearance characteristics (such as license plates, vehicle model identification, etc.) and in combination with the trained vehicle identification model. At the same time, the system will also identify the special vehicle quantity information, including the quantity of special vehicles such as dangerous goods transport vehicles and rescue vehicles. Through this classification, the system can effectively distinguish different types of vehicles and provide refined data for subsequent risk assessment.

[0080] Based on the existing vehicle information, the electric vehicle identification unit further identifies the specific electric vehicle model information and special vehicle type information through vehicle model analysis. The system uses a predefined vehicle model recognition library to accurately match the vehicle appearance features (such as body shape, logo, license plate, etc.) to determine the specific model of the electric vehicle and further classify the types of special vehicles. The model information of the electric vehicle is an important parameter for fire risk assessment because different models of electric vehicles may use different battery types and there are differences in fire risks. Similarly, the type information of special vehicles is also an important input parameter, which can help the system identify special vehicles (such as vehicles transporting dangerous chemicals) and adjust the fire prevention and control strategies accordingly.

[0081] Finally, based on the existing vehicle information, the electric vehicle identification unit combines the speed information and the picture shooting time information when the existing vehicle enters the tunnel to determine the position information of each electric vehicle and special vehicle in the tunnel. By analyzing the time point when the vehicle enters the tunnel, the driving speed and the change of vehicle position, the system can accurately estimate the specific position of the current vehicle in the tunnel. This step is particularly important for the subsequent area division of fire prevention and control and fire extinguishing response.

[0082] Speed information: Obtain the speed data of the vehicle when it enters the tunnel through vehicle sensors or external monitoring systems as the basis for position estimation.

[0083] Picture shooting time information: Combining the time point of picture shooting, the system can calculate the distance traveled by the vehicle according to the speed information to determine the accurate position of the vehicle in the tunnel.

[0084] In summary, the electric vehicle identification unit can accurately obtain the quantity, type, model of electric vehicles and special vehicles in the tunnel and their distribution position information in the tunnel. These information provide key input data for the system's fire monitoring and early warning and linkage fire extinguishing systems, ensuring that the system can respond accurately and in a timely manner according to the vehicle distribution and fire risks.

[0085] In some examples, as Figure 3 shown, S120 determines the above-mentioned electric vehicle quantity information, the above-mentioned special vehicle quantity information, the above-mentioned electric vehicle model information and the above-mentioned special vehicle type information based on the above-mentioned existing vehicle information, including:

[0086] S1201. Perform preprocessing operations on the picture information corresponding to the existing vehicle information to obtain preprocessed picture information;

[0087] S1202. Input the above-mentioned preprocessed picture information into the first recognition module to obtain the above-mentioned electric vehicle quantity information and the above-mentioned special vehicle quantity information, where the above-mentioned first recognition module is established based on the Faster R-CNN model;

[0088] S1203. Input the above preprocessed picture information into the second recognition module to obtain the above electric vehicle model information and the above special vehicle type information, where the above second recognition module is established based on the SVM model.

[0089] Exemplarily, first, perform preprocessing operations on the picture information corresponding to the existing vehicle information to obtain preprocessed picture information. This operation aims to improve the recognition accuracy and efficiency of the subsequent vehicle recognition module. The preprocessing steps may include image denoising, image enhancement, size adjustment, etc. By preprocessing the picture information, it can ensure that the quality of the image data input to the subsequent recognition module meets the requirements, thus guaranteeing the accuracy of recognition.

[0090] After completing the preprocessing operation, the system inputs the preprocessed picture information into the first recognition module to obtain the electric vehicle quantity information and the special vehicle quantity information. Specifically, the first recognition module is established based on the Faster R-CNN model. By processing the preprocessed picture information, the Faster R-CNN model can detect the vehicle targets appearing in the picture and perform a preliminary recognition of the vehicle categories. During this recognition process, the model will identify the electric vehicle quantity information in the current picture, that is, detect the number of targets belonging to electric vehicles in the picture. At the same time, the model will also identify the special vehicle quantity information, that is, detect the number of targets belonging to special vehicles in the picture. Through the object detection and classification capabilities of the Faster R-CNN model, the system can automatically analyze the vehicle pictures and generate the quantity information of electric vehicles and special vehicles. This step is the basis for the system to further refine the recognition.

[0091] After completing the recognition of the vehicle quantity information, the system further inputs the same preprocessed picture information into the second recognition module to obtain the electric vehicle model information and the special vehicle type information. The second recognition module is established based on the SVM model (Support Vector Machine). The system further extracts the model information of each electric vehicle by inputting the vehicle images in the preprocessed picture information into the SVM model. Different models of electric vehicles may use different battery types and have different safety levels. Therefore, accurately identifying the model information is particularly important for fire risk assessment. At the same time, the SVM model will also identify the type information of special vehicles, such as identifying different types of special vehicles such as dangerous goods transport vehicles and rescue vehicles. The type information of special vehicles will also be an important reference for fire risk prediction.

[0092] Specifically, the purpose of image preprocessing is to improve the accuracy of the recognition module, which generally includes but is not limited to the following operations: removing noise in the image using methods such as Gaussian filtering, mean filtering, or median filtering. Improving the image quality by adjusting brightness, contrast, or applying histogram equalization. Resizing the image to an appropriate size to meet the input requirements of the subsequent recognition module, such as 600×600 pixels. Through these preprocessing steps, the system generates high-quality preprocessed image information to ensure the stability of the image data input to the recognition module.

[0093] The structure of Faster R-CNN is divided into the following parts:

[0094] The convolutional neural network (CNN) is used to extract features of the image. Commonly used backbone networks include ResNet, VGG, etc.

[0095] The calculation formula of the convolutional layer:

[0096] ;

[0097] where I is the input image, W is the convolutional kernel, ∗ represents the convolution operation, b is the bias term, σ is the activation function (such as ReLU), and F is the extracted feature.

[0098] The Region Proposal Network (RPN) is used to generate candidate regions, predict whether each region contains an object, and generate its bounding box.

[0099] The loss function of RPN includes the classification loss L cls and the bounding box regression loss L reg :

[0100] ;

[0101] where pi is the predicted classification probability, pi ∗ is the true class, ti is the predicted bounding box, is the true bounding box, and λ is the balance coefficient.

[0102] ROI Pooling is used to perform ROI pooling operations on the candidate regions generated by RPN to fix the output feature size for subsequent classification and regression.

[0103] Perform class classification and fine adjustment of the bounding box for the candidate regions. Faster R-CNN classifies and predicts the number of vehicles and gives the corresponding class (electric vehicle or special vehicle).

[0104] Classification loss:

[0105] ;

[0106] Among them, is the true category, and p i is the predicted probability.

[0107] Bounding box regression loss:

[0108] ;

[0109] Among them, t i is the predicted box, and

[0110] is the true box.

[0111] Through the above structure, the Faster R-CNN model can identify the number of electric vehicles and special vehicles in the picture.

[0112] Feature extraction: Extract features from the preprocessed picture information through image processing algorithms, such as HOG (Histogram of Oriented Gradients) features.

[0113] HOG feature calculation:

[0114] ;

[0115] Among them, Gx and Gy are the gradients of the image.

[0116] SVM (Support Vector Machine) is a model used to identify the models of electric vehicles and the types of special vehicles. The structure and recognition process of SVM are as follows:

[0117] SVM classifier: SVM classifies by maximizing the margin between categories.

[0118] Classification function:

[0119] Among them, w is the weight vector, x is the feature vector, and b is the bias term.

[0120] Training objective: Train the SVM model by optimizing the following loss function:

[0121] Loss function (hinge loss):

[0122] ;

[0123] Among them, y i is the label of the sample, and f(xi) is the predicted value.

[0124] Through this process, SVM can accurately identify the models of electric vehicles and the types of special vehicles according to the features of the vehicles.

[0125] By performing image preprocessing on the existing vehicle information and using a first recognition module based on the Faster R-CNN model and a second recognition module based on the SVM model respectively, the automatic recognition and extraction of the number information of electric vehicles in the tunnel, the number information of special vehicles, the electric vehicle model information, and the special vehicle type information are realized. This method can effectively improve the recognition accuracy of the system and provide accurate basic data support for subsequent fire risk prediction and fire extinguishing system response.

[0126] In some examples, such as Figure 4 shown, S130 determines the position information of the above-mentioned electric vehicles in the tunnel and the position information of the above-mentioned special vehicles in the tunnel based on the above-mentioned existing vehicle information, the speed information of the existing vehicle information entering the tunnel, and the above-mentioned picture shooting time information, including:

[0127] S1301. Send the first existing vehicle information with the above-mentioned speed information less than the preset speed to the first prediction module to obtain the first predicted position corresponding to the first existing vehicle information, where the above-mentioned first prediction module is established based on the Kalman filter algorithm;

[0128] S1302. Send the second existing vehicle information with the above-mentioned speed information greater than or equal to the above-mentioned preset speed to the second prediction module to obtain the second predicted position corresponding to the second existing vehicle information, where the above-mentioned second prediction module is established based on the long short-term memory network;

[0129] S1303. Establish integrated position information according to the first predicted position and the second predicted position;

[0130] S1304. Determine the position information of the above-mentioned electric vehicles in the tunnel and the position information of the above-mentioned special vehicles in the tunnel according to the integrated position information and the vehicle type lane and order information of the existing vehicle information entering the tunnel.

[0131] Exemplarily, first, the system performs speed analysis on the existing vehicle information of all vehicles and compares the speed information. If the speed information of some vehicles is less than the preset speed, these vehicles will be classified as low-speed vehicles, that is, the first existing vehicle information. This vehicle information will be sent to the first prediction module for predicting its position.

[0132] The first prediction module is established based on the Kalman filter algorithm. The Kalman filter is a linear estimation method for continuous time series, especially suitable for predicting the position of low-speed or smoothly moving vehicles. The Kalman filter performs recursive estimation through the existing vehicle speed information, time information, and the previous position of the vehicle, and finally obtains the first predicted position corresponding to the first existing vehicle information.

[0133] For vehicles whose speed information is greater than or equal to the preset speed, the system classifies them as high-speed vehicles, i.e., the second existing vehicle information. This vehicle information will be sent to the second prediction module. The second prediction module is based on the Long Short-Term Memory network (LSTM). LSTM is a recurrent neural network that can handle long-term dependencies and is particularly suitable for vehicles with non-linear and complex motion trajectories. The LSTM model can predict the future motion trajectory of high-speed vehicles through historical time series data (such as the speed change of the vehicle and the position at the previous moment). The system uses the LSTM model to predict the future positions of these high-speed vehicles, and then obtains the second predicted positions corresponding to the second existing vehicle information.

[0134] The system obtains the results from the first predicted position predicted by the Kalman filter algorithm and the second predicted position predicted by the LSTM algorithm respectively. Next, the system combines the first predicted position (the predicted position of low-speed vehicles) with the second predicted position (the predicted position of high-speed vehicles) to establish integrated position information.

[0135] Finally, the system will further accurately determine the position information of electric vehicles and special vehicles in the tunnel by combining the integrated position information with the vehicle type, lane, and order information of the existing vehicles entering the tunnel.

[0136] The vehicle type information enables the system to identify electric vehicles and special vehicles and determine their specific positions in the tunnel respectively. The lane information is used to determine the driving trajectory of the vehicle to ensure that the system can accurately locate the vehicle according to the specific lane where the vehicle is located. The order information is used to determine the order in which vehicles enter the tunnel. Through this information, the system can dynamically track the movement path of the vehicle and further correct the predicted position to ensure that the position information determined by the system matches the actual situation.

[0137] Specifically, the system first obtains the speed information of all vehicles and classifies the vehicles into two categories:

[0138] The first category: low-speed vehicles, with speed information vi < vth (where vth is the preset speed threshold), i.e., the first existing vehicle information.

[0139] The second category: high-speed vehicles, with speed information vi ≥ vth, i.e., the second existing vehicle information.

[0140] Low-speed vehicle position prediction (based on Kalman filter) Kalman filter is a recursive algorithm used to estimate the current position of low-speed vehicles. The steps are as follows:

[0141] The state equation is:

[0142] xk+1 = Ax k + Bu k + wk

[0143] where x k is the state vector of the vehicle at a certain moment, including position and velocity, A is the state transition matrix, B is the control input matrix, and u k is the control input, and w k is the process noise.

[0144] Observation equation:

[0145] zk = Hx k + v k

[0146] where z k is the observation value obtained at time k (such as the current vehicle position and velocity information), H is the observation matrix, and v k is the observation noise.

[0147] The Kalman filtering process includes two steps: prediction and update:

[0148] Prediction step:

[0149] ;

[0150] where P is the state covariance matrix and Q is the process noise covariance.

[0151] Update step:

[0152] ;

[0153] where K k is the Kalman gain and R is the observation noise covariance.

[0154] Finally, the system obtains the predicted position of the low-speed vehicle through the Kalman filtering algorithm .

[0155] For high-speed vehicles, the LSTM model is used to capture complex speed changes and non-linear motion trajectories. The input of the LSTM includes the historical speed information and position information of the vehicle, and the output is the predicted position of the vehicle in the future.

[0156] The update equations of the LSTM are as follows:

[0157] Input gate:

[0158] ;

[0159] Forget gate:

[0160] ;

[0161] Candidate memory cell:

[0162] ;

[0163] Output gate:

[0164] ;

[0165] Memory cell update:

[0166] ;

[0167] Hidden state update:

[0168]

[0169] where x t is the input (historical speed and position), h t is the hidden state, and C t is the memory cell state.

[0170] By training the LSTM model, the system can predict the future position of high-speed vehicles .

[0171] In this solution, by separately using the first prediction module based on the Kalman filter algorithm and the second prediction module based on the long short-term memory network (LSTM), the position of vehicles with different speeds is predicted, and by integrating the predicted positions and combining vehicle type, lane, and sequence information, the position information of electric vehicles and special vehicles in the tunnel is finally accurately determined. This method can not only make personalized predictions according to the speed characteristics of vehicles, but also provide more accurate vehicle positioning information on the basis of integration, providing reliable data support for subsequent fire monitoring and early warning.

[0172] In some examples, the above risk prediction model includes a density risk prediction module, an environmental risk prediction module, and a risk fusion module;

[0173] As Figure 5 shown, the specific steps for the above risk prediction unit to determine the risk prediction level based on the risk prediction model according to the above electric vehicle distribution information, the above other vehicle distribution information, and the tunnel environment information include:

[0174] S210. Input the above electric vehicle distribution information and the above other vehicle distribution information into the above density risk prediction module to obtain a density risk factor;

[0175] S220. Input the above tunnel environment information into the above environmental risk prediction module to obtain an environmental risk factor;

[0176] S230. Input the above density risk factor and the above environmental risk factor into the above risk fusion module to obtain the above risk prediction level.

[0177] Exemplarily, the system inputs the electric vehicle distribution information and other vehicle distribution information into the density risk prediction module to calculate the density-related fire risk factor. The electric vehicle distribution information includes the number, model, and specific location of the electric vehicles in the target tunnel. The other vehicle distribution information includes the number, type, and location of the special vehicles in the tunnel. The density risk prediction module evaluates the vehicle density in the tunnel based on the vehicle distribution, especially the aggregation of electric vehicles and special vehicles. If the vehicles are too dense in a certain area, or high-risk vehicles such as electric vehicles and dangerous goods transport vehicles are concentrated in a specific area, the system will calculate the corresponding density risk factor according to the degree of density.

[0178] The system inputs the tunnel environmental information into the environmental risk prediction module to evaluate the impact of the tunnel environmental conditions on the fire risk. The tunnel environmental information usually includes but is not limited to:

[0179] Temperature information: The temperature change in the tunnel, especially the area with abnormal temperature rise.

[0180] Humidity information: The humidity data in the tunnel. Too low humidity may increase the likelihood of fire occurrence.

[0181] Ventilation condition information: The operation of the ventilation system in the tunnel. Good ventilation can quickly discharge harmful gases and heat, reducing the fire risk.

[0182] Air quality information: Data such as smoke and carbon dioxide concentration that reflect air pollution or potential fire hazards.

[0183] The environmental risk prediction module evaluates the safety of the tunnel based on these environmental parameters and generates the corresponding environmental risk factor through comprehensive calculation of data such as temperature, humidity, ventilation, and air quality. For example, when the temperature in the tunnel rises abnormally and the ventilation is poor, the environmental risk factor will be higher.

[0184] Finally, the system inputs the density risk factor obtained from the density risk prediction module and the environmental risk factor obtained from the environmental risk prediction module into the risk fusion module. The function of the risk fusion module is to weight and fuse the risk factors from different modules to generate a comprehensive fire risk assessment result. This module weights the density risk factor and the environmental risk factor respectively according to the weights set based on historical data or through real-time adjustment. The specific weights can be dynamically adjusted according to historical fire events in the tunnel, vehicle types, and current environmental conditions.

[0185] Finally, the risk integration module generates a comprehensive risk prediction level, which reflects the severity of the current fire risk in the tunnel and is used to guide the response of the interlock fire extinguishing system.

[0186] Through the mutual cooperation of three modules (density risk prediction module, environmental risk prediction module, and risk integration module), the system first calculates the density risk factors related to vehicle density according to the distribution information of electric vehicles and other vehicles. Then, based on the tunnel environmental information such as the temperature, humidity, ventilation condition, and air quality of the tunnel, the system generates corresponding environmental risk factors. Finally, the system inputs the two risk factors into the risk integration module and obtains the risk prediction level of the current tunnel through weighted processing. This risk prediction level is used to judge the level of fire risk and help the system make reasonable fire extinguishing responses under different fire risk levels.

[0187] In some examples, as Figure 6 shown, S210 inputs the above-mentioned electric vehicle distribution information and the above-mentioned other vehicle distribution information into the above-mentioned density risk prediction module to obtain density risk factors, including:

[0188] S2101. Grid the target tunnel to obtain the gridded information of the target tunnel;

[0189] S2102. Based on the above-mentioned gridded information of the target tunnel, the above-mentioned electric vehicle distribution information, and the above-mentioned other vehicle distribution information, establish the gridded vehicle distribution information;

[0190] S2103. Determine the density weight information of different vehicle types;

[0191] S2104. Determine the above-mentioned density risk factors according to the above-mentioned gridded vehicle distribution information and the above-mentioned density weight information.

[0192] Exemplarily, the system grids the target tunnel to obtain the gridded information of the target tunnel. The purpose of gridding is to divide the tunnel into multiple grid areas of equal size, and each grid represents a position segment in the tunnel. Through such grid division, the vehicle distribution in the tunnel can be refined, facilitating subsequent risk analysis. For example, the system can divide the tunnel length and width according to a predetermined resolution, and the size of each grid is determined according to the actual size of the tunnel and the monitoring accuracy. The gridded information of the target tunnel describes the positions, numbers, and their distributions in the tunnel of these grids.

[0193] After obtaining the grid information of the target tunnel, the system then establishes the vehicle distribution grid information based on this information, the electric vehicle distribution information, and the other vehicle distribution information. The electric vehicle distribution information includes the number, model, and specific location of electric vehicles in the tunnel. The system will classify electric vehicles into corresponding grids according to their location information. The other vehicle distribution information includes the number, type, and location of special vehicles in the tunnel. Similarly, the system will distribute these vehicles to the corresponding grids according to the location information of special vehicles. The vehicle distribution grid information is based on the specific distribution of vehicles in each grid in the tunnel after grid processing. This information includes the number of electric vehicles and special vehicles in each grid and is distinguished according to different vehicle types.

[0194] Next, the system sets the density weight information for different types of vehicles. The setting of the density weight information is determined according to the degree of influence of different vehicle types on the fire risk.

[0195] Due to their different types of lithium batteries, electric vehicles have different safety levels.

[0196] Special vehicles (such as dangerous goods transport vehicles) also have a relatively high density weight because the goods carried by such vehicles may increase the fire or explosion risk.

[0197] By setting the corresponding density weight information for each type of vehicle, the system can more accurately evaluate the impact of different vehicle types on the overall risk in the tunnel. For example, the weight of electric vehicles can be higher than that of ordinary vehicles, and the weight of dangerous goods transport vehicles may be even higher to better reflect their risks in density analysis.

[0198] In some examples, the density weight information of electric vehicles is obtained by weighted fusion based on the battery type information and collision test data information of electric vehicles.

[0199] Exemplarily, first, the system sets a preliminary weight value for electric vehicles according to the battery type information of electric vehicles. Different battery types play different roles in the fire risk. Some batteries are more susceptible to environmental temperature or collision impact and have a higher risk of ignition.

[0200] Lithium-ion battery: Lithium-ion batteries are widely used in electric vehicles. They have a high energy density but are prone to thermal runaway under high temperature or overcharging conditions. Therefore, electric vehicles equipped with lithium-ion batteries usually have a relatively high risk weight.

[0201] Lithium iron phosphate battery: This type of battery has high thermal stability and is relatively less prone to thermal runaway. Therefore, its weight will be lower than that of lithium-ion batteries.

[0202] Solid-state battery: Since solid-state batteries do not have liquid electrolytes, the risk of fire is relatively low. Therefore, the system assigns a lower risk weight to electric vehicles using solid-state batteries.

[0203] By identifying the battery type information of each electric vehicle, the system can set an initial weight value for electric vehicles with different types of batteries, reflecting their impact on the fire risk.

[0204] It should be noted that the battery type information is determined based on the fixed matching relationship between the electric vehicle type information and the battery type.

[0205] Next, the system will further adjust the weight value of the electric vehicle according to the collision test data information. Collision test data is an important indicator for evaluating the safety of electric vehicles in collision events. The structural strength and battery protection performance of electric vehicles in collisions directly affect their fire risk.

[0206] Collision level classification: According to the collision test data of different electric vehicles, the system classifies them into different collision levels. A higher collision level indicates that the vehicle has better safety performance in the event of a collision, while a lower collision level means that the battery may be more easily damaged in an accident, increasing the fire risk.

[0207] An electric vehicle with a higher collision level will have a lower weight value because the vehicle structure and battery protection measures are relatively perfect and the fire risk is smaller.

[0208] An electric vehicle with a lower collision level will have a higher weight value because the battery may be more easily damaged and cause a fire in an accident.

[0209] Source of collision data: Collision test data can come from standardized vehicle safety test institutions, such as safety test rating institutions like NCAP, to ensure the reliability and scientific nature of the data.

[0210] Based on the collision test data information of electric vehicles, the system can further adjust the weight of electric vehicles to accurately reflect their risk contribution in vehicle density.

[0211] In some examples, the above tunnel environment information includes temperature information, humidity information, ventilation condition information, and air quality information;

[0212] Inputting the above tunnel environment information into the above environmental risk prediction module to obtain environmental risk factors, including:

[0213] Performing weighted fusion based on the above temperature information, the above humidity information, the above ventilation condition information, and the above air quality information to obtain the above environmental risk factors.

[0214] Exemplarily, first, the system evaluates the fire risk in the tunnel based on the real-time monitored temperature information. High temperature may be a direct sign of fire hazard. Especially in a closed environment like a tunnel, an abnormal increase in temperature is usually a risk factor that requires key attention.

[0215] The system obtains the temperature information at different positions in the tunnel through multi-point temperature sensors, including the current temperature and the temperature change trend. For areas with too high temperature or rapid temperature rise, the system will set a high-risk weight because these areas may have battery overheating or other high-temperature hazards.

[0216] The risk weight Wtemp of the temperature information is adjusted according to the degree to which the temperature exceeds the safety threshold and the temperature change rate.

[0217] Humidity plays an important role in the occurrence and spread of fires. Monitoring the humidity information in the tunnel helps to evaluate the possibility of a fire. The system obtains the humidity information of the air in the tunnel through humidity sensors. Usually, too low humidity is likely to cause dry air, thus increasing the probability of a fire.

[0218] If the humidity in the tunnel is too low, the system will set a high-risk weight Whumidity for the humidity information. Conversely, if the humidity is high, the fire risk is considered relatively low. By setting the risk threshold of humidity, the system can automatically adjust the humidity risk factor to ensure that the fire risk under different humidity conditions is properly evaluated.

[0219] Ventilation condition information is another important factor in evaluating fire risk. Good ventilation can reduce the gas accumulation and high-temperature buildup in the tunnel, reducing the fire risk. The system monitors the ventilation condition information in real time through wind speed monitors installed in the tunnel. If the ventilation is poor, the system will set a high-risk weight Wvent for this area. Poor ventilation will accelerate the accumulation of heat and smoke, increasing the possibility of fire spread. Therefore, in the case of poor ventilation conditions, the system will increase the risk level of the corresponding area.

[0220] Air quality information includes the smoke concentration, carbon dioxide content, and other harmful gas concentrations in the tunnel. The deterioration of air quality may be a precursor to an impending or ongoing fire. The system monitors the air composition in the tunnel through air quality sensors, such as smoke concentration, carbon dioxide concentration, and other possible pollutants. If the air quality deteriorates sharply, especially when the concentration of smoke or harmful gases rises significantly, the system will assign a high-risk weight Wair to the air quality information. The change in air quality can directly reflect the potential fire situation. Therefore, through real-time monitoring, the system can identify the fire risk at an early stage.

[0221] Finally, the system will perform weighted fusion on the respective risk weights extracted from the temperature information, humidity information, ventilation condition information, and air quality information to generate an environmental risk factor. The specific weighted fusion formula is as follows:

[0222] ;

[0223] Fenv represents the finally calculated environmental risk factor.

[0224] w1, w2, w3, and w4 are the weight coefficients of each factor. These weight coefficients can be adjusted according to factors such as historical data and current environmental conditions to reflect the relative impact of different factors on the fire risk.

[0225] In some examples, S230 inputs the above density risk factor and the above environmental risk factor into the above risk fusion module to obtain the above risk prediction level, including:

[0226] S2301. Determine the first weight information corresponding to the above density risk factor and the second weight information corresponding to the above environmental risk factor based on the historical fire risk information;

[0227] S2302. Calculate the risk score based on the above density risk factor, the above environmental risk factor, the above first weight information, and the above second weight information;

[0228] S2303. Determine the above risk level based on the above risk score.

[0229] Exemplarily, first, the system will refer to the historical fire risk information to determine the relative importance of the density risk factor and the environmental risk factor in the comprehensive risk assessment.

[0230] The density risk factor reflects the density of vehicle distribution in the tunnel, especially the aggregation of electric vehicles and special vehicles. Based on historical fire cases, the system can evaluate the correlation between vehicle density and the occurrence of fires, and then set the first weight information for the density risk factor. The environmental risk factor evaluates the impact of temperature, humidity, ventilation condition, and air quality in the tunnel on the fire risk. By analyzing the relationship between past fire events and environmental conditions, the system sets the second weight information for the environmental risk factor.

[0231] These weight information (the first weight information and the second weight information) are obtained through statistical analysis of historical fire data and can be dynamically adjusted to ensure that the density risk and environmental risk can be appropriately reflected in different scenarios.

[0232] After obtaining the density risk factor, the environmental risk factor, and their respective weight information, the system proceeds to the next step, which is to calculate the risk score. The system multiplies the density risk factor by its corresponding first weight information to obtain the risk contribution value for the density part. Similarly, the system multiplies the environmental risk factor by the second weight information to obtain the contribution value for the environmental risk part.

[0233] Through weighted summation, the system calculates the comprehensive risk score, and the formula is as follows:

[0234] ;

[0235] Where R represents the final risk score.

[0236] Fdensity is the density risk factor, and Fenv is the environmental risk factor.

[0237] w1 and w2 are the first weight information and the second weight information determined from the historical fire risk information, respectively.

[0238] Through this weighted calculation, the system can obtain the overall fire risk score in the current tunnel based on the combined effects of vehicle density and environmental conditions.

[0239] Finally, the system determines the risk prediction level in the tunnel according to the calculated risk score. The relationship between the risk score and the risk level is usually defined by a pre-set grading standard. The following is an exemplary risk level division:

[0240] In this solution, the system first sets the corresponding first weight information and second weight information for the density risk factor and the environmental risk factor based on the historical fire risk information. Then, the system combines the density and environmental risk factors with their respective weights to calculate the comprehensive risk score. Finally, the system maps the risk score to different risk levels through the pre-set standard, so as to provide an accurate fire risk assessment and response plan for the linkage fire extinguishing system.

[0241] In some examples, the specific steps for the above control unit to control the waiting response level of the linkage fire extinguishing system based on the above risk prediction level include:

[0242] Based on the above risk prediction level, control the wind speed information, temperature monitoring frequency information, fire detection detector sensitivity information, and fire extinguishing equipment status information of the ventilation system of the linkage fire extinguishing system, where the above fire extinguishing equipment status information includes the operation status information of the automatic fire extinguishing device, the fire extinguishing area division status information, and the fire extinguishing agent injection volume status information.

[0243] Exemplarily, according to the risk prediction level evaluated by the system in real time, the control unit dynamically adjusts the wind speed information of the ventilation system. The ventilation system plays an important role in aspects such as air circulation and smoke emission in the tunnel.

[0244] When the risk prediction level is low, the system maintains the normal operation of the ventilation system, and the wind speed is maintained at the standard level for normal air circulation. When the risk prediction level rises to medium or high risk, the system will increase the wind speed according to the actual demand to ensure smoother air circulation in the tunnel and reduce the accumulation of harmful gases or smoke. At this time, the wind speed information will be adjusted to a higher value to cope with potential fire risks.

[0245] The temperature monitoring frequency information is another important control parameter. As the risk prediction level increases, the system needs to monitor the temperature change in the tunnel more frequently to promptly detect any temperature anomalies that may trigger a fire. In the case of low risk, the system maintains the standard temperature monitoring frequency, such as taking a temperature sample every once in a while. When the risk prediction level rises, the system will increase the temperature monitoring frequency to ensure real-time updating of the temperature status in the tunnel. The frequency may be shortened from the regular once every 10 minutes to once every 1 minute to ensure a faster response to temperature changes. In high-risk or extremely high-risk situations, the system can be set to continuous real-time monitoring to ensure that any minor temperature changes can be quickly detected.

[0246] The fire detector sensitivity information is used to adjust the sensitivity of the fire detector to fire signs. According to different risk prediction levels, the system dynamically adjusts the sensitivity of the detector. When the risk prediction level is low, the system maintains the normal detection sensitivity, and the detector will respond to potential fire signals according to the regular sensitivity threshold. When the risk level increases, the system will increase the sensitivity of the detector and lower its response threshold to ensure earlier detection of potential fire hazards. In high-risk situations, the sensitivity of the detector reaches the highest, and any temperature increase, smoke increase or other fire signs will be immediately captured by the system, triggering further fire response measures.

[0247] The system ensures that the fire extinguishing equipment is in the corresponding standby state at different risk levels by controlling the fire extinguishing equipment status information. Specifically, it includes:

[0248] (1) The operation status information of the automatic fire extinguishing device:

[0249] In the case of lower risk, the fire extinguishing equipment is in the standby state, and the pre-operation is not started. The system only conducts routine monitoring. As the risk prediction level increases, the system will gradually start the automatic fire extinguishing pre-operation, including preparing the fire extinguishing agent, checking the pipeline pressure, preheating the fire extinguishing equipment, etc., to ensure that the fire extinguishing system can be quickly put into operation when needed.

[0250] (2) Fire extinguishing area division status information:

[0251] The system divides the fire extinguishing area according to the risk prediction level and vehicle distribution information. At a lower risk, the system only preliminarily divides the high-risk areas and allocates potential fire extinguishing equipment to each area. When the risk level increases, the system divides the fire extinguishing area in the tunnel in more detail to ensure that in case of a fire, fire extinguishing can be accurately carried out in specific areas, avoiding waste of fire extinguishing resources.

[0252] (3) Fire extinguishing agent injection volume status information:

[0253] According to the risk prediction level, the system also controls the status information of the fire extinguishing agent injection volume. In the case of a lower risk, the injection volume of the fire extinguishing agent is in the lowest standby state, and an appropriate amount of fire extinguishing agent is prepared. When the risk increases, especially in the case of a high risk, the system increases the reserve amount of the fire extinguishing agent according to the predicted fire probability to ensure that there is enough fire extinguishing agent for rapid extinguishment in case of a fire. In the case of an extremely high risk, the injection volume of the fire extinguishing agent reaches the preset maximum value to ensure that the fire extinguishing system can fully cover the entire high-risk area and quickly and effectively control the fire.

[0254] In this embodiment, the control unit dynamically adjusts the wind speed information, temperature monitoring frequency information, fire detector sensitivity information, and equipment status information of the fire extinguishing equipment based on the risk prediction level, so as to realize the intelligent and hierarchical response of the linked fire extinguishing system. The system can take corresponding fire prevention measures and fire extinguishing preparations according to different risk levels to ensure the optimal management and response to the fire risk in the tunnel.

[0255] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system, characterized in that, Including: An image acquisition unit, which is used to acquire the first vehicle picture information of a vehicle driving into a target tunnel, and the image acquisition unit is also used to acquire the second vehicle picture information of a vehicle driving out of the target tunnel; An electric vehicle identification unit, which is used to determine the distribution information of electric vehicles and the distribution information of other vehicles in the target tunnel based on a vehicle identification model according to the first vehicle picture information and the second vehicle picture information. The distribution information of electric vehicles includes the number information of electric vehicles, the model information of electric vehicles, and the position information of electric vehicles in the tunnel. The distribution information of other vehicles includes the number information of special vehicles, the type information of special vehicles, and the position information of special vehicles in the tunnel; A tunnel environment collection unit, which is used to acquire the tunnel environment information of the target tunnel; A risk prediction unit, which is used to determine a risk prediction level based on a risk prediction model according to the distribution information of electric vehicles, the distribution information of other vehicles, and the tunnel environment information; A control unit, which is used to control the waiting response level of a linkage fire extinguishing system based on the risk prediction level; The risk prediction model includes a density risk prediction module, an environment risk prediction module, and a risk fusion module; The specific steps for the risk prediction unit to determine the risk prediction level based on the risk prediction model according to the distribution information of electric vehicles, the distribution information of other vehicles, and the tunnel environment information include: Inputting the distribution information of electric vehicles and the distribution information of other vehicles into the density risk prediction module to obtain a density risk factor; Inputting the tunnel environment information into the environment risk prediction module to obtain an environment risk factor; Inputting the density risk factor and the environment risk factor into the risk fusion module to obtain the risk prediction level; The step of inputting the distribution information of electric vehicles and the distribution information of other vehicles into the density risk prediction module to obtain a density risk factor includes: Performing grid processing on the target tunnel to obtain target tunnel grid information; Based on the target tunnel grid information, the distribution information of electric vehicles, and the distribution information of other vehicles, establishing vehicle distribution grid information; Determining the density weight information of different vehicle types, wherein the density weight information of electric vehicles is obtained by weighted fusion based on the battery type information and collision test data information of electric vehicles; Determining the density risk factor according to the vehicle distribution grid information and the density weight information.

2. The tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system according to claim 1, wherein The specific steps for the electric vehicle identification unit to determine the distribution information of electric vehicles in the target tunnel include: Based on the first vehicle picture information, the second vehicle picture information, and the picture shooting time information, determining the existing vehicle information in the target tunnel; Based on the existing vehicle information, determining the number information of electric vehicles, the number information of special vehicles, the model information of electric vehicles, and the type information of special vehicles; Based on the existing vehicle information, the speed information of the existing vehicle entering the tunnel, and the picture shooting time information, determining the position information of electric vehicles in the tunnel and the position information of special vehicles in the tunnel.

3. The tunnel vehicle-mounted lithium battery fire monitoring, early warning and interlock fire extinguishing system according to claim 2, characterized in that, Determining the electric vehicle quantity information, the special vehicle quantity information, the electric vehicle model information, and the special vehicle type information based on the existing vehicle information includes: Performing a preprocessing operation on the picture information corresponding to the existing vehicle information to obtain preprocessed picture information; Inputting the preprocessed picture information into a first recognition module to obtain the electric vehicle quantity information and the special vehicle quantity information, wherein the first recognition module is established based on the Faster R-CNN model; Inputting the preprocessed picture information into a second recognition module to obtain the electric vehicle model information and the special vehicle type information, wherein the second recognition module is established based on the SVM model.

4. The tunnel vehicle-mounted lithium battery fire monitoring, early warning and interlocking fire extinguishing system according to claim 2, wherein, Determining the in-tunnel position information of the electric vehicles and the in-tunnel position information of the special vehicles based on the existing vehicle information, the speed information of the existing vehicles entering the tunnel, and the picture shooting time information includes: Sending the first existing vehicle information with a speed information less than a preset speed to a first prediction module to obtain a first predicted position corresponding to the first existing vehicle information, wherein the first prediction module is established based on the Kalman filtering algorithm; Sending the second existing vehicle information with a speed information greater than or equal to the preset speed to a second prediction module to obtain a second predicted position corresponding to the second existing vehicle information, wherein the second prediction module is established based on the long short-term memory network; Establishing integrated position information according to the first predicted position and the second predicted position; Determining the in-tunnel position information of the electric vehicles and the in-tunnel position information of the special vehicles according to the integrated position information and the vehicle type lane and sequence information of the existing vehicles entering the tunnel.

5. The tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system according to claim 1, characterized in that, The tunnel environment information includes temperature information, humidity information, ventilation condition information, and air quality information; Inputting the tunnel environment information into the environmental risk prediction module to obtain environmental risk factors includes: Performing weighted fusion based on the temperature information, the humidity information, the ventilation condition information, and the air quality information to obtain the environmental risk factors.

6. For the tunnel vehicle-mounted lithium battery fire monitoring, early warning and linkage fire extinguishing system according to claim 1, inputting the density risk factor and the environmental risk factor into the risk fusion module to obtain the risk prediction level includes: Determining first weight information corresponding to the density risk factor and second weight information corresponding to the environmental risk factor based on historical fire risk information; Calculating a risk score based on the density risk factor, the environmental risk factor, the first weight information, and the second weight information; Determining the risk level based on the risk score.

7. The on-vehicle lithium battery fire monitoring, early warning and interlocking fire extinguishing system for tunnels according to claim 1, characterized in that, The specific steps for the control unit to control the waiting response level of the linkage fire extinguishing system based on the risk prediction level include: Based on the risk prediction level, control the wind speed information of the ventilation system of the linked fire extinguishing system, the temperature monitoring frequency information, the sensitivity information of the fire detectors for fire detection, and the equipment status information of the fire extinguishing equipment. Among them, the equipment status information of the fire extinguishing equipment includes the pre-operation status information of automatic fire extinguishing, the status information of fire extinguishing area division, and the status information of the fire extinguishing agent injection volume.

Images