Tunnel fire location positioning and fire source risk analysis method and system

Abstract

The invention discloses a tunnel fire position positioning and fire source risk analysis method and system, and relates to the technical field of tunnel fire detection, and the method comprises the steps: periodically collecting image, temperature, smoke, harmful gas concentration, infrared radiation and humidity data at each monitoring point in a tunnel, and dynamically adjusting the monitoring frequency according to the humidity; feature extraction is performed on the collected data, a factor set and a fuzzy evaluation set for fire judgment are constructed, and a fuzzy algorithm is utilized to calculate the membership degree of each feature to fire evaluation; a comprehensive membership value vector is generated in combination with the feature weight and is used for judging whether a fire occurs or not and performing preliminary positioning; initial monitoring equipment and front and back adjacent equipment are set, and the fire position is further accurate through fuzzy analysis of data of the front and back adjacent equipment; performing risk assessment on the fire according to the fire position and the abnormal pixel point proportion; intelligent detection, precise positioning and risk analysis of the tunnel fire are realized, and the tunnel safety management level is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel fire detection, and in particular to a method and system for locating a tunnel fire position and analyzing a fire source risk. Background Art

[0002] Tunnel fire is one of the main threats to tunnel operation safety. Once a fire occurs, it is extremely harmful, not only causing serious casualties and property losses, but also causing serious impacts on the surrounding environment and traffic. Due to the limited space inside the tunnel and poor ventilation, once a fire occurs, smoke and toxic and harmful gases spread rapidly, making it extremely difficult to evacuate personnel and extinguish fires. Therefore, timely and accurate location positioning and fire source risk analysis of tunnel fires are of great significance to improving the emergency response speed and rescue efficiency of tunnel fires.

[0003] The monitoring and early warning technology of tunnel fires has made some progress at home and abroad. In terms of fire location technology, video monitoring, temperature sensing, smoke sensing and other means are mainly used for fire detection and location, but there are still some defects;

[0004] Among them, the existing tunnel fire detection methods often lack the fusion analysis of multi-source data and set a fixed monitoring cycle. This single monitoring method is prone to false alarms or missed alarms, and will still monitor according to a fixed cycle when the humidity is high and the fire risk is low, resulting in waste of resources and slow response speed; in addition, the existing tunnel fire positioning methods often rely on a sensor with abnormal monitoring data to locate the fire position, which is easily affected by measurement errors and leads to inaccurate positioning; finally, the existing tunnel fire risk assessment methods are often based on simple threshold judgments or empirical judgments, and lack refined risk assessment methods. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In response to the technical problems in the background technology, the present invention proposes a method and system for locating the location of a tunnel fire and analyzing the risk of a fire source, which obtains multi-source data and dynamically adjusts the monitoring period based on humidity; performs fire judgment and preliminary positioning of the fire location based on a fuzzy algorithm; further determines the location or interval of the fire based on the fuzzy analysis results of adjacent monitoring equipment; and more accurately analyzes the risk of fire based on the screening of suspected flame pixels in the image, thereby solving the technical problems recorded in the background technology.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] A method for locating a tunnel fire position and analyzing a fire source risk, comprising:

[0010] At each monitoring point in each tunnel, images, first monitoring data and humidity data at corresponding positions are periodically acquired, wherein the first monitoring data includes temperature, smoke, concentration of harmful gases and infrared radiation data; and the monitoring period is dynamically adjusted based on the humidity data;

[0011] Extract features from the acquired first monitoring data; establish a factor set and a fuzzy evaluation set for fire judgment, and calculate the membership value of each feature for each element in the fuzzy evaluation set based on a fuzzy algorithm at the end of each monitoring cycle;

[0012] The membership value vectors of all feature data are combined to obtain a membership value matrix, and combined with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; based on the element values ​​in the comprehensive membership value vector, it is determined whether a fire occurs and whether to locate the fire position;

[0013] An initial monitoring device and corresponding front adjacent monitoring devices and rear adjacent monitoring devices are set; fuzzy analysis is performed on the data monitored by the front adjacent monitoring device and the rear adjacent monitoring device to obtain corresponding comprehensive membership value vectors; the position of the fire is located based on the size of the specified element values ​​in the two comprehensive membership value vectors; and the risk of the fire is judged based on the located fire position.

[0014] Specifically, all fire monitoring devices monitor fire-related data according to a preset monitoring cycle, and compare the humidity data monitored at the end of each monitoring cycle with the first and second humidity thresholds. If the monitored humidity data is greater than the second humidity threshold, the monitoring cycle of the corresponding fire monitoring device is adjusted to 2T; if the monitored humidity is less than the first humidity threshold, the monitoring cycle of the corresponding fire monitoring device is adjusted to If the monitored humidity is between the first humidity threshold and the second humidity threshold, the monitoring period is not adjusted.

[0015] Specifically, a fire judgment factor set and a fuzzy evaluation set are established, wherein the factor set includes the temperature data, smoke concentration data, CO concentration data, CO 2 Concentration data and infrared radiation intensity values ​​are used as extracted feature data; the fuzzy evaluation set is {low probability, high probability}.

[0016] Specifically, obtain the membership value vector A corresponding to each feature data i , and A i =(A 1i A 2i ), where A 1i , A 2iRespectively represent the membership values ​​of each feature data for the probability of fire occurrence as low probability and high probability, i represents the i-th feature data in the factor set, and i∈{1,2,3,4,5};

[0017] Combine the membership value vectors corresponding to each feature data in the factor set to obtain the membership value matrix B, and

[0018] The weight vector K of the feature data is calculated by the entropy weight method, where K = (k 1 k 2 k 3 );

[0019] The membership value matrix B is combined with the weight vector K of the feature data to obtain the comprehensive membership value vector C, which is expressed as: C = K·B = (C 1 C 2 ), where C 1 , C 2 The C represents the comprehensive membership value of low probability and high probability of fire occurrence, and 1 +C 2 =1.

[0020] Furthermore, a first comprehensive membership value threshold TC is preset 1 and the second comprehensive membership value threshold TC 2 , when C 1 ≥TC 1 , it is judged that there is no fire around the corresponding fire monitoring equipment; when C 2 ≥TC 2 , it is judged that a fire has occurred around the corresponding fire monitoring equipment, and further positioning of the fire location is required; when C 1 <TC 1 And C 2 <TC 2 , further judgment is required in conjunction with the next monitoring cycle; if within three consecutive monitoring cycles, the probability of fire occurrence is high or medium through analysis of the data monitored by the fire monitoring equipment, it is judged that a fire has occurred around the corresponding fire monitoring equipment, and the fire location needs to be further located.

[0021] Furthermore, whenever a fire monitoring device performs fuzzy analysis at the end of a corresponding monitoring cycle, all fire monitoring devices that also perform fuzzy analysis at the end of the cycle and the corresponding fuzzy analysis results are obtained; if there are multiple adjacent and continuous fire monitoring devices whose fuzzy analysis results all indicate that a fire has occurred around the fire monitoring device, the fire monitoring device with the largest comprehensive membership value for a high probability of fire occurrence is used as the initial detection device; for non-continuous fire monitoring devices, the fire location is located independently;

[0022] The two fire monitoring devices adjacent to the initial monitoring device are respectively recorded as the front adjacent monitoring device and the rear adjacent monitoring device, which correspond to the fire monitoring devices in front of and behind the initial monitoring device respectively.

[0023] Further, the comprehensive membership value vector C obtained by analyzing the preceding adjacent monitoring device and the following adjacent monitoring device at the end of the same monitoring cycle is obtained respectively. f With C b If the monitoring cycle of the previous adjacent monitoring device or the next adjacent monitoring device is different from that of the initial monitoring device, immediately acquire and fuzzy analyze the monitoring data of the previous adjacent monitoring device or the next adjacent monitoring device to obtain the comprehensive membership value vector C after fuzzy analysis. f With C b ;

[0024] According to the order of the initial monitoring device, the preceding adjacent monitoring device, and the following adjacent monitoring device, the comprehensive membership value C with a high probability of fire occurrence in the three comprehensive membership value vectors is obtained respectively. 1 , C f1 With C b1 .

[0025] Furthermore, if |C f1 -C b1 |≤δ, it means that the location where the fire occurred is the location where the initial monitoring equipment is placed in the tunnel, where δ represents the preset error value;

[0026] If C f1 -C b1 >δ, it means that the location of the fire is in front of the initial monitoring device in the tunnel, that is, between the initial monitoring device and the previous adjacent monitoring device, and the corresponding position interval of the two detection devices in the tunnel is sent;

[0027] If C b1 -C f1 >δ, it means that the location where the fire occurred is behind the initial monitoring device in the tunnel, that is, between the initial monitoring device and the subsequent adjacent monitoring device, and the corresponding position interval of the two monitoring devices in the tunnel is sent.

[0028] Furthermore, when it is determined that the location where the fire occurred is the location where the initial monitoring device is placed in the tunnel, the image data taken by the initial monitoring device at the end of the same monitoring cycle is obtained, and the image is screened for suspected flame pixels based on the R, G, and B values, and adjacent suspected flame pixels are merged into a suspected flame pixel set;

[0029] A plane rectangular coordinate system is established on the acquired image data to determine whether there is a suspected flame pixel with a Y-axis coordinate of 0 in the image taken by the initial monitoring device. If so, the suspected flame pixel set where the suspected flame pixel is located is obtained, and the total number of pixels n in the suspected flame pixel set is counted. 1 , based on n 1 The fire risk is determined by the ratio of / n, where n represents the total number of pixels in the captured image. 1 / n<α 1 , a first-level fire alarm is sent; if α 1 ≤n 1 / n<α 2 , a second-level fire alarm is sent; if n 1 / n≥α 2 , a level 3 fire alarm is sent; where α 1 , α 2 are the first and second discrimination thresholds of fire risk respectively, and α 2 >α 1 ;

[0030] When it is determined that the location of the fire is between the initial monitoring device and the previous adjacent monitoring device in the tunnel, or between the initial monitoring device and the subsequent adjacent monitoring device, all suspected flame pixel points located in the corresponding area are also screened out, and the suspected flame pixel point set where these suspected flame pixel points are located is obtained. In accordance with the same method as when the location of the fire is determined to be the placement position of the initial monitoring device in the tunnel, a first, second or third level fire alarm is sent.

[0031] A tunnel fire location positioning and fire source risk analysis system, comprising:

[0032] The data monitoring and acquisition module periodically acquires the image, temperature, smoke, harmful gas concentration, infrared radiation and humidity data at the corresponding location at each monitoring point in each tunnel; dynamically adjusts the monitoring cycle based on the humidity data;

[0033] The fuzzy analysis module extracts features from the acquired data; establishes a factor set and a fuzzy evaluation set for fire judgment, and calculates the membership value of each feature for each element in the fuzzy evaluation set based on the fuzzy algorithm at the end of each monitoring cycle;

[0034] The fire judgment module combines the membership value vectors of all feature data to obtain a membership value matrix, and combines it with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; based on the element values ​​in the comprehensive membership value vector, it judges whether a fire occurs and whether to locate the fire location;

[0035] The fire location and risk assessment module sets up the initial monitoring equipment and the corresponding front adjacent monitoring equipment and rear adjacent monitoring equipment; performs fuzzy analysis on the data monitored by the front adjacent monitoring equipment and the rear adjacent monitoring equipment to obtain the corresponding comprehensive membership value vector; locates the fire based on the size of the specified element values ​​in the two comprehensive membership value vectors; and makes a risk assessment of the fire based on the located fire location.

[0036] (III) Beneficial effects

[0037] The present invention provides a method and system for locating a tunnel fire position and analyzing a fire source risk, which has the following beneficial effects:

[0038] 1. By integrating multiple sensors and dynamically adjusting the monitoring cycle, comprehensive and accurate monitoring of fire-related data in the tunnel is achieved, effectively improving the accuracy of fire location and the reliability of fire source risk analysis, providing strong support for timely detection and efficient response to tunnel fires;

[0039] 2. Through feature extraction and fuzzy algorithms, the monitoring data is deeply analyzed to achieve accurate assessment of fire risks. Not only a variety of fire-related factors are considered, but also the probability of fire occurrence is quantified through fuzzy evaluation sets, which effectively improves the accuracy of fire warning and the refinement of fire source risk analysis, providing a scientific basis for tunnel safety management.

[0040] 3. By constructing a membership value matrix and a feature data weight vector, and integrating multi-source information to form a comprehensive membership value vector, an accurate quantitative assessment of fire risk is achieved; by setting a comprehensive membership value threshold, a scientific basis is provided for the timely discovery and location of fires, effectively improving the efficiency and accuracy of emergency response to tunnel fires;

[0041] 4. Through the refined layout of monitoring equipment and fuzzy analysis technology, the precise positioning of the fire location and the effective assessment of the risk level are achieved; not only the timeliness and accuracy of fire response are improved, but also the judgment of fire risk is further refined through image analysis and flame recognition technology, providing comprehensive and scientific decision-making support for emergency handling and safety management of tunnel fires. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A general diagram of the steps of a method for locating a tunnel fire position and analyzing a fire source risk provided by the present invention;

[0043] Figure 2 A flow chart of the steps of a method for locating a tunnel fire position and analyzing a fire source risk provided by the present invention;

[0044] Figure 3A structural schematic diagram of a tunnel fire location positioning and fire source risk analysis system provided by the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] refer to Figure 1 and Figure 2 The present invention provides a method for locating a tunnel fire position and analyzing a fire source risk, comprising:

[0047] Step 1: Periodically obtain the image, temperature, smoke, harmful gas concentration, infrared radiation and humidity data at the corresponding position at each monitoring point in each tunnel; dynamically adjust the monitoring period based on the humidity data;

[0048] The step one includes the following steps:

[0049] Step 101: select one monocular camera, one temperature sensor, one gas sensor, one humidity sensor and one infrared sensor, respectively design the selected monocular camera, one temperature sensor, one gas sensor, one humidity sensor and one infrared sensor into independent modules, and integrate them into the same device based on the interface and protocol of each module to monitor whether there is a fire in the tunnel, and record this device as a fire monitoring device;

[0050] Among them, the monocular camera is used to monitor the images in the tunnel. By capturing the video images of the monitoring area, the image processing algorithm is used to analyze the images in real time. When the color features, shape features, texture features or motion features (such as flickering, diffusion, etc.) that are the same as those of the flame are detected, it means that there is a fire inside the tunnel. Each monocular camera shoots downward, and the shooting range is a circle.

[0051] Temperature sensors monitor fires by monitoring temperature changes around the location where they are placed. When the temperature around a certain location is abnormally high, it means that there may be a fire around that location. When a fire occurs, the flames will generate a lot of heat, causing the ambient temperature to rise rapidly. The temperature sensor can capture this change and monitor the fire.

[0052] Gas sensors monitor fires by monitoring the concentration of smoke and harmful gases in the air inside the tunnel. When a fire occurs, it is often accompanied by the generation of a large amount of smoke and toxic gases. Gas sensors can monitor the concentration of these gases in real time and compare them with the preset safety value. Once the concentration exceeds the safety value, it means that there is a high possibility of fire inside the tunnel.

[0053] Infrared sensors monitor fires by monitoring changes in infrared radiation inside the tunnel. Heat sources such as human bodies and flames emit infrared radiation, and infrared sensors can sense changes in these radiations. When a fire occurs, the intensity of the flame's infrared radiation increases significantly. When an abnormal increase in the intensity of infrared radiation is detected, it indicates that a fire may have occurred in the tunnel.

[0054] Record the monitored temperature, smoke, harmful gas concentration, and infrared radiation data as the first monitoring data;

[0055] The monitoring and identification of fire in the tunnel based on each of the above sensor devices alone is not accurate enough, so the above multiple sensors are integrated into the same fire monitoring device to monitor multiple fire-related data at the same time, thereby making the fire monitoring results more accurate;

[0056] Step 102: Starting from the tunnel entrance, fire monitoring devices are evenly placed at the top of the tunnel, and the last fire monitoring device is exactly at the tunnel exit; the number of fire monitoring devices is set so that the monocular cameras in all fire monitoring devices can just capture all positions in the tunnel, that is, the overlapping area between the shooting ranges of the monocular cameras in any two fire monitoring devices is minimized; the position where each fire monitoring device is placed is recorded as a monitoring point;

[0057] Step 103, set a monitoring period T, and at the end of each monitoring period, obtain the data monitored by each fire monitoring device; for the monocular camera module in the fire monitoring device, obtain the monitored image data; for the temperature sensor module in the fire monitoring device, obtain the monitored temperature data; for the gas sensor module in the fire monitoring device, obtain the monitored smoke concentration and each gas concentration; for the infrared sensor module in the fire monitoring device, obtain the monitored infrared radiation intensity data;

[0058] The acquired data are preprocessed separately, including image denoising, data cleaning and smoothing, data calibration and normalization, as well as background subtraction and radiation correction operations. Due to the complexity of the tunnel, various noises may exist in the image, such as background noise, artifacts, etc. Denoising the image can improve the clarity and visualization of the image. The temperature sensor data may contain outliers or noise, which can be removed by data cleaning, while data smoothing can reduce the impact of noise on temperature readings. The gas sensor data may be biased due to changes in environmental conditions, so data calibration is required. In order to facilitate subsequent data analysis, it is usually necessary to normalize the gas concentration data and convert it into a unified metric. The infrared sensor data may contain radiation interference from the background, so background subtraction is required. In order to accurately reflect the intensity of infrared radiation generated by the fire, the infrared sensor data also needs to be radiation corrected to eliminate the impact of the instrument itself and environmental factors on the measurement results.

[0059] Step 104: The humidity sensor in the fire monitoring equipment evaluates the possibility of fire by monitoring the humidity inside the tunnel, and adjusts the monitoring period length of the fire monitoring equipment at the corresponding position based on the monitored humidity data. Specifically, all fire monitoring equipment monitors fire-related data according to a preset monitoring period, and the tunnel fire safety management personnel preset a first humidity threshold and a second humidity threshold, and the first humidity threshold is less than the second humidity threshold. The humidity data monitored each time is compared with the first and second humidity thresholds. If the monitored humidity data is greater than the second humidity threshold, it means that the humidity around the fire monitoring equipment is high, that is, the possibility of fire around the fire monitoring equipment is low, and the monitoring period of the corresponding fire monitoring equipment is adjusted to 2T; if the monitored humidity is less than the first humidity threshold, it means that the humidity around the fire monitoring equipment is low, that is, the possibility of fire around the fire monitoring equipment is high, and the monitoring period of the corresponding fire monitoring equipment is adjusted to If the monitored humidity is between the first humidity threshold and the second humidity threshold, the monitoring period is not adjusted.

[0060] When using, combine the contents in steps 101 to 104:

[0061] By integrating multiple sensors and dynamically adjusting the monitoring cycle, comprehensive and accurate monitoring of fire-related data in the tunnel is achieved, which effectively improves the accuracy of fire location and the reliability of fire source risk analysis, and provides strong support for the timely detection and efficient response to tunnel fires.

[0062] Step 2: extract features from the acquired first monitoring data; establish a factor set and a fuzzy evaluation set for fire judgment, and calculate the membership value of each feature for each element in the fuzzy evaluation set based on the fuzzy algorithm at the end of each monitoring cycle;

[0063] The step 2 includes the following steps:

[0064] Step 201: extracting features from the monitored data;

[0065] For the acquired temperature data, the temperature data monitored at the end of each monitoring cycle is used as the corresponding feature data; for the acquired gas data, the monitored smoke concentration data, CO concentration data and CO 2 The concentration data is used as the corresponding characteristic data; for the acquired infrared radiation data, the monitored infrared radiation intensity value is used as the corresponding characteristic data;

[0066] Step 202: Establish a fire judgment factor set and a fuzzy evaluation set; wherein the factor set refers to a set of various factors that affect the judgment object, which are the temperature data, smoke concentration data, CO concentration data, CO 2 Concentration data and infrared radiation intensity value data;

[0067] The fuzzy evaluation set refers to the set of various possible evaluation results that the evaluator may make on the evaluation object. Here, the fuzzy evaluation set is {low probability, high probability}, that is, the probability of fire;

[0068] In the present invention, the probability of fire occurrence is evaluated, so the fuzzy evaluation set is set to "{low probability, high probability}", and "low probability" means that the possibility of fire occurrence is low, and "high probability" means that the possibility of fire occurrence is high. The evaluation set elements are organized into a framework, and then combined with the membership function in the subsequent fuzzy comprehensive evaluation; an example of an embodiment is given below:

[0069] Based on the fuzzy evaluation set set in the present invention, at the end of each monitoring cycle, the membership of each feature data to "low probability" and "high probability" in the fuzzy evaluation set is calculated according to the fuzzy algorithm described in subsequent steps 203 and 204;

[0070] For example, for temperature data, if the currently monitored temperature value is within the normal range or only slightly higher than the normal range, its membership to "low probability" is higher, while its membership to "high probability" is lower; if the temperature value rises sharply and significantly exceeds the normal range, its membership to "high probability" is higher; similarly, the same fuzzy evaluation is performed for data such as smoke concentration, CO concentration, CO2 concentration, and infrared radiation intensity;

[0071] Step 203: Obtain the historical fire record data of the tunnel, including the location of the fire and the corresponding monitoring data, obtain the data monitored by the fire monitoring equipment closest to each fire, filter out the data monitored from the time the fire occurs to the time the fire alarm is released, and obtain the abnormal monitoring interval [X] of each feature data based on the maximum and minimum values ​​of each feature data. 2i ,X 5i ]; obtain the monitoring data of each feature when no fire is detected within a week in history, and obtain the normal monitoring interval of each feature data based on the maximum and minimum values ​​of each feature data [X 1i ,X 4i ]; Among them, the abnormal monitoring interval [X 2i ,X 5i ] and the normal monitoring interval [X 1i ,X 4i ] There are intersecting intervals in the middle, specifically [X 2i ,X 4i ]; where i represents the i-th feature data in the factor set, and i∈{1,2,3,4,5};

[0072] Step 204: At the end of the monitoring cycle corresponding to each fire monitoring device, the corresponding feature data is extracted according to the above steps, and a fuzzy analysis operation is performed to determine whether a fire has occurred, specifically:

[0073] Each feature data is fuzzily marked as "small" and "large". When the fuzzy mark is "small", it means that the probability of fire is low under the value of the corresponding feature data; when the fuzzy mark is "large", it means that the probability of fire is high under the value of the corresponding feature data; among them, the membership function of the feature fuzzily marked as "small" is:

[0074]

[0075] The membership function of the feature fuzzily labeled "large" is:

[0076]

[0077] Put the above two membership functions into the same plane rectangular coordinate system to obtain the overall membership function, and divide the overall membership function into two intervals, [0, X 3i ) and [X 3i ,1], where X 3 i is the horizontal coordinate value corresponding to the intersection of two membership functions;

[0078] Each characteristic data is fuzzily marked as "small" and "large" respectively. This is based on the specific data of each characteristic data and is brought into the overall membership function. The size of the membership value of each characteristic data value under the two fuzzy marks is combined to determine the probability of fire occurrence corresponding to the characteristic data value;

[0079] For example: for the temperature data that has been monitored and preprocessed, if the membership value of the temperature data to the fuzzy mark "small" is calculated to be greater than the membership value to the fuzzy mark "large", it means that the probability of fire corresponding to the temperature data is small; if the membership value of the temperature data to the fuzzy mark "small" is calculated to be less than the membership value to the fuzzy mark "large", it means that the probability of fire corresponding to the temperature data is large; similarly, the same membership value calculation and fuzzy judgment are performed for data such as smoke concentration, CO concentration, CO2 concentration and infrared radiation intensity;

[0080] Step 205: Based on the monitoring value of each feature data, the membership values ​​of the two fuzzy marks corresponding to the feature data are respectively obtained from the overall membership function; when the feature data X i Belongs to the interval [0,X 3i ), its membership value for "small" is calculated to be A 1 , the membership value for "big" is A 2 , then the membership value for the low probability of fire is A 1 , the membership value for the high probability of fire is A 2 , and at this time A 1 >A 2 , that is, the fire monitoring result corresponding to the characteristic data at this time tends to be low probability;

[0081] When the feature data X i Belongs to the interval [X 3i ,+∞], its membership value for “small” is calculated to be A 3 , the membership value for "big" is A 4 , then the membership value for the low probability of fire is A 3 , the membership value for the high probability of fire is A 4 , and at this time A 3 ≤A 4 , that is, at this time, the fire monitoring result corresponding to the feature data tends to be high probability.

[0082] When using, combine the contents in steps 201 to 205:

[0083] Through feature extraction and fuzzy algorithms, in-depth analysis of monitoring data is carried out to achieve accurate assessment of fire risks. Not only a variety of fire-related factors are taken into account, but the probability of fire occurrence is also quantified through fuzzy evaluation sets, which effectively improves the accuracy of fire warnings and the refinement of fire source risk analysis, providing a scientific basis for tunnel safety management.

[0084] Step 3: Combine the membership value vectors of all feature data to obtain a membership value matrix, and combine it with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; determine whether a fire occurs and whether to locate the fire location based on the element values ​​in the comprehensive membership value vector;

[0085] The step three includes the following steps:

[0086] Step 301: Obtain the membership value vector A corresponding to each feature data i , and A i =(A 1i A 2i ), where A 1i , A 2i Respectively represent the membership values ​​of each characteristic data pair for the probability of fire occurrence as low probability and high probability;

[0087] Combine the membership value vectors corresponding to each feature data in the factor set to obtain the membership value matrix B, and

[0088] Step 302: Calculate the weight vector K of the feature data by using the entropy weight method, where K = (k 1 k 2 k 3 ); Entropy weight method measures the degree of variation or the amount of information by calculating the information entropy of each indicator, and then determines the weight of each indicator; the membership value matrix B is combined with the weight vector K of the feature data to obtain the comprehensive membership value vector C, which is expressed as: C = K·B = (C 1 C 2 ), where C 1 , C 2 The C represents the comprehensive membership value of low probability and high probability of fire occurrence, and 1 +C 2 =1;

[0089] Step 303: The tunnel fire safety management personnel sets the first comprehensive membership value threshold TC 1 and the second comprehensive membership value threshold TC 2 , when C 1 ≥TC 1 , it means that the probability of fire is very low, that is, it is judged that there is no fire around the corresponding fire monitoring equipment, and TC 1Generally, the value is not less than 0.9; when C 2 ≥TC 2 , it means that the probability of fire is high, that is, it is judged that a fire has occurred around the corresponding fire monitoring equipment, and the fire location needs to be further located, and TC 2 Generally, the value is not less than 0.5; when C 1 <TC 1 And C 2 <TC 2 , it means that the probability of fire is medium and further judgment is needed in combination with the next monitoring cycle; if within three consecutive monitoring cycles, the data analysis monitored by the fire monitoring equipment shows that the probability of fire is high or medium, it is judged that a fire has occurred around the corresponding fire monitoring equipment and the fire location needs to be further located.

[0090] When using, combine the contents in steps 301 to 303:

[0091] By constructing a membership value matrix and a feature data weight vector and fusing multi-source information to form a comprehensive membership value vector, an accurate quantitative assessment of fire risks can be achieved. By setting the comprehensive membership value threshold, a scientific basis is provided for the timely discovery and location of fires, effectively improving the efficiency and accuracy of emergency response to tunnel fires.

[0092] Step 4: Set up the initial monitoring device and the corresponding front adjacent monitoring device and rear adjacent monitoring device; perform fuzzy analysis on the data monitored by the front adjacent monitoring device and the rear adjacent monitoring device to obtain the corresponding comprehensive membership value vector; locate the position of the fire based on the size of the specified element value in the two comprehensive membership value vectors; and make a risk assessment of the fire based on the located fire position;

[0093] The step 4 includes the following steps:

[0094] Step 401: Whenever a fire monitoring device performs fuzzy analysis at the end of a corresponding monitoring cycle, all fire monitoring devices that also perform fuzzy analysis at the end of the cycle and the corresponding fuzzy analysis results are obtained; if there are multiple adjacent and continuous fire monitoring devices whose fuzzy analysis results all indicate that a fire has occurred around the fire monitoring device, then according to the magnitude of the comprehensive membership values ​​of these fire monitoring devices with a high probability of fire occurrence at the end of the corresponding monitoring cycle, the initial monitoring device for further positioning the fire position is determined, and specifically, the fire monitoring device with the largest comprehensive membership value of a high probability of fire occurrence is used as the initial detection device; for non-continuous fire monitoring devices, the fire position is positioned independently;

[0095] Step 402: two fire monitoring devices adjacent to the initial monitoring device are recorded as a front adjacent monitoring device and a rear adjacent monitoring device, which correspond to the fire monitoring devices in front of and behind the initial monitoring device respectively;

[0096] Obtain the comprehensive membership value vector C obtained by fuzzy analysis of the initial monitoring device; respectively obtain the comprehensive membership value vector C obtained by the analysis of the preceding adjacent monitoring device and the following adjacent monitoring device at the end of the same monitoring cycle f With C b If the preceding adjacent monitoring device or the succeeding adjacent monitoring device does not monitor data at the end of the same monitoring cycle, that is, the monitoring cycle length is different from that of the initial monitoring device, the preceding adjacent monitoring device or the succeeding adjacent monitoring device is immediately subjected to a monitoring data acquisition and fuzzy analysis, and the comprehensive membership value vector C after fuzzy analysis is obtained. f With C b ;

[0097] Step 403: Obtain the comprehensive membership value C of the three comprehensive membership value vectors with high probability for the probability of fire occurrence in the order of the initial monitoring device, the previous adjacent monitoring device, and the next adjacent monitoring device. 1 , C f1 With C b1 , and C 1 >C f1 , C 1 >C b1 ;

[0098] The fire location interval is determined based on the values ​​of the three comprehensive membership values: if the fire location is exactly at the placement position of the initial monitoring device, without considering the impact of the environment, the changes in the characteristic data monitored by the previous adjacent monitoring device and the subsequent adjacent monitoring device should be equal within a certain error, that is, if |C f1 -C b1 |≤δ, it means that the location where the fire occurred is the location where the initial monitoring device is placed in the tunnel, and the location where the initial monitoring device is placed includes the spatial areas to the left and right and above and below the placement point, and δ represents the error;

[0099] The setting of the δ value is used to reduce the error caused by the actual complex situation in the tunnel; after judging that a fire occurs around a monitoring device A, the location of the fire is determined based on the comprehensive membership values ​​of the two monitoring devices adjacent to the monitoring device. For example, when the location of the fire is exactly at the location of the monitoring device A, the temperature, smoke concentration and other characteristic data may change unevenly due to the airflow, humidity and other conditions inside the tunnel, that is, the comprehensive membership values ​​of the two adjacent monitoring devices calculated at a certain time point may change, but this change is generally within a range of a threshold, that is, the error δ; the error δ is specifically obtained by simulating the difference between the comprehensive membership values ​​of the two adjacent monitoring devices at a fixed fire point under various environmental conditions, but it can also be set by the system administrator based on the actual experience;

[0100] If C f1 -C b1 >δ, it means that the changes of the characteristic data monitored by the front adjacent monitoring device are faster than the changes of the characteristic data monitored by the rear adjacent monitoring device, that is, the location of the fire is in front of the initial monitoring device in the tunnel, that is, between the initial monitoring device and the front adjacent monitoring device, and the corresponding position intervals of the two detection devices in the tunnel are sent to the tunnel fire management system;

[0101] If C b1 -C f1 >δ, it means that the changes of the characteristic data monitored by the rear adjacent monitoring device are faster than the changes of the characteristic data monitored by the front adjacent monitoring device, that is, the location of the fire is behind the initial monitoring device in the tunnel, that is, between the initial monitoring device and the rear adjacent monitoring device, and the corresponding position intervals of the two monitoring devices in the tunnel are sent to the tunnel fire management system;

[0102] Step 404: assess the risk of the fire based on the location of the fire, and take different safety fire-fighting measures for fires of different risks;

[0103] When it is determined that the location where the fire occurred is the placement location of the initial monitoring device in the tunnel, the image data taken by the initial monitoring device at the end of the same monitoring cycle is obtained, and the end of the same monitoring cycle indicates that the source of the feature data for the above-mentioned fuzzy analysis is the same monitoring cycle; the R, G, and B values ​​of each pixel in the image are obtained, and the R value represents the red component of each pixel. When the value of the R channel exceeds the R channel threshold preset by the tunnel fire safety management personnel, and the component values ​​of the three channels satisfy the condition that R value>G value>B value, the pixel can be considered as a suspected flame pixel; the adjacent points of each suspected flame pixel are matched and merged. If another pixel adjacent to the suspected flame pixel is also a suspected flame pixel, the two pixels are merged into the same suspected flame pixel set, and after the merging, the adjacent pixels of each suspected flame pixel in the suspected flame pixel set are traversed to determine whether they can be merged, until all adjacent pixels have been determined;

[0104] A plane rectangular coordinate system is established on the acquired image data, with the origin being the location of each fire monitoring device, the forward direction along the tunnel being the positive direction of the X-axis, and the backward direction being the negative direction of the X-axis; it is determined whether there is a suspected flame pixel point with a Y-axis coordinate of 0 in the image taken by the initial monitoring device. If so, the suspected flame pixel point set where the suspected flame pixel point is located is obtained, and the total number of pixels n in the suspected flame pixel set is counted 1 , based on n 1 The fire risk is determined by the ratio of / n, where n represents the total number of pixels in the captured image. 1 / n<α 1 , it means that the scope of the fire is small, and a level 1 fire alarm is sent to the tunnel fire management system; if α 1 ≤n 1 / n<α 2 , it means that the scope of the fire is moderate, and a secondary fire alarm is sent to the tunnel fire management system; if n 1 / n≥α 2 , it means that the fire is large in scope, and a level 3 fire alarm is sent to the tunnel fire management system; among them, α 1 , α 2 are the first and second discrimination thresholds of fire risk, respectively, which are set by the tunnel fire safety management personnel, and α 2 >α 1 ;

[0105] If it does not exist, it means that no flame is found in the current monitoring cycle, but the probability of fire is high through fuzzy analysis, and a fire prevention alarm is sent to the tunnel fire management system;

[0106] When it is determined that the location of the fire is between the initial monitoring device and the previous adjacent monitoring device in the tunnel, all suspected flame pixels located in the area between the initial monitoring device and the previous adjacent monitoring device are screened out, and the suspected flame pixel set where these suspected flame pixels are located is obtained. The fire risk is judged based on the proportion of all non-repeated suspected flame pixels in the suspected flame pixel set to the total pixels in the image, and the first, second, and third level fire alarms are sent according to the same method as above;

[0107] When the location of the fire is determined to be between the initial monitoring device and the subsequent adjacent monitoring device in the tunnel, the same fire risk judgment is performed.

[0108] When using, combine the contents in steps 401 to 404:

[0109] Through the refined layout of monitoring equipment and fuzzy analysis technology, the precise positioning of the fire location and the effective assessment of the risk level are achieved; not only the timeliness and accuracy of fire response are improved, but also the judgment of fire risk is further refined through image analysis and flame recognition technology, providing comprehensive and scientific decision-making support for emergency handling and safety management of tunnel fires.

[0110] refer to Figure 3 The present invention also provides a tunnel fire location positioning and fire source risk analysis system, comprising:

[0111] The data monitoring and acquisition module periodically acquires the image, temperature, smoke, harmful gas concentration, infrared radiation and humidity data at the corresponding location at each monitoring point in each tunnel; dynamically adjusts the monitoring cycle based on the humidity data;

[0112] The fuzzy analysis module extracts features from the acquired data; establishes a factor set and a fuzzy evaluation set for fire judgment, and calculates the membership value of each feature for each element in the fuzzy evaluation set based on the fuzzy algorithm at the end of each monitoring cycle;

[0113] The fire judgment module combines the membership value vectors of all feature data to obtain a membership value matrix, and combines it with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; based on the element values ​​in the comprehensive membership value vector, it judges whether a fire occurs and whether to locate the fire location;

[0114] The fire location and risk assessment module sets up the initial monitoring equipment and the corresponding front adjacent monitoring equipment and rear adjacent monitoring equipment; performs fuzzy analysis on the data monitored by the front adjacent monitoring equipment and the rear adjacent monitoring equipment to obtain the corresponding comprehensive membership value vector; locates the fire based on the size of the specified element values ​​in the two comprehensive membership value vectors; and makes a risk assessment of the fire based on the located fire location.

[0115] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer storage medium or transmitted via a computer storage medium.

[0116] Computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. Computer storage media can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media. Available media can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid state drives (SSDs)).

[0117] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for locating a tunnel fire location and analyzing fire source risk, characterized in that: include: At each monitoring point in each tunnel, images, first monitoring data and humidity data at corresponding positions are periodically acquired, wherein the first monitoring data includes temperature, smoke, concentration of harmful gases and infrared radiation data; and the monitoring period is dynamically adjusted based on the humidity data; Extract features from the acquired first monitoring data; establish a factor set and a fuzzy evaluation set for fire judgment, and calculate the membership value of each feature for each element in the fuzzy evaluation set based on a fuzzy algorithm at the end of each monitoring cycle; The membership value vectors of all feature data are combined to obtain a membership value matrix, and combined with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; based on the element values ​​in the comprehensive membership value vector, it is determined whether a fire occurs and whether to locate the fire position; Setting up an initial monitoring device and corresponding preceding adjacent monitoring devices and following adjacent monitoring devices; Perform fuzzy analysis on the data monitored by the front adjacent monitoring device and the rear adjacent monitoring device to obtain the corresponding comprehensive membership value vector; The location of the fire is located based on the size of the specified element values ​​in the two comprehensive membership value vectors; and the risk of the fire is judged based on the located fire location.

2. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 1, characterized in that: All fire monitoring devices are used to monitor fire-related data according to a preset monitoring cycle. The humidity data monitored at the end of each monitoring cycle is compared with the first and second humidity thresholds. If the humidity data monitored is greater than the second humidity threshold, the monitoring cycle of the corresponding fire monitoring device is adjusted to 2T; if the humidity data monitored is less than the first humidity threshold, the monitoring cycle of the corresponding fire monitoring device is adjusted to If the monitored humidity is between the first humidity threshold and the second humidity threshold, the monitoring period is not adjusted.

3. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 1, characterized in that: A factor set and a fuzzy evaluation set for fire judgment are established, wherein the factor set includes the temperature data, smoke concentration data, CO concentration data, CO2 concentration data and infrared radiation intensity value monitored at the end of each monitoring cycle as the extracted feature data; the fuzzy evaluation set is {low probability, high probability}.

4. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 1, characterized in that: Get the membership value vector A corresponding to each feature data i , and Ai=(A1i A2i), where A1i and A2i represent the membership values ​​of each feature data for the probability of fire occurrence as low probability and high probability, respectively, i represents the i-th feature data in the factor set, and i∈{1,2,3,4,5}; Combine the membership value vectors corresponding to each feature data in the factor set to obtain the membership value matrix B, and The weight vector K of the feature data is calculated respectively by the entropy weight method, where K = (k1 k2 k3); The membership value matrix B is combined with the weight vector K of the feature data to obtain the comprehensive membership value vector C, which is expressed as: C=K·B=(C1 C2), where C1 and C2 represent the comprehensive membership values ​​for low and high probability of fire occurrence, respectively, and C1+C2=1.

5. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 4, characterized in that: A first comprehensive membership value threshold TC1 and a second comprehensive membership value threshold TC2 are preset. When C1≥TC1, it is judged that no fire has occurred around the corresponding fire monitoring device; when C2≥TC2, it is judged that a fire has occurred around the corresponding fire monitoring device, and the fire position needs to be further located; when C1<TC1 and C2<TC2, further judgment is required in combination with the next monitoring cycle; if within three consecutive monitoring cycles, the probability of fire occurrence is high or medium through analysis of the data monitored by the fire monitoring device, it is judged that a fire has occurred around the corresponding fire monitoring device, and the fire position needs to be further located.

6. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 5, characterized in that: Whenever a fire monitoring device performs fuzzy analysis at the end of a corresponding monitoring cycle, all fire monitoring devices that also perform fuzzy analysis at the end of the cycle and the corresponding fuzzy analysis results are obtained; if there are multiple adjacent and continuous fire monitoring devices whose fuzzy analysis results all indicate that a fire has occurred around the fire monitoring device, the fire monitoring device with the largest comprehensive membership value for a high probability of fire occurrence is used as the initial detection device; for non-continuous fire monitoring devices, the fire location is located independently; The two fire monitoring devices adjacent to the initial monitoring device are respectively recorded as the front adjacent monitoring device and the rear adjacent monitoring device, which correspond to the fire monitoring devices in front of and behind the initial monitoring device respectively.

7. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 6, characterized in that: The comprehensive membership value vector C obtained by analyzing the front adjacent monitoring device and the rear adjacent monitoring device at the end of the same monitoring cycle is obtained respectively f With C b If the monitoring cycle of the previous adjacent monitoring device or the next adjacent monitoring device is different from that of the initial monitoring device, immediately acquire and fuzzy analyze the monitoring data of the previous adjacent monitoring device or the next adjacent monitoring device to obtain the comprehensive membership value vector C after fuzzy analysis. f With C b ; According to the order of the initial monitoring device, the preceding adjacent monitoring device, and the following adjacent monitoring device, the comprehensive membership values ​​C1, C2, and C3 for the high probability of fire occurrence in the three comprehensive membership value vectors are obtained respectively. f1 With C b1 .

8. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 7, characterized in that: If |C f1 -C b1 |≤δ, it means that the location where the fire occurred is the location where the initial monitoring equipment is placed in the tunnel, where δ represents the preset error value; If C f1 -C b1 >δ, it means that the location of the fire is in front of the initial monitoring device in the tunnel, that is, between the initial monitoring device and the previous adjacent monitoring device, and the corresponding position interval of the two detection devices in the tunnel is sent; If C b1 -C f1 >δ, it means that the location where the fire occurred is behind the initial monitoring device in the tunnel, that is, between the initial monitoring device and the subsequent adjacent monitoring device, and the corresponding position interval of the two monitoring devices in the tunnel is sent.

9. A method for locating a tunnel fire position and analyzing a fire source risk according to claim 8, characterized in that: When it is determined that the location where the fire occurred is the location where the initial monitoring device is placed in the tunnel, the image data taken by the initial monitoring device at the end of the same monitoring cycle is obtained, and the suspected flame pixels of the image are screened based on the R, G, and B values, and the adjacent suspected flame pixels are merged into a suspected flame pixel set; A plane rectangular coordinate system is established on the acquired image data to determine whether there is a suspected flame pixel point with a Y-axis coordinate of 0 in the image taken by the initial monitoring device. If so, the suspected flame pixel point set where the suspected flame pixel point is located is obtained, and the total number of pixels n1 in the suspected flame pixel set is counted. The risk of fire is determined based on the ratio of n1 / n, where n represents the total number of pixels in the taken image. If n1 / n<α1, a first-level fire alarm is sent; if α1≤n1 / n<α2, a second-level fire alarm is sent; if n1 / n≥α2, a third-level fire alarm is sent; where α1 and α2 are the first and second discrimination thresholds of fire risk, respectively, and α2>α1; When it is determined that the location of the fire is between the initial monitoring device and the previous adjacent monitoring device in the tunnel, or between the initial monitoring device and the subsequent adjacent monitoring device, all suspected flame pixel points located in the corresponding area are also screened out, and the suspected flame pixel point set where these suspected flame pixel points are located is obtained. In accordance with the same method as when the location of the fire is determined to be the placement position of the initial monitoring device in the tunnel, a first, second or third level fire alarm is sent.

10. A tunnel fire location positioning and fire source risk analysis system, characterized in that: include: The data monitoring and acquisition module periodically acquires the image, temperature, smoke, harmful gas concentration, infrared radiation and humidity data at the corresponding location at each monitoring point in each tunnel; dynamically adjusts the monitoring cycle based on the humidity data; The fuzzy analysis module extracts features from the acquired data; establishes a factor set and a fuzzy evaluation set for fire judgment, and calculates the membership value of each feature for each element in the fuzzy evaluation set based on the fuzzy algorithm at the end of each monitoring cycle; The fire judgment module combines the membership value vectors of all feature data to obtain a membership value matrix, and combines it with the weight vector corresponding to the feature data to obtain a comprehensive membership value vector; based on the element values ​​in the comprehensive membership value vector, it judges whether a fire occurs and whether to locate the fire location; Fire location and risk assessment module, setting up the initial monitoring equipment and the corresponding front adjacent monitoring equipment and rear adjacent monitoring equipment; Perform fuzzy analysis on the data monitored by the front adjacent monitoring device and the rear adjacent monitoring device to obtain the corresponding comprehensive membership value vector; The location of the fire is located based on the size of the specified element values ​​in the two comprehensive membership value vectors; and the risk of the fire is judged based on the located fire location.

Images