A fire warning method and system based on drone

Abstract

The present invention discloses a fire warning method and system based on an unmanned aerial vehicle, which relates to the field of fire warning technology. The environmental data of the natural gas transmission pipeline is collected in real time by carrying a variety of sensors on the unmanned aerial vehicle, which solves the problem that traditional manual inspections and ground sensors cannot be fully monitored. The high maneuverability of the unmanned aerial vehicle can cover wider and difficult-to-reach areas, realize comprehensive and accurate monitoring of pipelines, and significantly improve the accuracy and timeliness of monitoring. The collected data is filtered, denoised and standardized by the data preprocessing module, which further improves the data quality and reduces the impact of environmental noise and interference. In addition, the data fusion and analysis module fuses and extracts features of the warning data through the particle filter data fusion algorithm, which can combine the data of multiple sensors to provide more accurate risk prediction.

Description

Technical Field

[0001] The present invention relates to the field of fire warning technology, and in particular to a fire warning method and system based on an unmanned aerial vehicle. Background Art

[0002] With the acceleration of industrialization and urbanization, pipeline networks for dangerous gases such as natural gas have become important infrastructure. This field belongs to the broad category of industrial safety and environmental monitoring, especially in natural gas pipeline monitoring and disaster warning. Specifically for fire warning systems, especially for effective monitoring and fire prediction in complex environments, the technologies involved cover drone technology, sensor technology, data fusion technology and intelligent algorithms. In these specific applications, real-time monitoring and fire warning for long-distance natural gas transmission pipelines rely on the joint work of multiple sensors, which are carried by drones to collect data in real time, and ultimately achieve more accurate warning functions.

[0003] Currently, the monitoring of natural gas pipelines mostly relies on manual inspections and traditional ground-based fixed sensors, which have obvious limitations in actual operations. For example, manual inspections are inefficient and pose high safety risks in dangerous areas; while ground sensors cannot cover the dynamic changes of all pipelines, especially it is difficult to monitor the abnormal status of pipelines in real time, accurately locate the source of gas leaks, and promptly warn of fires. The limitations of this traditional method make it difficult to detect and respond to emergencies such as leaks and fires in the pipeline in a timely manner, thereby increasing safety hazards and the risk of accidents.

[0004] Due to the inadequacy of existing monitoring methods, once a natural gas pipeline leaks or fires occur, they are often not detected immediately. Especially in the early stages of a gas leak, it is difficult to accurately locate the leak point, which leads to the accumulation of gas concentration and the rapid spread of the fire, causing losses far beyond expectations. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a fire warning method and system based on a drone, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a fire warning system based on drones, including a data acquisition module, a data preprocessing module, a data fusion and analysis module, a risk prediction module, a real-time warning and alarm module, and a data reporting and decision support module; The data acquisition module collects environmental data of the natural gas transmission pipeline through sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; The data preprocessing module performs filtering, denoising and standardization on the acquired original data set XC to obtain the early warning data set X; The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through a particle filter data fusion algorithm; The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains a risk index Risk; The real-time early warning and alarm module determines the potential risk according to the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

[0007] Preferably, the data acquisition module includes a data acquisition unit and a data aggregation unit; The data acquisition unit collects the gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS of the natural gas transmission pipeline through sensors; The gas concentration CG includes the methane concentration in the target area of ​​the natural gas transmission pipeline and is collected by a gas concentration sensor, the temperature T is collected by a temperature sensor, the humidity H is collected by a humidity sensor, the infrared radiation intensity IR is collected by an infrared thermal imager, and the ambient wind speed WS is collected by a wind speed sensor; The data aggregation unit is responsible for combining the acquired gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS to form a unified original data set XC, and transmits it to the data preprocessing module.

[0008] Preferably, the data preprocessing module includes a data denoising and filtering unit and a data standardization unit; The data denoising and filtering unit removes noise from the original data set XC by using a Kalman filter, and smoothes the original data set XC by using a filtering algorithm to obtain a denoised data set XQ; The data standardization unit performs standardization processing on the denoised data set XQ, compares and analyzes the data of different sensors at the same scale, and obtains the early warning data set X; The warning data set X is obtained by the following formula: ; Where Xd represents the d-th item of data in the warning data set X, XQd represents the d-th item of data in the denoised data set XQ, μd represents the mean of the d-th item of data, and σd represents the standard deviation of the d-th item of data.

[0009] Preferably, the data fusion and analysis module includes a data fusion unit and a feature extraction unit; The data fusion unit performs weighted fusion on the data in the warning data set X through a particle filter data fusion algorithm to obtain a fused data set XR; Specifically, weights are assigned to different types of sensor data based on their reliability, accuracy and relevance to risk prediction, and the advantages of each type of data are combined. For spatiotemporal information such as infrared image data, the spatial information of the image data is fused with the temporal information of other sensors in combination with the spatiotemporal data processing algorithm to form a more accurate risk judgment.

[0010] The fused data set XR is obtained by the following formula: ; Where XR(t) represents the fused data set at time t, i represents the measurement value of the i-th sensor, and n represents the total number of sensors. represents the preset weight value of the i-th sensor, and , Xi(t) represents the measurement value of the i-th sensor in the warning data set X at time t.

[0011] Preferably, the feature extraction unit extracts features from the fused data set XR, including the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH, and the covariance ΔCW of the gas wind speed, and forms a risk feature set F; The infrared radiation intensity change rate ΔIR is obtained by the following formula: ; In the formula, IR(t) represents the infrared radiation intensity at time t, IR(t-1) represents the infrared radiation intensity at time t-1, and Δt represents the time interval; The temperature and humidity combined fluctuation rate ΔTH is obtained by the following formula: ; Where Tj represents the temperature value of the jth data point, Hj represents the humidity value of the jth data point, m represents the total number of data points, μT and μH represent the mean of temperature and humidity, σT and σH represent the standard deviation of temperature and humidity, respectively; The covariance ΔCW of the gas wind speed is obtained by the following formula: ; Where CGj represents the gas concentration of the jth data point, WSj represents the ambient wind speed of the jth data point, μCG and μWS represent the means of the gas concentration and ambient wind speed, respectively, and σT and σH represent the standard deviations of the gas concentration and ambient wind speed, respectively.

[0012] Preferably, the risk prediction module includes a trend analysis unit and a risk index calculation unit; The trend analysis unit analyzes the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F, and predicts the risks that may occur in the future; Analyze the trend of the infrared radiation intensity change rate ΔIR to identify potential fire risks. When the infrared radiation intensity change rate ΔIR shows a continuous upward trend, it indicates the risk of fire spreading. Analyze the changing trend of the temperature and humidity combined fluctuation rate ΔTH to reveal the risk of fire and gas leakage caused by abnormal temperature and humidity fluctuations; when the temperature and humidity combined fluctuation rate ΔTH shows a continuous upward trend, it indicates that there are drastic changes in the environment, which is a precursor to fire and leakage; when the temperature and humidity combined fluctuation rate ΔTH shows irregular changes around the natural gas transmission pipeline, it indicates that gas leakage has occurred; Analyze the changing trend of the covariance ΔCW of the gas wind speed to predict the diffusion trend and risk of gas leakage; when the covariance ΔCW of the gas wind speed shows a continuous upward trend, it means that the concentration of the gas leakage is changing, and the change in wind speed leads to rapid diffusion of the gas.

[0013] Preferably, the risk index calculation unit performs a comprehensive calculation on the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F to obtain the risk index Risk; The risk index Risk is obtained by the following formula: ; In the formula, k1 represents the intensity coefficient, k2 represents the fluctuation coefficient, and k3 represents the diffusion coefficient; The strength coefficient k1 is obtained by the following formula: ; In the formula, C1 represents a constant, IR(t) represents the infrared radiation intensity at time t, and IPmax represents the peak value of the infrared radiation intensity; The fluctuation coefficient k2 is obtained by the following formula: ; In the formula, C2 represents a constant, T(t) represents the temperature at time t, H(t) represents the humidity at time t, Tmax represents the peak value of temperature, and Hmax represents the peak value of humidity; The diffusion coefficient k3 is obtained by the following formula: ; In the formula, C3 represents a constant, CG(t) represents the gas concentration at time t, WS(t) represents the ambient wind speed at time t, CGmax represents the peak value of the gas concentration, and WSmax represents the peak value of the ambient wind speed.

[0014] Preferably, the real-time early warning and alarm module includes a risk judgment alarm unit and an alarm information transmission unit; The risk judgment alarm unit compares the acquired risk index Risk with a preset risk threshold TRk to judge the state of the natural gas transmission pipeline; The state of the natural gas transmission pipeline is obtained by matching the following formula: When the risk index Risk < risk threshold TRk, it means that the natural gas transmission pipeline is in normal condition and no alarm is triggered; When the risk index Risk ≥ risk threshold TRk, it means that the natural gas pipeline is in an abnormal state, triggering an alarm, and starting the alarm program to send alarm information to the ground control center and relevant personnel; The alarm information transmission unit generates alarm information according to the triggered alarm, including the leakage source location, fire area and spreading path, and transmits the alarm information to relevant personnel through the UAV communication system.

[0015] The location of the leak source is calculated based on the gas concentration distribution and diffusion model, and the location of the leak source is predicted through multiple sensor data collection and gas diffusion model. The leakage source position Pleak is obtained by the following formula: ; Where CG(P) represents the gas concentration at position P, PEa represents the set of all possible leakage positions, argmax represents the maximum possible position, and Problem(P) represents the leakage probability at position P, which is calculated according to the gas diffusion model; The fire area is predicted based on the change rate of infrared radiation intensity ΔIR and the fire spread model. The fire spread area Afire is predicted by analyzing the change of infrared radiation intensity. The fire spread area Afire is obtained by the following formula: ; Where Afire(t) represents the fire occurrence area at time t, ΔIR(t1) represents the rate of change of infrared radiation intensity at time t1, and SpMo(t1) represents the fire spread model, which is used to estimate the expansion range of the fire at time t1; The spread path is obtained through predictions based on wind speed and gas dispersion models, and the spread path of leaks and fires is inferred by using diffusion models.

[0016] Preferably, the data report and decision support module integrates the acquired early warning data set X, risk feature set F and risk index Risk to obtain a real-time monitoring data report; generates a risk prediction result report using the risk feature set F and risk index Risk, indicating potential fire and gas leakage risks; generates an alarm event analysis report based on the status of the natural gas transmission pipeline, reporting the cause of the alarm, the location of the leakage source, the fire area and the spread path.

[0017] A fire warning method based on drone includes the following steps: Step 1: The data acquisition module collects environmental data of the natural gas pipeline through the sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; Step 2: The data preprocessing module filters, denoises and standardizes the acquired original data set XC to obtain the warning data set X; Step 3: The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through the particle filter data fusion algorithm; Step 4: The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains the risk index Risk; Step 5: The real-time early warning and alarm module determines the potential risk based on the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; Step 6: The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

[0018] The present invention provides a fire warning method and system based on drone, which has the following beneficial effects: (1) When the system is running, drones equipped with multiple sensors collect environmental data of natural gas pipelines in real time, solving the problem that traditional manual inspections and ground sensors cannot fully monitor the pipeline. The high mobility of drones can cover wider and more difficult-to-reach areas, achieving comprehensive and accurate monitoring of pipelines, significantly improving the accuracy and timeliness of monitoring.

[0019] The data preprocessing module filters, denoises and standardizes the collected data, further improving the data quality and reducing the impact of environmental noise and interference. In addition, the data fusion and analysis module fuses and extracts features from the warning data through the particle filter data fusion algorithm, which can combine data from multiple sensors to provide more accurate risk predictions. This intelligent data processing method can significantly improve the system's ability to adapt to complex environments and dynamic changes, and solve the problems of data inconsistency, false alarms and missed alarms in traditional monitoring methods.

[0020] (2) Through the standardized data processing process, including data collection, aggregation, denoising, and standardization, each link from data collection to final warning is strictly controlled. This integrated process reduces manual intervention and potential errors, and improves the reliability and stability of the entire system. In complex environments, the system can automatically adjust and optimize the data processing process to provide real-time and accurate monitoring and warning information. Through standardized processing, the output of each sensor is unified into a standard scale, which not only improves the compatibility of data, but also provides a clear data foundation for subsequent analysis and decision-making. The standardized warning data set X will provide accurate and reliable data support for subsequent risk assessment, trend analysis, etc., thereby providing decision makers with a more scientific and accurate basis for decision-making.

[0021] (3) The data from different sensors are weighted and fused through the particle filter data fusion algorithm to generate a fused data set XR. This weighted fusion method can dynamically adjust the weight of each sensor data according to the accuracy and reliability of each sensor, ensuring that the final fused data is more accurate and reliable. Compared with traditional single sensor or simple data fusion methods, this solution processes dynamically changing sensor data through the particle filter algorithm, which can effectively eliminate noise, improve the accuracy of the overall data, and provide more stable data support for subsequent risk prediction and alarm.

[0022] After data fusion, the feature extraction unit extracts key features from the fused data set XR, including the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH, and the gas wind speed covariance ΔCW, to form the risk feature set F. These features can accurately reveal early signs of risks such as fire and gas leakage.

[0023] (4) By comprehensively calculating the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH, and the gas wind speed covariance ΔCW in the risk feature set F, a more accurate risk index Risk is obtained. This risk index calculation method based on multi-dimensional data can comprehensively consider multiple factors such as fire, gas leakage, and environmental fluctuations to provide a more comprehensive risk assessment. Compared with traditional single factor analysis, comprehensive calculation can reduce the possibility of missed judgments and misjudgments, ensuring early warning and accurate judgment of abnormal environmental changes.

[0024] The intensity coefficient k1, fluctuation coefficient k2 and diffusion coefficient k3 are calculated by formulas, and can be adaptively adjusted according to the changes in infrared radiation intensity, temperature and humidity fluctuations, gas concentration and wind speed. This coefficient calculation method based on dynamic changes enables the system to adjust the contribution of each sensor data to risk assessment according to actual conditions, thereby improving the accuracy and flexibility of the prediction. Especially in complex and changing environments, it can respond quickly and provide appropriate risk assessments. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of a flow chart of a fire warning system based on a drone according to the present invention; Figure 2 The present invention is a schematic diagram of the steps of a fire warning method based on a drone. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] Example 1 The present invention provides a fire warning system based on drones, please refer to Figure 1 , including data acquisition module, data preprocessing module, data fusion and analysis module, risk prediction module, real-time warning and alarm module and data reporting and decision support module; The data acquisition module collects environmental data of the natural gas transmission pipeline through sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; The data preprocessing module performs filtering, denoising and standardization on the acquired original data set XC to obtain the early warning data set X; The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through a particle filter data fusion algorithm; The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains a risk index Risk; The real-time early warning and alarm module determines the potential risk according to the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

[0028] In this embodiment, the environmental data of the natural gas transmission pipeline is collected in real time by using a variety of sensors equipped with drones, which solves the problem that traditional manual inspections and ground sensors cannot fully monitor the pipeline. The high mobility of drones can cover wider and hard-to-reach areas, achieve comprehensive and accurate monitoring of pipelines, and significantly improve the accuracy and timeliness of monitoring.

[0029] The data preprocessing module filters, denoises and standardizes the collected data, further improving the data quality and reducing the impact of environmental noise and interference. In addition, the data fusion and analysis module fuses and extracts features from the warning data through the particle filter data fusion algorithm, which can combine data from multiple sensors to provide more accurate risk predictions. This intelligent data processing method can significantly improve the system's ability to adapt to complex environments and dynamic changes, and solve the problems of data inconsistency, false alarms and missed alarms in traditional monitoring methods.

[0030] The risk prediction module performs trend analysis based on the extracted risk feature set, predicts the occurrence of potential risks, and generates a risk index Risk. Combined with the real-time warning and alarm module, the system can identify potential threats before risks occur, trigger alarms in a timely manner, and provide key information about the location of the leak source, the area where the fire occurred, and the path of spread. This mechanism can effectively reduce the probability of accidents, initiate emergency response in a timely manner, prevent the spread of disasters, and avoid serious casualties and property losses.

[0031] The data reporting and decision support module integrates real-time monitoring data, risk feature set F and risk index Risk to generate a detailed comprehensive report. This provides decision makers with real-time, comprehensive risk assessment and early warning information, supporting rapid decision-making and emergency response.

[0032] The present invention can automatically monitor, analyze and warn, greatly reducing the need for human intervention. Drones and automated systems can quickly capture abnormal signals when danger occurs, provide key data support in a timely manner, and ensure the speed and accuracy of emergency response. This automated and intelligent processing method, especially in high-risk environments, can provide reliable decision-making support for relevant personnel, thereby improving the response speed and efficiency of the overall system.

[0033] Example 2 This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the data acquisition module includes a data acquisition unit and a data aggregation unit; The data acquisition unit collects the gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS of the natural gas transmission pipeline through sensors; The gas concentration CG includes the methane concentration in the target area of ​​the natural gas transmission pipeline and is collected by a gas concentration sensor, the temperature T is collected by a temperature sensor, the humidity H is collected by a humidity sensor, the infrared radiation intensity IR is collected by an infrared thermal imager, and the ambient wind speed WS is collected by a wind speed sensor; The data aggregation unit is responsible for combining the acquired gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS to form a unified original data set XC, and transmits it to the data preprocessing module.

[0034] The data preprocessing module includes a data denoising and filtering unit and a data standardization unit; The data denoising and filtering unit removes noise from the original data set XC by using a Kalman filter, and smoothes the original data set XC by using a filtering algorithm to obtain a denoised data set XQ; The data standardization unit performs standardization processing on the denoised data set XQ, compares and analyzes the data of different sensors at the same scale, and obtains the early warning data set X; The warning data set X is obtained by the following formula: ; Where Xd represents the d-th item of data in the warning data set X, XQd represents the d-th item of data in the denoised data set XQ, μd represents the mean of the d-th item of data, and σd represents the standard deviation of the d-th item of data.

[0035] In this embodiment, multiple sensors are used to collect environmental data of the natural gas transmission pipeline at the same time. This method of multi-sensor collaborative work can obtain comprehensive and real-time monitoring data, covering the gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS of the pipeline. Compared with the traditional single sensor or manual inspection method, it can provide higher accuracy and a wider monitoring range, especially in complex and dangerous environments, and can avoid the blind spots and monitoring lag problems existing in traditional methods.

[0036] The data aggregation unit integrates the multi-dimensional data collected by each sensor into a unified raw data set XC and transmits it to the data preprocessing module. This automated data aggregation process reduces manual intervention and data transmission errors, while achieving integrated data management. Through this integrated data collection and aggregation method, the system can ensure that the data of each sensor is processed synchronously, avoiding data inconsistency or lag problems.

[0037] The data denoising and filtering unit in the data preprocessing module uses the Kalman filtering algorithm to remove noise, and smoothes the original data set XC through the filtering algorithm to obtain the denoised data set XQ. By removing environmental noise, the system can analyze data more accurately, avoid false alarms or missed alarms caused by noise interference, and ensure the stability and credibility of the monitoring system.

[0038] The data standardization unit standardizes the denoised data set XQ and unifies the data collected by different sensors to the same scale for comparison and analysis. This standardization process can effectively solve the problem of different dimensions of the output data of each sensor, so that different types of monitoring data can be effectively compared and fused on the same platform. This is crucial for multi-sensor data fusion and complex data analysis, and can ensure that the system can more accurately assess the potential risks of fire or leakage.

[0039] This embodiment ensures that every link from data collection to final warning is strictly controlled through a standardized data processing process, including data collection, aggregation, denoising, and standardization. This integrated process reduces manual intervention and potential errors, and improves the reliability and stability of the entire system. In complex environments, the system can automatically adjust and optimize the data processing process to provide real-time and accurate monitoring and early warning information. Through standardized processing, the output of each sensor is unified to a standard scale, which not only improves the compatibility of the data, but also provides a clear data basis for subsequent analysis and decision-making. The standardized early warning data set X will provide accurate and reliable data support for subsequent risk assessment, trend analysis, etc., thereby providing decision makers with a more scientific and accurate basis for decision-making.

[0040] Example 3 This embodiment is explained in Example 2. Please refer to Figure 1 ,Specifically: the data fusion and analysis module includes a data fusion unit and a feature extraction unit; The data fusion unit performs weighted fusion on the data in the warning data set X through a particle filter data fusion algorithm to obtain a fused data set XR; The fused data set XR is obtained by the following formula: ; Where XR(t) represents the fused data set at time t, i represents the measurement value of the i-th sensor, and n represents the total number of sensors. represents the preset weight value of the i-th sensor, and , Xi(t) represents the measurement value of the i-th sensor in the warning data set X at time t.

[0041] The feature extraction unit extracts features from the fused data set XR, including the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH, and the covariance ΔCW of the gas wind speed, and forms a risk feature set F; The infrared radiation intensity change rate ΔIR is obtained by the following formula: ; In the formula, IR(t) represents the infrared radiation intensity at time t, IR(t-1) represents the infrared radiation intensity at time t-1, and Δt represents the time interval; The temperature and humidity combined fluctuation rate ΔTH is obtained by the following formula: ; Where Tj represents the temperature value of the jth data point, Hj represents the humidity value of the jth data point, m represents the total number of data points, μT and μH represent the mean of temperature and humidity, σT and σH represent the standard deviation of temperature and humidity, respectively; The covariance ΔCW of the gas wind speed is obtained by the following formula: ; Where CGj represents the gas concentration of the jth data point, WSj represents the ambient wind speed of the jth data point, μCG and μWS represent the means of the gas concentration and ambient wind speed, respectively, and σT and σH represent the standard deviations of the gas concentration and ambient wind speed, respectively.

[0042] The risk prediction module includes a trend analysis unit and a risk index calculation unit; The trend analysis unit analyzes the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F, and predicts the risks that may occur in the future; Analyze the trend of the infrared radiation intensity change rate ΔIR to identify potential fire risks. When the infrared radiation intensity change rate ΔIR shows a continuous upward trend, it indicates the risk of fire spreading. Analyze the changing trend of the temperature and humidity combined fluctuation rate ΔTH to reveal the risk of fire and gas leakage caused by abnormal temperature and humidity fluctuations; when the temperature and humidity combined fluctuation rate ΔTH shows a continuous upward trend, it indicates that there are drastic changes in the environment, which is a precursor to fire and leakage; when the temperature and humidity combined fluctuation rate ΔTH shows irregular changes around the natural gas transmission pipeline, it indicates that gas leakage has occurred; Analyze the changing trend of the covariance ΔCW of the gas wind speed to predict the diffusion trend and risk of gas leakage; when the covariance ΔCW of the gas wind speed shows a continuous upward trend, it means that the concentration of the gas leakage is changing, and the change in wind speed leads to rapid diffusion of the gas.

[0043] In this embodiment, the data from different sensors are weighted and fused through the particle filter data fusion algorithm to generate a fused data set XR. This weighted fusion method can dynamically adjust the weight of each sensor data according to the accuracy and reliability of each sensor, ensuring that the final fused data is more accurate and reliable. Compared with traditional single sensor or simple data fusion methods, this solution processes dynamically changing sensor data through a particle filter algorithm, which can effectively eliminate noise, improve the accuracy of the overall data, and provide more stable data support for subsequent risk prediction and alarm.

[0044] After data fusion, the feature extraction unit extracts key features from the fused data set XR, including the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH, and the gas wind speed covariance ΔCW, to form a risk feature set F. These features can accurately reveal early signs of risks such as fire and gas leakage. The analysis of the infrared radiation intensity change rate ΔIR can effectively capture the risk of fire spread, the change in the temperature and humidity combined volatility ΔTH can reveal the potential risk of fire or gas leakage, and the gas wind speed covariance ΔCW can help predict the diffusion trend of gas leakage. The extraction of these features enables the system to identify and analyze potential dangers in a timely manner and improve early warning capabilities.

[0045] The risk prediction module analyzes the data in the risk feature set F through the trend analysis unit to predict future risk occurrences. When the infrared radiation intensity change rate ΔIR continues to rise, the system can judge the risk of fire spread in advance; when the temperature and humidity combined volatility ΔTH shows irregular changes, the system can identify the potential risk of gas leakage. By predicting the changing trend of the gas wind speed covariance ΔCW, the system can judge the diffusion trend of gas leakage and predict changes in the leakage area. This dynamic risk prediction mechanism can greatly improve the system's response speed to emergencies and provide strong support for the early deployment of emergency response.

[0046] Through multi-dimensional data fusion and real-time analysis, it can efficiently process a large amount of environmental data and make intelligent decisions through precise algorithms. The intelligent features of this system enable drones to autonomously monitor and analyze environmental changes, capture potential risks in a timely manner, and provide decision makers with detailed risk assessment reports.

[0047] In this embodiment, the system integrates multiple sensor data and combines complex data fusion algorithms and feature extraction methods to achieve comprehensive monitoring of natural gas transmission pipelines. This comprehensive monitoring solution can not only detect abnormal conditions such as fires and gas leaks in real time, but also conduct a comprehensive assessment of potential risks, providing a more accurate decision-making basis for emergency response. Through early warning, the system can identify potential threats before accidents occur, reduce the probability of accidents and the losses caused, and improve the safety of pipeline operations.

[0048] Example 4 This embodiment is explained in Example 3, please refer to Figure 1 Specifically: the risk index calculation unit performs a comprehensive calculation on the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F to obtain the risk index Risk; The risk index Risk is obtained by the following formula: ; In the formula, k1 represents the intensity coefficient, k2 represents the fluctuation coefficient, and k3 represents the diffusion coefficient; The strength coefficient k1 is obtained by the following formula: ; In the formula, C1 represents a constant, IR(t) represents the infrared radiation intensity at time t, and IPmax represents the peak value of the infrared radiation intensity; The fluctuation coefficient k2 is obtained by the following formula: ; In the formula, C2 represents a constant, T(t) represents the temperature at time t, H(t) represents the humidity at time t, Tmax represents the peak value of temperature, and Hmax represents the peak value of humidity; The diffusion coefficient k3 is obtained by the following formula: ; In the formula, C3 represents a constant, CG(t) represents the gas concentration at time t, WS(t) represents the ambient wind speed at time t, CGmax represents the peak value of the gas concentration, and WSmax represents the peak value of the ambient wind speed.

[0049] The real-time early warning and alarm module includes a risk judgment alarm unit and an alarm information transmission unit; The risk judgment alarm unit compares the acquired risk index Risk with a preset risk threshold TRk to judge the state of the natural gas transmission pipeline; The state of the natural gas transmission pipeline is obtained by matching the following formula: When the risk index Risk < risk threshold TRk, it means that the natural gas transmission pipeline is in normal condition and no alarm is triggered; When the risk index Risk ≥ risk threshold TRk, it means that the natural gas pipeline is in an abnormal state, triggering an alarm, and starting the alarm program to send alarm information to the ground control center and relevant personnel; The alarm information transmission unit generates alarm information according to the triggered alarm, including the leakage source location, fire area and spreading path, and transmits the alarm information to relevant personnel through the UAV communication system.

[0050] The data reporting and decision support module integrates the acquired early warning data set X, risk feature set F and risk index Risk to obtain a real-time monitoring data report; generates a risk prediction result report using the risk feature set F and risk index Risk, indicating potential fire and gas leakage risks; generates an alarm event analysis report based on the status of the natural gas transmission pipeline, reporting the cause of the alarm, the location of the leakage source, the fire area and the spread path.

[0051] In this embodiment, a more accurate risk index Risk is obtained by comprehensively calculating the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH, and the gas wind speed covariance ΔCW in the risk feature set F. This risk index calculation method based on multi-dimensional data can comprehensively consider multiple factors such as fire, gas leakage, and environmental fluctuations to provide a more comprehensive risk assessment. Compared with traditional single factor analysis, comprehensive calculation can reduce the possibility of missed judgments and misjudgments, and ensure early warning and accurate judgment of abnormal environmental changes.

[0052] The intensity coefficient k1, fluctuation coefficient k2 and diffusion coefficient k3 are calculated by formulas, and can be adaptively adjusted according to the changes in infrared radiation intensity, temperature and humidity fluctuations, gas concentration and wind speed. This coefficient calculation method based on dynamic changes enables the system to adjust the contribution of each sensor data to risk assessment according to actual conditions, thereby improving the accuracy and flexibility of the prediction. Especially in complex and changing environments, it can respond quickly and provide appropriate risk assessments.

[0053] The risk judgment alarm unit automatically judges the status of the natural gas transmission pipeline and triggers the alarm procedure by comparing the risk index Risk with the preset risk threshold TRk. This automated judgment mechanism can effectively reduce the need for manual intervention and improve the speed and accuracy of system response. Through this intelligent alarm triggering mechanism, the system can respond immediately when potential risks occur, reducing the delay of accidents and ensuring the timeliness of emergency response.

[0054] When the system triggers an alarm, the alarm information transmission unit will generate alarm information including the location of the leak source, the fire area and the spread path, and transmit it to relevant personnel in a timely manner through the drone communication system. This mechanism greatly improves the accuracy and timeliness of alarm information. It is not just a simple alarm trigger, but also includes the specific location of the incident and the path of potential risk spread, thereby providing relevant personnel with sufficient decision-making information to help them make a more scientific and efficient emergency response.

[0055] The data reporting and decision support module generates real-time monitoring data reports, risk prediction result reports and alarm event analysis reports by integrating real-time monitoring data, risk feature set F and risk index Risk. This comprehensive and intelligent report generation mechanism enables decision makers to quickly obtain detailed information about the accident, including key information such as leakage source, fire spread path, predicted risks, etc. Through these reports, decision makers can respond faster, optimize emergency response strategies, and reduce losses after the accident.

[0056] Example 5 A fire warning method based on drones, please refer to Figure 2 , specifically: including the following steps: Step 1: The data acquisition module collects environmental data of the natural gas pipeline through the sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; Step 2: The data preprocessing module filters, denoises and standardizes the acquired original data set XC to obtain the warning data set X; Step 3: The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through the particle filter data fusion algorithm; Step 4: The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains the risk index Risk; Step 5: The real-time early warning and alarm module determines the potential risk based on the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; Step 6: The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

[0057] In this embodiment, the environmental data of the natural gas transmission pipeline is comprehensively collected by using a variety of sensors carried by drones to form an original data set XC. The collaborative work of these multiple sensors enables the system to obtain gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS in real time, providing richer and more accurate monitoring data. Compared with the traditional monitoring method that relies on a single sensor, it can greatly improve the perception of environmental changes and monitoring accuracy. Through the high mobility of drones, the system can cover pipeline areas that are difficult to reach with traditional monitoring methods, ensuring comprehensive monitoring of the pipeline network.

[0058] The data preprocessing module optimizes the original data set XC through filtering, denoising and standardization to generate the warning data set X. This process uses advanced filtering algorithms to remove noise, ensuring the accuracy and reliability of the data. Standardization processing unifies the data of different sensors to the same scale for analysis, avoiding the difficulty of comparison caused by different dimensions. The data fusion and analysis module performs weighted fusion on the warning data set X through the particle filter data fusion algorithm to obtain the risk feature set F.

[0059] The risk prediction module performs trend analysis based on the fused risk feature set F, predicts possible future risks, and generates a risk index Risk. By analyzing features such as the infrared radiation intensity change rate ΔIR, the temperature and humidity combined volatility ΔTH, and the gas wind speed covariance ΔCW, the system can predict the occurrence of potential risks such as fire spread and gas leakage. This data trend prediction method can identify risks in advance and take countermeasures without relying on traditional methods that can only respond to events that have already occurred, thereby providing early warnings before accidents occur and significantly improving disaster prevention capabilities.

[0060] The real-time early warning and alarm module determines the potential risk according to the risk index Risk, triggers the alarm system, and transmits the alarm information to the relevant personnel in a timely manner. Compared with traditional manual monitoring and manual alarm, the automated alarm system in this embodiment can respond quickly when potential risks are detected, ensuring the timely transmission of alarm information. Through the communication system between the drone and the ground control center, the alarm information can be transmitted to the relevant personnel in real time, and includes important risk information, such as the location of the leak source, the fire area, and the spread path, etc., to help emergency response personnel quickly locate and take action, greatly improving the efficiency and accuracy of emergency response.

[0061] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fire warning system based on drones, characterized by: It includes data acquisition module, data preprocessing module, data fusion and analysis module, risk prediction module, real-time early warning and alarm module and data reporting and decision support module; The data acquisition module collects environmental data of the natural gas transmission pipeline through sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; The data preprocessing module performs filtering, denoising and standardization on the acquired original data set XC to obtain the early warning data set X; The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through a particle filter data fusion algorithm; The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains a risk index Risk; The real-time early warning and alarm module determines the potential risk according to the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

2. The fire warning system based on drone according to claim 1, characterized in that: The data acquisition module includes a data acquisition unit and a data aggregation unit; The data acquisition unit collects the gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS of the natural gas transmission pipeline through sensors; The gas concentration CG includes the methane concentration in the target area of ​​the natural gas transmission pipeline and is collected by a gas concentration sensor, the temperature T is collected by a temperature sensor, the humidity H is collected by a humidity sensor, the infrared radiation intensity IR is collected by an infrared thermal imager, and the ambient wind speed WS is collected by a wind speed sensor; The data aggregation unit is responsible for combining the acquired gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS to form a unified original data set XC, and transmits it to the data preprocessing module.

3. The fire warning system based on drone according to claim 1, characterized in that: The data preprocessing module includes a data denoising and filtering unit and a data standardization unit; The data denoising and filtering unit removes noise from the original data set XC by using a Kalman filter, and smoothes the original data set XC by using a filtering algorithm to obtain a denoised data set XQ; The data standardization unit performs standardization processing on the denoised data set XQ, compares and analyzes the data of different sensors at the same scale, and obtains the early warning data set X; The warning data set X is obtained by the following formula: ； Where Xd represents the d-th item of data in the warning data set X, XQd represents the d-th item of data in the denoised data set XQ, μd represents the mean of the d-th item of data, and σd represents the standard deviation of the d-th item of data.

4. The fire warning system based on drone according to claim 1, characterized in that: The data fusion and analysis module includes a data fusion unit and a feature extraction unit; The data fusion unit performs weighted fusion on the data in the warning data set X through a particle filter data fusion algorithm to obtain a fused data set XR; The fused data set XR is obtained by the following formula: ； Where XR(t) represents the fused data set at time t, i represents the measurement value of the i-th sensor, and n represents the total number of sensors. represents the preset weight value of the i-th sensor, and , Xi(t) represents the measurement value of the i-th sensor in the warning data set X at time t.

5. The fire warning system based on drone according to claim 4 is characterized in that: The feature extraction unit extracts features from the fused data set XR, including the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH, and the covariance ΔCW of the gas wind speed, and forms a risk feature set F; The infrared radiation intensity change rate ΔIR is obtained by the following formula: ； In the formula, IR(t) represents the infrared radiation intensity at time t, IR(t-1) represents the infrared radiation intensity at time t-1, and Δt represents the time interval; The temperature and humidity combined fluctuation rate ΔTH is obtained by the following formula: ； Where Tj represents the temperature value of the jth data point, Hj represents the humidity value of the jth data point, m represents the total number of data points, μT and μH represent the mean of temperature and humidity, σT and σH represent the standard deviation of temperature and humidity, respectively; The covariance ΔCW of the gas wind speed is obtained by the following formula: ； Where CGj represents the gas concentration of the jth data point, WSj represents the ambient wind speed of the jth data point, μCG and μWS represent the means of the gas concentration and ambient wind speed, respectively, and σT and σH represent the standard deviations of the gas concentration and ambient wind speed, respectively.

6. The fire warning system based on drone according to claim 5, characterized in that: The risk prediction module includes a trend analysis unit and a risk index calculation unit; The trend analysis unit analyzes the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F, and predicts the risks that may occur in the future; Analyze the trend of the infrared radiation intensity change rate ΔIR to identify potential fire risks. When the infrared radiation intensity change rate ΔIR shows a continuous upward trend, it indicates the risk of fire spreading. Analyze the changing trend of the temperature and humidity combined fluctuation rate ΔTH to reveal the risk of fire and gas leakage caused by abnormal temperature and humidity fluctuations; when the temperature and humidity combined fluctuation rate ΔTH shows a continuous upward trend, it indicates that there are drastic changes in the environment, which is a precursor to fire and leakage; when the temperature and humidity combined fluctuation rate ΔTH shows irregular changes around the natural gas transmission pipeline, it indicates that gas leakage has occurred; Analyze the changing trend of the covariance ΔCW of the gas wind speed to predict the diffusion trend and risk of gas leakage; when the covariance ΔCW of the gas wind speed shows a continuous upward trend, it means that the concentration of the gas leakage is changing, and the change in wind speed leads to rapid diffusion of the gas.

7. The fire warning system based on drone according to claim 6, characterized in that: The risk index calculation unit performs a comprehensive calculation on the infrared radiation intensity change rate ΔIR, the temperature and humidity combined fluctuation rate ΔTH and the covariance ΔCW of the gas wind speed in the risk feature set F to obtain the risk index Risk; The risk index Risk is obtained by the following formula: ； In the formula, k1 represents the intensity coefficient, k2 represents the fluctuation coefficient, and k3 represents the diffusion coefficient; The strength coefficient k1 is obtained by the following formula: ； In the formula, C1 represents a constant, IR(t) represents the infrared radiation intensity at time t, and IPmax represents the peak value of the infrared radiation intensity; The fluctuation coefficient k2 is obtained by the following formula: ； In the formula, C2 represents a constant, T(t) represents the temperature at time t, H(t) represents the humidity at time t, Tmax represents the peak value of temperature, and Hmax represents the peak value of humidity; The diffusion coefficient k3 is obtained by the following formula: ； In the formula, C3 represents a constant, CG(t) represents the gas concentration at time t, WS(t) represents the ambient wind speed at time t, CGmax represents the peak value of the gas concentration, and WSmax represents the peak value of the ambient wind speed.

8. The fire warning system based on drone according to claim 7 is characterized in that: The real-time early warning and alarm module includes a risk judgment alarm unit and an alarm information transmission unit; The risk judgment alarm unit compares the acquired risk index Risk with a preset risk threshold TRk to judge the state of the natural gas transmission pipeline; The state of the natural gas transmission pipeline is obtained by matching the following formula: When the risk index Risk < risk threshold TRk, it means that the natural gas transmission pipeline is in normal condition and no alarm is triggered; When the risk index Risk ≥ risk threshold TRk, it means that the natural gas pipeline is in an abnormal state, triggering an alarm, and starting the alarm program to send alarm information to the ground control center and relevant personnel; The alarm information transmission unit generates alarm information according to the triggered alarm, including the leakage source location, fire area and spreading path, and transmits the alarm information to relevant personnel through the UAV communication system.

9. The fire warning system based on drone according to claim 8, characterized in that: The data reporting and decision support module integrates the acquired early warning data set X, risk feature set F and risk index Risk to obtain a real-time monitoring data report; Using the risk feature set F and the risk index Risk, a risk prediction result report is generated to point out the potential risks of fire and gas leakage; an alarm event analysis report is generated based on the status of the natural gas transmission pipeline, reporting the cause of the alarm, the location of the leakage source, the fire area and the spread path.

10. A fire warning method based on a drone, applied to a fire warning system based on a drone as claimed in any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: The data acquisition module collects environmental data of the natural gas pipeline through the sensors carried by the drone, including gas concentration CG, temperature T, humidity H, infrared radiation intensity IR and ambient wind speed WS, and fits them into the original data set XC; Step 2: The data preprocessing module filters, denoises and standardizes the acquired original data set XC to obtain the warning data set X; Step 3: The data fusion and analysis module fuses and extracts features from the warning data set X, and obtains the risk feature set F through the particle filter data fusion algorithm; Step 4: The risk prediction module performs trend analysis on the risk feature set F, predicts future risks, and obtains the risk index Risk; Step 5: The real-time early warning and alarm module determines the potential risk based on the acquired risk index Risk. When the risk is determined, the alarm system will be triggered and the alarm information will be transmitted to relevant personnel through the communication system between the drone and the ground control center; Step 6: The data reporting and decision support module integrates the acquired warning data set X, risk feature set F and risk index Risk to generate a comprehensive report.

Images