Methods, devices, storage media and electronic equipment for fire risk assessment of power transmission lines

Abstract

This application discloses a method, apparatus, storage medium, and electronic device for assessing the fire risk of transmission lines. The method includes: acquiring multi-source data of a target area; extracting the values ​​of multiple fire risk assessment indicators for transmission lines based on the multi-source data; determining the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation and values ​​between the indicators within the same time slice; determining the temporal impact of the indicator values ​​from the previous time slice on the fire risk assessment of the current time slice based on the temporal dependency between the indicators in adjacent time slices; and determining the fire risk level of the transmission lines in the target area based on the immediate impact and the temporal impact. This application determines the temporal impact through the temporal dependency of indicators in adjacent time slices, tracks the temporal transmission of risk, compensates for the shortcomings of traditional dynamic analysis, and achieves spatiotemporal dynamic assessment by combining immediate impact, providing accurate basis for early warning and decision-making regarding fire risks of transmission lines.

Description

Technical Field

[0001] This application relates to the field of transmission line fire risk assessment technology, and in particular to a method, apparatus, storage medium and electronic equipment for transmission line fire risk assessment. Background Technology

[0002] Due to their long transmission distances and complex terrain crossings, power transmission lines have long faced threats from various natural and human factors, among which forest fires are one of the most prominent risks. Mountainous areas have dense vegetation cover and varied terrain, making them prone to forest fires. When a fire spreads to the vicinity of transmission lines, it can significantly reduce the insulation performance of the lines, and even cause short-circuit faults or tripping, seriously endangering the stable operation of the power grid.

[0003] Currently, risk assessment for transmission line fires typically employs partial least squares (PLS) techniques to construct static linear models. These models extract features from historical data through dimensionality reduction and establish multiple regression equations for risk level assessment. Existing technologies simply assume a linear correlation between assessment indicators and risk levels. However, in real-world scenarios, factors such as terrain slope and vegetation moisture content often interact in a synergistic manner. This linear assumption fundamentally differs from the actual environment. For example, high temperatures and low humidity can synergistically exacerbate vegetation dryness, while steep terrain accelerates fire spread. The combined effects of these multiple factors are not simply a linear superposition. Because static linear models cannot effectively represent such complex nonlinear interaction mechanisms, they struggle to accurately reflect the combined effects of multiple factors in real-world scenarios. Ultimately, this results in insufficient alignment between assessment results and actual on-site conditions, failing to provide precise decision support for transmission line fire prevention and control. Summary of the Invention

[0004] In view of the above problems, this application provides a method, apparatus, storage medium and electronic equipment for assessing fire risk of power transmission lines.

[0005] To solve the above-mentioned technical problems, this application proposes the following solution:

[0006] Firstly, this application provides a method for assessing the fire risk of transmission lines. The method includes: acquiring multi-source data of a target area, wherein the target area is a mountainous area with transmission lines; extracting the values ​​of multiple fire risk assessment indicators for transmission lines based on the multi-source data; determining the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation between the assessment indicators and their values ​​within the same time slice; determining the temporal impact of the indicator values ​​of the previous time slice on the fire risk assessment of transmission lines in the current time slice based on the temporal dependency between the assessment indicators in adjacent time slices; determining the fire risk level of transmission lines in the target area based on the immediate impact and temporal impact, and describing the correlation between the fire risk assessment indicators for transmission lines within the same time slice, wherein a transfer network is used to indicate the temporal dependency between the fire risk assessment indicators for transmission lines in adjacent time slices; and inputting the fire risk assessment indicators for transmission lines into a fire risk assessment model for transmission lines to determine the fire risk level of transmission lines in the target area.

[0007] Secondly, this application provides a fire risk assessment device for power transmission lines, which includes:

[0008] The acquisition module is used to acquire multi-source data for the target area, which is a mountainous area with power transmission lines.

[0009] The extraction module is used to extract the index values ​​of multiple transmission line fire risk assessment indicators based on multi-source data;

[0010] The assessment module is used to determine the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation and indicator values ​​among the assessment indicators within the same time slice, to determine the temporal impact of the indicator values ​​of the previous time slice on the fire risk assessment of transmission lines in the current time slice based on the temporal dependency between the assessment indicators of adjacent time slices, and to determine the fire risk level of transmission lines in the target area based on the immediate impact and temporal impact.

[0011] To achieve the above objectives, according to a third aspect of this application, a storage medium is provided, the storage medium including a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute the power transmission line fire risk assessment method of the first aspect.

[0012] To achieve the above objectives, according to a fourth aspect of this application, an electronic device is provided, the device including at least one processor, and at least one memory and bus connected to the processor; wherein the processor and memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the power transmission line fire risk assessment method of the first aspect described above.

[0013] By employing the above-described technical solution, the technical solution provided in this application has at least the following advantages:

[0014] This application acquires multi-source data from the target area, comprehensively covering diverse information about the surrounding environment of power transmission lines in mountainous areas. This avoids assessment bias caused by single data sources and provides multi-dimensional data support for risk analysis. Based on multi-source data, assessment index values ​​are extracted, transforming complex environmental factors into quantifiable parameters, achieving a leap from qualitative judgment to quantitative analysis and improving the objectivity of the assessment. The immediate impact is determined by the correlation and values ​​of indicators within the same time slice, capturing the current impact of multiple factors synergistically on fire risk and overcoming the limitations of traditional static models in representing nonlinear coupling effects. The temporal impact is determined by the temporal dependence of indicators in adjacent time slices, tracking the transmission patterns of risk factors over time and compensating for the lack of dynamic evolution analysis in traditional methods. Finally, the risk level is determined by combining the immediate and temporal impact levels, achieving a spatiotemporal dynamic assessment of fire risk for power transmission lines in mountainous areas. This ensures that the assessment results reflect both the current risk status and predict development trends, providing accurate basis for risk warning and operation and maintenance decisions, and effectively improving the scientific rigor, timeliness, and environmental adaptability of the assessment.

[0015] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0017] Figure 1 A flowchart illustrating a method for assessing fire risk in power transmission lines according to an embodiment of this application is shown.

[0018] Figure 2 This paper shows a schematic diagram of the structure of a power transmission line fire risk assessment device provided in an embodiment of this application;

[0019] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0020] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0021] In the embodiments of this application, the terms "first," "second," etc., do not have a logical or temporal dependency, nor do they limit the quantity or execution order. It should also be understood that although the following description uses the terms "first," "second," etc., to describe various elements, these elements should not be limited by the terms. These terms are merely used to distinguish one element from another.

[0022] In this application, the term "at least one" means one or more, and the term "multiple" means two or more.

[0023] It should also be understood that the term “if” can be interpreted as “when” or “upon”, or “in response to determination” or “in response to detection”. Similarly, depending on the context, the phrase “if determination…” or “if detection [the stated condition or event]” can be interpreted as “when determination…” or “in response to determination…” or “when detection [the stated condition or event]” or “in response to detection [the stated condition or event]”.

[0024] Due to their long transmission distances and complex terrain crossings, power transmission lines have long faced threats from various natural and human factors, among which forest fires are one of the most prominent risks. Mountainous areas have dense vegetation cover and varied terrain, making them prone to forest fires. When a fire spreads to the vicinity of transmission lines, it can significantly reduce the insulation performance of the lines, and even cause short-circuit faults or tripping, seriously endangering the stable operation of the power grid.

[0025] Current transmission line fire risk assessment employs a static linear model based on partial least squares (PLS) techniques. This model extracts key feature parameters from historical data through dimensionality reduction and constructs a multiple regression equation based on linear assumptions to assess risk levels. However, existing technologies simply assume a linear relationship between assessment indicators and risk levels, neglecting the fact that in actual transmission line fire risk scenarios, multiple factors such as terrain slope, vegetation moisture content, and meteorological factors often interact in a nonlinear and network-like manner. For example, high temperatures and low humidity can synergistically exacerbate vegetation dryness, while steep terrain can accelerate fire spread. The combined effects of these multiple factors are not simply a linear superposition. Because the model cannot effectively represent such complex nonlinear interaction mechanisms, it struggles to accurately reflect the combined effects of multiple factors in actual risk scenarios, ultimately leading to insufficient matching between the assessment results and the actual on-site risk conditions, thus failing to provide accurate decision support for transmission line fire prevention and control.

[0026] Based on this, the method for assessing fire risk in power transmission lines is explained in detail with reference to the attached diagram. Figure 1 This application provides a flowchart illustrating a method for assessing fire risks in power transmission lines. The method specifically includes the following steps:

[0027] Step 110: Obtain multi-source data for the target area.

[0028] First, remote sensing image data of the target area is collected using satellite remote sensing platforms, drone aerial photography, and other means. In this embodiment, the target area is a mountainous area with power transmission lines. Ground-based lidar and airborne LiDAR equipment are used to acquire laser point cloud data. Meteorological data is extracted from meteorological monitoring stations, sensor networks, or meteorological databases. Record data is also collected through field surveys and document reviews.

[0029] For remote sensing image data, radiometric correction is performed to eliminate the influence of factors such as atmospheric scattering and sensor response errors, and to restore the true reflective radiation information of ground objects. At the same time, image enhancement processing, such as contrast stretching and filtering sharpening, is used to improve the recognizability of ground object features in the image.

[0030] During the preprocessing of laser point cloud data, noise points and outliers are first removed by algorithms such as statistical filtering and radius filtering. Then, the iterative nearest point algorithm or feature-based registration method is used to unify the point cloud data acquired from different stations and at different times into the same coordinate system. Finally, high-density point clouds are downsampled by grid filtering and simplification algorithms to reduce the amount of data while preserving the geometric features of ground features.

[0031] In the meteorological data preprocessing stage, data cleaning is first performed through outlier detection and missing value imputation to eliminate the impact of observation errors and missing data on subsequent analysis. Then, standardization or normalization methods are used to convert data with different dimensions and value ranges into a unified format. At the same time, time series data are processed by interpolation smoothing and periodic analysis to ensure the temporal continuity and consistency of the data, laying the foundation for subsequent fusion and analysis of multi-source data.

[0032] Step 120: Extract the index values ​​of multiple transmission line fire risk assessment indicators based on multi-source data.

[0033] After completing the multi-source data preprocessing, the fire assessment indicators for the first transmission line were extracted based on the preprocessed data, specifically covering the extraction of basic environmental factors such as topography, combustible characteristics, combustible moisture content, and meteorological factors.

[0034] For topographic indicators, a Digital Elevation Model (DEM) is constructed based on preprocessed laser point cloud data. Elevation, slope, and aspect parameters are obtained through spatial analysis of the DEM data. The slope is determined by calculating the spatial rate of change of elevation data in the east-west (X) and north-south (Y) directions, using the formula: Slope = The slope aspect is determined by the arctangent function, with the formula: Slope Aspect = In the above calculation, Z represents the elevation value, and X and Y are the spatial coordinates of the east-west and north-south directions, respectively. This calculation can accurately characterize the slope and orientation of the terrain.

[0035] In terms of combustible material characteristic extraction, based on preprocessed satellite remote sensing image data, the differences in spectral reflectance characteristics of different ground features in visible light, near-infrared and other bands are utilized to accurately classify ground features using the support vector machine classification method. This method constructs an optimal classification hyperplane to map spectral feature vectors to a high-dimensional space to identify combustible material types. Combined with the data from field surveys, information such as vegetation types and distribution range can be further determined, providing a basis for combustible material load assessment.

[0036] For the retrieval of combustible material moisture content, the spectral information of the short-wave infrared band (such as around 1.6μm and 2.2μm) in satellite remote sensing imagery is also used. This band is highly sensitive to vegetation moisture content, and its reflectance decreases regularly with increasing moisture content. The retrieval is achieved by constructing the vegetation moisture content index (NDMI), and the specific formula is as follows: The index : where NIR is the reflectance in the near-infrared band and SWIR is the reflectance in the short-wave infrared band. This index can effectively reflect the water content in vegetation by quantifying the difference in reflectance between the near-infrared and short-wave infrared bands.

[0037] Meteorological factor indicators rely on preprocessed meteorological data to extract conventional elements such as average temperature, relative humidity, wind speed, wind direction, and total precipitation within a preset time period. At the same time, they combine the time series characteristics of historical meteorological data to analyze the driving effects of temperature and humidity change trends and wind conditions on the occurrence of transmission line fires. The extraction process of the above indicators is based on the accuracy of the preprocessed data. The scientific nature of the assessment indicators is ensured through collaborative analysis of multi-source data, providing core parameter support for the construction of the transmission line fire risk assessment model.

[0038] After completing the analysis of basic environmental factors such as topography, combustible material characteristics, and meteorological factors involved in the first transmission line fire assessment index, the focus is further on the impact of the transmission line's own discharge risk on the causes of transmission line fires. The second transmission line fire assessment index quantifies the risk of gap discharge in transmission lines, establishing a key correlation between the condition of line facilities and the potential for fire hazards. This index uses gap distance extracted from laser point cloud data, insulator usage time weighted from ledger data, and humidity influence coefficient converted from meteorological data as core parameters. Through multi-factor coupling analysis, it achieves a dynamic assessment of the line insulation condition, providing key inputs reflecting the potential discharge hazards in transmission lines for the transmission line fire risk model.

[0039] In this embodiment, the gap distance is a key parameter affecting the discharge voltage of transmission lines, and its quantitative assessment is of great significance for predicting the fire risk of transmission lines. Generally, the larger the gap distance, the higher the breakdown voltage and the stronger the line insulation safety. Specifically, the gap distance of the target area is obtained through the following technical solution: First, the laser point cloud data is processed using a semantic segmentation algorithm (such as support vector machine). By constructing vectors containing features such as three-dimensional coordinates, reflection intensity, and neighborhood curvature, the point cloud is accurately classified into transmission line conductor point cloud clusters and obstacle point cloud clusters (such as vegetation point clouds and tower point clouds). Then, based on the three-dimensional spatial positional relationship between the two types of point cloud clusters, the spatial region between the conductor point cloud cluster and the obstacle point cloud cluster is identified as the gap candidate region. This region covers the potential gap range that may cause discharge risk. Subsequently, the point cloud of the gap candidate area was clustered using the Euclidean distance clustering algorithm. By setting a spatial distance threshold, the point cloud was divided into different clusters. Combined with the safety distance standards in the transmission line design specifications (e.g., the safety distance between a 110kV line conductor and a tree is 4 meters), gap areas with actual distances less than the safety threshold were selected, thus clarifying the target area with potential safety hazards. For the selected gap areas, the transmission line conductor point cloud clusters were mapped from three-dimensional space to a two-dimensional plane using a projection transformation algorithm. The centerline of the power line was extracted using the least squares method, and this centerline was obtained using parametric equations. Characterizing the spatial orientation of the conductor, where For online purposes, Let be the direction vector. Finally, for each point in the obstacle point cloud cluster, calculate its perpendicular distance to the centerline of the power line using the three-dimensional spatial distance formula. After removing outliers exceeding the safety threshold, the arithmetic mean of the effective distance values ​​is taken. The final gap distance is obtained, and this value can be directly used to assess the insulation status of the line and the fire risk level of the transmission line.

[0040] As a key insulating component of transmission lines, insulators' insulation performance gradually degrades over time even without physical damage. During long-term operation, dust, salt spray, and other pollutants in the air gradually accumulate on the insulator surface, forming a contamination layer. This contamination layer, under humid conditions, reduces the insulation resistance of the insulator surface, leading to increased leakage current and consequently affecting the discharge characteristics of the transmission line gap (such as reducing breakdown voltage and increasing the probability of partial discharge). The specific process for obtaining the insulator service time is as follows: First, the installation date of the transmission line insulator is extracted from the ledger data, and the actual service time of the insulator is calculated by combining it with the current date. This time parameter directly reflects the insulator's operating years and the duration of pollution accumulation. Then, the actual service time of the insulator is normalized and mapped to a preset time weight range (such as 0-1) to obtain the relative weight value. Further, the correlation analysis between the insulator aging degree and gap discharge fault in historical fault data (such as the conditional probability distribution obtained through Bayesian network training) is combined to correct the normalized weight value, resulting in the final insulator service time weight used for risk assessment. This weight can quantitatively characterize the impact of insulator pollution accumulation on gap discharge risk and participate in the comprehensive assessment of the transmission line fire risk level in conjunction with the gap distance parameter.

[0041] The dielectric strength of air is significantly negatively correlated with humidity. In high-humidity environments, water molecules are uniformly distributed in the air in a polar molecular form. When an external electric field is applied, water molecules easily polarize and form dipole orientations, lowering the ionization energy barrier of gas molecules and thus altering the breakdown characteristics of gas gaps. Specifically, as relative humidity increases, the air breakdown voltage exhibits a nonlinear decreasing trend. For example, under standard atmospheric pressure, when the relative humidity increases from 20% to 80%, the breakdown voltage in a uniform electric field may decrease by 10%-15%. To quantify this effect, relative humidity parameters for a preset time period are extracted from preprocessed meteorological monitoring data (such as hourly observation data from regional meteorological stations or numerical weather prediction model outputs). Specifically, this includes: acquiring measured humidity sensor values ​​from meteorological stations in the target area; spatially interpolating multi-station data to generate a humidity distribution grid; analyzing the degree of anomaly in the current humidity value by combining the time series characteristics of historical humidity data for the same period; and further converting the measured humidity value into an influence coefficient on the dielectric properties of air using an empirical model of humidity breakdown voltage (such as the humidity correction factor recommended in the IEEE standard). This coefficient can be directly coupled to the gap discharge risk assessment model and used in conjunction with parameters such as gap distance and insulator aging to calculate the probability of discharge induced by transmission line fires.

[0042] After completing the quantitative analysis of individual factors such as gap distance, insulator service time weight, and humidity influence coefficient, a comprehensive fire impact index for gap discharge transmission lines (the second transmission line fire assessment index) is constructed through polynomial fitting to achieve a quantitative assessment of the synergistic effect of multiple factors. Specifically, the gap distance D, insulator service time weight W, and humidity influence coefficient H are used as input variables, and a linear polynomial fitting model is used to construct the second transmission line fire assessment index. ,in These are the weighting coefficients for the corresponding parameters. The weighting process is as follows: First, measured data of each parameter and corresponding discharge fault records from historical transmission line fire accidents are collected to form a training dataset containing gap distance, insulator service time, humidity value, and accident level. Then, the least squares method is used to optimize the polynomial weights, with the objective function being to minimize the fitting error between the discharge risk probability in historical data and the second transmission line fire assessment index. Simultaneously, a regularization term is introduced to avoid overfitting and improve the model's generalization ability. The final second transmission line fire assessment index can quantitatively characterize the potential risk of transmission line fires induced by gap discharge. A larger value for the second transmission line fire assessment index indicates poorer gap insulation performance and a higher discharge risk. This index, through the weighted coupling of multiple parameters, transforms physical quantities (distance), state quantities (weights), and environmental quantities (humidity coefficient) into unified risk measurement parameters, providing core input variables for the transmission line fire risk assessment model and achieving a technological leap from single-factor analysis to multi-factor collaborative assessment.

[0043] After quantifying the risk of transmission line gap discharge involved in the second transmission line fire assessment index, this study further explores the potential impact of the spatial relationship between vegetation and transmission lines, as well as the characteristics of vegetation itself, on the induction of transmission line fires. The third transmission line fire assessment index establishes a key correlation between natural vegetation factors and line fire risk by evaluating the disaster-causing potential of vegetation and transmission lines. This index uses parameters such as the distance from vegetation to the line calculated from laser point cloud data, the flammability level of vegetation analyzed from remote sensing imagery, and the vegetation normalized index as its core. Through multi-dimensional comprehensive analysis, it achieves a dynamic assessment of the disaster-causing risk of vegetation, providing key inputs reflecting the interaction between vegetation and the line for the transmission line fire risk model.

[0044] First, using preprocessed laser point cloud data, a deep learning model is used to perform semantic segmentation on the 3D point cloud. This model captures local and global features of the point cloud to achieve accurate classification of targets such as power lines, vegetation, and towers. Based on this, a projection-based fitting algorithm is designed for the power line point cloud. After projecting the point cloud data onto a 2D plane, the centerline of the power line is extracted using the least squares method. Vegetation areas are extracted using height threshold segmentation (setting a tree canopy height threshold). Subsequently, the vertical distance between the power line centerline and the highest point of the vegetation (e.g., the maximum elevation of the point cloud at the top of the tree canopy), and the horizontal distance between the power line centerline and the vegetation outline (extracted using a convex hull algorithm or edge detection to extract tree trunk and canopy edge points) are calculated. By averaging the vertical and horizontal distances, a comprehensive distance parameter between the vegetation and the power transmission line is obtained.

[0045] For the classification of vegetation species and assessment of flammability near power transmission lines, a support vector machine (SVM) classifier is used: First, feature vectors are constructed based on the spectral features of preprocessed satellite remote sensing images (such as reflectance in red, near-infrared, and short-wave infrared bands) and the three-dimensional structural features of laser point clouds (such as vegetation height and canopy width). The SVM classifier is then trained using labeled samples to achieve accurate classification of vegetation species in the target area. Simultaneously, by combining parameters such as ignition point and heat release rate of different vegetation species in the forest ledger data, a mapping relationship between vegetation species and flammability is established to obtain the disaster risk level of each vegetation type.

[0046] In addition, the Normalized Difference Vegetation Index (NDVI) is introduced as an indicator to assess vegetation density. This index is calculated by dividing the difference between the reflectance of the near-infrared (NIR) band and the sum of the reflectance of the red band by the following formula: The higher the value, the higher the vegetation coverage and the richer the biomass.

[0047] Finally, a fire assessment index for the third transmission line was constructed using a polynomial fitting method. ,in, The weights for different parameters are determined through correlation analysis between historical transmission line fire accident data and vegetation parameters. The weights can be optimized using the least squares method, enabling the third transmission line fire assessment index to accurately quantify the disaster-causing potential of vegetation and transmission lines, providing core parameter support for transmission line fire risk assessment. The specific implementation method of the weights for the third transmission line fire assessment index is the same as that for the second transmission line fire assessment index, and will not be elaborated here.

[0048] Step 130: Determine the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation and indicator values ​​among the assessment indicators within the same time frame.

[0049] First, for the fire risk assessment index values ​​and historical index values ​​of transmission lines obtained within the same time window, a method combining Granger causality test and information entropy is used to identify the nonlinear correlations between the indicators, and a directed acyclic graph is constructed to represent the index correlation network. For example, an increase in the frequency of lightning strikes will significantly increase the probability of insulator flashover, while the combination of vegetation flammability level and terrain slope will accelerate the spread of fire to the transmission line.

[0050] Secondly, after determining the correlation structure of the indicators, a Bayesian network conditional probability inference algorithm is applied to calculate the immediate impact weight of each assessment indicator on fire risk, combined with real-time indicator values. Specifically, the preprocessed indicator values ​​are input into a trained Bayesian network model, and the posterior probability of each node is updated through a joint probability distribution to extract the contribution of each indicator to the fire risk node in the current state. For example, when the horizontal distance between vegetation and power transmission lines is less than the safety threshold and the vegetation flammability level is highly flammable, the impact weight of these two indicators on fire risk will increase significantly.

[0051] Furthermore, to address the dynamic changes in mountainous environments, a time-varying parameter Kalman filter algorithm is introduced to correct the impact level in real time. This algorithm dynamically adjusts the time decay factor of indicator weights by fusing historical data with current observations, enabling it to adaptively capture changes in risk caused by sudden weather conditions or equipment malfunctions. For example, during thunderstorms, the impact weight of lightning activity indicators will be dynamically increased, while during the vegetation growing season, the weight of vegetation coverage indicators will be adjusted accordingly.

[0052] Step 140: Determine the temporal impact of the indicator values ​​of the previous time slice on the current time slice's transmission line fire risk assessment based on the temporal dependencies between the assessment indicators of adjacent time slices.

[0053] To analyze the temporal dependency of evaluation metrics for adjacent time slices, a data-driven quantitative model of temporal impact needs to be constructed. The specific implementation process is as follows:

[0054] First, according to Calculate the time slice parameters, where, Let be the set of indicators for assessing the fire risk of transmission lines at time s. For the i-th indicator in the indicator set at time s+1, To Preceding indicators with direct causal influence. For example, when calculating the risk of discharge during the gap at time s+1, the parent nodes are the gap distance at time s and the humidity at time s. Conditional probability parameters can be obtained through historical data statistics. For example, the probability of "high" discharge risk when the gap distance is less than 5m and the humidity is less than 30% is 0.9.

[0055] Next, dynamic node selection is performed. The dispersion of each node over the time series is calculated, which can be quantified using the coefficient of variation, as shown in the formula: ,in, As an indicator standard deviation As an indicator The mean. For example, when the standard deviation of the "wind speed" indicator over the past 30 days is 2.5 m / s and the mean is 5 m / s, the CV = 0.5. Then, the random forest algorithm is used to calculate the indicator's importance and assess its risk impact, using the formula: ,in, Let AUC be the area under the curve of the model when the i-th metric is included in the t-th experiment. To determine the AUC of the model after removing the i-th indicator in the t-th experiment, for example, the FI value of the "humidity" indicator is 0.32. A first threshold is set. =0.2, second threshold =0.2, when the index satisfies CV≥ And FI≥ The nodes are determined to be dynamic. For example, the "wind speed" index with CV=0.5 and FI=0.25 is determined to be a dynamic node, while the "terrain slope" index with CV=0.05 and FI=0.1 is determined to be a non-dynamic node.

[0056] For dynamic nodes, the state transition relationships are explicitly modeled using time slice parameters. For example, the transition probabilities of adjacent time slices for the "wind speed" node are defined using a conditional probability table. The states of non-dynamic nodes remain unchanged between time slices and are only used as static inputs. Finally, the temporal transition relationships of dynamic nodes are integrated with the static associations of non-dynamic nodes to form a transition network composed of multi-layer Bayesian networks, with each layer corresponding to a time slice and the edges between layers defined by time slice parameters.

[0057] In practical applications, the degree of dispersion can be calculated using parameters such as standard deviation and range. The first threshold can be dynamically adjusted according to the season. Risk impact assessment can be performed using methods such as SHAP value and causal effect strength. Time slice parameters can also be dynamically fitted through deep learning models. This transfer network can accurately model the time-series changes of dynamic indicators, reduce the computational complexity of non-dynamic indicators, and effectively improve the accuracy of risk prediction and optimize computational efficiency by using data-driven dynamic nodes to adaptively select transmission line fire risk characteristics for different scenarios.

[0058] Step 150: Determine the fire risk level of the transmission lines in the target area based on the degree of immediate impact and the degree of temporal impact.

[0059] In the actual operation of the transmission line fire risk assessment model, when abnormal fluctuations are detected in the temporal characteristics of the transmission line fire risk assessment indicators (such as the coefficient of variation of indicators such as wind speed and humidity exceeding the historical threshold), or when the geographical environment of the target area (such as significant changes in topography due to earthquakes or engineering construction) or the facility status (such as sudden changes in transmission line gap distance or accelerated aging of insulators) undergo substantial changes, the original network structure may not be able to accurately reflect the current risk transmission logic. At this time, it is necessary to trigger the adjustment of the network structure describing the correlation and temporal dependency.

[0060] Specifically, the network structure is adjusted at the node level based on the type of triggering factor, including but not limited to adding, deleting, or merging nodes in the network structure. For example, if the "lightning strike frequency" indicator needs to be included due to the addition of new meteorological monitoring equipment, a corresponding node is added to the network structure, and a directed edge is established between it and the "probability of transmission line fires" node. If the feature importance of the "historical transmission line fire records" node is less than 0.1, the node is deleted to reduce model complexity. If the Pearson correlation coefficient between the "vegetation moisture content" and "surface humidity" nodes is greater than 0.8, they are merged into a "combustible material humidity" node.

[0061] After node adjustments are completed, a structural causal model (SCM) is used to assess the strength of causal effects among indicators. First, a causal graph structure of the indicators is constructed based on historical observation data to clarify the directed dependencies between indicators (such as meteorological factors, vegetation characteristics, and topographic parameters). For example, causal paths are determined such as "rising temperature" leading to "decreased vegetation moisture content" and "lightning strike" leading to "transmission line fire." Next, intervention operations are used to simulate scenarios of "removing the influence of a certain indicator" or "forcibly changing the state of a certain indicator." Taking "wind speed" and "fire spread rate" indicators in transmission line fire risk assessment as examples, the SCM is intervened in through do-calculus to calculate the distribution changes of fire spread rate under wind speed. The conditional probability differences of fire spread rate before and after intervention are compared to quantify the strength of the causal effect of "wind speed" on "fire spread rate." Simultaneously, by combining counterfactual reasoning analysis with the changes in causal relationships between indicators under different hypothetical conditions (such as "how would the fire spread range change if the wind speed increased by 2 m / s at a certain time in the past"), and integrating the results of multiple intervention experiments and counterfactual analysis, a causal effect strength matrix is ​​constructed. The elements in the matrix correspond to the quantitative values ​​of the causal effects between two indicators. For example, the causal effect value of "sustained high temperature" on "vegetation flammability" is calculated to be 0.65 through intervention experiments. Based on this, the conditional probability parameter is adjusted: the original P(flammability = high | high temperature = yes) = 0.7 is corrected to 0.85, and the node connection weights are updated simultaneously to generate candidate network structures.

[0062] Finally, according to The optimal network structure is determined from the candidate network structures, where... To calculate the log-likelihood of transmission line fire index data D based on the conditional probability table of candidate network structures, where N is the number of training samples for the transmission line fire risk assessment model and n is the number of conditional probability parameters, the network structure with the largest BIC value is selected as the optimal solution. For example, if candidate network structure A has a BIC of -235.6 and candidate network structure B has a BIC of -212.3, then candidate network structure B is determined as the final model.

[0063] After completing the construction of the dynamic adjustment mechanism of the network structure, it is necessary to further determine the initial conditional probability parameters of the transmission line fire risk assessment model through data statistical methods to support the probabilistic reasoning of the correlation between indicators. Specifically, firstly, a frequency table of transmission line fire risk assessment indicators is constructed based on historical observation data. This table provides a data foundation for probabilistic modeling by statistically analyzing the frequency of occurrence of each indicator under different states. For example, meteorological, vegetation, and line indicator data of the target area over the past three years are collected. The "humidity" indicator is divided into three states: "low," "medium," and "high" according to the numerical range. Statistically, in 1095 days of observation, the number of days with "humidity = low" is 320 days, so its marginal frequency is 320 / 1095≈0.292. For indicators with parent nodes, such as "vegetation moisture content" whose parent node is "humidity," the joint frequency needs to be calculated. For example, the number of days with "humidity = low and moisture content = low" is 280 days, corresponding to a joint frequency of 280 / 1095≈0.256.

[0064] Based on the frequency table, an initial conditional probability table is constructed using the maximum likelihood estimation method, approximating the frequencies as conditional probability parameters. This is for the child node Y and its set of parent nodes. conditional probability The conditional probability is determined by the ratio of the corresponding joint frequency to the edge frequency of the parent node. Taking "vegetation moisture content" as an example, P(moisture content = low | humidity = low) = 0.2256 / 0.292, which is stored in the conditional probability table as the initial conditional probability parameter. For indicators with multiple parent nodes, such as "gap discharge risk" whose parent nodes are "gap distance" and "humidity", it is necessary to count the three-dimensional joint frequency (e.g., the number of samples of "gap distance < 5m and humidity = low and discharge risk = high"), and then calculate the conditional probability through the frequency ratio, such as P(discharge risk = high | gap distance, humidity = low) = joint probability / parent node joint edge frequency, thereby constructing a complete conditional probability table and providing quantitative correlation parameters for the initial inference of the model.

[0065] The process of constructing the frequency table and the initial conditional probability table enables the model to establish probabilistic relationships between indicators based on historical data characteristics. This lays the foundation for subsequent parameter updates through sliding windows and adaptation to dynamic data changes, ensuring that the transmission line fire risk assessment model has causal reasoning capabilities that conform to historical patterns from the initialization stage.

[0066] After determining the initial conditional probability parameters, considering the rapidly changing and complex non-steady-state characteristics of transmission line fire risk data, a dynamic parameter update mechanism is needed to ensure the model continuously adapts to real-time risk characteristics. Specifically, a sliding window method is used to iteratively optimize the conditional probability parameters: First, a suitable window size (e.g., 30 time steps) is selected. At each time step, new observation data for transmission line fire risk assessment indicators (e.g., the latest meteorological monitoring values, line inspection data) are acquired and added to the sliding window. Simultaneously, the earliest data of the same amount within the window is removed to maintain a constant amount of data within the window. For example, the window initially stores data from day 1 to day 30. After acquiring new data on day 31, the data from day 1 is removed, forming a window dataset from day 2 to day 31.

[0067] Using the latest observational data within the window, the frequency table is reconstructed and conditional probability parameters are calculated to achieve dynamic updates of model parameters. Taking the relationship between "humidity" and "vegetation moisture content" as an example, if the new data within the window shows that the proportion of samples with "humidity = low" and "moisture content = low" increases from 85% to 90% compared to the historical level, then P(moisture content = low | humidity = low) is updated from 0.85 to 0.90 through maximum likelihood estimation. For indicators with multiple parent nodes, such as "gap discharge risk," the conditional probability table is recalculated based on the joint frequency of parent and child nodes such as "gap distance" and "humidity" within the window. For example, when the proportion of samples with "gap distance < 5m and humidity = low" within the window decreases from 90% to 80%, P(discharge risk = high | gap distance, humidity = low) is adjusted accordingly to 0.80.

[0068] This sliding window parameter update mechanism continuously incorporates the latest observation data and removes outdated information, enabling the conditional probability parameters to reflect the strength of the correlation between indicators under the current environment in real time. This effectively addresses the dynamic changes in transmission line fire risk factors (such as seasonal weather pattern changes and vegetation growth cycle replacements), ensuring that the model maintains its assessment accuracy during non-steady-state processes and providing reliable parameter support for the real-time prediction and prevention of transmission line fire risks.

[0069] The following example uses a hilly area traversed by a 110kV transmission line. First, point cloud data was collected using an airborne LiDAR. After statistical filtering to remove outliers, a 5m×5m resolution DEM was constructed, and the average slope of the area was calculated to be 25° with an aspect of 135° (southeast). Simultaneously, remote sensing images were acquired, and the reflectance of the red band (0.6), near-infrared band (0.8), and shortwave infrared band (0.7) was extracted. Feature vectors were constructed and classified using a support vector machine, determining that 70% of the area consisted of pine trees (ignition point 320℃, moisture content 12%) and 30% consisted of shrubs (ignition point 280℃, moisture content 8%). The vegetation moisture content index was calculated using the NDMI formula as (0.8-0.7) / (0.8+0.7)=0.067. Combined with actual meteorological station data (daily average temperature 32℃, relative humidity 22%, wind speed 6m / s, precipitation 0mm), the fire assessment indicators for the first transmission line were extracted. For the fire assessment indicators of the second transmission line, semantic segmentation of laser point clouds was used to identify candidate gaps between conductor point cloud clusters and pine tree point cloud clusters. Euclidean clustering (distance threshold 1.5m) was used to filter out three potential hazard areas. The parametric equations of the power line centerline were fitted using RANSAC. The calculated vertical distances from the obstacle point cloud to the centerline were 3.2m, 3.8m, and 4.1m, respectively. After removing the outlier of 4.1m, the average gap distance was 3.5m (less than the 110kV line safety threshold of 4m). The insulator installation date was extracted from the ledger as June 13, 2017, and the current time is June 13, 2025, with a usage time of 8 years. Normalized to the 0-1 interval, a weight of 0.4 was obtained. Combined with a humidity of 22%, the influence coefficient was calculated to be 0.85 using the IEEE humidity correction model. Substituting these values ​​into the polynomial... The fire assessment index for the second transmission line is 0.5×3.5+0.3×0.4+0.2×0.85=2.02. For the fire assessment index of the third transmission line, vegetation point clouds are segmented using a height threshold (>2m). The highest point elevation in the pine area is 125m, the centerline elevation of the power line is 121m, the vertical distance is 4m, the average horizontal distance of the contour points is 3.2m, and the comprehensive distance is (4+3.2) / 2=3.6m. Based on remote sensing spectral characteristics, the pine trees are mapped to level 4 (out of 5) in the flammability database. NDVI=(0.8-0.6) / (0.8+0.6)=0.143, which is then substituted into the polynomial. The fire assessment index for the third transmission line is 0.4×3.6+0.5×4+0.1×0.143=3.45.

[0070] After inputting the above indicators into the dynamic Bayesian network, the initial network constructs conditional probabilities based on historical data. For example, by collecting humidity and vegetation moisture content data of the target mountain area over the past 5 years, the network divides the states and counts the frequencies, obtaining a conditional probability of 0.85 for "humidity = low" and "vegetation moisture content = low". This is used to construct a directed acyclic graph representing the relationship between the indicators. The transition network first filters dynamic nodes, calculates the coefficients of variation of wind speed and humidity (wind speed CV = 0.5, humidity CV = 0.35, both greater than the first threshold of 0.2), and determines them as dynamic nodes. Nodes with CVs less than the threshold, such as terrain slope, are non-dynamic nodes, and their states remain unchanged between time slices. Then, time slice parameters are constructed based on historical data. For example, analyzing 200 samples with "gap distance < 4m and humidity = low", 180 samples have "discharge risk = high", resulting in P(discharge risk = high | gap distance < 4m, humidity = low) = 0.9.

[0071] Next, when optimizing the network structure using the BIC criterion, multiple candidate network structures are first constructed based on historical observation data. Each structure corresponds to different index connection methods and conditional probability parameters. For each candidate structure, its log-likelihood value is calculated, which reflects the degree of fit of the model to the observed data D (such as humidity, line temperature, and other index values). At the same time, the number of conditional probability parameters n is counted, and the model complexity penalty term (n / 2)logN is calculated in combination with the number of training samples N, thus obtaining the BIC value. For example, candidate structure A has a log-likelihood of -235.6, a number of parameters n=15, and a number of samples N=1000. Its BIC value is -235.6-(15 / 2)log1000≈-235.6-22.47≈-258.07; candidate structure B has a log-likelihood of -212.3, a number of parameters n=12, and its BIC value is -212.3-(12 / 2)log1000≈-212.3-18.06≈-230.36. By comparing the two, candidate structure B with the larger BIC value is selected as the optimal network structure to balance the model's fitting ability and complexity.

[0072] After determining the optimal structure, joint probabilistic inference is performed using a dynamic Bayesian network. The conditional probability of 0.85 for "vegetation moisture content = low" when "humidity = low" in the initial network is combined with the time slice parameter P(discharge risk = high | gap distance < 4m, humidity = low) = 0.9 in the transition network. Wind speed (CV = 0.5) and humidity (CV = 0.35) are considered as the temporal transition relationships of dynamic nodes, while the states of non-dynamic nodes remain unchanged. After weighted fusion of immediate influence weights (such as line temperature contributing 35% and humidity contributing 25%) and temporal influence degrees (such as wind speed contributing 20% ​​in the previous time slice), the joint probability of transmission line fire risk in this area is finally calculated to be 0.72 through probability propagation, corresponding to the "high risk" level.

[0073] In summary, this application integrates technologies from multiple fields such as remote sensing, meteorology, and power engineering to construct a complete system from data acquisition to risk assessment. By extracting indicators such as topography, line discharge, and vegetation interaction from multi-source data, it breaks away from the reliance on expert experience in traditional methods and achieves objective quantification of assessment criteria. By leveraging the synergy of static causal networks and dynamic temporal networks, it can not only characterize the spatial correlation of factors affecting transmission line fires but also capture the temporal evolution of risks, solving the problem of traditional models lacking spatiotemporal analysis. Utilizing point cloud semantic segmentation and dynamic node selection technologies, it accurately identifies line hazards and focuses on key risk factors, improving the model's adaptability to complex environments. Through Bayesian network structure optimization and sliding window parameter update mechanisms, it calibrates model parameters in real time according to environmental changes, ensuring that the assessment results closely match the actual situation on site.

[0074] It is understood that, in order to achieve the functions in the above embodiments, the computer device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and method steps described in conjunction with the embodiments disclosed in this application, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0075] Furthermore, as a response to the above Figure 1 The implementation of the method embodiment shown in this application provides a power transmission line fire risk assessment device. The embodiment of this device corresponds to the foregoing method embodiment. For ease of reading, this embodiment will not repeat the details of the foregoing method embodiment, but it should be understood that the device in this embodiment can implement all the contents of the foregoing method embodiment. Specifically, as shown... Figure 2 As shown, the power transmission line fire risk assessment device 200 includes:

[0076] The acquisition module 210 is used to acquire multi-source data of the target area, which is a mountainous area with power transmission lines.

[0077] Extraction module 220 is used to extract the index values ​​of multiple transmission line fire risk assessment indicators based on multi-source data;

[0078] The assessment module 230 is used to determine the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation and indicator values ​​among the assessment indicators within the same time slice, to determine the temporal impact of the indicator values ​​of the previous time slice on the fire risk assessment of transmission lines in the current time slice based on the temporal dependency between the assessment indicators of adjacent time slices, and to determine the fire risk level of transmission lines in the target area based on the immediate impact and temporal impact.

[0079] Furthermore, such as Figure 2 As shown, the evaluation module 230 is specifically used to evaluate based on... Determine the time slice parameters, where, Let be the set of indicators for assessing the fire risk of transmission lines at time s. This refers to the i-th transmission line fire risk assessment indicator in the indicator set at time s+1. To Pre-existing indicators with direct causal impact; the dispersion of each transmission line fire risk assessment indicator in the time series is calculated based on the correlation, and the impact of each transmission line fire risk assessment indicator on the transmission line fire risk is evaluated. Transmission line fire risk assessment indicators with dispersion greater than a first threshold and impact greater than a second threshold are used as dynamic indicators; the temporal impact is determined based on time slice parameters and dynamic indicators, wherein the state transition relationship of dynamic indicators between adjacent time slices is calculated by time slice parameters.

[0080] Furthermore, such as Figure 2 As shown, the evaluation module 230 is specifically used to trigger adjustments to the network structure describing correlations and temporal dependencies when the temporal characteristics of transmission line fire risk assessment indicators show abnormal fluctuations, or when the geographical environment or facility status of the target area changes. Based on the type of triggering factor, the network structure is adjusted, including but not limited to adding, deleting, or merging transmission line fire risk assessment indicators in the network structure, evaluating the strength of correlations between indicators, and adjusting the conditional probability parameters of the network structure based on the strength of correlations to optimize the connection weights of transmission line fire risk assessment indicators in the network structure, thereby obtaining candidate network structures. The optimal network structure is determined from the candidate network structures, where... For the fire risk assessment index D of transmission lines, in terms of conditional probability parameters The log-likelihood is given by N, where N is the number of training samples and n is the number of conditional probability parameters.

[0081] Furthermore, such as Figure 2 As shown, the evaluation module 230 is specifically used to construct a frequency table of fire risk assessment indicators for transmission lines. The frequency table is used to indicate the frequency of occurrence of each fire risk assessment indicator for transmission lines under different conditions. Based on the frequency table, an initial conditional probability table is constructed. The initial conditional probability table is used to determine the initial conditional probability parameters of the fire risk assessment model for transmission lines.

[0082] Furthermore, such as Figure 2As shown, the evaluation module 230 is specifically used to add the newly acquired observation data of the transmission line fire risk assessment indicators at each time step into the window and remove the earliest identical data in the window; and to redetermine the conditional probability parameters using the observation data in the window in order to update the parameters of the transmission line fire risk assessment model.

[0083] Furthermore, such as Figure 2 As shown, the extraction module 220 is specifically used to extract the first parameter affecting the discharge of the transmission line gap based on multi-source data. The first parameter includes the gap distance, the insulator's service life, and the humidity. Based on the first parameter, a second transmission line fire assessment index is constructed. The module also extracts the second parameter affecting the interaction between vegetation and the transmission line based on multi-source data. The second parameter includes the distance from the vegetation to the transmission line, the vegetation classification, and the vegetation normalization index. Based on the second parameter, a third transmission line fire assessment index is constructed.

[0084] Furthermore, such as Figure 2 As shown, the extraction module 220 is specifically used to process laser point cloud data using a semantic segmentation algorithm to identify power line conductor point cloud clusters and obstacle point cloud clusters; determine gap candidate areas based on the three-dimensional spatial positional relationship between the power line conductor point cloud clusters and obstacle point cloud clusters; cluster the point clouds of the gap candidate areas using an Euclidean distance clustering algorithm, and filter out gap areas with potential safety hazards based on a safe distance threshold; use a projection transformation algorithm to map the power line conductor point cloud clusters in the gap area onto a two-dimensional plane to extract the power line centerline; calculate the vertical distance between the power line centerline and each point in the obstacle point cloud cluster, and use the average vertical distance as the gap distance.

[0085] Furthermore, such as Figure 2 As shown, the extraction module 220 is specifically used to perform semantic segmentation on the laser point cloud data to identify power line point cloud clusters and vegetation point cloud clusters; it uses a projection transformation algorithm to map the power line point cloud clusters onto a two-dimensional plane, and then extracts the center line of the power line through curve fitting; it performs point cloud segmentation on the vegetation point cloud clusters by setting a height threshold to identify tree areas in the target area; it calculates the vertical distance between the center line of the power line and the highest point of the tree area, and the horizontal distance between the center line of the power line and the outline points of the tree area; and it uses the average of the vertical and horizontal distances as the distance from the vegetation to the power transmission line.

[0086] Furthermore, such as Figure 2As shown, the extraction module 220 is specifically used to extract the spectral features of remote sensing images and construct a feature vector containing reflectance of the red band, near-infrared band, and short-wave infrared band. The feature vector is classified using a support vector machine to identify vegetation types within a preset range of the transmission line. Based on a pre-constructed vegetation flammability database, vegetation types are mapped to corresponding flammability levels, which are determined according to the ignition point, moisture content, and volatile organic compound content of the plants.

[0087] Optionally, the transmission line fire risk assessment device may be an electronic device with data processing capabilities, or a functional module within such electronic device, without limitation.

[0088] For example, the electronic device can be a server, which can be a single server or a server cluster consisting of multiple servers. As another example, the electronic device can be a mobile phone, tablet computer, desktop computer, laptop computer, handheld computer, notebook computer, ultra-mobile personal computer (UMPC), netbook, as well as cellular phone, personal digital assistant (PDA), augmented reality (AR), virtual reality (VR) device, and other terminal devices. Furthermore, the electronic device can also be a recording device, video surveillance device, etc. This application does not impose any special limitations on the specific form of the electronic device.

[0089] The following example uses an electronic device for assessing fire risks in power transmission lines. Figure 3 As shown, Figure 3 The hardware structure of an electronic device 300 provided in this application.

[0090] like Figure 3 As shown, the electronic device 300 includes a processor 310, a communication line 320, and a communication interface 330.

[0091] Optionally, the electronic device 300 may also include a memory 340. The processor 310, memory 340, and communication interface 330 can be connected via a communication line 320.

[0092] The processor 310 can be a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 310 can also be any other device with processing capabilities, such as a circuit, device, or software module, without limitation.

[0093] In one example, processor 310 may include one or more CPUs, for example Figure 3 CPU0 and CPU1 in the CPU.

[0094] As an optional implementation, the electronic device 300 may include multiple processors, for example, in addition to processor 310, it may also include processor 370. A communication line 320 is used to transmit information between the components included in the electronic device 300.

[0095] Communication interface 330 is used for communication with other devices or other communication networks. These other communication networks can be Ethernet, Radio Access Network (RAN), Wireless Local Area Networks (WLAN), etc. Communication interface 330 can be a module, circuit, transceiver, or any device capable of enabling communication.

[0096] The memory 340 is used to store instructions. These instructions can be computer programs.

[0097] The memory 340 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and / or instructions; it may also be a random access memory (RAM) or other type of dynamic storage device capable of storing information and / or instructions; it may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, etc., without limitation.

[0098] It should be noted that the memory 340 can exist independently of the processor 310, or it can be integrated with the processor 310. The memory 340 can be used to store instructions, program code, or some data, etc. The memory 340 can be located inside or outside the electronic device 300, without restriction.

[0099] The processor 310 is configured to execute instructions stored in the memory 340 to implement the communication method provided in the following embodiments of this application. For example, when the electronic device 300 is a terminal or a chip in a terminal, the processor 310 can execute instructions stored in the memory 340 to implement the steps performed by the sending end in the following embodiments of this application.

[0100] As an optional implementation, the electronic device 300 also includes an output device 350 and an input device 360. The output device 350 can be a display screen, speaker, or other device capable of outputting data from the electronic device 300 to the user. The input device 360 ​​can be a keyboard, mouse, microphone, joystick, or other device capable of inputting data into the electronic device 300.

[0101] It should be pointed out that, Figure 3 The structure shown does not constitute a limitation on the electronic device, except... Figure 3 In addition to the components shown, the electronic device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0102] The transmission line fire risk assessment device and application scenarios described in this application are intended to more clearly illustrate the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of transmission line fire risk assessment devices and the emergence of new business scenarios, the technical solutions provided in this application are also applicable to similar technical problems.

[0103] This application provides a storage medium storing a program that, when executed by a processor, implements the power transmission line fire risk assessment method.

[0104] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program with the following method steps: acquiring multi-source data of a target area, wherein the target area is a mountainous area with transmission lines; extracting the index values ​​of multiple transmission line fire risk assessment indicators based on the multi-source data; determining the immediate impact of each assessment indicator on the transmission line fire risk based on the correlation between the assessment indicators and the index values ​​within the same time slice; determining the temporal impact of the index values ​​of the previous time slice on the transmission line fire risk assessment of the current time slice based on the temporal dependency between the assessment indicators of adjacent time slices; determining the transmission line fire risk level of the target area based on the immediate impact level and the temporal impact level, and describing the correlation between the transmission line fire risk assessment indicators within the same time slice, wherein a transfer network is used to indicate the temporal dependency between the transmission line fire risk assessment indicators of adjacent time slices; and inputting the transmission line fire risk assessment indicators into the transmission line fire risk assessment model to determine the transmission line fire risk level of the target area.

[0105] Furthermore, according to Determine the time slice parameters, where, Let be the set of indicators for assessing the fire risk of transmission lines at time s. This refers to the i-th transmission line fire risk assessment indicator in the indicator set at time s+1. To Pre-existing indicators with direct causal impact; the dispersion of each transmission line fire risk assessment indicator in the time series is calculated based on the correlation, and the impact of each transmission line fire risk assessment indicator on the transmission line fire risk is evaluated. Transmission line fire risk assessment indicators with dispersion greater than a first threshold and impact greater than a second threshold are used as dynamic indicators; the temporal impact is determined based on time slice parameters and dynamic indicators, wherein the state transition relationship of dynamic indicators between adjacent time slices is calculated by time slice parameters.

[0106] Furthermore, when the temporal characteristics of transmission line fire risk assessment indicators exhibit abnormal fluctuations, or when the geographical environment or facility status of the target area changes, adjustments to the network structure describing correlations and temporal dependencies are triggered. Based on the type of triggering factor, the network structure is adjusted, including but not limited to adding, deleting, or merging transmission line fire risk assessment indicators in the network structure. The strength of the correlation between indicators is assessed, and the conditional probability parameters of the network structure are adjusted based on the strength of the correlation to optimize the connection weights of the transmission line fire risk assessment indicators in the network structure, thereby obtaining candidate network structures. The optimal network structure is determined from the candidate network structures, where... For the fire risk assessment index D of transmission lines, in terms of conditional probability parameters The log-likelihood is given by N, where N is the number of training samples and n is the number of conditional probability parameters.

[0107] Furthermore, a frequency table of fire risk assessment indicators for transmission lines is constructed, which indicates the frequency of occurrence of each fire risk assessment indicator under different conditions. An initial conditional probability table is constructed based on the frequency table, which is used to determine the initial conditional probability parameters of the fire risk assessment model for transmission lines.

[0108] Furthermore, the newly acquired observation data of transmission line fire risk assessment indicators at each time step are added to the window, and the earliest identical data in the window is removed; the conditional probability parameters are re-determined using the observation data in the window to update the parameters of the transmission line fire risk assessment model.

[0109] Furthermore, based on multi-source data, a first parameter affecting the discharge of transmission line gaps is extracted. This first parameter includes gap distance, insulator usage time, and humidity. A second transmission line fire assessment index is constructed based on this first parameter. Based on multi-source data, a second parameter affecting the interaction between vegetation and transmission lines is extracted. This second parameter includes distance from vegetation to transmission lines, vegetation classification, and vegetation normalization index. A third transmission line fire assessment index is constructed based on this second parameter.

[0110] Furthermore, semantic segmentation algorithms are used to process the laser point cloud data to identify power line conductor point cloud clusters and obstacle point cloud clusters. Candidate gap areas are determined based on the three-dimensional spatial relationship between the power line conductor point cloud clusters and obstacle point cloud clusters. Euclidean distance clustering algorithms are used to cluster the point clouds in the candidate gap areas, and gap areas with potential safety hazards are selected based on a safety distance threshold. A projection transformation algorithm is used to map the power line conductor point cloud clusters within the gap area onto a two-dimensional plane to extract the power line centerline. The vertical distance between the power line centerline and each point in the obstacle point cloud cluster is calculated, and the average vertical distance is used as the gap distance.

[0111] Furthermore, semantic segmentation is performed on the laser point cloud data to identify power line point cloud clusters and vegetation point cloud clusters; a projection transformation algorithm is used to map the power line point cloud clusters onto a two-dimensional plane, and then the center line of the power line is extracted through curve fitting; the vegetation point cloud clusters are segmented by setting a height threshold to identify tree areas in the target area; the vertical distance between the center line of the power line and the highest point of the tree area, and the horizontal distance between the center line of the power line and the outline points of the tree area are calculated; the average of the vertical and horizontal distances is taken as the distance from vegetation to the power transmission line.

[0112] Furthermore, spectral features of remote sensing images are extracted to construct feature vectors containing reflectance in the red band, near-infrared band, and short-wave infrared band. Support vector machines are used to classify the feature vectors to identify vegetation types within a preset range of the transmission line. Based on a pre-constructed vegetation flammability database, vegetation types are mapped to corresponding flammability levels, which are determined based on the ignition point, moisture content, and volatile organic compound content of the plants.

[0113] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0114] In a typical configuration, the device includes one or more processors (CPUs), memory, and a bus. The device may also include input / output interfaces, network interfaces, etc.

[0115] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM, and memory includes at least one memory chip. Memory is an example of computer-readable media.

[0116] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0117] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0118] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0119] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for assessing the fire risk of power transmission lines, characterized in that, The method includes: Acquire multi-source data for a target area, which is a mountainous area with power transmission lines. Based on the multi-source data, the index values ​​of multiple transmission line fire risk assessment indicators are extracted; The immediate impact of each assessment indicator on the risk of fire on transmission lines is determined based on the correlation between the assessment indicators within the same time frame and the values ​​of the indicators. The temporal influence of the indicator values ​​of the previous time slice on the fire risk assessment of the transmission line in the current time slice is determined based on the temporal dependencies between the assessment indicators of adjacent time slices. The fire risk level of the transmission lines in the target area is determined based on the immediate impact level and the temporal impact level. The immediate impact of each assessment indicator on the risk of transmission line fires is determined based on the correlation between the assessment indicators within the same time frame and the indicator values, including: For the fire risk assessment index values ​​and historical index values ​​of transmission lines obtained within the same time window, a method combining Granger causality test and information entropy is used to identify the nonlinear correlation between the indicators and construct a directed acyclic graph to represent the index correlation network. The Bayesian network conditional probability inference algorithm is applied to calculate the weight of the immediate impact of each assessment indicator on fire risk by combining real-time indicator values.

2. The method according to claim 1, characterized in that, The temporal impact of the indicator values ​​from the previous time slot on the current time slot's transmission line fire risk assessment is determined based on the temporal dependencies between assessment indicators in adjacent time slots, including: according to Determine the time slice parameters, where, Let s be the set of fire risk assessment indicators for transmission lines at time s. This refers to the i-th transmission line fire risk assessment indicator in the indicator set at time s+1. To Pre-existing indicators that have a direct causal effect; Based on the aforementioned correlation, the dispersion of each transmission line fire risk assessment index in the time series is calculated, and the impact of each transmission line fire risk assessment index on the transmission line fire risk is evaluated. Transmission line fire risk assessment indices with dispersion greater than a first threshold and impact greater than a second threshold are used as dynamic indices. The degree of temporal impact is determined based on the time slice parameters and the dynamic indicators, wherein the state transition relationship of the dynamic indicators between adjacent time slices is calculated by displaying the time slice parameters.

3. The method according to claim 2, characterized in that, The method further includes: When the temporal characteristics of the fire risk assessment indicators of the transmission line show abnormal fluctuations, or when the geographical environment and facility status of the target area change, the network structure describing the correlation and the temporal dependency relationship is adjusted. The network structure is adjusted according to the type of triggering factors, including adding, deleting or merging transmission line fire risk assessment indicators in the network structure, evaluating the strength of the correlation between the indicators, and adjusting the conditional probability parameters of the network structure according to the strength of the correlation to optimize the connection weights of the transmission line fire risk assessment indicators in the network structure, thereby obtaining a candidate network structure. according to The optimal network structure is determined from the candidate network structures, wherein... The fire risk assessment index D for the transmission line is given by the conditional probability parameter. The log-likelihood is given by N, where N is the number of training samples and n is the number of conditional probability parameters.

4. The method according to claim 3, characterized in that, The method further includes: A frequency table of fire risk assessment indicators for transmission lines is constructed, wherein the frequency table is used to indicate the frequency of occurrence of each fire risk assessment indicator for transmission lines under different conditions; An initial conditional probability table is constructed based on the frequency table, and the initial conditional probability table is used to determine the initial conditional probability parameters of the network structure.

5. The method according to claim 4, characterized in that, The method further includes: Add the newly acquired observation data of the transmission line fire risk assessment indicators at each time step to the window, and remove the earliest data of the same value in the window; The conditional probability parameters are redefined using the observation data within the window to update the parameters of the transmission line fire risk assessment model.

6. The method according to claim 1, characterized in that, Based on the multi-source data, the index values ​​of multiple transmission line fire risk assessment indicators are extracted, including: Based on the multi-source data, a first transmission line fire assessment index is extracted. The first transmission line fire assessment index is used to quantify the basic environmental factors that directly affect the occurrence and spread of transmission line fires within the target area. Based on the multi-source data, a first parameter affecting the discharge of gaps in transmission lines is extracted. The first parameter includes gap distance, insulator usage time, and humidity. A second transmission line fire assessment index is constructed based on the first parameter. The second transmission line fire assessment index is used to quantify the risk of spark discharge caused by insulation gap breakdown in transmission lines. Based on the multi-source data, a second parameter affecting the interaction between vegetation and power transmission lines is extracted. The second parameter includes the distance from vegetation to power transmission lines, vegetation classification, and vegetation normalization index. A third power transmission line fire assessment index is constructed based on the second parameter. The third power transmission line fire assessment index is used to quantify the potential risk of power transmission line fires caused by the spatial interaction between vegetation and power transmission lines.

7. The method according to claim 6, characterized in that, Based on the multi-source data, the first parameter affecting the discharge of transmission line gaps is extracted, including: The laser point cloud data in the multi-source data is processed using a semantic segmentation algorithm to identify point cloud clusters of transmission line conductors and obstacle point cloud clusters; The gap candidate region is determined based on the three-dimensional spatial positional relationship between the point cloud cluster of the power transmission line conductor and the point cloud cluster of the obstacle. The point cloud of the gap candidate region is clustered using the Euclidean distance clustering algorithm, and gap regions with potential safety hazards are selected based on the safety distance threshold. A projection transformation algorithm is used to map the point cloud clusters of transmission line conductors in the gap area onto a two-dimensional plane in order to extract the center line of the power line; Calculate the vertical distance between the centerline of the power line and each point in the obstacle point cloud cluster, and use the average value of the vertical distance as the gap distance.

8. The method according to claim 6, characterized in that, Based on the multi-source data, a second parameter affecting the interaction between vegetation and transmission lines is extracted, including: Semantic segmentation is performed on the laser point cloud data in the multi-source data to identify power line point cloud clusters and vegetation point cloud clusters; The power line point cloud clusters are mapped to a two-dimensional plane using a projection transformation algorithm, and then the center line of the power line is extracted by curve fitting. The vegetation point cloud clusters are segmented by setting a height threshold to identify tree areas in the target area; Calculate the vertical distance between the centerline of the power line and the highest point of the tree area, and the horizontal distance between the centerline of the power line and the outline point of the tree area; The average of the vertical distance and the horizontal distance is taken as the distance from the vegetation to the power transmission line.

9. The method according to claim 6, characterized in that, The second parameter affecting the interaction between vegetation and transmission lines, extracted based on the multi-source data, also includes: The spectral features of the remote sensing images in the multi-source data are extracted to construct a feature vector containing the reflectance of the red band, near-infrared band, and short-wave infrared band. The feature vectors are classified using a support vector machine to identify vegetation types within a preset range of the transmission line; Based on a pre-constructed vegetation flammability database, the vegetation species are mapped to corresponding flammability levels, which are determined according to the plant's ignition point, moisture content, and volatile organic compound content.

10. A fire risk assessment device for power transmission lines, characterized in that, The device includes: The acquisition module is used to acquire multi-source data of a target area, which is a mountainous area with power transmission lines. The extraction module is used to extract the index values ​​of multiple transmission line fire risk assessment indicators based on the multi-source data; The assessment module is used to determine the immediate impact of each assessment indicator on the fire risk of transmission lines based on the correlation between assessment indicators within the same time slice and the indicator values; to determine the temporal impact of the indicator values ​​of the previous time slice on the fire risk assessment of transmission lines in the current time slice based on the temporal dependency between assessment indicators in adjacent time slices; and to determine the fire risk level of transmission lines in the target area based on the immediate impact and the temporal impact. The assessment module is specifically used to identify the nonlinear relationships between indicators by combining Granger causality test and information entropy for the fire risk assessment index values ​​and historical index values ​​of transmission lines obtained within the same time window, and to construct a directed acyclic graph to represent the index relationship network; and to calculate the immediate impact weight of each assessment index on fire risk by applying the Bayesian network conditional probability inference algorithm in combination with the real-time index values.

11. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the storage medium to perform the transmission line fire risk assessment method as described in any one of claims 1-9.

12. An electronic device, characterized in that, The device includes at least one processor, at least one memory connected to the processor, and a bus; wherein the processor and the memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the transmission line fire risk assessment method as described in any one of claims 1-9.

Images