A transmission line risk prediction method based on multi-modal fusion

Abstract

The present invention relates to the field of power transmission line technology, specifically a power transmission line risk prediction method based on multimodal fusion, comprising the following steps: collecting static images and dynamic videos of the power transmission line, extracting geographic location information embedded in the image and video file names, comparing the extracted geographic location information with the geographic coordinates in the geographic information, and matching the image and video to the corresponding geographic coordinates. The present invention integrates static images and dynamic video data of the power transmission line to achieve analysis of its environmental characteristics and risk prediction. At the same time, the extraction of geographic location information and the fusion of image data are introduced to improve the accuracy of risk area positioning and the reliability of prediction. By analyzing the relative position and morphological parameters of the power transmission line and the surrounding objects, the potential threat of environmental changes to the line is monitored in real time. Not only the timeliness of fault prediction is improved, but also the ability to predict future risk trends is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission lines, and in particular to a power transmission line risk prediction method based on multi-modal fusion. Background Art

[0002] The field of transmission line technology involves the design, construction, maintenance and monitoring of transmission networks in power systems. Key tasks in this field include ensuring the safe and efficient transmission of electricity from power stations to various distribution networks and end users through high-voltage transmission lines. Technologies include line layout planning, physical and chemical properties of wires, weather resistance and safety assessments. In addition, modern transmission line technology also emphasizes the application of intelligent monitoring and fault prediction systems.

[0003] Existing technologies mainly rely on traditional monitoring and historical data, which makes it difficult to capture risks caused by rapid environmental changes, such as sudden weather events, in real time. This technology limits the timeliness and comprehensiveness of risk prediction and cannot provide comprehensive data analysis support. At the same time, it lacks dynamic update capabilities in data processing and risk warning, and is inefficient in processing large-scale real-time data, which leads to the inability to respond quickly in emergencies, increasing operational risks and potential economic losses. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings existing in the prior art and propose a transmission line risk prediction method based on multi-modal fusion.

[0005] In order to achieve the above object, the present invention adopts the following technical solution, a transmission line risk prediction method based on multimodal fusion, comprising the following steps:

[0006] Collect static images and dynamic videos of the power transmission line, extract geographic location information embedded in the image and video file names, compare the extracted geographic location information with the geographic coordinates in the geographic information, match the images and videos to the corresponding geographic coordinates, and obtain a matched multimedia data set; perform coordinate calculations on the images and videos based on the matched multimedia data set, calculate the positions of the power transmission line and surrounding objects, analyze pixel values ​​to detect the edges and shapes of the power transmission line and surrounding objects, calculate the distance and angle between the power transmission line and surrounding objects, and obtain recognized spatial features;

[0007] Based on the identified spatial features, morphological parameters of the transmission line and surrounding objects are extracted, the spatial relationship of proximity and arrangement is calculated, the geometric configuration is analyzed to identify the area where the transmission line interacts with the surrounding objects, and spatial relationship data is obtained. The spatial relationship data is analyzed to identify key areas and potential danger points, and the identified risk areas are obtained.

[0008] Preferably, based on the identified risk area, time series data of environmental parameters of the risk area are collected, including meteorological conditions and vegetation growth conditions, the time series data are merged with the spatial relationship data to obtain a spatiotemporal data set, and a spatiotemporal analysis is performed on the spatiotemporal data set to calculate the rate of change of the environmental parameters over time, calculate the impact of the change on the distance and angle between the transmission line and surrounding objects, use time extrapolation to predict the spatial configuration in the future time period, and obtain the risk trend prediction result;

[0009] Based on the risk trend prediction results, the predicted risk level is compared with the safety threshold set for the transmission line, the locations where the predicted risk level exceeds the threshold are identified, the predicted fault points are obtained, the predicted fault points are sorted with the geographic coordinates and the associated risk levels, and a risk prediction report is generated.

[0010] Preferably, the steps of acquiring the matched multimedia data set are:

[0011] Collect static images and dynamic videos of the transmission line, extract the embedded geographic location information from the file name of each file, and generate a geographic location information list;

[0012] Based on the geographic location information list, the matching degree between the geographic location of each file and the geographic coordinates in the geographic information is calculated, and the calculation formula is:

[0013] ;

[0014] in, is the geographic location coordinates in the file, is the coordinate in GIS, Indicates the geographic matching degree of the i-th file;

[0015] Based on the matching degree, multimedia contents with matching degrees higher than a set threshold are screened out, and the multimedia contents are integrated into a new set to obtain a matched multimedia data set.

[0016] Preferably, the steps of obtaining the identified spatial features are:

[0017] Extracting images and videos from the multimedia data set to perform coordinate calculation, determine the location of the power transmission line and surrounding objects, analyze pixel values ​​of the images, identify edges and shapes of the power transmission line and surrounding objects, and obtain edge and shape data;

[0018] Based on the edge and shape data, the distance and angle between the transmission line and the surrounding objects are calculated using the following formula:

[0019] ;

[0020] and

[0021] ;

[0022] in, are the coordinates of the transmission line, are the coordinates of the surrounding objects, Indicates distance, Indicates angle;

[0023] Based on the distance and angle, the spatial relationship between the power transmission line and surrounding objects is determined to obtain the identified spatial features.

[0024] Preferably, the steps of acquiring the spatial relationship data are:

[0025] Based on the identified spatial features, extract the size, shape and position relative to the transmission line of objects around the transmission line to construct morphological parameter data;

[0026] Based on the morphological parameter data, the spatial relationship score between the transmission line and the surrounding objects is calculated using the following formula:

[0027] ;

[0028] in, represents the spatial relationship score, is the weight of the score of the ith object, is the distance from the transmission line to the i-th object;

[0029] Based on the spatial relationship score, the geometric configuration of the object relative to the power transmission line is analyzed, the interaction area is identified, and the spatial relationship data is obtained.

[0030] Preferably, the steps of obtaining the identified risk area are:

[0031] Analyze the spatial relationship data to determine the spatial distribution characteristics of the power transmission line and surrounding objects, screen areas with potential safety risk characteristics, and obtain preliminary identified key areas;

[0032] Based on the initially identified key areas, the geometric shape of the object, the distance from the transmission line, and the similarity with known risk events are analyzed to mark potential danger points and obtain potential danger point data;

[0033] Based on the potential danger point data, the risk area that poses a threat to the operational safety of the power transmission line is identified and verified to obtain an identified risk area.

[0034] Preferably, the steps of acquiring the spatiotemporal dataset are:

[0035] Based on the identified risk area, collecting time series data of environmental parameters of the identified risk area to obtain environmental parameter data;

[0036] Based on the environmental parameter data, preprocessing and cleaning the environmental parameter data are performed, and the environmental parameter data are calibrated to match the time point and date to obtain preprocessed time series data;

[0037] The preprocessed time series data is fused with the spatial relationship data to form data including time and space dimensions, thereby obtaining a spatiotemporal data set.

[0038] Preferably, the steps for obtaining the risk trend prediction result are:

[0039] Performing spatiotemporal analysis on the spatiotemporal data set to identify changes in temperature, humidity and wind speed over time and obtain environmental parameter change trend data;

[0040] Based on the environmental parameter change trend data, the change in the distance and angle between the transmission line and the surrounding objects is calculated using the following formula:

[0041] ;

[0042] in, Represents the change in distance or angle. is the current distance or angle, is the rate of change of environmental parameters, is the time variation;

[0043] Based on the variation, a time extrapolation method is used to predict the spatial configuration in a future time period, and the risk development trend in the future time period is analyzed to obtain a risk trend prediction result.

[0044] Preferably, the steps for obtaining the risk prediction report are:

[0045] Based on the risk trend prediction results, the predicted risk level of each location is compared with the safety threshold set for the transmission line, locations where the predicted risk level exceeds the safety threshold are identified, and a list of locations with over-threshold risk is obtained;

[0046] Based on the list of super-threshold risk locations, the geographic coordinates and associated risk levels of each location are collected to obtain a risk prediction report.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are:

[0048] The present invention integrates static images and dynamic video data of power transmission lines to achieve analysis of their environmental characteristics and risk prediction. At the same time, the extraction of geographic location information and image data fusion are introduced to improve the accuracy of risk area positioning and the reliability of prediction. By analyzing the relative position and morphological parameters of the power transmission line and surrounding objects, the potential threat of environmental changes to the line is monitored in real time. Not only does it improve the timeliness of fault prediction, but it also enhances the ability to predict future risk trends, making preventive measures more targeted and effective, and reducing the probability of unexpected failures and risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] See also Figure 1 The present invention provides a technical solution, a transmission line risk prediction method based on multi-modal fusion, comprising the following steps:

[0052] Collect static images and dynamic videos of the power transmission line, extract geographic location information embedded in the image and video file names, compare the extracted geographic location information with the geographic coordinates in the geographic information, match the images and videos to the corresponding geographic coordinates, and obtain a matched multimedia data set; perform coordinate calculations on the images and videos based on the matched multimedia data set, calculate the positions of the power transmission line and surrounding objects, analyze pixel values ​​to detect the edges and shapes of the power transmission line and surrounding objects, calculate the distance and angle between the power transmission line and surrounding objects, and obtain recognized spatial features;

[0053] Based on the identified spatial features, morphological parameters of the transmission line and surrounding objects are extracted, the spatial relationship of proximity and arrangement is calculated, the geometric configuration is analyzed to identify the area where the transmission line interacts with the surrounding objects, spatial relationship data is obtained, the spatial relationship data is analyzed to identify key areas and potential danger points, and the identified risk areas are obtained;

[0054] Based on the identified risk area, collect time series data of environmental parameters of the risk area, including meteorological conditions and vegetation growth, fuse the time series data with spatial relationship data, obtain a spatiotemporal data set, perform spatiotemporal analysis on the spatiotemporal data set, calculate the rate of change of environmental parameters over time, calculate the impact of the change on the distance and angle between the transmission line and surrounding objects, use time extrapolation to predict the spatial configuration in the future time period, and obtain risk trend prediction results;

[0055] Based on the risk trend prediction results, the predicted risk level is compared with the safety threshold set for the transmission line, the locations where the predicted risk level exceeds the threshold are identified, the predicted fault points are obtained, the predicted fault points are sorted with the geographic coordinates and the associated risk levels, and a risk prediction report is generated.

[0056] The steps of obtaining the matched multimedia data set are:

[0057] Collect static images and dynamic videos of the transmission line, extract the embedded geographic location information from the file name of each file, and generate a geographic location information list;

[0058] Based on the geographic location information list, the matching degree between the geographic location of each file and the geographic coordinates in the geographic information is calculated, and the calculation formula is:

[0059] ;

[0060] in, is the geographic location coordinates in the file, is the coordinate in GIS, Indicates the geographic matching degree of the i-th file;

[0061] Based on the matching degree, multimedia contents with matching degrees higher than a set threshold are screened out, and the multimedia contents are integrated into a new set to obtain a matched multimedia data set.

[0062] Specifically, static images and dynamic videos of the transmission lines are collected, and the embedded geographic location information is extracted from the file name of each file. Combined with the timestamp and GPS coordinates, the shooting date and location of each image or video file is extracted by parsing the file naming rules to generate a geographic location information list.

[0063] The benefit of the formula is that by applying the sigmoid function to process the Euclidean distance, a nonlinear conversion from distance to similarity is achieved, which makes the matching process more robust and less susceptible to outliers.

[0064] Calculation process: The geographic location coordinates in the file are , the coordinates in GIS are , first calculate the squared part of the Euclidean distance:

[0065] ;

[0066] Then take the square root to get the distance , put it into the formula and calculate:

[0067] ;

[0068] The result shows that the geographic match of the file is 0.4975, which means that the location of the file is very close to the location of the GIS record. If this value is greater than a certain threshold, such as 0.5, the match can be considered valid.

[0069] Based on the matching degree, multimedia content with a matching degree higher than a set threshold is screened out. The matching degree of each file is compared through database query and algorithm screening. The threshold is set to 0.5. The threshold can be adjusted according to actual needs and effects. All files above this threshold are regarded as valid data with successful geographic location matching. These data are further integrated into a new multimedia content collection to ensure that only accurately matched data is used for subsequent analysis to obtain a matched multimedia data set.

[0070] The steps of obtaining the identified spatial features are:

[0071] Extracting images and videos from the multimedia data set to perform coordinate calculation, determine the location of the power transmission line and surrounding objects, analyze pixel values ​​of the images, identify edges and shapes of the power transmission line and surrounding objects, and obtain edge and shape data;

[0072] Based on the edge and shape data, the distance and angle between the transmission line and the surrounding objects are calculated using the following formula:

[0073] ;

[0074] and

[0075] ;

[0076] in, are the coordinates of the transmission line, are the coordinates of the surrounding objects, Indicates distance, Indicates angle;

[0077] Based on the distance and angle, the spatial relationship between the power transmission line and surrounding objects is determined to obtain the identified spatial features.

[0078] Specifically, images and videos in multimedia data sets are extracted for coordinate calculation, images and videos are separated from multimedia data sets, and the approximate locations of transmission lines and surrounding objects are identified using machine learning models. Then, the pixel values ​​of the images are analyzed, the boundaries and shapes of the transmission lines and surrounding objects are marked, and based on these shape data, a detailed edge and shape data table is established.

[0079] The benefits of the formula are that by calculating the distance D and the angle θ, the relative position of the transmission line and the surrounding objects in space can be described, which is crucial for risk assessment and avoiding potential accidents. The steps for obtaining D and θ are as follows: first, the coordinates of the transmission line (x1, y1) and the coordinates of the surrounding objects (x2, y2) are obtained through the geographic information system. These coordinates are obtained by multiple measurements at different time points using the on-site GPS positioning equipment to ensure accuracy.

[0080] Calculation process: The coordinates of the transmission line are (100, 200), and the coordinates of the surrounding objects are (150, 250). The distance D is calculated as follows:

[0081] ;

[0082] The angle θ is calculated as follows:

[0083] ;

[0084] The results show that the distance is about 71 meters and the angle is 45 degrees, indicating that the surrounding objects are located northeast of the transmission line and are relatively close. There may be potential risks and further monitoring and management are required.

[0085] Based on the distances and angles, determine the spatial relationship between the transmission line and surrounding objects, including using the calculated distance and angle data to analyze and evaluate the potential for interaction between the two; in particular, simulate the impact under different weather and environmental conditions, evaluate the stability of the transmission line in extreme weather such as strong winds or ice and snow, and the physical impact that surrounding objects may have on the transmission line.

[0086] The steps for obtaining the spatial relationship data are as follows:

[0087] Based on the identified spatial features, extract the size, shape and position relative to the transmission line of objects around the transmission line to construct morphological parameter data;

[0088] Based on the morphological parameter data, the spatial relationship score between the transmission line and the surrounding objects is calculated using the following formula:

[0089] ;

[0090] in, represents the spatial relationship score, is the weight of the score of the ith object, is the distance from the transmission line to the i-th object;

[0091] Based on the spatial relationship score, the geometric configuration of the object relative to the power transmission line is analyzed, the interaction area is identified, and the spatial relationship data is obtained.

[0092] Specifically, the process of extracting the size, shape and relative position of objects around the transmission line based on the identified spatial features involves the Geographic Information System (GIS). First, high-resolution static images and real-time video data are used in combination with GIS to calibrate the position of each object. Then, the shape of the object is identified, the size of the object is measured, and the three-dimensional size and relative position of each object are obtained, providing accurate basic data for the safety monitoring of the transmission line.

[0093] The benefit of the formula is that it takes into account the inverse square of the distance between each object and the transmission line in the form of a weighted sum, which increases the influence of the distance factor in the score and makes the score more sensitive to changes in distance; The steps to obtain the parameters are to analyze the potential danger of each object through field investigation and historical accident data, and assign different weights. The steps of obtaining the parameters are: using a laser rangefinder to measure the straight-line distance from the transmission line to each object;

[0094] Calculation process: There are three objects on the scene. The distances of objects A, B, and C are 10m, 20m, and 30m respectively. The corresponding weights are set to 0.5, 1.5, and 0.1 respectively. Substitute them into the formula for calculation:

[0095] ;

[0096] The results show that the overall spatial relationship score is 0.008861, which means that objects close to the transmission lines have a greater impact on the total score. If the score exceeds the preset threshold, it may indicate a higher risk level and corresponding risk mitigation measures need to be taken. If past accidents mostly occurred in areas where the score exceeded a certain value, then this value can be set as the preset threshold.

[0097] Based on the spatial relationship score, after analyzing the spatial relationship score between the object and the transmission line, determining the interaction area involves comparing the score data with the preset threshold. For example, areas with a score higher than 0.01 require further analysis. The score of each object is mapped to the map, and different colors represent different risk levels. For areas with higher scores, on-site surveys are conducted, including direct measurement of the distance between the object and the transmission line, and checking the physical state and environmental conditions of the object. Help determine the specific interaction area, point out possible sources of interference and safety hazards, and thus obtain spatial relationship data.

[0098] The steps for obtaining the identified risk area are:

[0099] Analyze the spatial relationship data to determine the spatial distribution characteristics of the power transmission line and surrounding objects, screen areas with potential safety risk characteristics, and obtain preliminary identified key areas;

[0100] Based on the initially identified key areas, the geometric shape of the object, the distance from the transmission line, and the similarity with known risk events are analyzed to mark potential danger points and obtain potential danger point data;

[0101] Based on the potential danger point data, the risk area that poses a threat to the operational safety of the power transmission line is identified and verified to obtain an identified risk area.

[0102] Specifically, the spatial relationship data is analyzed to determine the accurate spatial layout of the transmission lines and surrounding objects, measure the distance and angle between objects and lines, and identify areas that may cause faults due to abnormal proximity or arrangement. The location and size of each object are recorded to assess whether these areas increase safety risks due to their specific location or layout characteristics, and screen out areas with potential risks, such as trees or buildings close to high-voltage lines. These are considered to be critical areas that may trigger safety alarms due to their proximity. Next, determine which areas are less than the safety threshold from the transmission line. This threshold is determined based on previous accident data and safety standards, and is usually set to a safety distance of no less than 10 meters between objects and lines, and obtain preliminary identification of key areas.

[0103] Continue to analyze the initially identified key areas in depth, evaluate the shape of the objects and the actual distance from the transmission line. Compare the high-risk events recorded in the historical data to identify geometric configurations similar to accident cases. This step involves analyzing the shape, size and layout of objects in each key area, using laser ranging to determine the location and shape parameters of the objects. Compare these data with known accident cases. For example, when the distance between the object and the line is less than 5 meters and it has flammable materials, it is marked as high risk. Based on these details, potential danger points are marked and potential danger point data is obtained.

[0104] Based on the data of potential danger points, identify which objects may pose a direct threat to the safety of transmission line operations. Use the collected data, such as the size, shape and material of the object, to conduct a comprehensive risk assessment, with special attention to objects that may cause contact or short circuits due to their location, material or volume. In this process, determining the risk area not only relies on distance measurement, but also considers the impact of the object's material on safety, such as metal objects that are too close to the transmission line may cause arcing. High-risk areas that may actually cause accidents are marked, and these marked risk areas ultimately form detailed risk area identification results.

[0105] The steps for obtaining the spatiotemporal dataset are as follows:

[0106] Based on the identified risk area, collecting time series data of environmental parameters of the identified risk area to obtain environmental parameter data;

[0107] Based on the environmental parameter data, preprocessing and cleaning the environmental parameter data are performed, and the environmental parameter data are calibrated to match the time point and date to obtain preprocessed time series data;

[0108] The preprocessed time series data is fused with the spatial relationship data to form data including time and space dimensions, thereby obtaining a spatiotemporal data set.

[0109] Specifically, environmental parameter data is collected, including temperature, humidity and wind speed data obtained from temperature sensors, humidity sensors and anemometers every hour. First, the data accuracy of each sensor output is verified to confirm that the data value is within the set safety range, such as temperature between -10°C and 50°C, humidity between 10% and 90%, and wind speed not exceeding 100 km / h. Then, the timestamp of each data point is calibrated to ensure the consistency and accuracy of the time record. This step is achieved by comparing the standard time of the synchronized time server. For those data points that are out of the normal range, they are eliminated or corrected to ensure the cleanliness and consistency of the data set. The purpose of this process is to generate a detailed data set that reflects the actual environmental conditions in the risk area.

[0110] Preprocess and clean the environmental parameter data to ensure the neatness and usability of the data set. First, use the moving average method and median filter to remove noise and occasional outliers in the data, such as the temperature suddenly jumping to an illogically high value. Next, standardize the data and adjust all measurements to a uniform scale for subsequent analysis. Calibrate the data points to match the exact collection time and date, and use linear interpolation to fill in missing data points to ensure the consistency of the data in the time series. Through these steps, ensure that the preprocessed data has a high degree of reliability and availability, laying a solid foundation for spatiotemporal data fusion.

[0111] The preprocessed time series data is fused with the spatial relationship data to construct a spatiotemporal dataset. The core of this fusion process is to ensure that each time series data point is accurately aligned with its corresponding spatial position. By checking the matching degree of the data points and the spatial coordinates one by one, it is ensured that there is no misalignment in time or space. Using these fused data, the environmental changes of each risk area at different time points and their potential impact on the surrounding areas can be observed and analyzed in detail. Such a spatiotemporal dataset can help us accurately identify and predict how environmental factors affect the safety of risk areas in time and space.

[0112] The steps for obtaining the risk trend prediction results are:

[0113] Performing spatiotemporal analysis on the spatiotemporal data set to identify changes in temperature, humidity and wind speed over time and obtain environmental parameter change trend data;

[0114] Based on the environmental parameter change trend data, the change in the distance and angle between the transmission line and the surrounding objects is calculated using the following formula:

[0115] ;

[0116] in, Represents the change in distance or angle. is the current distance or angle, is the rate of change of environmental parameters, is the time variation;

[0117] Based on the variation, a time extrapolation method is used to predict the spatial configuration in a future time period, and the risk development trend in the future time period is analyzed to obtain a risk trend prediction result.

[0118] Specifically, we conduct spatiotemporal analysis of spatiotemporal data sets, including time series analysis of parameters such as temperature, humidity, and wind speed. We first obtain the original time series data, ensure data availability through data cleaning and preprocessing operations, and then apply statistical analysis to identify the main trends and periodic changes in the data. We then perform regression analysis on the data to determine the change patterns of each parameter over time and obtain the trend data of environmental parameter changes.

[0119] The usefulness of the formula is that it directly relates to the dynamic interaction of environmental changes with physical space. By considering the time variation and the rate of change of environmental parameters, the expected change in the distance or angle between the transmission line and surrounding objects can be calculated; The parameter acquisition step is to obtain the distance or angle measurement value at the current time point from the on-site monitoring system. The parameter acquisition step is to calculate the rate of change from the environmental monitoring data; for example, if the temperature increases by 2 degrees from the beginning of the year to the end of the year, and the temperature at the beginning is 10 degrees, then the annual temperature change rate is It can be expressed as or 20% annual growth rate, The steps of obtaining the parameters are as follows: obtaining the parameters according to the set time interval;

[0120] Calculation process: The distance between the current transmission line and an object is 100 meters, the annual change rate of environmental parameters is 0.05, and the time interval is set to 1 year. According to the formula,

[0121] ;

[0122] This shows that the expected distance after one year is 105 meters; this result shows that the distance will increase by 5 meters after one year. This changing trend is very important information for the maintenance of transmission lines and the management of the surrounding environment.

[0123] Based on the changes in environmental parameters and the known distance or angle data between the transmission lines and surrounding objects, the time extrapolation method is used to predict the future spatial configuration. This prediction process first relies on the time series model established through historical data. The model can reveal the long-term impact of environmental changes on the location of objects. Through model extrapolation, it is possible to foresee the changes in spatial position caused by environmental changes in the future.

[0124] The steps for obtaining the risk prediction report are as follows:

[0125] Based on the risk trend prediction results, the predicted risk level of each location is compared with the safety threshold set for the transmission line, locations where the predicted risk level exceeds the safety threshold are identified, and a list of locations with over-threshold risk is obtained;

[0126] Based on the list of super-threshold risk locations, the geographic coordinates and associated risk levels of each location are collected to obtain a risk prediction report.

[0127] Specifically, we start to analyze the risk trend prediction results and compare the risk level of each location with the set safety threshold. The safety threshold is pre-set based on historical accident data and safety standards. For example, if historical data shows that an area with an accident rate of more than 10% is considered high risk, this is set as the threshold. On this basis, we check whether the risk level of each predicted location exceeds this standard. Using an iterative method, for each location, we compare its risk value with the threshold, identify the locations that exceed the standard, and update this information in real time to reflect the latest risk status, so as to obtain a clear list of risk locations that exceed the threshold.

[0128] Continue to conduct in-depth analysis of each location in the list of super-threshold risk locations, collect and collate the geographical coordinates of these locations and the associated risk level. The accuracy of the coordinates of each location is critical, ensuring that they are located through a geographic information system (GIS). For each location, its coordinates are recorded and linked to the corresponding risk level. Generate a comprehensive data set that details the geographical coordinates of each location and the risk level.

[0129] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A transmission line risk prediction method based on multimodal fusion, characterized in that: The following steps are involved: Collect static images and dynamic videos of the transmission line, extract the geographic location information embedded in the image and video file names, compare the extracted geographic location information with the geographic coordinates in the geographic information, match the images and videos to the corresponding geographic coordinates, and obtain the matched multimedia data set; Based on the matched multimedia data set, coordinate calculation is performed on the image and video, the position of the power transmission line and the surrounding objects is calculated, the pixel values ​​are analyzed to detect the edge and shape of the power transmission line and the surrounding objects, the distance and angle between the power transmission line and the surrounding objects are calculated, and the recognized spatial features are obtained; Based on the identified spatial features, morphological parameters of the transmission line and surrounding objects are extracted, the spatial relationship of proximity and arrangement is calculated, the geometric configuration is analyzed to identify the area where the transmission line interacts with the surrounding objects, spatial relationship data is obtained, the spatial relationship data is analyzed to identify key areas and potential danger points, and the identified risk areas are obtained; The steps of obtaining the identified spatial features are: Extracting images and videos from the multimedia data set to perform coordinate calculation, determine the location of the power transmission line and surrounding objects, analyze pixel values ​​of the images, identify edges and shapes of the power transmission line and surrounding objects, and obtain edge and shape data; Based on the edge and shape data, the distance and angle between the transmission line and the surrounding objects are calculated using the following formula: ； and ； in, are the coordinates of the transmission line, are the coordinates of the surrounding objects, Indicates distance, Indicates angle; Based on the distance and angle, determine the spatial relationship between the transmission line and surrounding objects to obtain the identified spatial features; The steps for obtaining the spatial relationship data are as follows: Based on the identified spatial features, extract the size, shape and position of objects around the transmission line relative to the transmission line to construct morphological parameter data; Based on the morphological parameter data, the spatial relationship score between the transmission line and the surrounding objects is calculated using the following formula: ； in, represents the spatial relationship score, is the weight of the score of the ith object, is the distance from the transmission line to the i-th object; Based on the spatial relationship score, analyzing the geometric configuration of the object relative to the power transmission line, identifying the interaction area, and obtaining spatial relationship data; The steps of obtaining the matched multimedia data set are: Collect static images and dynamic videos of the transmission line, extract the embedded geographic location information from the file name of each file, and generate a geographic location information list; Based on the geographic location information list, the matching degree between the geographic location of each file and the geographic coordinates in the geographic information is calculated, and the calculation formula is: ； in, is the geographic location coordinates in the file, is the coordinate in GIS, Indicates the geographic matching degree of the i-th file; Based on the matching degree, multimedia contents with matching degrees higher than a set threshold are screened out, and the multimedia contents are integrated into a new set to obtain a matched multimedia data set.

2. The transmission line risk prediction method based on multimodal fusion according to claim 1 is characterized in that: Based on the identified risk area, collect time series data of environmental parameters of the risk area, including meteorological conditions and vegetation growth, fuse the time series data with spatial relationship data, obtain a spatiotemporal data set, perform spatiotemporal analysis on the spatiotemporal data set, calculate the rate of change of environmental parameters over time, calculate the impact of the change on the distance and angle between the transmission line and surrounding objects, use time extrapolation to predict the spatial configuration in the future time period, and obtain risk trend prediction results; Based on the risk trend prediction results, the predicted risk level is compared with the safety threshold set for the transmission line, the locations where the predicted risk level exceeds the threshold are identified, the predicted fault points are obtained, the predicted fault points are sorted with the geographic coordinates and the associated risk levels, and a risk prediction report is generated.

3. The transmission line risk prediction method based on multimodal fusion according to claim 1 is characterized in that: The steps for obtaining the identified risk area are: Analyze the spatial relationship data to determine the spatial distribution characteristics of the power transmission line and surrounding objects, screen areas with potential safety risk characteristics, and obtain preliminary identified key areas; Based on the initially identified key areas, the geometric shape of the object, the distance from the transmission line, and the similarity with known risk events are analyzed to mark potential danger points and obtain potential danger point data; Based on the potential danger point data, the risk area that poses a threat to the operational safety of the power transmission line is identified and verified to obtain an identified risk area.

4. The transmission line risk prediction method based on multimodal fusion according to claim 2 is characterized in that: The steps for obtaining the spatiotemporal dataset are as follows: Based on the identified risk area, collecting time series data of environmental parameters of the identified risk area to obtain environmental parameter data; Based on the environmental parameter data, preprocessing and cleaning the environmental parameter data are performed, and the environmental parameter data are calibrated to match the time point and date to obtain preprocessed time series data; The preprocessed time series data is fused with the spatial relationship data to form data including time and space dimensions, thereby obtaining a spatiotemporal data set.

5. The transmission line risk prediction method based on multimodal fusion according to claim 2 is characterized in that: The steps for obtaining the risk trend prediction results are: Performing spatiotemporal analysis on the spatiotemporal data set to identify changes in temperature, humidity and wind speed over time and obtain environmental parameter change trend data; Based on the environmental parameter change trend data, the change in the distance and angle between the transmission line and the surrounding objects is calculated using the following formula: ; in, Represents the change in distance or angle, is the current distance or angle, is the rate of change of environmental parameters, is the time variation; Based on the variation, a time extrapolation method is used to predict the spatial configuration in a future time period, and the risk development trend in the future time period is analyzed to obtain a risk trend prediction result.

6. The transmission line risk prediction method based on multimodal fusion according to claim 2 is characterized in that: The steps for obtaining the risk prediction report are as follows: Based on the risk trend prediction results, the predicted risk level of each location is compared with the safety threshold set for the transmission line, locations where the predicted risk level exceeds the safety threshold are identified, and a list of locations with over-threshold risk is obtained; Based on the list of super-threshold risk locations, the geographic coordinates and associated risk levels of each location are collected to obtain a risk prediction report.

Images