Mountain fire and tree line discharge fault monitoring method based on intelligent video identification

Abstract

The invention relates to the technical field of fault monitoring, and discloses a forest fire and tree line discharge fault monitoring method based on intelligent video recognition, and the method comprises the steps: determining a to-be-monitored wire, collecting the configuration information of the to-be-monitored wire, and determining a camera installation strategy based on the configuration information; determining a to-be-monitored area, and constructing a tree line monitoring data set based on the number, the terrain information and the growth information; obtaining a discharge fault risk value of the to-be-monitored electric wire and each to-be-monitored area according to the tree line monitoring data set; determining the acquisition frequency of a camera located in the to-be-monitored area according to the discharge fault risk value; judging whether the acquisition frequency of the camera is optimized or not based on the real-time environment information; determining an optimization coefficient according to the environmental influence score, and obtaining an optimized camera acquisition frequency; judging whether forest fire and tree line discharge faults exist or not based on the image features; and if yes, determining a fault level according to the image features. According to the invention, efficient and accurate monitoring of forest fire and tree line discharge faults can be realized.

Description

Technical Field

[0001] This invention relates to the field of fault monitoring technology, and more specifically, to a method for monitoring wildfire and treeline discharge faults based on intelligent video recognition. Background Technology

[0002] Traditional monitoring methods for wildfires and tree-line discharge faults have several limitations. Some methods rely on manual inspections, which are not only costly in terms of manpower, resources, and time, but also have long inspection cycles, making it difficult to detect potential faults in a timely manner. In severe weather or complex terrain, the difficulty and risk of manual inspections increase significantly, potentially rendering the work impossible. Other methods employ simple sensor monitoring technologies, which can acquire relevant data to some extent, but these sensors have limited monitoring ranges and are easily affected by environmental factors, resulting in low accuracy and reliability of the monitoring data. For example, in severe weather conditions such as strong winds and heavy rain, sensors may misjudge or lose data, affecting the timely detection and handling of faults. Furthermore, traditional monitoring methods struggle to effectively identify and analyze complex fault characteristics. Wildfires and tree-line discharge faults are often accompanied by a variety of complex phenomena and characteristics that traditional methods cannot comprehensively and accurately capture. This makes it difficult to scientifically assess and predict the severity and development trend of the faults, posing significant challenges to fault prevention and management.

[0003] Therefore, it is necessary to design a wildfire and treeline discharge fault monitoring method based on intelligent video recognition to solve the problems existing in the current technology. Summary of the Invention

[0004] In view of this, the present invention proposes a method for monitoring wildfire and tree line discharge faults based on intelligent video recognition, aiming to solve the problem that traditional methods cannot fully and accurately capture these features, making it difficult to make scientific assessments and predictions of the severity and development trend of the faults, which brings great difficulties to the prevention and handling of the faults.

[0005] This invention proposes a method for monitoring wildfires and treeline discharge faults based on intelligent video recognition, comprising: Identify the power line to be monitored, collect the configuration information of the power line to be monitored, and determine the camera installation strategy based on the configuration information; The area to be monitored is determined and numbered. Topographic information and growth information of trees to be monitored are collected within the area to be monitored. A treeline monitoring dataset is constructed based on the number, topographic information and growth information. Based on the treeline monitoring dataset, obtain the discharge fault risk value of the monitored power line and each monitored area; The camera acquisition frequency in the area to be monitored is determined based on the discharge fault risk value. Collect real-time environmental information of the area to be monitored, and determine whether to optimize the camera's acquisition frequency based on the real-time environmental information; If so, determine the environmental impact score based on the real-time environmental information, and determine the optimization coefficient based on the environmental impact score to obtain the optimized camera acquisition frequency; The image information of the area to be monitored is acquired at the optimized camera acquisition frequency, and the image information is analyzed to extract image features related to wildfire and treeline discharge faults. Based on the image features, it is determined whether wildfire and treeline discharge faults exist. If so, the fault level is determined based on the image features.

[0006] Furthermore, when determining the camera installation strategy based on the configuration information, the following steps are included: The configuration information is parsed to obtain the wire type of the wire to be monitored, and the camera installation strategy is determined based on the wire type. The types of wires mentioned include high-voltage wires, low-voltage wires, and ultra-high-voltage wires; When the type of power line is a high-voltage power line, the camera installation strategy is determined to be the first camera installation strategy; When the type of wire is low-voltage wire, the camera installation strategy is determined to be the second camera installation strategy; When the type of power line is an ultra-high voltage power line, the camera installation strategy is determined to be a third camera installation strategy.

[0007] Further, when obtaining the discharge fault risk value of the monitored power line and each monitored area based on the tree-line monitoring dataset, the process includes: The age, height, crown morphology characteristics of each tree to be monitored within the monitoring area, the slope, altitude, soil type of the monitoring area, and the erection height of the power line to be monitored are obtained. The contact risk assessment value between each tree to be monitored and the power line to be monitored is determined based on the age, height, canopy morphology characteristics and the erection height. The terrain impact risk assessment values ​​of the trees to be monitored and the power lines to be monitored are determined based on the slope, altitude, and soil type. A failure risk vector is constructed based on the contact risk assessment value and the terrain impact risk assessment value. The fault risk vector is compared with historical data. If there is a historical fault risk vector in the historical data that matches the fault risk vector, then the historical discharge fault risk value corresponding to the historical fault risk vector is used as the discharge fault risk value. Otherwise, the discharge fault risk value is determined based on the fault risk vector.

[0008] Further, determining the discharge fault risk value based on the fault risk vector includes: The failure risk impact score is determined based on the failure risk vector. The fault risk impact score is compared with the first fault risk impact score and the second fault risk impact score, and the discharge fault risk value is determined based on the comparison result; wherein, the first fault risk impact score is less than the second fault risk impact score. When the fault risk impact score is less than or equal to the first fault risk impact score, the discharge fault risk value is determined to be the first discharge fault risk value. When the fault risk impact score is greater than the first fault risk impact score and less than or equal to the second fault risk impact score, the discharge fault risk value is determined to be the second discharge fault risk value. When the fault risk impact score is greater than the second fault risk impact score, the discharge fault risk value is determined to be the third discharge fault risk value.

[0009] 5. The method for monitoring wildfires and treeline discharge faults based on intelligent video recognition according to claim 4, characterized in that, when determining the camera acquisition frequency in the area to be monitored based on the discharge fault risk value, it includes: The discharge fault risk value is compared with a preset acquisition frequency mapping table, and the acquisition frequency of the camera is determined based on the comparison result.

[0010] Furthermore, when determining whether to optimize the camera's acquisition frequency based on the real-time environmental information, the process includes: Feature extraction is performed on the real-time environment feature values ​​to obtain several real-time environment feature values; Obtain the real-time environmental standard value corresponding to each of the aforementioned real-time environmental feature values; If any of the real-time environmental feature values ​​exceed the corresponding real-time environmental standard value, then it is determined that the camera acquisition frequency should be optimized. Otherwise, it is determined that the camera's acquisition frequency will not be optimized.

[0011] Furthermore, when determining the environmental impact score based on the real-time environmental information, the process includes: Obtain each real-time environmental feature value that exceeds the corresponding real-time environmental standard value, and construct a problem environment feature dataset; Obtain the significance score of each real-time environmental feature value in the problem environment feature dataset out of all real-time environmental feature values; Obtain the deviation index between each real-time environmental feature value and the corresponding real-time environmental standard value in the problem environment feature dataset; Based on the importance score and deviation index, calculate the environmental impact sub-score for each real-time environmental feature value; The environmental impact score is obtained by summing the environmental impact sub-scores of all real-time environmental feature values.

[0012] Further, when determining the optimization coefficient based on the environmental impact score to obtain the optimized camera acquisition frequency, the process includes: The environmental impact score is compared with the first environmental impact score and the second environmental impact score, and the optimization coefficient is determined based on the comparison result; wherein the first environmental impact score is less than the second environmental impact score. When the environmental impact score is less than or equal to the first environmental impact score, the optimization coefficient is determined to be the first optimization coefficient; When the environmental impact score is greater than the first environmental impact score and less than or equal to the second environmental impact score, the optimization coefficient is determined to be the second optimization coefficient. When the environmental impact score is greater than the second environmental impact score, the optimization coefficient is determined to be the third optimization coefficient; The optimized camera acquisition frequency is obtained by multiplying the optimization coefficient by the camera acquisition frequency.

[0013] Furthermore, when determining whether wildfires and treeline discharge faults exist based on the image features, the following steps are included: The image features include flame features, smoke features, and electric arc features; When any one or more of the flame features, smoke features, and electric arc features are detected in the image information, it is determined that there is a wildfire and tree line discharge fault. If none of the flame characteristics, smoke characteristics, and arc characteristics are identified, it is determined that there is no wildfire or tree line discharge fault.

[0014] Further, when determining the fault level based on the image features, the process includes: When the flame feature is identified, the flame area is extracted, and the flame area is compared with a flame area threshold. The fault level is determined based on the comparison result. If the flame area is less than or equal to the lower limit of the flame area threshold, the fault level is determined to be Level 1; If the flame area is within the flame area threshold, the fault level is determined to be level two; If the flame area is greater than or equal to the upper limit of the flame area threshold, the fault level is determined to be level three; When the smoke feature is detected, the smoke concentration is extracted, and the smoke concentration is compared with a smoke concentration threshold. The fault level is determined based on the comparison result. If the smoke concentration is less than or equal to the lower limit of the smoke concentration threshold, the fault level is determined to be Level 1; If the smoke concentration is within the smoke concentration threshold, the fault level is determined to be Level 2; if the smoke concentration is greater than or equal to the upper limit of the smoke concentration threshold, the fault level is determined to be Level 3. When the arc feature is identified, the arc intensity is extracted, the arc intensity is compared with the arc intensity threshold, and the fault level is determined based on the comparison result. If the arc intensity is less than or equal to the lower limit of the arc intensity threshold, the fault level is determined to be Level 1; If the arc intensity is within the arc intensity threshold, the fault level is determined to be level two; If the arc intensity is greater than or equal to the upper limit of the arc intensity threshold, the fault level is determined to be level three; When multiple features are identified simultaneously, the fault level corresponding to each image feature is combined, and the highest fault level is taken as the final fault level.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: The intelligent video recognition-based method for monitoring wildfires and tree-line discharge faults provided by this invention can achieve efficient and accurate monitoring of wildfires and tree-line discharge faults. By comprehensively considering the configuration information of the power lines to be monitored, the terrain of the monitored area, and tree growth information, a tree-line monitoring dataset is constructed to accurately assess the risk value of discharge faults. The camera acquisition frequency is dynamically adjusted according to the risk value, and the acquisition frequency is further optimized by combining real-time environmental information to ensure that key image information can be captured in a timely manner under different environmental conditions. This method also utilizes intelligent video recognition technology to analyze the acquired image information, accurately extract image features related to wildfires and tree-line discharge faults, and achieve rapid fault judgment. Furthermore, it can accurately determine the fault level based on different image features, such as flame area, smoke concentration, and arc intensity, providing a strong basis for subsequent fault handling and emergency response. Compared with traditional monitoring methods, this invention greatly improves the accuracy and timeliness of monitoring, reduces the safety hazards of power systems caused by wildfires and tree-line discharge faults, reduces potential economic losses, and ensures the stability and reliability of power supply, resulting in significant social and economic benefits. Meanwhile, the method is highly intelligent and automated, reducing the workload and errors of manual monitoring, improving work efficiency, and providing a more scientific and effective guarantee for the safe operation of the power system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of a wildfire and treeline discharge fault monitoring method based on intelligent video recognition provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] See Figure 1 As shown in some embodiments of this application, this embodiment provides a method for monitoring wildfire and treeline discharge faults based on intelligent video recognition, including the following steps: S100: Determine the power line to be monitored, collect the configuration information of the power line to be monitored, and determine the camera installation strategy based on the configuration information; S200: Determine the area to be monitored and number the area to be monitored; collect the terrain information and the growth information of the trees to be monitored within the area to be monitored; and construct a treeline monitoring dataset based on the number, terrain information and growth information. S300: Obtain the discharge fault risk value of the monitored power line and each monitored area based on the tree line monitoring dataset; S400: Determine the camera acquisition frequency in the area to be monitored based on the discharge fault risk value; S500: Collect real-time environmental information of the area to be monitored, and determine whether to optimize the camera's acquisition frequency based on the real-time environmental information; S600: If so, determine the environmental impact score based on the real-time environmental information, and determine the optimization coefficient based on the environmental impact score to obtain the optimized camera acquisition frequency; S700: Acquire image information of the area to be monitored at the optimized camera acquisition frequency, analyze the image information, extract image features related to wildfire and treeline discharge faults, and determine whether wildfire and treeline discharge faults exist based on the image features; S800: If so, determine the fault level based on the image features.

[0021] It is understood that the intelligent video recognition-based wildfire and treeline discharge fault monitoring method provided in this embodiment can achieve efficient and accurate monitoring of wildfire and treeline discharge faults. By comprehensively considering the configuration information of the power lines to be monitored, the terrain of the monitored area, and tree growth information, a treeline monitoring dataset is constructed to accurately assess the risk value of discharge faults. The camera acquisition frequency is dynamically adjusted based on the risk value, and the acquisition frequency is further optimized by combining real-time environmental information to ensure that key image information can be captured in a timely manner under different environmental conditions. This method also utilizes intelligent video recognition technology to analyze the acquired image information, accurately extract image features related to wildfire and treeline discharge faults, and achieve rapid fault judgment. Furthermore, it can accurately determine the fault level based on different image features, such as flame area, smoke concentration, and arc intensity, providing a strong basis for subsequent fault handling and emergency response. Compared with traditional monitoring methods, this invention greatly improves the accuracy and timeliness of monitoring, reduces the power system safety hazards caused by wildfires and treeline discharge faults, reduces potential economic losses, and ensures the stability and reliability of power supply, with significant social and economic benefits. Meanwhile, the method is highly intelligent and automated, reducing the workload and errors of manual monitoring, improving work efficiency, and providing a more scientific and effective guarantee for the safe operation of the power system.

[0022] Specifically, determining the camera installation strategy based on the configuration information includes: The configuration information is parsed to obtain the wire type of the wire to be monitored, and the camera installation strategy is determined based on the wire type. The types of wires mentioned include high-voltage wires, low-voltage wires, and ultra-high-voltage wires; When the type of power line is a high-voltage power line, the camera installation strategy is determined to be the first camera installation strategy; When the type of wire is low-voltage wire, the camera installation strategy is determined to be the second camera installation strategy; When the type of power line is an ultra-high voltage power line, the camera installation strategy is determined to be a third camera installation strategy.

[0023] Understandably, the first camera installation strategy involves a spacing of 400 meters and a height of 10 meters on the tower / pole; the second camera strategy involves a spacing of 600 meters and a height of 8 meters on the tower / pole; and the third camera strategy involves a spacing of 200 meters and a height of 12 meters on the tower / pole. Different camera installation strategies are used for different types of power lines because power lines of different voltage levels generate different electric and magnetic fields during operation and exhibit different characteristics when faults occur. High-voltage power lines have higher voltages, resulting in stronger electric and magnetic fields. When a line discharge fault occurs, the resulting arc and spark energy is larger, and the impact on the surrounding environment is relatively wider. Therefore, a relatively dense and higher camera layout is needed to ensure comprehensive capture of potential fault situations. Low-voltage power lines have lower voltages, and the impact range when a fault occurs is relatively smaller. Therefore, the camera installation spacing can be appropriately increased and the installation height lowered. However, ultra-high-voltage power lines, due to their extremely high voltage, can cause serious consequences if a fault occurs. Therefore, an even denser and higher camera installation method must be used to detect any possible signs of fault in a timely and accurate manner.

[0024] Specifically, when obtaining the discharge fault risk value of the monitored power line and each monitored area based on the tree line monitoring dataset, the process includes: The age, height, crown morphology characteristics of each tree to be monitored within the monitoring area, the slope, altitude, soil type of the monitoring area, and the erection height of the power line to be monitored are obtained. The contact risk assessment value between each tree to be monitored and the power line to be monitored is determined based on the age, height, canopy morphology characteristics and the erection height. The terrain impact risk assessment values ​​of the trees to be monitored and the power lines to be monitored are determined based on the slope, altitude, and soil type. A failure risk vector is constructed based on the contact risk assessment value and the terrain impact risk assessment value. The fault risk vector is compared with historical data. If there is a historical fault risk vector in the historical data that matches the fault risk vector, then the historical discharge fault risk value corresponding to the historical fault risk vector is used as the discharge fault risk value. Otherwise, the discharge fault risk value is determined based on the fault risk vector.

[0025] The calculation process for contact risk assessment is understandable: First, considering the age of the tree to be monitored and its growth rate and trend at different growth stages, its future height and growth direction are estimated. Generally, young trees grow rapidly and are more likely to increase in height, thus posing a higher risk of contact with the power lines. The current height of the tree is directly measured and compared with the height of the power line; if it approaches or exceeds the height of the power line, the contact risk increases significantly. Canopy morphology is also important; a large, lush canopy has a higher probability of contact with the power line, and its impact can be quantified by measuring the canopy diameter, shape, and foliage density. The contact risk is comprehensively assessed by combining the height of the power line with the tree height, growth trend, and canopy morphology. In the calculation, age, height, canopy morphology characteristics, and installation height are assigned weights, such as age 0.2, height 0.4, canopy morphology characteristics 0.3, and installation height 0.1. Different indicators have different risk coefficients. Regarding age, the risk coefficient is 0.8 for trees aged 1-5 years, 0.6 for 6-10 years, and 0.4 for trees over 10 years old. In terms of height, the risk coefficient is 0.9 for a height difference of less than 1 meter between the tree and the power line, 0.7 for 1-3 meters, and 0.5 for more than 3 meters. Regarding canopy morphology, a wide and dense canopy has a risk coefficient of 0.8, a medium to moderate canopy has 0.6, and a narrow and sparse canopy has 0.4. The risk coefficient is 0.7 for a height below 10 meters, 0.5 for 10-15 meters, and 0.3 for more than 15 meters. The contact risk assessment value is obtained by multiplying each risk coefficient by its corresponding weight and then summing them. For example, a 3-year-old tree with a height difference of 0.5 meters from the power line, a wide and dense canopy, and a power line erected at a height of 12 meters has an assessment value of 0.2×0.8+0.4×0.9+0.3×0.8+0.1×0.5=0.77.

[0026] Understandably, the calculation process for the risk assessment value of terrain impact is as follows: First, slope significantly affects the risk of contact between trees and power lines. A shallow slope has a small impact, while a steep slope increases the risk of tilting. For slopes less than 10 degrees, the risk coefficient is 0.2; for 10-30 degrees, it's 0.5; and for greater than 30 degrees, it's 0.8. Second, altitude affects tree growth and the risk of electrical faults. High-altitude areas have harsh climates, limiting tree growth or causing them to fall over. Below 500 meters, the risk coefficient is 0.3; between 500-1500 meters, it's 0.6; and above 1500 meters, it's 0.9. Third, soil type affects tree stability. Different soils have different bearing capacity and water permeability, affecting root growth. Clay retains water but has poor aeration, sandy soil is aerated but has poor water retention, and loam has moderate aeration and water retention. The risk coefficient is 0.7 for clay, 0.6 for sandy soil, and 0.4 for loam. Then, weights are assigned to slope, altitude, and soil type, such as slope 0.3, altitude 0.3, and soil type 0.4. Each risk coefficient is multiplied by its corresponding weight and then summed to obtain the assessment value. For example, in an area with a slope of 20 degrees, an altitude of 1000 meters, and clay soil, the assessment value is 0.3×0.5+0.3×0.6+0.4×0.7=0.61. This calculation process accurately assesses the impact of terrain on tree and power line discharge faults, providing support for subsequent monitoring and prevention.

[0027] Specifically, determining the discharge fault risk value based on the fault risk vector includes: The failure risk impact score is determined based on the failure risk vector. The fault risk impact score is compared with the first fault risk impact score and the second fault risk impact score, and the discharge fault risk value is determined based on the comparison result; wherein, the first fault risk impact score is less than the second fault risk impact score. When the fault risk impact score is less than or equal to the first fault risk impact score, the discharge fault risk value is determined to be the first discharge fault risk value. When the fault risk impact score is greater than the first fault risk impact score and less than or equal to the second fault risk impact score, the discharge fault risk value is determined to be the second discharge fault risk value. When the fault risk impact score is greater than the second fault risk impact score, the discharge fault risk value is determined to be the third discharge fault risk value.

[0028] It is understandable that the order of the fault risk scores is: first discharge fault risk value < second discharge fault risk value < third discharge fault risk value. Different discharge fault risk values ​​correspond to different monitoring strategies and emergency measures.

[0029] Specifically, when determining the camera acquisition frequency in the monitored area based on the discharge fault risk value, the following steps are included: The discharge fault risk value is compared with a preset acquisition frequency mapping table, and the acquisition frequency of the camera is determined based on the comparison result.

[0030] Understandably, a preset acquisition frequency mapping table is a pre-defined table that records the mapping relationship between different discharge fault risk values ​​and corresponding camera acquisition frequencies. In this table, discharge fault risk values ​​are divided into multiple ranges, each corresponding to a specific camera acquisition frequency. For example, when the discharge fault risk value is in a low range, the corresponding camera acquisition frequency might be relatively low, such as acquiring image information once every 10 minutes; when the discharge fault risk value is in a medium range, the camera acquisition frequency will increase accordingly, perhaps to once every 5 minutes; and when the discharge fault risk value is in a high range, the camera acquisition frequency will further increase, perhaps reaching once every 1 minute. Through this mapping relationship, the camera acquisition frequency can be dynamically adjusted according to different discharge fault risk values, thereby ensuring timely capture of image information related to potential wildfires and treeline discharge faults while avoiding unnecessary resource waste, improving the efficiency and economy of the monitoring system. Furthermore, this preset acquisition frequency mapping table will be continuously optimized and adjusted based on actual monitoring conditions and historical data to ensure its accuracy and practicality.

[0031] Specifically, when determining whether to optimize the camera's acquisition frequency based on the real-time environmental information, the process includes: Feature extraction is performed on the real-time environment feature values ​​to obtain several real-time environment feature values; Obtain the real-time environmental standard value corresponding to each of the aforementioned real-time environmental feature values; If any of the real-time environmental feature values ​​exceed the corresponding real-time environmental standard value, then it is determined that the camera acquisition frequency should be optimized. Otherwise, it is determined that the camera's acquisition frequency will not be optimized.

[0032] Understandably, the preferred real-time environmental characteristics are temperature, humidity, wind speed, and light intensity. These environmental factors all influence the probability of wildfires and tree-line discharge faults. For example, hot and dry environments increase the flammability of trees, while strong winds accelerate the spread of fires and the occurrence of discharge faults. Light intensity may also affect the camera's image quality, thus impacting fault monitoring.

[0033] Specifically, determining the environmental impact score based on the real-time environmental information includes: Obtain each real-time environmental feature value that exceeds the corresponding real-time environmental standard value, and construct a problem environment feature dataset; Obtain the significance score of each real-time environmental feature value in the problem environment feature dataset out of all real-time environmental feature values; Obtain the deviation index between each real-time environmental feature value and the corresponding real-time environmental standard value in the problem environment feature dataset; Based on the importance score and deviation index, calculate the environmental impact sub-score for each real-time environmental feature value; The environmental impact score is obtained by summing the environmental impact sub-scores of all real-time environmental feature values.

[0034] Understandably, the deviation index is a quantitative indicator of the degree of deviation between a real-time environmental characteristic value and its corresponding standard value. The calculation method for the deviation index differs for different real-time environmental characteristics. Taking temperature as an example, the absolute value of the difference between the real-time and standard temperatures is first calculated, then divided by the standard temperature to obtain the relative deviation. This relative deviation is then converted into a deviation index according to rules. For example, if the real-time temperature is 35℃ and the standard temperature is 25℃, the relative deviation is 0.4, and the deviation index is 2. For humidity, the difference between the real-time and standard humidity is considered. A larger difference results in a higher deviation index; for example, if the standard humidity is 60% and the real-time humidity is 30%, the deviation index may be 3. A smaller difference results in a lower deviation index; for example, if the real-time humidity is 55%, the deviation index may be 1. The wind speed deviation index is calculated by considering its impact on wildfire spread and electrical discharge faults. When the wind speed exceeds the standard by a certain extent, the deviation index increases significantly; for example, if the standard wind speed is 5 m / s and the real-time wind speed is 15 m / s, the deviation index may be 4. The light intensity deviation index is determined based on its impact on the camera's shooting effect. Excessive or insufficient light affects image quality. For example, if the standard light intensity is 500 lux and the real-time light intensity is 100 lux, the deviation index may be 2.

[0035] Understandably, the calculation process for the environmental impact sub-score is as follows: First, the importance score is determined. Different real-time environmental characteristics have different levels of importance in influencing the occurrence of wildfires and tree-line discharge faults. For example, in dry seasons, temperature and humidity are relatively more important and can be assigned higher importance scores, such as a temperature importance score of 0.4 and a humidity importance score of 0.3; while wind speed and light intensity are relatively less important and can be assigned lower importance scores, such as a wind speed importance score of 0.2 and a light intensity importance score of 0.1. Then, the importance score of each real-time environmental characteristic is multiplied by the deviation index corresponding to that characteristic value to obtain the environmental impact sub-score for that real-time environmental characteristic value. For example, if the importance score for temperature is 0.4 and the deviation index is 2, then the environmental impact sub-score for temperature is 0.4 × 2 = 0.8; if the importance score for humidity is 0.3 and the deviation index is 3, then the environmental impact sub-score for humidity is 0.3 × 3 = 0.9; if the importance score for wind speed is 0.2 and the deviation index is 4, then the environmental impact sub-score for wind speed is 0.2 × 4 = 0.8; and if the importance score for light intensity is 0.1 and the deviation index is 2, then the environmental impact sub-score for light intensity is 0.1 × 2 = 0.2. Through this calculation process, the impact of each real-time environmental characteristic value exceeding the standard value on wildfires and tree-line discharge faults can be accurately measured. Furthermore, by adding all the environmental impact sub-scores, the environmental impact score is obtained, providing strong data support for subsequent decisions such as optimizing camera acquisition frequencies.

[0036] Specifically, determining the optimization coefficient based on the environmental impact score to obtain the optimized camera acquisition frequency includes: The environmental impact score is compared with the first environmental impact score and the second environmental impact score, and the optimization coefficient is determined based on the comparison result; wherein the first environmental impact score is less than the second environmental impact score. When the environmental impact score is less than or equal to the first environmental impact score, the optimization coefficient is determined to be the first optimization coefficient; When the environmental impact score is greater than the first environmental impact score and less than or equal to the second environmental impact score, the optimization coefficient is determined to be the second optimization coefficient. When the environmental impact score is greater than the second environmental impact score, the optimization coefficient is determined to be the third optimization coefficient; The optimized camera acquisition frequency is obtained by multiplying the optimization coefficient by the camera acquisition frequency.

[0037] It's understandable that the optimization coefficients follow the order: First optimization coefficient < Second optimization coefficient < Third optimization coefficient. Different optimization coefficients correspond to different levels of adjustment in camera acquisition frequency. When the optimization coefficient is small, it means the real-time environment has a relatively small impact on wildfires and tree-line discharge faults. In this case, the increase in camera acquisition frequency is not significant, maintaining a reasonable and economical level. For example, with a first optimization coefficient of 1.2, if the original camera acquisition frequency is once every 5 minutes, the optimized acquisition frequency becomes once every 4.17 minutes (5 ÷ 1.2 ≈ 4.17). When the optimization coefficient is at a medium level, it indicates that the real-time environment has an increased impact on fault occurrence, requiring an appropriate increase in camera acquisition frequency to capture potential fault information more promptly. For example, with a second optimization coefficient of 1.5, the original acquisition frequency remains once every 5 minutes, but after optimization, it becomes once every 3.33 minutes (5 ÷ 1.5 ≈ 3.33). When the optimization coefficient is large, it indicates that the real-time environment has a significant impact on the occurrence of wildfires and tree-line discharge faults. In this case, the camera acquisition frequency needs to be significantly increased to ensure that relevant image information can be acquired as soon as possible. Assuming the third optimization coefficient is 2, and the original acquisition frequency is once every 5 minutes, then the optimized acquisition frequency becomes once every 2.5 minutes.

[0038] Specifically, when determining whether wildfires and treeline discharge faults exist based on the image features, the following steps are included: The image features include flame features, smoke features, and electric arc features; When any one or more of the flame features, smoke features, and electric arc features are detected in the image information, it is determined that there is a wildfire and tree line discharge fault. If none of the flame characteristics, smoke characteristics, and arc characteristics are identified, it is determined that there is no wildfire or tree line discharge fault.

[0039] Understandably, flame characteristics can be identified through color, shape, and dynamic changes. Wildfire flames typically appear bright orange-red or yellow, with irregular shapes and constantly flickering. In video images, color models can be used to extract potential flame areas, and combined with dynamic characteristics such as flicker frequency and area changes, further confirmation can be made as to whether it is a real flame. Smoke characteristics are mainly judged based on its color, concentration, and diffusion pattern. Generally, smoke from wildfires is mostly gray, black, or white, and its concentration and diffusion range change as the fire develops. Smoke concentration can be analyzed through image grayscale values ​​and texture features, and morphological algorithms can be used to identify the diffusion pattern of the smoke. The identification of electric arc characteristics is relatively complex. Electric arcs typically manifest as instantaneous flashes of intense light, with high brightness and specific spectral characteristics. The presence of electric arcs can be detected through brightness changes and spectral analysis of images. In practical applications, to improve the accuracy of identification, multiple frames of images can be analyzed, and the feature changes over time can be used to further confirm the presence of wildfires and tree-line discharge faults. At the same time, machine learning and deep learning algorithms can be introduced to train on a large number of sample images to improve the recognition accuracy of flame, smoke and electric arc features.

[0040] Specifically, determining the fault level based on the image features includes: When the flame feature is identified, the flame area is extracted, and the flame area is compared with a flame area threshold. The fault level is determined based on the comparison result. If the flame area is less than or equal to the lower limit of the flame area threshold, the fault level is determined to be Level 1; If the flame area is within the flame area threshold, the fault level is determined to be level two; If the flame area is greater than or equal to the upper limit of the flame area threshold, the fault level is determined to be level three; When the smoke feature is detected, the smoke concentration is extracted, and the smoke concentration is compared with a smoke concentration threshold. The fault level is determined based on the comparison result. If the smoke concentration is less than or equal to the lower limit of the smoke concentration threshold, the fault level is determined to be Level 1; If the smoke concentration is within the smoke concentration threshold, the fault level is determined to be Level 2; if the smoke concentration is greater than or equal to the upper limit of the smoke concentration threshold, the fault level is determined to be Level 3. When the arc feature is identified, the arc intensity is extracted, the arc intensity is compared with the arc intensity threshold, and the fault level is determined based on the comparison result. If the arc intensity is less than or equal to the lower limit of the arc intensity threshold, the fault level is determined to be Level 1; If the arc intensity is within the arc intensity threshold, the fault level is determined to be level two; If the arc intensity is greater than or equal to the upper limit of the arc intensity threshold, the fault level is determined to be level three; When multiple features are identified simultaneously, the fault level corresponding to each image feature is combined, and the highest fault level is taken as the final fault level.

[0041] Understandably, different fault levels correspond to different emergency response measures. Level 1 fault indicates that the wildfire and tree-line discharge fault are in their initial stages, with a small impact area. At this stage, small-scale monitoring and early warning measures can be implemented. For example, nearby patrol personnel can be notified to strengthen patrols of the area, closely monitor the development of the fault, and relevant early warning systems can be activated to send low-level warning information to potentially affected individuals. Level 2 fault means the fault has developed to a certain extent, requiring more proactive measures. At this stage, sufficient fire-fighting resources and maintenance personnel can be deployed to the scene to prepare for firefighting and fault repair, while expanding the warning area, raising the warning level, and alerting more people to be aware of safety. Level 3 fault indicates that the fault is already quite serious and may have a significant impact on the surrounding environment and facilities. At this stage, the emergency plan should be activated immediately, mobilizing a large number of fire-fighting, rescue, and maintenance forces to the scene to carry out firefighting and fault repair work. Simultaneously, nearby residents and workers should be evacuated promptly to ensure their safety, and the situation should be reported to higher authorities to seek further support and assistance. Furthermore, corresponding resource allocation plans and follow-up recovery schemes can be developed for different fault levels to ensure that normal production and daily life can be restored as quickly as possible after the fault is resolved. Throughout the monitoring and handling process, a comprehensive information feedback mechanism should be established to promptly report the development of the fault and the progress of its resolution to higher-level departments and relevant personnel, enabling more scientific and rational decision-making.

[0042] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0043] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for monitoring wildfires and treeline discharge faults based on intelligent video recognition, characterized in that, include: Identify the power line to be monitored, collect the configuration information of the power line to be monitored, and determine the camera installation strategy based on the configuration information; The area to be monitored is determined and numbered. Topographic information and growth information of trees to be monitored are collected within the area to be monitored. A treeline monitoring dataset is constructed based on the number, topographic information and growth information. The discharge fault risk value of the monitored power line and each monitored area is obtained based on the tree line monitoring dataset. The camera acquisition frequency in the area to be monitored is determined based on the discharge fault risk value. Collect real-time environmental information of the area to be monitored, and determine whether to optimize the camera's acquisition frequency based on the real-time environmental information; If so, determine the environmental impact score based on the real-time environmental information, and determine the optimization coefficient based on the environmental impact score to obtain the optimized camera acquisition frequency; The image information of the area to be monitored is acquired at the optimized camera acquisition frequency, and the image information is analyzed to extract image features related to wildfire and treeline discharge faults. Based on the image features, it is determined whether wildfire and treeline discharge faults exist. If so, the fault level is determined based on the image features.

2. The method for monitoring wildfires and treeline discharge faults based on intelligent video recognition according to claim 1, characterized in that, When determining the camera installation strategy based on the configuration information, the following are included: The configuration information is parsed to obtain the wire type of the wire to be monitored, and the camera installation strategy is determined based on the wire type. The types of wires mentioned include high-voltage wires, low-voltage wires, and ultra-high-voltage wires; When the type of power line is a high-voltage power line, the camera installation strategy is determined to be the first camera installation strategy; When the type of wire is low-voltage wire, the camera installation strategy is determined to be the second camera installation strategy; When the type of power line is an ultra-high voltage power line, the camera installation strategy is determined to be a third camera installation strategy.

3. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 2, characterized in that, When obtaining the discharge fault risk value of the monitored power line and each monitored area based on the treeline monitoring dataset, the following are included: The age, height, crown morphology characteristics of each tree to be monitored within the monitoring area, the slope, altitude, soil type of the monitoring area, and the erection height of the power line to be monitored are obtained. The contact risk assessment value between each tree to be monitored and the power line to be monitored is determined based on the age, height, canopy morphology characteristics and the erection height. The terrain impact risk assessment values ​​of the trees to be monitored and the power lines to be monitored are determined based on the slope, altitude, and soil type. A failure risk vector is constructed based on the contact risk assessment value and the terrain impact risk assessment value. The fault risk vector is compared with historical data. If there is a historical fault risk vector in the historical data that matches the fault risk vector, then the historical discharge fault risk value corresponding to the historical fault risk vector is taken as the discharge fault risk value. Otherwise, the discharge fault risk value is determined based on the fault risk vector.

4. The method for monitoring wildfires and treeline discharge faults based on intelligent video recognition according to claim 3, characterized in that, When determining the discharge fault risk value based on the fault risk vector, the following are included: The failure risk impact score is determined based on the failure risk vector. The fault risk impact score is compared with the first fault risk impact score and the second fault risk impact score, and the discharge fault risk value is determined based on the comparison result; wherein, the first fault risk impact score is less than the second fault risk impact score. When the fault risk impact score is less than or equal to the first fault risk impact score, the discharge fault risk value is determined to be the first discharge fault risk value. When the fault risk impact score is greater than the first fault risk impact score and less than or equal to the second fault risk impact score, the discharge fault risk value is determined to be the second discharge fault risk value. When the fault risk impact score is greater than the second fault risk impact score, the discharge fault risk value is determined to be the third discharge fault risk value.

5. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 4, characterized in that, When determining the camera acquisition frequency in the monitored area based on the discharge fault risk value, the following steps are included: The discharge fault risk value is compared with a preset acquisition frequency mapping table, and the acquisition frequency of the camera is determined based on the comparison result.

6. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 5, characterized in that, When determining whether to optimize the camera's acquisition frequency based on the real-time environmental information, the following are included: Feature extraction is performed on the real-time environment feature values ​​to obtain several real-time environment feature values; Obtain the real-time environmental standard value corresponding to each of the aforementioned real-time environmental feature values; If any of the real-time environmental feature values ​​exceed the corresponding real-time environmental standard value, then it is determined that the camera acquisition frequency should be optimized. Otherwise, it is determined that the camera's acquisition frequency will not be optimized.

7. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 6, characterized in that, When determining the environmental impact score based on the real-time environmental information, the following are included: Obtain each real-time environmental feature value that exceeds the corresponding real-time environmental standard value, and construct a problem environment feature dataset; Obtain the significance score of each real-time environmental feature value in the problem environment feature dataset among all real-time environmental feature values; Obtain the deviation index between each real-time environmental feature value and the corresponding real-time environmental standard value in the problem environment feature dataset; Based on the importance score and deviation index, calculate the environmental impact sub-score for each real-time environmental feature value; The environmental impact score is obtained by summing the environmental impact sub-scores of all real-time environmental feature values.

8. The method for monitoring wildfires and treeline discharge faults based on intelligent video recognition according to claim 7, characterized in that, When determining the optimization coefficients based on the environmental impact score and obtaining the optimized camera acquisition frequency, the following steps are included: The environmental impact score is compared with the first environmental impact score and the second environmental impact score, and the optimization coefficient is determined based on the comparison result; wherein the first environmental impact score is less than the second environmental impact score. When the environmental impact score is less than or equal to the first environmental impact score, the optimization coefficient is determined to be the first optimization coefficient; When the environmental impact score is greater than the first environmental impact score and less than or equal to the second environmental impact score, the optimization coefficient is determined to be the second optimization coefficient. When the environmental impact score is greater than the second environmental impact score, the optimization coefficient is determined to be the third optimization coefficient; The optimized camera acquisition frequency is obtained by multiplying the optimization coefficient by the camera acquisition frequency.

9. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 8, characterized in that, When determining whether wildfires and treeline discharge faults exist based on the image features, the following are included: The image features include flame features, smoke features, and electric arc features; When any one or more of the flame features, smoke features, and electric arc features are detected in the image information, it is determined that there is a wildfire and tree line discharge fault. If none of the flame characteristics, smoke characteristics, and arc characteristics are identified, it is determined that there is no wildfire or tree line discharge fault.

10. The method for monitoring wildfire and treeline discharge faults based on intelligent video recognition according to claim 9, characterized in that, Determining the fault level based on the image features includes: When the flame feature is identified, the flame area is extracted, and the flame area is compared with a flame area threshold. The fault level is determined based on the comparison result. If the flame area is less than or equal to the lower limit of the flame area threshold, the fault level is determined to be Level 1; If the flame area is within the flame area threshold, the fault level is determined to be level two; If the flame area is greater than or equal to the upper limit of the flame area threshold, the fault level is determined to be level three; When the smoke feature is detected, the smoke concentration is extracted, and the smoke concentration is compared with a smoke concentration threshold. The fault level is determined based on the comparison result. If the smoke concentration is less than or equal to the lower limit of the smoke concentration threshold, the fault level is determined to be Level 1; If the smoke concentration is within the smoke concentration threshold, the fault level is determined to be Level 2; if the smoke concentration is greater than or equal to the upper limit of the smoke concentration threshold, the fault level is determined to be Level 3. When the arc feature is identified, the arc intensity is extracted, the arc intensity is compared with the arc intensity threshold, and the fault level is determined based on the comparison result. If the arc intensity is less than or equal to the lower limit of the arc intensity threshold, the fault level is determined to be Level 1; If the arc intensity is within the arc intensity threshold, the fault level is determined to be level two; If the arc intensity is greater than or equal to the upper limit of the arc intensity threshold, the fault level is determined to be level three; When multiple features are identified simultaneously, the fault level corresponding to each image feature is combined, and the highest fault level is taken as the final fault level.

Images