A power transmission channel forest fire danger positioning method and system based on multi-source fusion

Abstract

The application discloses a power transmission channel mountain fire danger positioning method and system based on multi-source fusion, adopts multi-source fusion of images, infrared imaging and three-dimensional point clouds, monitors the environment around the power transmission line channel all day long, adopts a hierarchical monitoring and early warning mechanism, can accurately position a mountain fire hidden danger, and can meet the system requirements of low power consumption and high real-time performance. Meanwhile, the application can perform danger evaluation and dynamic early warning on different mountain fire hidden dangers, so that accurate and forward-looking guidance is given to power transmission line patrol personnel.

Description

Technical Field

[0001] This invention belongs to the field of online monitoring of power transmission and transformation, and in particular relates to a method and system for locating wildfire hazards in power transmission channels based on multi-source fusion. Background Technology

[0002] Wildfire hazards along power transmission lines have always been a thorny issue affecting the safe and stable operation of the power grid. In recent years, with the rapid development of image AI recognition technology, image tracking and recognition methods have gradually been applied to power transmission line corridors, offering some early warning capabilities for wildfire hazards. However, their actual effectiveness is limited by the following fundamental constraints, making it difficult to meet the needs of accurate early warning:

[0003] 1. Insufficient reliability in wildfire identification, with a high false alarm rate. Relying solely on image cameras, visual recognition is highly susceptible to complex external factors such as image quality, hazard distance, weather, and day / night cycles. Especially in long-distance scenarios, wildfire targets occupy a small percentage of pixels in the overall image, resulting in weak features and frequent missed and false alarms.

[0004] 2. It is impossible to accurately determine the distance and severity of wildfire hazards. Furthermore, images alone cannot assess the severity of a fire, making it difficult to provide data support for response strategies.

[0005] 3. It is impossible to continuously and accurately track the spread and direction of wildfire hazards. Early warning information is seriously lagging behind the dynamic development of the fire situation, and cannot provide accurate and forward-looking guidance for power line patrol personnel. Summary of the Invention

[0006] Purpose of the invention: The purpose of this invention is to provide a method and system for locating wildfire hazards in power transmission channels based on multi-source fusion. Through the organic coordination and deep fusion of multi-source information, it can simultaneously complete the reliable identification, accurate location, hazard assessment and dynamic early warning of wildfire hazards. At the same time, it can meet the system requirements of low power consumption and high real-time performance, and strongly support the safe operation of power transmission lines.

[0007] Technical solution: The method for locating wildfire hazards in power transmission channels based on multi-source fusion includes:

[0008] Based on the three-dimensional point cloud image of the power transmission channel, establish the power transmission line equation for each span that can be scanned;

[0009] A preliminary assessment of wildfire hazards is made based on smoke recognition algorithms in the monitored area images. If a wildfire hazard is preliminarily identified, a three-dimensional temperature field is generated based on the infrared scanning results of the suspected wildfire area. Effective hazard points are identified from the three-dimensional temperature field, and several real wildfire hazard datasets are constructed based on the effective hazard points. The range of each wildfire hazard target is calculated. If no wildfire hazard is found in the preliminary assessment, image monitoring of the area continues.

[0010] Locate the three-dimensional coordinates of suspected wildfire hazards and calculate the distance from the actual wildfire hazard to the power transmission channel based on the aforementioned power transmission line equation;

[0011] Based on the scope of the stated wildfire hazard targets and the distance from the actual wildfire hazard to the power transmission corridor, different risk levels of wildfire hazards to the power transmission corridor are comprehensively given.

[0012] Optionally, the three-dimensional point cloud image of the power transmission channel is obtained by scanning the power transmission channel with a lidar.

[0013] Optionally, the image of the monitored area is obtained using an image camera.

[0014] Optionally, when a preliminary assessment of potential wildfire hazards is made, infrared cameras and lidar can be activated to complete infrared scanning and location of suspected wildfire areas.

[0015] Optionally, the method further includes: providing a corresponding warning level based on the rate of development of the danger level.

[0016] Optionally, the generation of the early warning level includes: periodically updating the range of wildfire hazards and its distance to power transmission channels, calculating the wildfire spread speed and direction, and issuing a wildfire danger warning based on the range of wildfire hazards, its distance to power transmission channels, the wildfire spread speed, and the direction of spread.

[0017] Optionally, the step of identifying effective hazard points from the three-dimensional temperature field and constructing several real wildfire hazard datasets based on the effective hazard points includes:

[0018] Several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are designated as effective hazard points, and the current highest temperature point is designated as the hazard center point. The distance between each effective hazard point and the hazard center point is calculated, and all effective hazard points whose distance is less than the set reference threshold are added to the current target data subset. If there are multiple current highest temperature points, one of them is randomly selected as the hazard center point, and the other highest temperature points are assigned to the same target data subset, or a new target data subset is established with another highest temperature point as the center point.

[0019] Repeat the above steps for the remaining potential hazards to construct multiple target data subsets;

[0020] The presence of interference is determined based on the number of valid points of potential hazards in the target data subset. If interference exists, the interfering data subset is removed; otherwise, the target data subset is deemed valid and is considered a real wildfire hazard target dataset.

[0021] Optionally, the calculation of the range of each wildfire hazard target includes: calculating the contour feature data of each real wildfire hazard target dataset, and generating a contour volume based on the contour feature data to characterize the range of the wildfire hazard target; wherein, the contour feature data includes the center coordinates and the longest radius of each real wildfire hazard target dataset.

[0022] Optionally, the three-dimensional coordinates for locating suspected wildfire hazards are achieved in the following way:

[0023] The image camera, infrared camera, and lidar are calibrated and registered to establish a mapping relationship between two-dimensional image coordinates and three-dimensional point cloud coordinates. When all three scan a suspicious area at the same time, the position of the corresponding three-dimensional point cloud image is located according to the two-dimensional image coordinates, which is the three-dimensional coordinate of the suspected wildfire hazard.

[0024] Optionally, the calculation of the distance from the actual wildfire hazard to the power transmission corridor based on the power transmission line equation includes:

[0025] Outskirts of wildfire hazard points Substituting the equations of the power transmission lines for the multiple power transmission channels that may be threatened by wildfires, we obtain...

[0026]

[0027] in, , For the first The coefficients of the straight line equation for a transmission line span of a certain length. For constant terms;

[0028] The shortest distance among them was selected as the distance from the actual wildfire hazard to the power transmission line.

[0029] The power transmission channel wildfire hazard location system based on multi-source fusion includes:

[0030] The sensing module includes an image camera, an infrared camera, a lidar, and a hierarchical trigger control unit. The image camera, infrared camera, and lidar are integrated on a gimbal with a fixed relative spatial relationship and are driven by the gimbal to perform common field-of-view scanning. The hierarchical trigger control unit is used to control the start and stop of the infrared camera and lidar based on the preliminary wildfire hazard assessment results from the image camera.

[0031] The data fusion and analysis module is communicatively connected to the perception triggering module and is used to process multi-source fusion data collected after the infrared camera and lidar are activated, in order to identify and locate the fire scene and assess the risk.

[0032] The infrared camera can provide temperature values ​​at various locations, and by combining these values ​​with the coordinates of the laser point cloud, the spatial location of potential wildfire hazards can be obtained.

[0033] Optionally, the data fusion and analysis module includes:

[0034] The fire quantification unit is used to generate a three-dimensional temperature field based on the scanning results of the infrared camera on the suspected wildfire area, identify several real wildfire hazards based on the three-dimensional temperature field, and obtain the quantification results of each target.

[0035] The precise positioning unit is used to calculate the distance from the actual wildfire hazard to the power transmission channel based on the three-dimensional coordinates of the suspected wildfire hazard obtained after calibration and registration by the sensing module and a pre-stored power transmission line model.

[0036] The risk assessment unit is used to determine the hazard level based on the quantitative results of the actual wildfire hazards and their distance to the power transmission channel.

[0037] Optionally, the fire quantification unit, after identifying several real wildfire hazards based on the three-dimensional temperature field and obtaining the quantification results of each target, performs the following steps:

[0038] Several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are designated as effective points of potential hazards, and the current point with the highest temperature is designated as the center point of potential hazards.

[0039] Calculate the distance between each valid point of a hazard and the center point of the hazard, and add all valid points of hazard whose distance is less than the set reference threshold to the current target data subset;

[0040] Repeat the above steps for the remaining potential hazards to construct multiple target data subsets;

[0041] The presence of interference is determined based on the number of target data subsets, and the valid target data subsets are used as the target dataset for real wildfire hazards.

[0042] Optionally, the system also includes an early warning module;

[0043] After the real wildfire hazard identification stage, the hierarchical trigger control unit controls the infrared camera and lidar to be in working state, and together with the image camera, they perform timed scanning.

[0044] The early warning module is used to calculate the coordinates of the wildfire hazard and its distance to the power transmission channel based on the timed scanning results, compare it with the previous calculation data, and calculate the spread speed and direction of the wildfire hazard with reference to the scanning time; at the same time, it updates the range of the wildfire hazard; and based on the range of the wildfire hazard and its distance to the power transmission channel, the spread speed and direction of the wildfire, it issues a wildfire danger warning.

[0045] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0046] 1. Based on the multi-source fusion of monitoring images, infrared imaging data and three-dimensional point cloud images, this invention can monitor the surrounding environment of power transmission line channels around the clock and in all weathers, and promptly and accurately detect potential wildfire hazards, especially the identification of potential wildfire hazards at a distance.

[0047] 2. This invention employs a tiered monitoring and early warning mechanism. Initially, it assesses potential wildfire hazards solely based on images. Only when a hazard is confirmed will infrared scanning and laser ranging be performed, significantly reducing energy consumption for real-time monitoring. Infrared scanning results confirm the hazard's authenticity, and the location data obtained from infrared and laser calibration determines the hazard's extent and distance from power transmission lines, thus revealing the wildfire's spread (including direction and range). Even if some small targets are initially undetectable due to limitations in image cameras and existing smoke recognition algorithms, once a valid hazard point is identified, this invention can quickly and accurately locate and provide tiered early warnings of wildfire danger, keeping the fire under control and promptly notifying relevant departments to take action.

[0048] 3. This invention can comprehensively provide the danger level and early warning level of different wildfire hazards, and provide accurate hazard locations, thereby strongly supporting the safe operation of power transmission lines. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the method flow of the present invention in one embodiment;

[0050] Figure 2 This is a schematic diagram showing the distribution of effective points of potential hazards in one embodiment;

[0051] Figure 3 This is a schematic diagram of a power transmission line in one embodiment. Detailed Implementation

[0052] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0053] A method for locating wildfire hazards along power transmission corridors based on multi-source fusion, characterized in that the method includes:

[0054] Based on the three-dimensional point cloud image of the power transmission channel, establish the power transmission line equation for each span that can be scanned;

[0055] Preliminary assessment of wildfire hazards based on smoke recognition algorithms in images of the monitored area;

[0056] If a preliminary assessment indicates the presence of a wildfire hazard, a three-dimensional temperature field is generated based on the infrared scanning results of the suspected wildfire area. Effective points of the hazard are identified from the three-dimensional temperature field, and several real wildfire hazard datasets are constructed based on these effective points. The range of each wildfire hazard target is then calculated.

[0057] Locate the three-dimensional coordinates of suspected wildfire hazards and calculate the distance from the actual wildfire hazard to the power transmission channel based on the aforementioned power transmission line equation;

[0058] Based on the scope of the stated wildfire hazard targets and the distance from the actual wildfire hazard to the power transmission corridor, different risk levels of wildfire hazards to the power transmission corridor are comprehensively given.

[0059] In one embodiment, the method further includes: periodically updating the range of wildfire hazards and its distance to power transmission channels, calculating the wildfire spread speed and direction, and issuing wildfire danger warnings based on the range of wildfire hazards, its distance to power transmission channels, the wildfire spread speed, and the direction of spread.

[0060] In one implementation, identifying effective hazard points from the three-dimensional temperature field and constructing several real wildfire hazard datasets based on these effective hazard points includes:

[0061] Several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are designated as effective hazard points, and the current highest temperature point is designated as the hazard center point. The distance between each effective hazard point and the hazard center point is calculated, and all effective hazard points whose distance is less than the set reference threshold are added to the current target data subset. If there are multiple current highest temperature points, one of them is randomly selected as the hazard center point, and the other highest temperature points are assigned to the same target data subset, or a new target data subset is established with another highest temperature point as the center point.

[0062] Repeat the above steps for the remaining potential hazards to construct multiple target data subsets;

[0063] The presence of interference is determined based on the number of valid points of potential hazards in the target data subset. If interference exists, the interfering data subset is removed; otherwise, the target data subset is deemed valid and is considered a real wildfire hazard target dataset.

[0064] In one specific implementation, the calculation of the range of each wildfire hazard target includes: calculating the contour feature data of each real wildfire hazard target dataset, and generating a contour volume based on the contour feature data to characterize the range of the wildfire hazard target; wherein, the contour feature data includes, but is not limited to, the center coordinates and the longest radius of each real wildfire hazard target dataset.

[0065] In one specific implementation, the three-dimensional coordinates for locating suspected wildfire hazards are achieved in the following way:

[0066] The image camera, infrared camera, and lidar are calibrated and registered to establish a mapping relationship between two-dimensional image coordinates and three-dimensional point cloud coordinates. When all three scan a suspicious area at the same time, the position of the corresponding three-dimensional point cloud image is located according to the two-dimensional image coordinates, which is the three-dimensional coordinate of the suspected wildfire hazard.

[0067] In a preferred embodiment, calculating the distance from the actual wildfire hazard to the power transmission corridor based on the power transmission straight line equation includes:

[0068] Outskirts of wildfire hazard points Substituting the equations of the power transmission lines for the multiple power transmission channels that may be threatened by wildfires, we obtain...

[0069]

[0070] in, , For the first The coefficients of the straight line equation for a transmission line span of a certain length. For constant terms;

[0071] The shortest distance among them was selected as the distance from the actual wildfire hazard to the power transmission line.

[0072] This invention also provides a multi-source fusion-based power transmission channel wildfire hazard location system, mainly comprising:

[0073] The sensing module includes an image camera, an infrared camera, a lidar, and a hierarchical trigger control unit. The image camera, infrared camera, and lidar are integrated on a gimbal with a fixed relative spatial relationship and are driven by the gimbal to perform common field-of-view scanning. The hierarchical trigger control unit is used to control the start and stop of the infrared camera and lidar based on the preliminary wildfire hazard assessment results from the image camera.

[0074] The data fusion and analysis module is communicatively connected to the perception triggering module and is used to process multi-source fusion data collected after the infrared camera and lidar are activated, in order to identify and locate the fire scene and assess the risk.

[0075] Furthermore, the data fusion and analysis module includes:

[0076] The fire quantification unit is used to generate a three-dimensional temperature field based on the scanning results of the infrared camera on the suspected wildfire area, identify several real wildfire hazards based on the three-dimensional temperature field, and obtain the quantification results of each target.

[0077] The precise positioning unit is used to calculate the distance from the actual wildfire hazard to the power transmission channel based on the three-dimensional coordinates of the suspected wildfire hazard obtained after calibration and registration by the sensing module and a pre-stored power transmission line model.

[0078] The risk assessment unit is used to determine the hazard level based on the quantitative results of the actual wildfire hazards and their distance to the power transmission channel.

[0079] In some specific implementations, the fire quantification unit identifies several real wildfire hazards based on the three-dimensional temperature field and performs the following steps after obtaining the quantification results for each target:

[0080] Several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are designated as effective hazard points, and the current highest temperature point is designated as the hazard center point. The distance between each effective hazard point and the hazard center point is calculated, and all effective hazard points whose distance is less than the set reference threshold are added to the current target data subset. If there are multiple current highest temperature points, one of them is randomly selected as the hazard center point, and the other highest temperature points are assigned to the same target data subset, or a new target data subset is established with another highest temperature point as the center point.

[0081] Repeat the above steps for the remaining potential hazards to construct multiple target data subsets;

[0082] The presence of interference is determined based on the number of valid points of potential hazards in the target data subset. If interference exists, the interfering data subset is removed; otherwise, the target data subset is deemed valid and is considered a real wildfire hazard target dataset.

[0083] Furthermore, the system also includes an early warning module;

[0084] After the real wildfire hazard identification stage, the hierarchical trigger control unit controls the infrared camera and lidar to be in working state, and together with the image camera, they perform timed scanning.

[0085] The early warning module is used to calculate the coordinates of the wildfire hazard and its distance to the power transmission channel based on the timed scanning results, compare it with the previous calculation data, and calculate the spread speed and direction of the wildfire hazard with reference to the scanning time; at the same time, it updates the range of the wildfire hazard; and based on the range of the wildfire hazard and its distance to the power transmission channel, the spread speed and direction of the wildfire, it issues a wildfire danger warning.

[0086] Example 1

[0087] like Figure 1 The method for locating wildfire hazards along power transmission corridors based on multi-source fusion, as shown, includes the following steps:

[0088] Step 1: Initialize the image camera, infrared camera and lidar for calibration and registration, and establish the mapping relationship between two-dimensional image coordinates and three-dimensional point cloud coordinates, so that the position of the corresponding three-dimensional point cloud image can be located according to the two-dimensional image coordinates.

[0089] Image cameras, infrared cameras, and lidar can be fixedly mounted together on a pan-tilt unit to achieve synchronous rotation, thereby ensuring that after the equipment coordinates are calibrated, they can scan the same target synchronously from the same angle.

[0090] Step 2: As Figure 3 As shown, a lidar scanner scans the power transmission channel to establish the straight-line equations for each span of the transmission line. , , ..., ;in , For the first The coefficients of the straight line equation for a transmission line span of a certain length. This is a constant term.

[0091] Step 3: The camera performs a wide-angle scan to generate images of the monitored local area, and uses an AI image smoke recognition algorithm to preliminarily determine whether there is a risk of wildfire.

[0092] This invention does not limit the choice of smoke recognition algorithm. The YOLO5 model can be used and trained with a large number of wildfire samples. This model is currently maturely used in power transmission monitoring.

[0093] Step 4: After confirming the potential for wildfire, activate the infrared camera to scan the suspected wildfire area, generate 3D infrared temperature data of the area, provide the range of high-temperature areas, and determine whether it is a real wildfire hazard.

[0094] Step 4.1: Select multiple points with temperatures exceeding a threshold from the entire infrared scan data as valid hazard points. Then, select the current highest temperature point as the hazard center point. Calculate the distance between each valid hazard point and the center point. All points whose distance from the center point is less than a set reference threshold are considered points within the hazard range and added to the target data subset. Continue this process, identifying points whose distance from any point in the target dataset is less than the reference threshold, until all valid hazard points have been calculated. Similarly, construct multiple target data subsets from the remaining valid hazard points. If multiple current highest temperature points exist, randomly select one as the hazard center point, and the other highest temperature points will be assigned to the same target data subset, or a new target data subset will be established with another highest temperature point as the center point.

[0095] Then, it is determined whether it is interference based on the number of valid hidden danger points in the target data subset. In this embodiment, the valid target data subset is assumed to contain more than 10 valid hidden danger points. Figure 2As shown in the figure, data subset 1 has a relatively large number of valid points, so it is considered a valid target data subset. Data subsets 2 and 3 are constructed as independent data subsets using the above method, but because they have fewer valid points (less than 10 points), they are judged as interference (interference is generally relatively scattered). If they are interference data subsets, they are directly removed and do not participate in subsequent algorithm judgment. Thus, the existence of valid target data subsets determines whether there is a real wildfire hazard.

[0096] Furthermore, the temperature threshold in step 4.1 is automatically configured based on whether a micro-meteorological sensing unit is configured, specifically including:

[0097] If a micro-meteorological sensor unit is configured, the temperature threshold is set to the ambient temperature collected by the micro-meteorological sensor unit plus a certain fixed value (usually 5°). If no micro-meteorological sensor unit is configured, the pre-prepared historical temperature data for the whole year is called, and the historical monthly average value of the current time plus a certain fixed value (usually 10°) is selected as the temperature threshold.

[0098] Since there is no direct power source on the transmission tower, only photovoltaic cells can be used for power supply. In order to reduce the overall power consumption of the device, a scheme is adopted to activate and wake up the infrared camera and lidar after image monitoring and recognition, and then perform scanning, monitoring and judgment. Under normal circumstances, the infrared camera and lidar are in sleep mode and do not work.

[0099] Step 4.2: Based on the target data subset, calculate the contour feature data such as the center coordinates and the longest radius of each target data subset, that is, give the range of wildfire hazard (high temperature area range).

[0100] In one embodiment, the center coordinates are calculated as follows:

[0101] Calculate the horizontal coordinates of the leftmost point in the subset. and the horizontal coordinates of the rightmost point , set the horizontal coordinate of its midpoint Using the horizontal coordinate of the center point as an example, calculate the vertical coordinates of the nearest and farthest points respectively. and And the vertical coordinates of the highest and lowest points and Then the center coordinates are .

[0102] In one implementation, the longest radius is calculated as follows:

[0103] Calculate the relative distance between each coordinate. , , The maximum value among them is taken as the longest radius.

[0104] Step 5: Activate the lidar to scan the suspected area. Based on the coordinate mapping between the calibrated 2D image and the 3D point cloud, locate the 3D coordinates of the suspected wildfire hazard. Based on the multiple transmission line equations constructed in Step 2, calculate the peripheral points of the wildfire hazard. Distance to multiple transmission lines The shortest distance among them is selected as the distance from the wildfire hazard to the power transmission channel.

[0105] Step 6: Based on the target dataset determined by infrared in Step 4, calculate the contour volume data of the potential hazard target, and combine it with the wildfire hazard distance calculated in Step 5 to give different danger levels of wildfire hazards in the power transmission channel: the closer the distance and the faster the speed, the more urgent the warning level, requiring notification of maintenance personnel for emergency handling, and even requiring power outage of the line to ensure the safety of the power grid.

[0106] In one embodiment, the hazard level classification is shown in Table 1 below.

[0107] In one embodiment, the method for locating wildfire hazards along power transmission channels based on multi-source fusion further includes:

[0108] Step 7: Activate the timed scanning mechanism. At this time, the infrared camera and lidar are no longer in sleep mode, and they periodically scan and calculate the coordinates of the wildfire, as well as the extent, speed, and direction of the wildfire hazard. Based on the extent, speed, and direction of the wildfire, a corresponding warning level is given. Generally speaking, the larger the wildfire area, the more dangerous it is; the faster it spreads towards power lines, the more dangerous it is.

[0109] In one embodiment, the warning level is set based on experience, as shown in Table 2 below.

[0110]

[0111] Among these, when the wildfire area is small (<10m) or the wildfire distance is large (>5000m), and it spreads away from the transmission line, it is considered a general warning, reminding operation and maintenance personnel to pay attention; when the wildfire area expands to (10m-100m) and the wildfire distance is <5000m, and the spread direction is close to the transmission channel, with a spread speed of 100m / minute, it is considered to threaten the safe operation of the transmission channel, and is considered an emergency level three, reminding operation and maintenance personnel to arrange personnel to handle the situation on-site; when the wildfire area expands to (100m-200m) and the wildfire distance is <5000m, and the wildfire distance is <5000m, it is considered to threaten the safe operation of the transmission channel, and is considered to be an emergency level three, reminding operation and maintenance personnel to arrange personnel to handle the situation on-site; when the wildfire area expands to (100m-200m) and the wildfire distance is <5000m, it is considered to be close to the transmission line, and the spread direction is close to the transmission line, with a spread speed of 100m / minute, it is considered to threaten the safe operation of the transmission line, and is considered to be an emergency level three, reminding operation and maintenance personnel to arrange personnel to handle the situation on-site. If the fire is less than 2000m away and spreads in a direction close to the power transmission line at a speed of 40m / minute, it poses a serious threat to the safe operation of the power transmission line and is considered an emergency level II. Maintenance personnel are advised to immediately notify and arrange for all relevant personnel to handle the situation urgently. If the fire expands to more than 100m or is less than 2000m away and spreads in a direction close to the power transmission line at a speed of 41m / minute, the fire hazard is extremely urgent and is considered an emergency level I. Maintenance personnel are advised to not only notify and arrange for all relevant personnel to handle the situation on-site urgently, but also to shut off power to the power lines.

[0112] In a specific implementation, the spread rate can be calculated based on the center distance, by comparing the current data with the previously calculated data and referring to the scan time. Similarly, the spread direction is the primary spread direction, i.e., the fastest spread direction.

Claims

1. A method for locating wildfire hazards along power transmission channels based on multi-source fusion, characterized in that, The method includes: Based on the three-dimensional point cloud image of the power transmission channel, establish the power transmission line equation for each span that can be scanned; A preliminary assessment of wildfire hazards is made based on smoke recognition algorithms in the monitored area images. If a wildfire hazard is preliminarily identified, a three-dimensional temperature field is generated based on the infrared scanning results of the suspected wildfire area. Effective hazard points are identified from the three-dimensional temperature field, and several real wildfire hazard datasets are constructed based on the effective hazard points. The range of each wildfire hazard target is calculated. If no wildfire hazard is found in the preliminary assessment, image monitoring of the area continues. Locate the three-dimensional coordinates of suspected wildfire hazards and calculate the distance from the actual wildfire hazard to the power transmission channel based on the aforementioned power transmission line equation; Based on the scope of the stated wildfire hazard targets and the distance from the actual wildfire hazard to the power transmission corridor, different risk levels of wildfire hazards to the power transmission corridor are comprehensively given.

2. The method for locating wildfire hazards in power transmission channels based on multi-source fusion according to claim 1, characterized in that, The method also includes: periodically updating the range of wildfire hazards and their distance to power transmission channels, calculating the wildfire spread speed and direction, and issuing wildfire danger warnings based on the range of wildfire hazards, their distance to power transmission channels, the wildfire spread speed, and the direction of spread.

3. The method for locating wildfire hazards in power transmission channels based on multi-source fusion according to claim 1, characterized in that, The process of identifying effective hazard points from the three-dimensional temperature field and constructing several real wildfire hazard datasets based on these effective hazard points includes: Several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are designated as effective hazard points, and the current highest temperature point is designated as the hazard center point. The distance between each effective hazard point and the hazard center point is calculated, and all effective hazard points whose distance is less than the set reference threshold are added to the current target data subset. If there are multiple current highest temperature points, one of them is randomly selected as the hazard center point, and the other highest temperature points are assigned to the same target data subset, or a new target data subset is established with another highest temperature point as the center point. Repeat the above steps for the remaining potential hazards to construct multiple target data subsets; The presence of interference is determined based on the number of valid points of potential hazards in the target data subset. If interference exists, the interfering data subset is removed; otherwise, the target data subset is deemed valid and is considered a real wildfire hazard target dataset.

4. The method for locating wildfire hazards in power transmission channels based on multi-source fusion according to claim 1, characterized in that, The calculation of the range of each wildfire hazard target includes: calculating the contour feature data of each real wildfire hazard target dataset, and generating a contour volume based on the contour feature data to characterize the range of the wildfire hazard target; wherein, the contour feature data includes the center coordinates and the longest radius of each real wildfire hazard target dataset.

5. The method for locating wildfire hazards in power transmission channels based on multi-source fusion according to claim 1, characterized in that, The three-dimensional coordinates for locating suspected wildfire hazards are achieved in the following way: The image camera, infrared camera, and lidar are calibrated and registered to establish a mapping relationship between two-dimensional image coordinates and three-dimensional point cloud coordinates. When all three scan a suspicious area at the same time, the position of the corresponding three-dimensional point cloud image is located according to the two-dimensional image coordinates, which is the three-dimensional coordinate of the suspected wildfire hazard.

6. The method for locating wildfire hazards in power transmission channels based on multi-source fusion according to claim 1, characterized in that, The calculation of the distance from the actual wildfire hazard to the power transmission corridor based on the power transmission straight line equation includes: Outskirts of wildfire hazard points Substituting the equations of the power transmission lines for the multiple power transmission channels that may be threatened by wildfires, we obtain... in, , For the first The coefficients of the straight line equation for a transmission line span of a certain length. For constant terms; The shortest distance among them was selected as the distance from the actual wildfire hazard to the power transmission line.

7. A wildfire hazard location system for power transmission channels based on multi-source fusion, characterized in that, The system includes: The sensing module includes an image camera, an infrared camera, a lidar, and a hierarchical trigger control unit. The image camera, infrared camera, and lidar are integrated on a gimbal with a fixed relative spatial relationship and are driven by the gimbal to perform common field-of-view scanning. The hierarchical trigger control unit is used to control the start and stop of the infrared camera and lidar based on the preliminary wildfire hazard assessment results from the image camera. The data fusion and analysis module is communicatively connected to the perception triggering module and is used to process multi-source fusion data collected after the infrared camera and lidar are activated, in order to identify and locate the fire scene and assess the risk.

8. The multi-source fusion-based power transmission channel wildfire hazard location system according to claim 7, characterized in that, The data fusion and analysis module includes: The fire quantification unit is used to generate a three-dimensional temperature field based on the scanning results of the infrared camera on the suspected wildfire area, identify several real wildfire hazards based on the three-dimensional temperature field, and obtain the quantification results of each target. The precise positioning unit is used to calculate the distance from the actual wildfire hazard to the power transmission channel based on the three-dimensional coordinates of the suspected wildfire hazard obtained after calibration and registration by the sensing module and a pre-stored power transmission line model. The risk assessment unit is used to determine the hazard level based on the quantitative results of the actual wildfire hazards and their distance to the power transmission channel.

9. The multi-source fusion-based power transmission channel wildfire hazard location system according to claim 8, characterized in that, The fire quantification unit identifies several real wildfire hazards based on the three-dimensional temperature field and obtains the quantification results of each target, and performs the following steps: several points in the three-dimensional temperature field whose temperature exceeds the temperature threshold are taken as valid hazard points, and the current highest temperature point is taken as the hazard center point. Calculate the distance between each valid point of a hazard and the center point of the hazard, and add all valid points of hazard whose distance is less than the set reference threshold to the current target data subset; If there are multiple points with the highest current temperature, one of them will be randomly selected as the center point of the potential hazard. Other points with the highest temperature will be assigned to the same subset of target data, or a new subset of target data will be established with another point with the highest temperature as the center point. Repeat the above steps for the remaining potential hazards to construct multiple target data subsets; The presence of interference is determined based on the number of valid points of potential hazards in the target data subset. If interference exists, the interfering data subset is removed. Otherwise, the target data subset is deemed valid and is considered a real wildfire hazard target dataset.

10. The multi-source fusion-based power transmission channel wildfire hazard location system according to claim 7, characterized in that, The system also includes an early warning module; After the real wildfire hazard identification stage, the hierarchical trigger control unit controls the infrared camera and lidar to be in working state, and together with the image camera, they perform timed scanning. The early warning module is used to calculate the coordinates of the wildfire hazard and its distance to the power transmission channel based on the timed scanning results, compare it with the previous calculation data, and calculate the spread speed and direction of the wildfire hazard with reference to the scanning time; at the same time, it updates the range of the wildfire hazard; and based on the range of the wildfire hazard and its distance to the power transmission channel, the spread speed and direction of the wildfire, it issues a wildfire danger warning.

Images