A fire point identification method and system for power grid fire prevention

Abstract

The application relates to the technical field of power grid forest fire identification, and discloses a fire point identification method and system for power grid forest fire prevention, which comprises the following steps: obtaining candidate fire point pixel coordinates based on low threshold preliminary screening, and mapping the candidate fire point pixel coordinates to high-resolution land cover classification grid data; extracting an industrial identification window for representing a global land class and an edge shift detection window for representing edge features; performing pre-identification of an industrial dense area based on the industrial identification window; calculating a local normalized vegetation index for representing a real-time combustible material state of a real ground surface; combining the edge shift detection window and the local normalized vegetation index to perform elimination of registration drift, and outputting a fire point for power grid forest fire prevention after elimination. The application solves the technical problems that power grid forest fire prevention is prone to introducing urban and industrial heat source interference in the prior art, and that a conventional static mask is easily affected by satellite positioning deviation, and significantly improves the real-time performance and high confidence of power grid forest fire warning.

Description

Technical Field

[0001] This application relates to the field of power grid wildfire identification technology, specifically a method and system for identifying fire points for power grid wildfire prevention. Background Technology

[0002] Wildfires pose a significant threat to the safe operation of ultra-high voltage (UHV) power transmission lines and transmission lines at all levels. Wildfire prevention for power grids has highly unique characteristics: on the one hand, to prevent line tripping due to high temperatures and dense smoke, real-time and very early detection of fires is essential, requiring appropriately lowered thresholds at the satellite end to avoid missing small-scale fires; on the other hand, power grids focus on heat sources with the potential for combustion and spread, such as those surrounded by flammable vegetation. In response to these unique characteristics, it is necessary to design a technical solution for fire point identification in power grid wildfire prevention. Summary of the Invention

[0003] The purpose of this application is to provide a fire point identification method and system for power grid wildfire prevention, in order to solve the technical problems in the prior art where power grid wildfire prevention is easily affected by urban and industrial heat sources, and conventional static masks are easily affected by satellite positioning deviations.

[0004] To achieve the above objectives, this application provides a fire point identification method for power grid wildfire prevention, comprising: Obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low-threshold thermal radiation anomalies, and map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data; Based on the mapping results, and combining the ratio between the spatial resolution of satellite sensors and the spatial resolution of high-resolution land cover classification raster data, an industrial identification window for characterizing global land cover types and an edge drift detection window for characterizing edge features are extracted. Pre-identification of densely populated industrial areas is performed based on industrial identification windows to eliminate interference from static artificial heat sources caused by frequent human activities and lack of combustibles. After removing static artificial heat sources, the local normalized vegetation index is calculated for the coordinates of the remaining candidate fire point pixels to characterize the real-time combustible state of the actual ground surface. By combining the ground features and local normalized vegetation index within the edge drift detection window, the remaining candidate fire point pixel coordinates are eliminated due to registration drift. After elimination, the fire points facing the power grid for wildfire prevention are output.

[0005] To achieve the above objectives, this application also provides a fire point identification system for power grid wildfire prevention, which applies the fire point identification method for power grid wildfire prevention described above, including: The low threshold mapping module is used to obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low threshold thermal radiation anomalies, and to map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data. A multi-scale window construction module is used to extract an industrial identification window for representing global land cover types and an edge drift detection window for representing edge features based on the mapping results and the ratio of the spatial resolution of satellite sensors to the spatial resolution of high-resolution land cover classification raster data. The pre-identification module is used to perform pre-identification of densely populated industrial areas based on the industrial identification window, so as to eliminate interference from static artificial heat sources due to frequent human activities and lack of combustibles. The local flammability analysis module is used to calculate the local normalized vegetation index, which represents the real-time flammability state of the actual ground surface, based on the coordinates of the remaining candidate fire point pixels after removing static artificial heat sources. The joint identification module is used to combine the ground features within the edge drift detection window and the local normalized vegetation index to remove the remaining candidate fire point pixel coordinates due to registration drift. After removal, the fire points facing the power grid for wildfire prevention are output.

[0006] To achieve the above objectives, this application also provides an electronic device, including at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which calls the program instructions to execute the fire point identification method for power grid wildfire prevention as described above.

[0007] To achieve the above objectives, this application also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the fire point identification method for power grid wildfire prevention as described above.

[0008] Beneficial Effects: The fire point identification method and system for power grid wildfire prevention proposed in this application address the technical shortcomings of power grid wildfire prevention, which is easily affected by urban and industrial heat sources and conventional static masks are susceptible to satellite positioning deviations. It proposes a hierarchical error correction mechanism that combines physical spatial resolution mapping with real-time combustible material status. After acquiring candidate fire points, it prioritizes the construction of a multi-scale spatial semantic window with adaptive satellite pixel resolution. False fire points in densely populated industrial areas are filtered out based on the complete pixel-level window to reduce subsequent computational resource consumption. Subsequently, the local normalized vegetation index is calculated for the remaining fire points, and boundary windows are used to identify edge spatial positioning drift. This effectively integrates power grid business logic with multi-dimensional spatial semantics, resulting in high computational efficiency, effective filtering of artificial heat sources with no risk of spread, and significantly improving the real-time performance and high confidence level of power grid wildfire early warning. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating a fire point identification method for power grid wildfire prevention provided in an embodiment of this application; Figure 2 The diagram shows the structure of a fire point identification system for power grid wildfire prevention provided in this application embodiment; in the diagram: 10, low threshold mapping module; 20, multi-scale window construction module; 30, pre-identification module; 40, local flammability analysis module; 50, joint identification module.

[0011] The implementation, functional features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0012] The technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0013] In this document, the term "comprising" is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0014] This embodiment discloses a fire point identification method and system for power grid wildfire prevention. In application, it belongs to the technical field of transmission line safety operation and maintenance and natural disaster remote sensing monitoring. The technical solution mainly involves fire point identification technology for power grid wildfire prevention based on satellite positioning drift correction and multi-dimensional semantics.

[0015] The technical challenges of the fire point identification method and system for power grid wildfire prevention in this embodiment will now be explained. Due to the aforementioned specificities, the practical application of solutions for power grid wildfire prevention presents the following technical challenges: When a low threshold is introduced, industrial heat sources such as photovoltaic power plants and blast furnaces bring a large number of false alarms. However, the traditional point-to-point calibration method that relies on static land use and land cover data (LULC) has two major drawbacks: First, satellite-level products generally have positioning deviations (registration drift), and heat sources from factories located in urban-rural fringe areas are very likely to drift to adjacent vegetation pixels and thus be missed. Second, blindly calculating real-time NDVI (Normalized Difference Vegetation Index) and surrounding environmental parameters for all acquired pixels not only consumes a lot of computing resources, but also ignores the characteristics of power grid wildfire business itself. Specifically, in densely populated industrial areas with large areas and high-frequency population flow, firstly, there are very few objective conditions for large-scale forest fires, and secondly, even if an alarm occurs, there are very few cases where it is not detected by humans or not alerted in time.

[0016] In response to the above analysis, for power grid wildfire prevention, there is a need for a multi-dimensional intelligent identification technology that can effectively overcome satellite positioning edge drift based on spatial resolution mapping, and can also intercept artificial heat sources in advance by combining power grid business logic and reduce computing power overhead.

[0017] In response to the above requirements, this embodiment proposes a fire point identification method and system for power grid wildfire prevention. This embodiment is summarized below: To address the technical challenges of massive interference from man-made heat sources caused by lowered thresholds in power grid wildfire monitoring, and the large-scale failure of conventional static land cover masks due to satellite positioning errors, this embodiment provides a fire point identification technology for power grid wildfire prevention based on satellite positioning drift correction and multi-dimensional semantics. The technical problem this fire point identification technology aims to solve is: To achieve early detection and prevention of wildfires, existing remote sensing monitoring for power grid wildfire prevention typically lowers the threshold for thermal anomalies, inevitably introducing a large number of artificial false heat sources such as photovoltaic power plants and industrial blast furnaces. The traditional method of elimination is to directly compare the coordinates of satellite pixels with land cover (LULC) data. However, due to the hundreds of meters of pixel registration drift commonly found in satellite Level 1 products (L1B data), artificial heat sources located in forest-industrial ecotones are easily misidentified as real wildfires. In addition, blindly performing multispectral band reviews and real-time normalized vegetation index (NDVI) calculations on all candidate pixels would consume a great deal of system computing power and ignore the business logic of a well-established artificial early warning mechanism for power grid wildfire prevention in densely populated industrial areas.

[0018] In response to the aforementioned technical issues, the core design idea of ​​this embodiment is to construct a dynamic hierarchical error correction model that corroborates the multi-scale spatial semantics and real-time combustible characteristics based on adaptive resolution mapping.

[0019] The fire point identification method and system for power grid wildfire prevention in this embodiment will now be described.

[0020] Reference Figure 1 , Figure 1 A flowchart illustrating a fire point identification method for power grid wildfire prevention provided in an embodiment of this application.

[0021] like Figure 1 As shown, this embodiment discloses a fire point identification method for power grid wildfire prevention, including: S10: Obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low-threshold thermal radiation anomalies, and map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data.

[0022] In the application, low-threshold candidate fire points are acquired, and the geometric centers of pixels on satellite imagery are mapped to a high spatial resolution LULC raster system using a coordinate mapping function. Subsequently, a multi-scale spatial semantic extraction model adaptive to satellite pixel resolution is established. This model is based on the physical spatial resolution of the satellite sensor. With LULC spatial resolution The proportional relationship is used to dynamically calculate the side lengths of the two key semantic windows.

[0023] S20: Based on the mapping results, and combining the ratio of the spatial resolution of satellite sensors to the spatial resolution of high-resolution land cover classification raster data, extract the industrial identification window for representing global land cover types and the edge drift detection window for representing edge features.

[0024] In the application, the side lengths of the two key semantic windows are dynamically calculated, including: An industrial identification window with the entire physical swath of a satellite pixel Its side length The calculation formula is: Precise edge drift detection window of half a pixel physical width Its side length The calculation formula is: in, The pixel spatial resolution of satellite thermal anomaly products; Spatial resolution for land cover classification raster data; This represents the floor function.

[0025] S30: Perform pre-identification of densely populated industrial areas based on the industrial identification window to eliminate interference from static artificial heat sources caused by frequent human activities and lack of combustibles.

[0026] In application, S30 actually performs pre-interception in densely populated industrial areas based on power grid business logic. This is for the extracted... Window, through the impermeable surface density characteristic function Assess its clustering degree: in, This is an indicator function for impermeable surfaces / urban built land categories, when the coordinates are located at... The value is 1 when the cell belongs to the impermeable surface or urban built-up area category, and 0 otherwise; Greater than the set industrial agglomeration determination threshold If the percentage is 80%, it indicates that the candidate point is deeply embedded in a heavy industrial area lacking conditions for wildfire spread. It is directly eliminated as a static artificial heat source, thereby blocking the downward transmission of ineffective computing power.

[0027] S40: After removing static artificial heat sources, calculate the local normalized vegetation index (NUD) of the remaining candidate fire point pixel coordinates to characterize the real-time combustible state of the actual ground surface.

[0028] In application, S40 aims to further develop real-time combustible abundance data based on local areas. For surviving candidate points that were not intercepted at the previous stage, the satellite's visible light red band reflectance is retrieved. Near-infrared band reflectivity The real-time combustible abundance of the local area is extracted. In this embodiment, the real-time combustible abundance is quantified based on the local normalized vegetation index (NUV). The NUV is: in, Reflectance in the visible red band; Reflectivity in the near-infrared band; Minimum constants to prevent division by zero anomalies; Centered on candidate fire points Satellite pixel space window; This represents the total number of valid pixels within the window after removing invalid data such as cloud cover. This indicates whether the actual land surface at this time truly possesses the vegetation material basis to ignite a wildfire and spread to power grid lines.

[0029] S50: Combining the ground features and local normalized vegetation index within the edge drift detection window, the remaining candidate fire point pixel coordinates are eliminated due to registration drift. After elimination, the fire points facing the power grid for wildfire prevention are output.

[0030] In application, the S50 aims to complete the final joint identification based on positioning drift logic. Specifically, even if the center of the fire point's coordinates falls within a relatively natural vegetation environment, it is still necessary to invoke the edge drift detection window in conjunction with the fire point. and local vegetation index Perform condition-based decision expression Cross-validation: in, For logical OR operation; For logical AND operation; An indicator function for bare land categories; This is a lower limit threshold for vegetation abundance that characterizes the risk of fire spread; if the left spatial neighborhood (within half a pixel width) of the formula covers hardened or bare terrain, and the real-time combustible abundance of the right pixel is lower than the set threshold. If the deviation is due to satellite-level product positioning error, it is determined to be a false industrial heat source that has drifted to the edge of the natural surface. Such misjudgments are precisely eliminated, ultimately leaving only high-confidence, genuine, spreading fire points.

[0031] Thus, the fire point identification method for power grid wildfire prevention in this embodiment solves the technical problems of power grid wildfire prevention being easily affected by urban and industrial heat sources, and the susceptibility of conventional static masks to satellite positioning deviations.

[0032] Reference Figure 2 , Figure 2 The diagram shows the structure of a fire point identification system for power grid wildfire prevention provided in this application embodiment; in the diagram: 10, low threshold mapping module; 20, multi-scale window construction module; 30, pre-identification module; 40, local flammability analysis module; 50, joint identification module.

[0033] Secondly, such as Figure 2 As shown, this embodiment also discloses a fire point identification system for power grid wildfire prevention, which applies the fire point identification method for power grid wildfire prevention described above, including: The low threshold mapping module 10 is used to obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low threshold thermal radiation anomalies, and to map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data. The multi-scale window construction module 20 is used to extract an industrial identification window for representing global land cover types and an edge drift detection window for representing edge features based on the mapping results and the ratio of the spatial resolution of satellite sensor to the spatial resolution of high-resolution land cover classification raster data. The pre-identification module 30 is used to perform pre-identification of densely populated industrial areas based on the industrial identification window, so as to eliminate interference from static artificial heat sources due to frequent human activities and lack of combustibles. The local flammability analysis module 40 is used to calculate the local normalized vegetation index, which represents the real-time flammability state of the real ground surface, based on the coordinates of the remaining candidate fire point pixels after removing static artificial heat sources. The joint identification module 50 is used to combine the ground features within the edge drift detection window and the local normalized vegetation index to remove the remaining candidate fire point pixel coordinates due to registration drift, and output the fire points facing the power grid for wildfire prevention after removal.

[0034] Thirdly, this embodiment also discloses an electronic device, including at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which calls the program instructions to execute the fire point identification method for power grid wildfire prevention as described above.

[0035] Fourthly, this embodiment also discloses a storage medium on which a computer program is stored, which, when executed by a processor, implements the fire point identification method for power grid wildfire prevention as described above.

[0036] It should be noted that the fire point identification system, electronic equipment, and storage medium for power grid wildfire prevention in this embodiment correspond to the aforementioned fire point identification method for power grid wildfire prevention. Therefore, any content not specifically described in the fire point identification system, electronic equipment, and storage medium for power grid wildfire prevention in this embodiment, including but not limited to functional definitions, working principles, and technical effects, can be referred to the descriptions in the aforementioned fire point identification system, electronic equipment, and storage medium method for power grid wildfire prevention, and will not be repeated here.

[0037] In summary, the fire point identification method and system for power grid wildfire prevention in this embodiment addresses the technical shortcomings of power grid wildfire prevention, such as the susceptibility to interference from urban and industrial heat sources and the vulnerability of conventional static masks to satellite positioning errors. It proposes a hierarchical error correction mechanism that combines physical spatial resolution mapping with real-time combustible material status. After acquiring candidate fire points, it prioritizes the construction of a multi-scale spatial semantic window with adaptive satellite pixel resolution. False fire points in densely populated industrial areas are filtered out based on the complete pixel-level window to reduce subsequent computational resource consumption. Subsequently, the local normalized vegetation index is calculated for the remaining fire points, and edge spatial positioning drift is identified using boundary windows. This effectively integrates power grid business logic with multi-dimensional spatial semantics, resulting in high computational efficiency, effective filtering of artificial heat sources with no risk of spread, and significantly improving the real-time performance and high confidence level of power grid wildfire early warning.

[0038] The application and technical effects of the fire point identification method and system for power grid wildfire prevention in this embodiment will now be explained in conjunction with specific implementation.

[0039] Satellite data equipped with VIIRS sensor (I-band spatial resolution) (meters) and land cover products (spatial resolution) This embodiment will be described in detail using a meter as an example.

[0040] First, high-precision coordinate mapping of low-threshold candidate fire points is performed. A relaxed dynamic environment threshold is used to detect initial small fire candidate pixels. High-precision land cover classification raster files are read. A coordinate projection transformation transformer is used to convert the latitude and longitude of candidate fire points into row and column numbers of the LULC raster projection matrix. .

[0041] Secondly, dynamic construction of multi-scale windows adapted to satellite resolution. (For VIIRS) Resolution and LULC Differences in objective resolution attributes, dynamically calculating the semantic window side length: Building an industrial identification window This window represents the absolute physical area covered by an entire satellite pixel. Calculation That is, extract the area surrounding the center point. (Approximately covered) , exactly equal to one A LULC matrix of pixels.

[0042] Construct an edge drift detection window Used to capture micro-level land cover changes within a half-pixel width range caused by satellite pixel bias. Calculation That is, extract the area surrounding the center point. matrix.

[0043] Then, parallel dimensionality reduction and pre-interception based on dense regions are performed. This is for the acquired... The LULC matrix is ​​calculated, and the impermeable surface area is evaluated using the density characterization formula. The business logic behind this step is that when the calculated proportion of impermeable surface pixels is extremely high, it is determined from the perspective of power grid wildfire prevention to be a densely populated industrial area or central urban area. Such areas do not have sprawling, continuous flammable materials and have frequent population activity; in the event of a fire, the municipal fire-fighting system will respond promptly. Therefore, these areas are directly intercepted and excluded, eliminating the need for complex and time-consuming multispectral detection, thus greatly saving computational costs.

[0044] Next, the real-time state of natural surface combustibles is extracted. Only for the remaining non-dense suburban areas, woodlands, and forestry-industrial boundary candidate points, the I1 (visible red band) and I2 (near-infrared band) band data of the VIIRS sensor are retrieved to obtain the real-time disaster vegetation abundance, i.e., the local normalized vegetation index.

[0045] Finally, manual edge positioning is used for precise identification and removal of drift. The edge drift detection window matrix. Although the fire point coordinates center falls on vegetation, but if If a cell within the window matrix is ​​classified as an impermeable surface, and the calculated local normalized vegetation index shows extremely low vegetation cover, indicating a non-combustible terrain, then based on the drift tolerance conditional judgment formula, it is strictly determined that the latitude and longitude of the fire point has location drifted, and the actual heat source is the adjacent factory building. This type of false fire is then eliminated. At this point, the real fire points that meet the requirements and possess extremely high threat are pushed to the power grid operation and maintenance terminal.

[0046] In summary, the fire point identification method and system for power grid wildfire prevention in this embodiment achieves at least the following technical effects: It breaks through the bottleneck of spatial drift misjudgment. Specifically, it introduces dual verification of half-width multi-scale window and refined local NDVI, and uses the mathematical mapping law of physical spatial resolution to overcome the irregular registration offset of satellite system, and completely solves the industry problem of frequent failure of conventional static mask in urban-rural fringe areas and forest-industrial ecotones.

[0047] The computing resources are highly concentrated. Specifically, the innovative approach combines the needs of wildfire prevention with the proposed pre-interception logic for dense industrial areas. This logic directly cuts off the massive amount of background noise falling into heavy industrial areas, avoiding requests for a large amount of invalid spectral data and matrix operations, and significantly improving the concurrent processing efficiency and timeliness of fire point warnings.

[0048] In the embodiments provided in this application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the associated hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media may be any available medium accessible to a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.

[0049] Finally, it should be noted that the above description is only a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A fire point identification method for power grid wildfire prevention, characterized in that, include: Obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low-threshold thermal radiation anomalies, and map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data; Based on the mapping results, and combined with the ratio of the spatial resolution of satellite sensors to the spatial resolution of high-resolution land cover classification raster data, an industrial identification window for characterizing global land cover types and an edge drift detection window for characterizing edge features are extracted. Pre-identification of densely populated industrial areas is performed based on industrial identification windows to eliminate interference from static artificial heat sources caused by frequent human activities and lack of combustibles. After removing static artificial heat sources, the local normalized vegetation index is calculated for the coordinates of the remaining candidate fire point pixels to characterize the real-time combustible state of the actual ground surface. By combining the ground features and local normalized vegetation index within the edge drift detection window, the remaining candidate fire point pixel coordinates are eliminated due to registration drift. After elimination, the fire points facing the power grid for wildfire prevention are output.

2. The fire point identification method for power grid wildfire prevention according to claim 1, characterized in that, The industrial identification window Window side length for: in, For satellite sensor spatial resolution, The spatial resolution corresponding to high-resolution land cover classification raster data; This represents the floor function.

3. The fire point identification method for power grid wildfire prevention according to claim 1, characterized in that, The edge drift detection window Window side length for: in, For satellite sensor spatial resolution, The spatial resolution corresponding to high-resolution land cover classification raster data; This represents the floor function.

4. The fire point identification method for power grid wildfire prevention according to claim 1, characterized in that, The pre-identification of densely populated industrial areas is based on the calculation of the impermeable surface density feature within the industrial identification window. When the calculated impermeable surface density feature is greater than the preset impermeable surface density feature threshold, rejection is triggered.

5. The fire point identification method for power grid wildfire prevention according to claim 4, characterized in that, The impermeable surface density characteristic is as follows: in: The impermeable surface density feature within the calculated industrial identification window; The window side length for industrial identification; For indicator functions, when the coordinates are located The value is 1 when the pixel belongs to the impermeable surface or urban building land category, and 0 otherwise.

6. The fire point identification method for power grid wildfire prevention according to claim 1, characterized in that, The local normalized vegetation index is: in: The local normalized vegetation index is calculated for an industrial identification window. This represents the total number of valid pixels within the industrial identification window after removing invalid data, including cloud cover. Centered on the coordinates of the remaining candidate fire pixels Satellite pixel space window; For the industrial identification window point Near-infrared reflectivity; For the industrial identification window point Reflectivity in the visible red band; To prevent minimum constants from division by zero.

7. The fire point identification method for power grid wildfire prevention according to claim 1, characterized in that, The removal of registration drift is based on a conditional decision made using a linked edge drift detection window and a local normalized vegetation index. The conditional decision is as follows: in: The output judgment result is as follows: if there is registration drift, the output is 1 and rejection is performed; if there is no registration drift, the output is 0 and rejection is not performed. For logical OR operation; For logical AND operation; Indicates position Edge drift detection window Inside; For indicator functions, when position The value is 1 when the cell belongs to the impermeable surface or urban building land category, and 0 otherwise; An indicator function for bare land category, when location The value is 1 when the cell is bare ground, and 0 otherwise; For local normalized vegetation index, The threshold value is the preset local normalized vegetation index.

8. A fire point identification system for power grid wildfire prevention, employing the fire point identification method for power grid wildfire prevention as described in any one of claims 1 to 7, characterized in that, include: The low threshold mapping module is used to obtain the coordinates of candidate fire point pixels obtained from the initial screening based on low threshold thermal radiation anomalies, and to map the coordinates of candidate fire point pixels to the pre-constructed high-resolution land cover classification raster data. A multi-scale window construction module is used to extract an industrial identification window for representing global land cover types and an edge drift detection window for representing edge features based on the mapping results and the ratio of the spatial resolution of satellite sensors to the spatial resolution of high-resolution land cover classification raster data. The pre-identification module is used to perform pre-identification of densely populated industrial areas based on the industrial identification window, so as to eliminate interference from static artificial heat sources due to frequent human activities and lack of combustibles. The local flammability analysis module is used to calculate the local normalized vegetation index, which represents the real-time flammability state of the actual ground surface, based on the coordinates of the remaining candidate fire point pixels after removing static artificial heat sources. The joint identification module is used to combine the ground features within the edge drift detection window and the local normalized vegetation index to remove the remaining candidate fire point pixel coordinates due to registration drift. After removal, the fire points facing the power grid for wildfire prevention are output.

9. An electronic device, characterized in that, Includes at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the fire point identification method for power grid wildfire prevention as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the fire point identification method for power grid wildfire prevention as described in any one of claims 1 to 7.

Images