Method, device and equipment for identifying abnormal hidden danger of overhead line based on unmanned aerial vehicle

Abstract

The invention provides an unmanned aerial vehicle-based overhead line abnormal hidden danger identification method, device and equipment, and relates to the technical field of overhead line safety monitoring. The method comprises the following steps: dividing a target overhead line into a plurality of initial detection areas based on distribution parameters and design parameters of the target overhead line; and then, selecting a target detection area from the plurality of initial detection areas, and formulating a target inspection path for the target detection area by adopting a preset path design algorithm based on the parameter environment of the target detection area. And meanwhile, controlling the target unmanned aerial vehicle to inspect the target detection area according to the target inspection path. Furthermore, an inspection image shot by the target unmanned aerial vehicle in the inspection process is obtained, the inspection image is input to a pre-trained object detection model, and whether the target detection area has an abnormal hidden danger or not is identified according to a detection result output by the object detection model. According to the invention, the efficiency and accuracy of abnormal hidden danger identification of the overhead line can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of overhead line safety monitoring, and in particular to a method, device and equipment for identifying abnormal hidden dangers of overhead lines based on drones. Background Art

[0002] Overhead lines are crucial for power transmission, and their safety is directly linked to the stability of the power system and the security of its supply. However, during operation, overhead lines are susceptible to damage from external forces, particularly intrusion by unusual objects, leading to short circuits and outages. Therefore, it is essential to inspect overhead lines to eliminate potential anomalies and hidden dangers.

[0003] Existing technologies still rely on manual inspections or fixed-mounted surveillance cameras to identify potential anomalies and hidden dangers in overhead lines. Manual inspections typically require personnel to physically inspect the overhead lines, while surveillance cameras capture real-time images of the overhead lines and use image processing technology to process them to identify potential anomalies and hidden dangers.

[0004] However, the existing technology has relatively low efficiency and accuracy in identifying abnormal hidden dangers of overhead lines. Summary of the Invention

[0005] The present application provides a method, device and equipment for identifying abnormal hidden dangers of overhead lines based on drones, which is used to solve the technical problem that the existing technology has relatively low efficiency and accuracy in identifying abnormal hidden dangers of overhead lines.

[0006] In a first aspect, the present application provides a method for identifying abnormal hidden dangers of overhead lines based on a drone, comprising:

[0007] Based on distribution parameters and design parameters of a target overhead line, the target overhead line is divided into a plurality of initial detection areas; wherein the target overhead line is an overhead line for which abnormal hidden dangers need to be identified; the distribution parameters represent the layout of the target overhead line in geographic space, and the design parameters indicate the structural characteristics of the target overhead line;

[0008] Selecting a target detection area from the multiple initial detection areas; formulating a target inspection path for the target detection area using a preset path design algorithm based on environmental parameters of the target detection area;

[0009] Sending the target inspection path to the target UAV, and controlling the target UAV to inspect the target detection area according to the target inspection path;

[0010] Obtaining inspection images taken by the target UAV during the inspection process, and inputting the inspection images into a pre-trained object detection model; wherein the object detection model is used to detect abnormal objects in the inspection images;

[0011] Determine the detection result output by the object detection model, and based on the detection result, identify whether there are any abnormal hidden dangers in the target detection area.

[0012] In one possible design, the target inspection path for the target detection area is formulated using a preset path design algorithm based on the environmental parameters of the target detection area, including:

[0013] Using a preset path design algorithm, a candidate inspection path is formulated for the target detection area according to the environmental parameters of the target detection area;

[0014] The candidate inspection path is optimized using a preset optimization algorithm, and the optimized candidate inspection path is determined as the target inspection path.

[0015] In one possible design, the preset path design algorithm is used to formulate a candidate inspection path for the target detection area according to the environmental parameters of the target detection area, including:

[0016] Based on the environmental parameters of the target detection area, a target inspection model is constructed; wherein the target inspection model is used to simulate the inspection path of the target UAV in the target detection area;

[0017] Based on the preset path design algorithm, an inspection starting point and an inspection end point are set in the target inspection model, and the inspection starting point and the inspection end point are connected according to a preset rule;

[0018] A connecting line between the inspection starting point and the inspection end point is determined as a candidate inspection path for the target detection area.

[0019] In one possible design, optimizing the candidate inspection path using a preset optimization algorithm and determining the optimized candidate inspection path as the target inspection path includes:

[0020] Determine the candidate inspection path as an initial solution of the preset optimization algorithm; and calculate a first objective function value of the initial solution based on a preset objective function;

[0021] Adding a preset disturbance to the initial solution to obtain a neighborhood solution; calculating a second objective function value of the neighborhood solution based on the preset objective function;

[0022] Comparing the first objective function value with the second objective function value, and determining the initial solution or the neighborhood solution as the current solution according to the comparison result;

[0023] Based on the preset optimization algorithm, the current solution is iteratively optimized; the optimized current solution is determined as the optimized candidate inspection path, and the optimized candidate inspection path is determined as the target inspection path.

[0024] In one possible design, selecting a target detection area from the multiple initial detection areas includes:

[0025] Based on preset dimension indicators, the multiple initial detection areas are divided into risk levels;

[0026] Determining a risk level for each initial detection area; and ranking the risk levels in descending order;

[0027] The initial detection area corresponding to the risk level ranked first is determined as the target detection area.

[0028] In one possible design, identifying whether there is an abnormal hidden danger in the target detection area based on the detection result includes:

[0029] If the detection result indicates that an abnormal object exists in the inspection image, determining the category of the abnormal object and the position of the abnormal object in the inspection image;

[0030] Determining the hazard level of the abnormal object based on the category of the abnormal object and the position in the inspection image in combination with a preset abnormal object hazard level database;

[0031] Calculate the actual position of the abnormal object within the target detection area based on the preset three-dimensional coordinates and the position of the abnormal object in the inspection image; calculate the intrusion range of the abnormal object into the target detection area based on the actual position and the hazard level; wherein the intrusion range includes the depth, width, and intrusion area of ​​the intrusion;

[0032] Based on the hazard level and the intrusion range, the target detection area is analyzed to determine whether there are any abnormal hidden dangers; wherein the abnormal hidden danger indicates that the abnormal object comes into contact with the target overhead line in the target detection area within the current time period or a preset time period in the future.

[0033] In one possible design, the method further includes:

[0034] If there is an abnormal hidden danger in the target detection area, an abnormal alarm signal is sent to the target staff; wherein the abnormal alarm signal is used to prompt the target staff to go to the target detection area to clean up the abnormal object;

[0035] Receive the abnormal object cleaning result fed back by the target staff; if the abnormal object cleaning result indicates that the abnormal object has been cleared, update the risk level of the target detection area, and the ranking result of the risk level of the initial detection area and the updated risk level of the target detection area; based on the ranking result, reselect a new target detection area for abnormal hidden danger identification.

[0036] In a second aspect, the present application provides a device for identifying abnormal hidden dangers of overhead lines based on a drone, comprising:

[0037] a partitioning module configured to partition a target overhead line into a plurality of initial detection areas based on distribution parameters and design parameters of the target overhead line; wherein the target overhead line is an overhead line requiring abnormal hidden danger identification; the distribution parameters characterize the layout of the target overhead line in geographic space, and the design parameters indicate the structural characteristics of the target overhead line;

[0038] A processing module is configured to select a target detection area from the multiple initial detection areas; and formulate a target inspection path for the target detection area using a preset path design algorithm based on environmental parameters of the target detection area;

[0039] An inspection module is used to send the target inspection path to the target UAV, and control the target UAV to inspect the target detection area according to the target inspection path;

[0040] The processing module is further configured to obtain inspection images captured by the target drone during the inspection process, and input the inspection images into a pre-trained object detection model; wherein the object detection model is configured to detect abnormal objects in the inspection images;

[0041] The recognition module is used to determine the detection results output by the object detection model and identify whether there are any abnormal hidden dangers in the target detection area based on the detection results.

[0042] In a possible design, the processing module further includes: a formulation module, an optimization module,

[0043] The formulation module is used to formulate candidate inspection paths for the target detection area based on the environmental parameters of the target detection area using a preset path design algorithm;

[0044] The optimization module is used to optimize the candidate inspection path by using a preset optimization algorithm, and determine the optimized candidate inspection path as the target inspection path.

[0045] In a possible design, the formulation module further includes: a construction module, a setting module, a connection module, and a determination module.

[0046] The construction module is used to construct a target inspection model based on the environmental parameters of the target detection area; wherein the target inspection model is used to simulate the inspection path of the target UAV in the target detection area;

[0047] The setting module is used to set the inspection starting point and inspection end point in the target inspection model based on the preset path design algorithm;

[0048] The connection module is used to connect the inspection starting point and the inspection end point according to a preset rule;

[0049] The determining module is configured to determine a connecting line between the inspection starting point and the inspection end point as a candidate inspection path for the target detection area.

[0050] In one possible design, the determining module is further configured to determine the candidate inspection path as an initial solution of the preset optimization algorithm;

[0051] The optimization module also includes: a calculation module, an addition module, and a comparison module.

[0052] The calculation module is used to calculate the first objective function value of the initial solution based on a preset objective function;

[0053] The adding module is used to add a preset disturbance to the initial solution to obtain a neighborhood solution;

[0054] The calculation module is further configured to calculate a second objective function value of the neighborhood solution based on the preset objective function;

[0055] The comparison module is configured to compare the first objective function value with the second objective function value, and determine the initial solution or the neighborhood solution as the current solution according to the comparison result;

[0056] The optimization module is further configured to iteratively optimize the current solution based on the preset optimization algorithm;

[0057] The determination module is further configured to determine the optimized current solution as the optimized candidate inspection path, and to determine the optimized candidate inspection path as the target inspection path.

[0058] In a possible design, the processing module further includes: a partitioning module and a sorting module.

[0059] The classification module is used to classify the risk levels of the multiple initial detection areas based on preset dimension indicators;

[0060] The determination module is further configured to determine the risk level of each initial detection area;

[0061] The ranking module is used to sort the risk levels in descending order;

[0062] The determination module is further configured to determine the initial detection area corresponding to the risk level ranked first as the target detection area.

[0063] In one possible design, the determining module is further configured to:

[0064] If the detection result indicates that an abnormal object exists in the inspection image, determining the category of the abnormal object and the position of the abnormal object in the inspection image;

[0065] Determining the hazard level of the abnormal object based on the category of the abnormal object and the position in the inspection image in combination with a preset abnormal object hazard level database;

[0066] The calculation module is further configured to calculate the actual position of the abnormal object within the target detection area based on the preset three-dimensional coordinates and the position of the abnormal object in the inspection image; and calculate the invasion range of the abnormal object into the target detection area based on the actual position and the hazard level; wherein the invasion range includes the depth, width, and invasion area of ​​the invasion;

[0067] The identification module further includes: an analysis module for analyzing whether there are abnormal hidden dangers in the target detection area based on the hazard level and the intrusion range; wherein the abnormal hidden danger indicates that the abnormal object comes into contact with the target overhead line in the target detection area within the current time period or a preset time period in the future.

[0068] In a possible design, the device for identifying abnormal hidden dangers of overhead lines based on drones further includes: a sending module, a receiving module, and an updating module.

[0069] The sending module is used to send an abnormality alarm signal to the target staff if there is an abnormal hidden danger in the target detection area; wherein the abnormality alarm signal is used to prompt the target staff to go to the target detection area to clean up the abnormal object;

[0070] The receiving module is used to receive the abnormal object cleaning result fed back by the target staff;

[0071] The updating module is configured to update the risk level of the target detection area and the ranking result of the risk level of the initial detection area and the updated risk level of the target detection area if the abnormal object clearing result indicates that the abnormal object has been cleared;

[0072] The identification module is further configured to reselect a new target detection area for abnormal hidden danger identification based on the sorting result.

[0073] In a third aspect, the present application provides an electronic device comprising: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the method described in the first aspect above and various possible designs.

[0074] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the method described in the first aspect and various possible designs is implemented.

[0075] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method described in the first aspect and various possible designs.

[0076] The present application provides a method, device, and apparatus for identifying abnormal hidden dangers in overhead lines based on drones. Based on the distribution parameters and design parameters of the target overhead line, the target overhead line is divided into multiple initial detection areas. A target detection area is then selected from the multiple initial detection areas, and a target inspection path is developed for the target detection area using a preset path design algorithm based on the environmental parameters of the target detection area. By dividing the target overhead line into multiple small detection areas, detailed inspections of each inspection area can be achieved, laying the foundation for improving the accuracy of identifying abnormal hidden dangers in overhead lines. Furthermore, the target inspection path is sent to the target drone, which is controlled to inspect the target inspection area according to the inspection path, and inspection images captured by the target drone during the inspection process are obtained. Because the environmental parameters of the target detection area are taken into account when developing the target inspection path, the resulting target inspection path can avoid routes in the target detection area that may affect the flight of the target drone, meaning that the target drone can efficiently complete the inspection task by conducting inspections according to the target inspection path. The inspection image is input into a pre-trained object detection model to obtain the detection results output by the object detection model. Based on the detection results, it is determined whether there are any abnormal hidden dangers in the target detection area. Since the target drone is inspecting within a small target detection area according to the target inspection path, its field of view is no longer limited. Therefore, the target drone can capture relatively clear inspection images, and the object detection model can also output relatively accurate detection results based on the relatively clear inspection images. Therefore, this application can improve the efficiency and accuracy of identifying abnormal hidden dangers in overhead lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0078] Figure 1 Schematic diagram of the process of identifying abnormal hidden dangers of overhead lines based on drones provided in the embodiment of the present application Figure 1 ;

[0079] Figure 2 Schematic diagram of the process of identifying abnormal hidden dangers of overhead lines based on drones provided in the embodiment of the present application Figure 2 ;

[0080] Figure 3 A schematic diagram of the structure of a device for identifying abnormal hidden dangers of overhead lines based on a drone provided in an embodiment of the present application;

[0081] Figure 4 This is a hardware structure diagram of the electronic device provided in an embodiment of the present application.

[0082] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0083] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0084] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate, such that the embodiments of the present invention described herein can, for example, be practiced in an order other than that illustrated or described herein.

[0085] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0086] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards, and corresponding operation entrances must be provided for users to choose to authorize or refuse.

[0087] Overhead lines, defined as power transmission lines suspended in the air via supporting structures, are a crucial component of the power system. They carry electricity from power plants to the various consuming areas, and their safety directly impacts the stable operation of the entire power system and the reliability of its power supply.

[0088] Failures in overhead lines can lead to power system instability, impacting power supply across a wide area. Therefore, ensuring the safety of overhead lines is a critical task in power system management. To prevent these failures, regular inspection and maintenance of overhead lines is necessary to identify and eliminate potential anomalies and potential hazards.

[0089] In most cases, existing technologies still rely on manual inspections or fixed-installed surveillance cameras to identify abnormal hidden dangers in overhead lines.

[0090] Manual inspections require personnel to physically visit overhead power lines and rely on observation and judgment to identify potential anomalies and hidden dangers. Manual inspections are often time-consuming and labor-intensive, especially in complex terrain or with long overhead lines. Furthermore, manual inspections can be limited by personnel experience and attention spans, resulting in low efficiency and accuracy in identifying potential anomalies and hidden dangers.

[0091] Surveillance cameras, installed near overhead lines, capture real-time images of the lines. They then use image processing technology to analyze these images and identify potential anomalies and hidden dangers. While surveillance cameras provide real-time monitoring, their coverage is limited and they cannot capture all potential problems with overhead lines. Furthermore, surveillance cameras are also affected by weather, lighting, and other factors, which may prevent them from capturing clear images.

[0092] Therefore, the existing technology has relatively low efficiency and accuracy in identifying abnormal hidden dangers of overhead lines.

[0093] In response to the above technical problems, the inventors took into account that it is easy to overlook some minor potential problems when monitoring the entire overhead line. Based on this, the inventors thought of first dividing the entire overhead line into multiple small detection areas, and then conducting detailed inspections of each small detection area. The inventors also took into account that the efficiency and accuracy of identifying abnormal hidden dangers in the divided small detection areas by relying on manual inspections or fixed-installed surveillance cameras are still not high, so they decided to use drones to replace manual inspections or surveillance cameras for inspections. At the same time, in order to avoid the impact of obstacles in the detection area on the flight process of the drone, the inventors also thought that when designing the inspection path for the drone, the environmental parameters of the detection area should be taken into account. It can be seen that the drone can efficiently inspect small detection areas and capture relatively clear inspection images at the same time, and then can more accurately identify abnormal hidden dangers of the overhead line based on the relatively clear inspection images.

[0094] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0095] An embodiment of the present application provides a method for identifying abnormal hidden dangers of overhead lines based on a drone. Figure 1 Schematic diagram of the process of identifying abnormal hidden dangers of overhead lines based on drones provided in the embodiment of the present application Figure 1 ,like Figure 1 As shown, the method for identifying abnormal hidden dangers of overhead lines based on drones includes:

[0096] S101. Divide the target overhead line into a plurality of initial detection areas based on the distribution parameters and design parameters of the target overhead line.

[0097] The execution subject of the embodiments of the present application is an electronic device, which can be a terminal device, such as a laptop computer, a desktop computer, a tablet computer, etc., or a server. In actual applications, whether the electronic device is a terminal device or a server can be determined based on actual conditions and is not specifically limited to this.

[0098] For explanation, overhead lines are a form of power lines, specifically those power transmission lines suspended in the air by supporting structures (such as poles, towers, etc.). The target overhead lines in this step refer to the overhead lines that need to be identified for abnormal hidden dangers.

[0099] The distribution parameters of target overhead lines refer to their geographic layout, including their orientation, path, and spacing. The orientation of a target overhead line refers to its direction of extension, the path refers to the actual route it takes from its starting point to its end point, and the spacing refers to the vertical or horizontal distance between different target overhead lines on the same support structure. These distribution parameters can be obtained using a Geographic Information System (GIS).

[0100] Design parameters refer to a series of technical parameters followed during the design and construction of the target overhead line, such as the target overhead line's support structure type, line height, span, and wiring method. The target overhead line's support structure type, such as steel pipe towers or iron towers, determines its stability. Line height affects the target drone's flight angle and viewing angle. Span refers to the horizontal distance between two adjacent support structures. The wiring method affects the target overhead line's inspection path, such as single-circuit inspection or multi-circuit inspection. Understandably, the target overhead line's design parameters can indicate its structural characteristics.

[0101] After obtaining the target overhead line's distribution parameters and design parameters, a rough distribution map of the target overhead line is drawn based on the target line's direction and path as determined by the distribution parameters. A reasonable detection range is then determined based on the target overhead line spacing as determined by the distribution parameters. It should be understood that this detection range should cover all key locations of the target overhead line, including the target overhead line itself, its supporting structure, and potential potential hazards.

[0102] Furthermore, the determined reasonable range and the design parameters of the target overhead line are input into professional simulation software for simulation. Through simulation, the target overhead line can be divided into multiple possible initial detection areas.

[0103] S102 , selecting a target detection area from a plurality of initial detection areas; and formulating a target inspection path for the target detection area using a preset path design algorithm based on environmental parameters of the target detection area.

[0104] In one possible implementation, multiple initial detection areas are classified into risk levels based on preset dimensional indicators. These include, but are not limited to, historical fault records of the target overhead line, environmental conditions (such as wind speed, temperature, and humidity), geographic characteristics (such as landslide and flood-prone areas), human activities (such as construction and agricultural areas), and design parameters of the target overhead line. Specifically, each preset dimensional indicator is assigned a different weight based on its impact on the safety of the target overhead line. By quantifying and weighting the scores and summing these dimensional indicators, the risk level of each initial detection area can be determined.

[0105] It should be noted that the risk level of the initial detection area is typically color-coded on the risk level distribution map. For example, red represents high-risk areas, yellow represents medium-risk areas, and green represents low-risk areas. Based on the risk level distribution map, the risk level of the initial detection area can be intuitively identified.

[0106] Furthermore, the risk levels of the initial detection areas are ranked from high to low, and the initial detection area corresponding to the highest risk level is determined as the target detection area. It is understood that the highest risk level is the highest risk level. By determining the initial detection area with the highest risk level as the target detection area, it is ensured that, given limited resources, the highest risk detection area is prioritized for abnormal hidden danger identification, thereby more effectively ensuring the safety of the target overhead line.

[0107] After determining the target inspection area, a target inspection route is further developed for the target inspection area. It should be noted that the developed target inspection route should be able to cover all key parts of the target inspection area, including the target overhead lines, supporting structures, and possible hidden danger points within the target inspection area.

[0108] Specifically, a preset path design algorithm is used to develop one or more candidate inspection paths for the target inspection area based on its environmental parameters, such as its terrain, obstacle distribution, and the direction of the target overhead lines. These candidate inspection paths should be able to avoid obstacles in the target inspection area. Furthermore, a preset optimization algorithm is used to optimize the candidate inspection paths, and the optimized candidate inspection paths are determined as the target inspection paths.

[0109] Explanatory note: Before developing a target inspection route for a target detection area, it is necessary to fully understand the environmental parameters of the target detection area (topography, obstacle distribution, and the direction of the target overhead line, etc.). The topography of the target detection area includes the undulations and slopes of the ground. The terrain is crucial to the flight altitude and obstacle avoidance strategy of the target drone. Obstacle distribution involves various fixed or moving objects that may affect the flight of the target drone, such as buildings, trees, or other power facilities. The direction of the target overhead line determines the direction in which the target drone needs to patrol, as well as the shooting angle and flight speed that may be required at different locations.

[0110] In one possible implementation, a preset path design algorithm is used to formulate candidate inspection paths for the target detection area based on the environmental parameters of the target detection area. The method specifically includes the following steps: Based on the environmental parameters of the target detection area, a three-dimensional target inspection model is constructed using tools such as GIS or three-dimensional modeling. The target inspection model is used to simulate the inspection path of the target UAV within the target detection area. Afterwards, based on the preset path design algorithm, an inspection starting point and an inspection end point are set in the target inspection model, wherein the inspection starting point corresponds to the take-off position of the target UAV, and the inspection end point corresponds to a preset inspection point. The inspection starting point and the inspection end point are connected according to preset rules, wherein the preset rules can be set to: the path between the inspection starting point and the inspection end point is the shortest, and no specific restrictions are imposed here. The connecting line between the inspection starting point and the inspection end point is determined as the candidate inspection path for the target detection area.

[0111] It should be understood that since the target detection area includes multiple preset inspection points, it is necessary to repeatedly perform the operation of setting the inspection start point and the inspection end point in the target inspection model to adjust the positions of the inspection start point and the inspection end point.

[0112] It should be noted that this embodiment relies on the target UAV to inspect the target detection area. Therefore, when using the preset path design algorithm to formulate the target inspection path for the target detection area, in addition to the comprehensive environmental parameters of the target detection area, the flight capabilities of the target UAV can also be combined, such as the maximum flight speed, maximum flight altitude, maximum flight distance, and minimum turning radius.

[0113] After obtaining the candidate inspection path formulated for the target detection area, it is also necessary to use a preset optimization algorithm to optimize the candidate inspection path, which specifically includes the following steps: determining the candidate inspection path as the initial solution of the preset optimization algorithm, and calculating the first objective function value of the initial solution based on the preset objective function. Obtain a neighborhood solution by adding a preset perturbation to the initial solution, and calculate the second objective function value of the neighborhood solution based on the preset objective function. Compare the first objective function value with the second objective function value, and determine the initial solution or the neighborhood solution as the current solution based on the comparison result. Further, based on the preset optimization algorithm, iteratively optimize the current solution, determine the optimized current solution as the optimized candidate inspection path, and determine the optimized candidate inspection path as the target inspection path.

[0114] Next, we'll use a specific example to explain the optimization process for the candidate inspection paths described above in detail. In this example, the preset optimization algorithm uses a simulated annealing algorithm, and the preset objective function uses a fitness function. The fitness function is used to measure the overall performance of the candidate inspection path, typically taking into account multiple metrics such as the target drone's flight time, flight distance, and image capture efficiency. Through the fitness function, the candidate inspection path is assigned an objective function value, which reflects the candidate's overall performance.

[0115] The simulated annealing algorithm is a random search algorithm based on the physical annealing process, capable of finding an optimal solution or a near-global optimal solution in space. It iteratively optimizes candidate inspection paths, adjusting their parameters until they gradually approach their optimal values.

[0116] Explanatory: During the optimization process, the simulated annealing algorithm will accept solutions that are worse than the current solution with a certain probability, thereby avoiding falling into local optimal solutions and increasing the possibility of finding the global optimal solution. By using the simulated annealing algorithm, an optimal inspection path can be optimized, namely the target inspection path.

[0117] Specifically, at the current temperature, the candidate inspection path is used as the initial solution of the simulated annealing algorithm, and the first objective function value of the initial solution is calculated based on the fitness function. A neighborhood solution is obtained by adding a preset perturbation to the initial solution, and the second objective function value of the neighborhood solution is calculated based on the preset objective function. Based on the current temperature, the first objective function value, and the second objective function value, the acceptance probabilities of the initial solution and the neighborhood solution are determined, where the path with a higher acceptance probability is more likely to be selected as the current solution. That is, if the acceptance probability of the initial solution is greater than that of the neighborhood solution, the initial solution is determined as the current solution; if the acceptance probability of the initial solution is less than that of the neighborhood solution, the neighborhood solution is determined as the current solution. Furthermore, it is determined whether the current solution meets the stopping condition. If so, the current solution is selected as the optimal inspection path, i.e., the target inspection path. If not, the current temperature is lowered, and the above steps for determining the current solution are repeated until a preset number of iterations is reached, or the current temperature drops below a preset threshold, to obtain the target inspection path.

[0118] S103: Send the target inspection path to the target UAV, and control the target UAV to inspect the target detection area according to the target inspection path.

[0119] It can be understood that when the target UAV inspects the target detection area according to the target inspection path, it can avoid obstacles in the target detection area, ensuring that the target UAV completes the inspection task efficiently while ensuring safety.

[0120] S104: Obtain inspection images taken by the target UAV during the inspection process, and input the inspection images into a pre-trained object detection model.

[0121] Explanatory note: The target drone flies along the target inspection route, ensuring a comprehensive and detailed inspection of all key areas within the target inspection area. During the inspection process, the target drone also serves as a filming tool, using its internal high-definition cameras, sensors and other equipment to capture inspection images of key areas within the target inspection area.

[0122] The inspection images are then fed into a pre-trained object detection model, which detects unusual objects within them. In one possible implementation, to reduce transmission latency and improve image processing efficiency, the object detection model can be deployed on a mobile edge computing platform located at a ground base station or edge node on the target drone.

[0123] After receiving the inspection image, the object detection model divides the inspection image into multiple grids and detects abnormal objects in each grid. This gridded processing method enables the object detection model to process the inspection image more efficiently, improving the speed and accuracy of abnormal object detection. In each grid, the object detection model sets a bounding box for the detected abnormal object, and determines the position and size of the bounding box, as well as the category of the abnormal object in the bounding box and the confidence of the bounding box. Among them, the bounding box is used to locate the specific position of the abnormal object in the inspection image, the category of the abnormal object indicates the type of abnormal object, and the confidence of the bounding box reflects the degree of certainty of the object detection model about the detected abnormal object. The confidence calculation formula is:

[0124]

[0125] Where, Indicates the The confidence of the bounding box, Indicates the The probability that the grid where the bounding box is located contains an abnormal object, Represents the intersection-over-union ratio between the predicted bounding box and the ground-truth bounding box. The predicted bounding box refers to the bounding box set by the object detection model for the abnormal object in the inspection image, and the ground-truth bounding box refers to the actual bounding box corresponding to the abnormal object category.

[0126] It's important to note that before inputting inspection images into the object detection model, they must be preprocessed. This preprocessing includes, but is not limited to, image enhancement, denoising, and resizing. Image enhancement enhances the contrast, brightness, and sharpness of inspection images, making details clearer. Denoising removes noise and interference from inspection images, improving their purity. Resizing adjusts the inspection images to a size suitable for the object detection model to ensure detection accuracy. Preprocessing inspection images further improves the object detection model's ability to detect inspection images, allowing for more accurate detection of abnormal objects within them.

[0127] It should be understood that the object detection model is pre-trained and has been trained and optimized using a large number of historical inspection images.

[0128] Specifically, a deep learning model with object detection capabilities was selected as the initial model. Historical inspection images of the target overhead line were obtained. These images, acquired through long-term inspections and recording, contain actual image data of various abnormal objects. These images were then carefully annotated to clearly indicate the presence, location, and category of abnormal objects. This ensured that the initial model could accurately learn the characteristics of abnormal objects during training.

[0129] Next, a dataset is constructed based on the annotated historical inspection images and divided into a training set and a test set according to a preset ratio. The training set is used for model training, while the test set is used to optimize model performance. The initial model is trained based on the training set, and its parameters are continuously adjusted to improve its accuracy and robustness in detecting abnormal objects, resulting in a trained initial model. The trained initial model is further tested on the test set to obtain its loss and accuracy. The model parameters are then optimized based on these values ​​until the preset conditions are met, resulting in an object detection model.

[0130] Among them, by calculating the loss value and accuracy, we can intuitively understand the detection effect of the model. The calculation formula of the loss value is:

[0131]

[0132] Where, 、 Respectively represent the center point of the predicted bounding box Axis and The coordinates on the axis, 、 Represents the center point of the true bounding box in Axis and Coordinates on the axis. 、 denote the width and height of the predicted bounding box respectively, 、 denote the width and height of the ground-truth bounding box respectively. 、 represent the predicted confidence and the true confidence respectively. represents the category probability of abnormal objects predicted by the object detection model, The label representing the true category of the abnormal object. and represents the weight coefficient, is the number of samples in the test set.

[0133] S105: Determine the detection result output by the object detection model, and identify whether there are any abnormal hidden dangers in the target detection area based on the detection result.

[0134] The object detection model outputs interpretable detection results, including the presence, category, and location of abnormal objects within the inspection image. Abnormal objects are objects that could cause damage to the target overhead line, including but not limited to tree branches, plastic bags, kites, and bird nests. The object detection model can quickly and accurately identify abnormal objects in inspection images, allowing for the timely detection and resolution of potential anomalies and hazards, effectively ensuring the safe operation of the target overhead line.

[0135] Specifically, if the detection result indicates the presence of an abnormal object in the inspection image, the category of the abnormal object and the location of the abnormal object in the inspection image are determined. Based on the category of the abnormal object and its location in the inspection image, the hazard level of the abnormal object is determined in combination with a preset abnormal object hazard level database. The hazard level of the abnormal object refers to the potential damage or risk level that the abnormal object may cause to the target overhead line. For example, tree branches may come into contact with the target overhead line due to growth, causing the target overhead line to short-circuit or break; plastic bags and kites may become entangled in the target overhead line due to wind, affecting the normal operation of the target overhead line; bird nests may cause a short circuit in the target overhead line due to bird activity.

[0136] It should be noted that the preset abnormal object hazard level database stores the hazard levels corresponding to different categories of abnormal objects. This hazard level information is determined based on a combination of factors such as historical data, expert experience, and industry standards. By querying the preset abnormal object hazard level database, the hazard level corresponding to abnormal objects detected by the object detection model can be quickly and accurately obtained.

[0137] The actual position of the abnormal object within the target detection area is then calculated based on the preset three-dimensional coordinates and the position of the abnormal object in the inspection image. This calculation of the actual position of the abnormal object within the target detection area may incorporate not only the three-dimensional coordinates and the position of the abnormal object in the inspection image, but also the target inspection path and the flight parameters of the target drone, without specific limitations here.

[0138] Among them, the flight parameters of the target UAV include flight altitude, shooting angle, etc. Based on the flight altitude and shooting angle of the target UAV, the vertical distance and horizontal position of the abnormal object in the inspection image relative to the target UAV can be determined. Further combined with the target inspection path, that is, the movement trajectory of the target UAV during the inspection process, the actual position of the abnormal object in the target detection area can be calculated. In addition, by combining the flight parameters of the target UAV and the image information of the abnormal object, such as the size and shape of the abnormal object in the inspection image, the three-dimensional coordinates of the abnormal object in the target detection area can be further calculated. The three-dimensional coordinates include the vertical height, horizontal position, and distance of the abnormal object relative to the target overhead line.

[0139] Next, based on the abnormal object's hazard level and its actual location within the target detection area, the intrusion range of the abnormal object into the target detection area is calculated. The intrusion range refers to the spatial extent or impact area occupied by the abnormal object within the target detection area, including the width, depth, and area of ​​the intrusion. The depth of the intrusion represents the shortest distance between the abnormal object and the target overhead line, while the width represents the horizontal range occupied by the abnormal object near the target overhead line. The intrusion range comprehensively considers the abnormal object's hazard level, its actual location within the target detection area, and environmental factors within the target detection area to determine the extent of the abnormal object's impact on the safe operation of the target overhead line.

[0140] After calculating the range of the abnormal object's intrusion into the target detection area, the system analyzes whether the target detection area contains any potential abnormalities based on the abnormal object's hazard level and intrusion range. A potential abnormality refers to the possibility that the abnormal object will come into contact with a target overhead power line within the target detection area, either within the current time period or within a preset future time period. For example, an abnormal object like a tree branch, if it is lush and close to the target overhead power line, could potentially come into contact with the target power line. Furthermore, abnormal objects like plastic bags or kites could be blown by the wind and become entangled in the target overhead power line.

[0141] If analysis determines that the target detection area contains anomalies or potential safety hazards, an anomaly alert signal is sent to the target personnel, prompting them to clear the abnormal objects from the target detection area. Furthermore, the target drone's inspection strategy and frequency can be adjusted to better monitor and prevent potential safety risks in the target detection area. Subsequently, the target personnel provide feedback on the abnormal object clearance results. If the results indicate that the abnormal objects have been cleared, the target detection area's risk level is updated, along with the ranking of the initial detection area's risk level and the updated target detection area's risk level. Based on this ranking, a new target detection area is selected for abnormality and potential safety risk identification.

[0142] Explanatory note: the target detection area is the area with the highest current risk level, and the target detection area is prioritized for abnormal hidden danger identification. When an abnormal hidden danger is identified in the target detection area, the target staff is notified to handle it, thereby ensuring the safety of the target overhead line in the target detection area. The target detection area processed by the target staff is marked, and the risk level of the target detection area is updated, and then the initial detection area of ​​the next risk level is used as the new target detection area. And so on, until all initial detection areas have completed the abnormal hidden danger identification. In this way, comprehensive coverage of the entire target overhead line can be achieved, ensuring that each area is effectively monitored and abnormal hidden dangers are identified.

[0143] In summary, the overall process of the method for identifying abnormal hidden dangers of overhead lines based on drones provided in this embodiment is described. Figure 2 Schematic diagram of the process of identifying abnormal hidden dangers of overhead lines based on drones provided in the embodiment of the present application Figure 2 ,like Figure 2 As shown in FIG, the overall process of the abnormal hidden danger identification method of overhead lines based on drones is as follows:

[0144] S201, dividing the target overhead line into multiple initial detection areas based on the distribution parameters and design parameters of the target overhead line, and selecting a target detection area from the multiple initial detection areas;

[0145] S202, obtaining environmental parameters of the target detection area, and formulating candidate inspection paths for the target detection area according to the environmental parameters based on a preset path design algorithm;

[0146] S203, optimizing the candidate inspection path using a preset optimization algorithm, and determining the optimized candidate inspection path as the target inspection path;

[0147] S204: Send the target inspection path to the target UAV, and control the target UAV to inspect the target inspection area according to the target inspection path;

[0148] S205: Obtain inspection images taken by the target UAV during the inspection process, send the inspection images to a pre-trained object detection model, and obtain detection results output by the object detection model;

[0149] S206. If the detection result indicates that an abnormal object exists in the inspection image, determine the hazard level of the abnormal object and the extent of the abnormal object's intrusion into the target detection area;

[0150] S207. Based on the hazard level and intrusion range, analyze whether there are any abnormal hidden dangers in the target detection area.

[0151] S208. If there is an abnormal hidden danger in the target detection area, an abnormal alarm signal is sent to the target staff to prompt the target staff to go to the target detection area to clean up the abnormal objects.

[0152] The present application provides a method for identifying abnormal hidden dangers of overhead lines based on drones. Based on the distribution parameters and design parameters of the target overhead lines, the target overhead lines are divided into multiple initial detection areas, and the target detection areas are selected from the initial detection areas according to the risk level of each initial detection area. Afterwards, a preset path design algorithm is used to formulate candidate inspection paths for the target detection areas according to the environmental parameters of the target detection areas, and a preset optimization algorithm is used to optimize the candidate inspection paths, and the optimized candidate inspection paths are determined as the target inspection paths. By dividing the target overhead lines into multiple small inspection areas, detailed inspections of each inspection area can be achieved, laying the foundation for improving the accuracy of identifying abnormal hidden dangers of overhead lines. Furthermore, the target inspection path is sent to the target drone, and the target drone is controlled to inspect the target inspection area according to the inspection path, and the inspection images taken by the target drone during the inspection process are obtained. Since the environmental parameters of the target detection area are taken into account when formulating the target inspection path, the target inspection path obtained can avoid the routes in the target detection area that may affect the flight of the target drone, which means that the target drone can efficiently complete the inspection task by inspecting according to the target inspection path. The inspection image is input into a pre-trained object detection model to obtain the detection result output by the object detection model. If the detection result indicates that there are abnormal objects in the inspection image, the target detection area is analyzed based on the hazard level of the abnormal object and the intrusion range of the abnormal object into the target detection area. Since the target drone inspects in a small target detection area according to the target inspection path, its field of view is no longer limited. Therefore, the target drone can capture relatively clear inspection images, and then the object detection model can also output relatively accurate detection results based on the relatively clear inspection images. Therefore, the present application can improve the efficiency and accuracy of identifying abnormal hidden dangers in overhead lines.

[0153] Figure 3 A schematic diagram of the structure of the abnormal hidden danger identification device for overhead lines based on drones provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the abnormal hidden danger identification device 300 of overhead lines based on drones includes: a division module 301, a processing module 302, an inspection module 303, and an identification module 304;

[0154] The division module 301 is configured to divide the target overhead line into a plurality of initial detection areas based on the distribution parameters and design parameters of the target overhead line; wherein the target overhead line is an overhead line for which abnormal hidden dangers need to be identified; the distribution parameters represent the layout of the target overhead line in geographic space, and the design parameters indicate the structural characteristics of the target overhead line;

[0155] The processing module 302 is configured to select a target detection area from the multiple initial detection areas; and formulate a target inspection path for the target detection area using a preset path design algorithm based on the environmental parameters of the target detection area;

[0156] The inspection module 303 is used to send the target inspection path to the target UAV and control the target UAV to inspect the target detection area according to the target inspection path;

[0157] The processing module 302 is further configured to obtain inspection images captured by the target UAV during the inspection process and input the inspection images into a pre-trained object detection model; wherein the object detection model is configured to detect abnormal objects in the inspection images;

[0158] The identification module 304 is used to determine the detection results output by the object detection model and identify whether there are any abnormal hidden dangers in the target detection area based on the detection results.

[0159] In one possible design, the processing module 302 further includes: a formulation module 305, an optimization module 306,

[0160] A formulation module 305 is configured to formulate a candidate inspection path for the target detection area based on environmental parameters of the target detection area using a preset path design algorithm;

[0161] The optimization module 306 is configured to optimize the candidate inspection path using a preset optimization algorithm, and determine the optimized candidate inspection path as the target inspection path.

[0162] In a possible design, the formulation module 305 further includes: a construction module 307, a setting module 308, a connection module 309, and a determination module 310.

[0163] A construction module 307 is used to construct a target inspection model based on the environmental parameters of the target detection area; wherein the target inspection model is used to simulate the inspection path of the target UAV in the target detection area;

[0164] A setting module 308 is used to set an inspection start point and an inspection end point in the target inspection model based on a preset path design algorithm;

[0165] A connection module 309 is used to connect the inspection starting point and the inspection end point according to a preset rule;

[0166] The determination module 310 is configured to determine a connecting line between an inspection start point and an inspection end point as a candidate inspection path for the target inspection area.

[0167] In one possible design, the determination module 310 is further configured to determine the candidate inspection path as an initial solution of a preset optimization algorithm;

[0168] The optimization module 306 further includes: a calculation module 311, an addition module 312, and a comparison module 313.

[0169] A calculation module 311 is configured to calculate a first objective function value of an initial solution based on a preset objective function;

[0170] An adding module 312 is used to add a preset perturbation to the initial solution to obtain a neighborhood solution;

[0171] The calculation module 311 is further configured to calculate a second objective function value of the neighborhood solution based on a preset objective function;

[0172] A comparison module 313 is configured to compare the first objective function value with the second objective function value, and determine the initial solution or the neighborhood solution as the current solution based on the comparison result;

[0173] The optimization module 306 is further configured to iteratively optimize the current solution based on a preset optimization algorithm;

[0174] The determination module 310 is further configured to determine the optimized current solution as the optimized candidate inspection path, and determine the optimized candidate inspection path as the target inspection path.

[0175] In a possible design, the classification module 301 is further configured to classify the risk levels of the multiple initial detection areas based on preset dimension indicators;

[0176] The determination module 310 is further configured to determine the risk level of each initial detection area;

[0177] The processing module 302 further includes: a sorting module 314 for sorting the risk levels in descending order;

[0178] The determination module 310 is further configured to determine the initial detection area corresponding to the risk level ranked first as the target detection area.

[0179] In one possible design, the determination module 310 is further configured to:

[0180] If the detection result indicates that an abnormal object exists in the inspection image, then determining the category of the abnormal object and the location of the abnormal object in the inspection image;

[0181] Based on the category of the abnormal object and its location in the inspection image, combined with the preset abnormal object hazard level database, the hazard level of the abnormal object is determined;

[0182] The calculation module 311 is further configured to calculate the actual position of the abnormal object within the target detection area based on the preset three-dimensional coordinates and the position of the abnormal object in the inspection image; and calculate the invasion range of the abnormal object into the target detection area based on the actual position and the hazard level; wherein the invasion range includes the depth, width, and invasion area of ​​the invasion;

[0183] The identification module 304 also includes: an analysis module 315, which is used to analyze whether there are abnormal hidden dangers in the target detection area based on the hazard level and the intrusion range; wherein the abnormal hidden danger indicates that an abnormal object comes into contact with the target overhead line in the target detection area within the current time period or a preset time period in the future.

[0184] In a possible design, the device 300 for identifying abnormal hidden dangers of overhead lines based on drones further includes: a sending module 316, a receiving module 317, and an updating module 318.

[0185] The sending module 316 is used to send an abnormality alarm signal to the target staff if there is an abnormal hidden danger in the target detection area; wherein the abnormality alarm signal is used to prompt the target staff to go to the target detection area to clear the abnormal object;

[0186] A receiving module 317 is used to receive the abnormal object cleaning result fed back by the target staff;

[0187] An updating module 318 for updating the risk level of the target detection area and the ranking result of the risk level of the initial detection area and the updated risk level of the target detection area if the abnormal object cleaning result indicates that the abnormal object has been cleared;

[0188] The identification module 304 is further configured to reselect a new target detection area for abnormal hidden danger identification based on the sorting result.

[0189] The device for identifying abnormal hidden dangers of overhead lines based on drones provided in the embodiments of the present application can be used to execute the method for identifying abnormal hidden dangers of overhead lines based on drones in any of the above embodiments. Its implementation principles and technical effects are similar and will not be repeated here.

[0190] It should be noted that it should be understood that the division of the various modules of the above device is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. Moreover, these modules can all be implemented in the form of software called by processing elements; they can also all be implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. In addition, these modules can be fully or partially integrated together or implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the hardware integrated logic circuit in the processor element or by instructions in the form of software.

[0191] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present application. Figure 4 As shown, the electronic device may include: a transceiver 41 , a processor 42 , and a memory 43 .

[0192] The processor 42 executes the computer-executable instructions stored in the memory, so that the processor 42 implements the solutions in the above-mentioned embodiments. The processor 42 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0193] The memory 43 is connected to the processor 42 via a system bus and communicates with the processor 42. The memory 43 is used to store computer program instructions.

[0194] The transceiver 41 can be used to communicate with other devices.

[0195] The system bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, among others. System buses can be divided into address buses, data buses, and control buses. For ease of illustration, the diagram uses only a single thick line, but this does not imply a single bus or type of bus. Transceivers enable communication between the database access device and other computers (such as clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and non-volatile memory.

[0196] The electronic device provided in the embodiments of the present application can be used to execute the method provided in any of the above embodiments. Its implementation principles and technical effects are similar and will not be repeated here.

[0197] An embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored. When the computer instructions are executed on a computer, the computer executes the method provided in any of the above embodiments.

[0198] An embodiment of the present application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when at least one processor executes the computer program, it can implement the method provided in any of the above embodiments.

[0199] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or module, which can be electrical, mechanical or other forms.

[0200] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of these modules may be selected to implement the solution of this embodiment based on actual needs.

[0201] In addition, the functional modules in the various embodiments of the present application may be integrated into a single processing unit, or each module may exist physically separately, or two or more modules may be integrated into a single unit. The above-mentioned modules may be implemented in the form of hardware or hardware plus software functional units.

[0202] The integrated modules implemented in the form of software function modules can be stored in a computer-readable storage medium. The software function modules stored in a storage medium include a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute some of the steps of the methods of various embodiments of the present application.

[0203] It should be understood that the processor described above may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0204] The memory may include a high-speed RAM memory, and may also include non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk.

[0205] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0206] The storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0207] An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic control unit or a main control device.

[0208] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0209] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying abnormal hidden dangers of overhead lines based on drones, characterized in that: include: Based on distribution parameters and design parameters of a target overhead line, the target overhead line is divided into a plurality of initial detection areas; wherein the target overhead line is an overhead line for which abnormal hidden dangers need to be identified; the distribution parameters represent the layout of the target overhead line in geographic space, and the design parameters indicate the structural characteristics of the target overhead line; Selecting a target detection area from the multiple initial detection areas; formulating a target inspection path for the target detection area using a preset path design algorithm based on environmental parameters of the target detection area; Sending the target inspection path to the target UAV, and controlling the target UAV to inspect the target detection area according to the target inspection path; Obtaining inspection images taken by the target UAV during the inspection process, and inputting the inspection images into a pre-trained object detection model; wherein the object detection model is used to detect abnormal objects in the inspection images; Determine the detection result output by the object detection model, and based on the detection result, identify whether there are any abnormal hidden dangers in the target detection area.

2. The method according to claim 1, characterized in that The method of formulating a target inspection path for the target detection area using a preset path design algorithm based on the environmental parameters of the target detection area includes: Using a preset path design algorithm, a candidate inspection path is formulated for the target detection area according to the environmental parameters of the target detection area; The candidate inspection path is optimized using a preset optimization algorithm, and the optimized candidate inspection path is determined as the target inspection path.

3. The method according to claim 2, characterized in that The method of using a preset path design algorithm to formulate a candidate inspection path for the target detection area according to the environmental parameters of the target detection area includes: Based on the environmental parameters of the target detection area, a target inspection model is constructed; wherein the target inspection model is used to simulate the inspection path of the target UAV in the target detection area; Based on the preset path design algorithm, an inspection starting point and an inspection end point are set in the target inspection model, and the inspection starting point and the inspection end point are connected according to a preset rule; A connecting line between the inspection starting point and the inspection end point is determined as a candidate inspection path for the target detection area.

4. The method according to claim 2, characterized in that The step of optimizing the candidate inspection path by using a preset optimization algorithm and determining the optimized candidate inspection path as the target inspection path includes: Determine the candidate inspection path as an initial solution of the preset optimization algorithm; and calculate a first objective function value of the initial solution based on a preset objective function; Adding a preset disturbance to the initial solution to obtain a neighborhood solution; calculating a second objective function value of the neighborhood solution based on the preset objective function; Comparing the first objective function value with the second objective function value, and determining the initial solution or the neighborhood solution as the current solution according to the comparison result; Based on the preset optimization algorithm, the current solution is iteratively optimized; the optimized current solution is determined as the optimized candidate inspection path, and the optimized candidate inspection path is determined as the target inspection path.

5. The method according to claim 1, wherein The selecting a target detection area from the multiple initial detection areas includes: Based on preset dimension indicators, the multiple initial detection areas are divided into risk levels; Determining a risk level for each initial detection area; and ranking the risk levels in descending order; The initial detection area corresponding to the risk level ranked first is determined as the target detection area.

6. The method according to any one of claims 1 to 5, characterized in that The step of identifying whether there is an abnormal hidden danger in the target detection area based on the detection result includes: If the detection result indicates that an abnormal object exists in the inspection image, determining the category of the abnormal object and the position of the abnormal object in the inspection image; Determining the hazard level of the abnormal object based on the category of the abnormal object and the position in the inspection image in combination with a preset abnormal object hazard level database; Calculate the actual position of the abnormal object within the target detection area based on the preset three-dimensional coordinates and the position of the abnormal object in the inspection image; calculate the intrusion range of the abnormal object into the target detection area based on the actual position and the hazard level; wherein the intrusion range includes the depth, width, and intrusion area of ​​the intrusion; Based on the hazard level and the intrusion range, the target detection area is analyzed to determine whether there are any abnormal hidden dangers; wherein the abnormal hidden danger indicates that the abnormal object comes into contact with the target overhead line in the target detection area within the current time period or a preset time period in the future.

7. The method according to claim 6, characterized in that The method further comprises: If there is an abnormal hidden danger in the target detection area, an abnormal alarm signal is sent to the target staff; wherein the abnormal alarm signal is used to prompt the target staff to go to the target detection area to clean up the abnormal object; Receive the abnormal object cleaning result fed back by the target staff; if the abnormal object cleaning result indicates that the abnormal object has been cleared, update the risk level of the target detection area, and the ranking result of the risk level of the initial detection area and the updated risk level of the target detection area; based on the ranking result, reselect a new target detection area for abnormal hidden danger identification.

8. A device for identifying abnormal hidden dangers of overhead lines based on drones, characterized in that: include: a partitioning module configured to partition a target overhead line into a plurality of initial detection areas based on distribution parameters and design parameters of the target overhead line; wherein the target overhead line is an overhead line requiring abnormal hidden danger identification; the distribution parameters characterize the layout of the target overhead line in geographic space, and the design parameters indicate the structural characteristics of the target overhead line; A processing module is configured to select a target detection area from the multiple initial detection areas; and formulate a target inspection path for the target detection area using a preset path design algorithm based on environmental parameters of the target detection area; An inspection module is used to send the target inspection path to the target UAV, and control the target UAV to inspect the target detection area according to the target inspection path; The processing module is further configured to obtain inspection images captured by the target drone during the inspection process, and input the inspection images into a pre-trained object detection model; wherein the object detection model is configured to detect abnormal objects in the inspection images; The recognition module is used to determine the detection results output by the object detection model and identify whether there are any abnormal hidden dangers in the target detection area based on the detection results.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the abnormal hidden danger identification method for overhead lines based on a drone as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method for identifying abnormal hidden dangers of overhead lines based on a drone according to any one of claims 1 to 7.

Images