Unmanned aerial vehicle flight path intelligent planning method and system for rural power grid inspection

Abstract

The application discloses a UAV flight path intelligent planning method and system for rural power grid inspection, and relates to the technical field of UAV control. The method comprises the following steps: based on the task association information of a target power grid area, drawing a power stealing risk distribution map, a terrain risk distribution map and a line risk distribution map; based on the power stealing risk distribution map, the terrain risk distribution map and the line risk distribution map, and in combination with a preset weight ratio, evaluating a task block to obtain a task priority factor; in a pre-constructed three-dimensional flight task planning space, based on the attribute information of a UAV to be executed and the task priority factor, performing UAV flight path optimization with the aim of minimizing the total task cost, and outputting an optimal path planning scheme; and controlling the UAV to perform power stealing detection and inspection on the target power grid area according to the optimal path planning scheme. The application effectively improves the rural power grid inspection efficiency and the precision of preventing and controlling power stealing behavior.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle control, in particular to an unmanned aerial vehicle flight path intelligent planning method and system for rural power grid inspection. BACKGROUND

[0002] With the wide application of unmanned aerial vehicle technology in the power industry, the autonomous flight mode based on preset waypoints has become the mainstream scheme for power grid inspection. This type of operation mode plans the flight route in advance through the ground station, and integrates GPS positioning and basic obstacle avoidance functions, which has good applicability in cities and flat areas.

[0003] However, this method still faces certain challenges in rural complex environments. The current inspection strategy mostly adopts a uniform coverage fixed flight line mode, without differentiated path planning in combination with factors such as electricity theft risk, resulting in insufficient coverage of high-risk areas for electricity theft, and difficulty in effectively discovering and preventing electricity theft behavior. At the same time, in the single machine operation mode, limited by the endurance and fixed flight line planning, it is difficult to achieve key coverage of large-scale scattered power grid facilities within limited resources, resulting in uneven allocation of inspection resources and limited overall inspection efficiency. SUMMARY

[0004] The present application provides an unmanned aerial vehicle flight path intelligent planning method and system for rural power grid inspection, aiming to solve the technical problems of low efficiency and insufficient accuracy of electricity theft behavior identification in the prior art.

[0005] In view of the above problems, the present application provides an unmanned aerial vehicle flight path intelligent planning method and system for rural power grid inspection.

[0006] In a first aspect, the present application provides an unmanned aerial vehicle flight path intelligent planning method for rural power grid inspection, comprising:

[0007] Based on the M task association information of the M task blocks of the target power grid area, draw a power theft risk distribution map, a terrain risk distribution map and a line risk distribution map;

[0008] Based on the power theft risk distribution map, the terrain risk distribution map and the line risk distribution map, evaluate the M task blocks in combination with a preset weight ratio to obtain M task priority factors;

[0009] In a pre-constructed three-dimensional flight task planning space, based on the attribute information of N unmanned aerial vehicles to be executed and the M task priority factors, the unmanned aerial vehicle flight path is optimized to minimize the total task cost, and an optimal path planning scheme is output;

[0010] According to the optimal path planning scheme, control the N unmanned aerial vehicles to perform power theft detection and inspection on the target power grid area.

[0011] In a second aspect, the application provides a UAV flight path intelligent planning system for rural power grid inspection, comprising:

[0012] a risk distribution map drawing module configured to draw a power theft risk distribution map, a terrain risk distribution map and a line risk distribution map based on M task association information of M task blocks of a target power grid area;

[0013] a task priority evaluation module configured to evaluate the M task blocks based on the power theft risk distribution map, the terrain risk distribution map and the line risk distribution map and in combination with a preset weight ratio to obtain M task priority factors;

[0014] an optimal path planning module configured to perform UAV flight path optimization in a pre-constructed three-dimensional flight task planning space based on attribute information of N UAVs to be executed and the M task priority factors, with the objective of minimizing total task cost, and output an optimal path planning scheme;

[0015] a UAV inspection execution module configured to control the N UAVs to perform power theft detection and inspection on the target power grid area according to the optimal path planning scheme.

[0016] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0017] The application provides a UAV flight path intelligent planning method and system for rural power grid inspection, which realizes precise coverage of high-risk areas of rural power grids and efficient use of limited inspection resources through risk-driven differentiated inspection path planning and multi-UAV collaborative operation, realizes the transition from indiscriminate full coverage to data-driven, risk-prioritized and resource-adapted precise inspection, and significantly improves the inspection efficiency of rural power grids and the precision of preventing and controlling power theft behavior. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort.

[0019] Figure 1 A flowchart of a UAV flight path intelligent planning method for rural power grid inspection provided by the embodiments of the application is shown in the figure.

[0020] Figure 2 A structural diagram of a UAV flight path intelligent planning system for rural power grid inspection provided by the embodiments of the application is shown in the figure.

[0021] The components represented by each number in the attached diagram are explained below:

[0022] Risk distribution map drawing module 11, task priority evaluation module 12, optimal path planning module 13, and UAV inspection execution module 14. Detailed Implementation

[0023] This application provides a method and system for intelligent flight path planning of unmanned aerial vehicles (UAVs) for rural power grid inspection, which is intended to address the technical problems of low efficiency in rural power grid inspection and insufficient accuracy in identifying electricity theft in existing technologies.

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0025] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0026] Example 1, as Figure 1 As shown, this application provides an intelligent flight path planning method for unmanned aerial vehicles (UAVs) used for rural power grid inspection, the method comprising:

[0027] S100: Based on the M task association information of M task blocks in the target power grid area, draw the electricity theft risk distribution map, terrain risk distribution map and line risk distribution map.

[0028] In this embodiment, based on the task association information of M task blocks in the target power grid area, a power theft risk distribution map, a terrain risk distribution map, and a line risk distribution map are drawn. This step aims to transform a large-scale, complex rural power grid inspection area into a series of quantifiable and visually analyzable risk layers. By constructing risk distribution maps, various heterogeneous information can be standardized, processed, and fused for evaluation, laying a data foundation for the subsequent generation of risk-oriented, precise inspection strategies.

[0029] Step S100 in the method provided in this application embodiment includes:

[0030] According to the preset line distance, the target rural power grid inspection area is divided into M task blocks, where the preset line distance is 5 kilometers and M is an integer greater than 1;

[0031] According to the electricity stealing associated information, the terrain associated information and the line associated information of the M task blocks, electricity stealing risk assessment, terrain risk assessment and line risk assessment are respectively performed, and electricity stealing risk coefficient distribution, terrain risk coefficient distribution and line risk coefficient distribution are output;

[0032] Based on the electricity stealing risk coefficient distribution, the terrain risk coefficient distribution and the line risk coefficient distribution, electricity stealing risk distribution map, terrain risk distribution map and line risk distribution map are respectively drawn.

[0033] The electricity stealing associated information includes historical electricity stealing records, user electricity consumption anomaly data and seasonal electricity stealing rules, the terrain associated information includes terrain complexity parameters extracted from satellite remote sensing data, and the line associated information includes line voltage grade and type, line aging degree and line importance, wherein the terrain complexity parameters at least include slope, slope direction and vegetation coverage.

[0034] Firstly, the target rural power grid inspection area is divided into M task blocks according to a preset line distance, wherein the preset line distance is 5 kilometers, and M is an integer greater than 1. The preset line distance is a pre-set line length standard for splitting the inspection area; the task block is an electric grid area unit with independent inspection boundary and evaluation attribute after splitting. The geographical scope of the target rural power grid inspection area and the total length of the transmission line covered are first determined, and then 5 kilometers is taken as the fixed splitting basis, and the M task blocks that do not overlap and completely cover the target area are divided along the line direction, so as to ensure that the line length in each task block does not exceed 5 kilometers, and M is determined according to the total line length and is greater than 1. For example, the target rural power grid inspection area covers a total length of 40 kilometers of transmission line, and is split according to the 5 kilometer standard along the line from the starting point to the ending point, and finally 8 task blocks are obtained, M = 8. Each task block contains 5 kilometers of line, and corresponds to a clear geographical boundary, covering the electricity consumption area and terrain range around the line section.

[0035] Secondly, according to the electricity stealing associated information, the terrain associated information and the line associated information of the M task blocks, the electricity stealing risk assessment, the terrain risk assessment and the line risk assessment are respectively carried out, and the electricity stealing risk coefficient distribution, the terrain risk coefficient distribution and the line risk coefficient distribution are output. The electricity stealing associated information is a set of feature data related to electricity stealing behavior, including historical electricity stealing records, user electricity abnormal data and seasonal electricity stealing rules; the terrain associated information is a set of parameters reflecting the terrain conditions, including the terrain complexity parameters extracted from satellite remote sensing data, and the terrain complexity parameters at least include slope, slope direction and vegetation coverage; the line associated information is a set of information describing the properties and state of the line itself, including line voltage grade and type, line aging degree and line importance; the risk coefficient distribution is the global distribution data in the form of quantitative coefficients presenting the risk assessment results of each task block. For each task block, the corresponding electricity stealing, terrain and line associated information is extracted, and a preset evaluation model such as fuzzy comprehensive evaluation method is used for quantitative processing and risk level determination, and the corresponding electricity stealing, terrain and line risk coefficients of each task block are output. The same coefficients of all task blocks are integrated to form three types of global risk coefficient distributions, and the risk coefficient value ranges from 0 to 1, and the larger the value is, the higher the risk is.

[0036] For example, taking 8 task blocks divided by the target area as an example, the evaluation of task block 3 is focused on. The electricity stealing associated information: 3 times of historical electricity stealing records, 3 households of user electricity fluctuation over 50%, and in line with seasonal electricity stealing rules in summer and winter; the terrain associated information: average slope 35°, shade slope proportion 70%, vegetation coverage 70%; the line associated information: 10kV overhead line, aging degree medium, importance high. Using fuzzy comprehensive evaluation method combined with index weight calculation, the electricity stealing risk coefficient of task block 3 is 0.82, the terrain risk coefficient is 0.75, and the line risk coefficient is 0.68. The remaining 7 task blocks are evaluated according to the same logic, and finally the global coefficient distribution is formed: the electricity stealing risk coefficient distribution: {0.35, 0.42, 0.82, 0.51, 0.63, 0.71, 0.48, 0.55}; the terrain risk coefficient distribution: {0.28, 0.33, 0.75, 0.41, 0.58, 0.62, 0.39, 0.46}; the line risk coefficient distribution: {0.40, 0.45, 0.68, 0.53, 0.61, 0.70, 0.50, 0.56}.

[0037] Finally, based on the electricity stealing risk coefficient distribution, the terrain risk coefficient distribution and the line risk coefficient distribution, an electricity stealing risk distribution map, a terrain risk distribution map and a line risk distribution map are drawn respectively. The risk distribution map is a visual chart that carries the risk coefficient distribution in a geographical space map and identifies the risk difference through gradient color or legend. Taking the geographical space map of the target area as the base map, the geographical boundaries of each task block are accurately mapped to the base map, and then according to the three types of risk coefficient distributions, gradient color is used for annotation, such as light red to dark red corresponding to low to high risk coefficient, each task block is filled with the corresponding color, and a legend is added to explain the corresponding relationship between the color and the risk coefficient, and finally three visual charts are generated, which respectively present the spatial distribution of electricity stealing risk, terrain risk and line risk.

[0038] For example, when drawing the electricity stealing risk distribution map based on the aforementioned coefficient distribution, the electricity stealing risk coefficient of task block 3 is 0.82, which is marked as dark red, the electricity stealing risk coefficient of task block 1 is 0.35, which is marked as light red, and the rest of the task blocks are marked according to the color gradient corresponding to the coefficient; in the terrain risk distribution map, the terrain risk coefficient of task block 3 is 0.75, which is marked as dark yellow, the terrain risk coefficient of task block 1 is 0.28, which is marked as light yellow, and the rest of the task blocks are marked according to the color gradient corresponding to the coefficient; in the line risk distribution map, the line risk coefficient of task block 3 is 0.68, which is marked as dark orange, the line risk coefficient of task block 1 is 0.40, which is marked as light orange, and the rest of the task blocks are marked according to the color gradient corresponding to the coefficient. The three charts clearly mark the task block number, geographical coordinate range and legend, and intuitively present the risk difference of different task blocks.

[0039] In the embodiments of the present application, the refinement of risk assessment is achieved by splitting the region, the comprehensiveness of risk analysis is ensured by independent evaluation of three types of factors, and the visualization of risk information is achieved by visual charts. The overall provides accurate, comprehensive and easy-to-apply basic data for subsequent weight ratio calculation task priority factor and unmanned aerial vehicle three-dimensional path optimization, ensuring the scientificity and feasibility of subsequent processes.

[0040] S200: Based on the electricity stealing risk distribution map, the terrain risk distribution map and the line risk distribution map, the M task blocks are evaluated to obtain M task priority factors according to the preset weight ratio.

[0041] In the embodiment of the present application, based on the electricity stealing risk distribution map, the terrain risk distribution map and the line risk distribution map, the M task blocks are evaluated in combination with a preset weight ratio to obtain M task priority factors. The three risk distribution maps generated in step S100 reflect relatively static and inherent risk conditions, but the actual inspection task execution is also strongly affected by seasonal temporary factors and real-time weather conditions. If only the static risk is fused based on the fixed weight, the formulated inspection plan may encounter a higher failure risk or low efficiency in actual execution. In this step, by introducing a dynamic influence coefficient, the weight ratio is optimized in real time, so that the finally calculated task priority factor can not only reflect the long-term inherent risk of the region, but also respond sensitively to short-term environmental changes, thereby guiding the generation of more robust and adaptive inspection paths.

[0042] The step S200 in the method provided in the embodiment of the present application comprises:

[0043] A first task block is randomly selected from the M task blocks, and based on historical unmanned aerial vehicle inspection records, a historical seasonal obstacle feature set of a region where the first task block is located corresponding to a current time zone is collected to perform seasonal obstacle influence evaluation to obtain a first seasonal obstacle influence coefficient;

[0044] Real-time environmental weather data of the region where the first task block is located in the current time zone is collected to perform line state influence evaluation to obtain a first line state influence coefficient;

[0045] The initial weight ratio is optimized based on the first seasonal obstacle influence coefficient and the first line state influence coefficient to obtain a preset weight ratio;

[0046] After the first electricity stealing risk coefficient, the first terrain risk coefficient and the first line risk coefficient of the first task block are dimensionless processed, the preset weight ratio is used for weighted calculation, a first task priority factor is output, and M task priority factors of the M task blocks are sequentially calculated.

[0047] First, a first task block is randomly selected from the M task blocks, a historical seasonal obstacle feature set of the region where the first task block is located in the current time zone is collected based on historical unmanned aerial vehicle inspection records, seasonal obstacle influence evaluation is performed, and a first seasonal obstacle influence coefficient is obtained. The seasonal obstacle feature set is a set of seasonal obstacle related data that affects the unmanned aerial vehicle inspection in the historical corresponding period in the region where the first task block is located; the seasonal obstacle influence coefficient is a parameter that quantifies the degree of interference of seasonal obstacles on the inspection, and the value range is 0-1, the larger the value, the stronger the interference. The terrain in rural areas is complex and changeable, and the preset fixed route cannot flexibly respond to seasonal changes in obstacles such as agricultural machinery during the harvesting period and temporary agricultural facilities, resulting in a task interruption rate of up to 30% or more. A first task block is randomly selected from M task blocks, historical unmanned aerial vehicle inspection records of the region where the task block is located are retrieved, seasonal obstacle data corresponding to the current time zone are extracted, the frequency of obstacles and the degree of interference on the inspection are statistically analyzed, and a first seasonal obstacle influence coefficient is calculated. For example, task block 3 is randomly selected from 8 task blocks as the first task block, it is summer in July, and it is also the early rice harvesting period in the local area. Historical inspection records are retrieved to find that in this region, during the harvesting period from July to August every year, field agricultural machinery operates an average of 12 times per month, and there are 8 temporary agricultural sheds that have caused a total of 4 inspection interruptions. The frequency weight is 0.4 and the interruption number weight is 0.6 by AHP, and the weighted calculation after statistical quantization gives the first seasonal obstacle influence coefficient 0.7.

[0048] Secondly, real-time environmental meteorological data of the area where the first task block is located within the current time zone is collected to assess the impact on line status and obtain the first line status impact coefficient. Real-time environmental meteorological data refers to meteorological monitoring data within the area of ​​the first task block in the current time zone, such as temperature, wind speed, humidity, and precipitation. The line status impact coefficient is a parameter that quantifies the degree of impact of real-time meteorological data on line status and inspection safety, ranging from 0 to 1; the larger the value, the more significant the impact. Real-time environmental meteorological data directly affects the operating status of rural power grid lines. For example, high temperatures can easily cause line overload, and strong winds may cause line swaying or affect the flight stability of drones, thus interfering with the inspection progress. Real-time environmental meteorological data of the area where the first task block is located is collected, and key indicators such as temperature, wind speed, and humidity are extracted. An assessment is conducted using a combination of statistical analysis and analytic hierarchy process (AHP). First, the AHP is used to determine the impact weight of each meteorological indicator on line status and inspection safety. Then, statistical analysis is used to quantify the data of each indicator into values ​​within the 0-1 range, and a weighted calculation is performed to obtain the first line status impact coefficient. For example, taking task block 3 as the first task block, currently in July (summer), real-time environmental meteorological data is collected through regional meteorological monitoring stations: temperature 35℃, wind speed 3.8m / s, relative humidity 60%, and no precipitation. The analytic hierarchy process (AHP) is used to determine the weights for temperature (0.4), wind speed (0.4), and humidity (0.2). Statistical analysis is used to quantify each indicator: 35℃ corresponds to 0.6, wind speed 3.8m / s to 0.5, and humidity 60% to 0.3. The weighted calculation yields the first line state influence coefficient: 0.6×0.4 + 0.5×0.4 + 0.3×0.2 = 0.24 + 0.2 + 0.06 = 0.5.

[0049] Furthermore, the initial weight allocation is optimized based on the first seasonal obstacle influence coefficient and the first line status influence coefficient to obtain a preset weight allocation.

[0050] Specifically, the initial weight allocation is optimized based on the first seasonal obstacle influence coefficient and the first line condition influence coefficient to obtain a preset weight allocation, including:

[0051] Obtain the initial weight ratio, wherein the initial weight ratio includes the initial electricity theft risk weight, the initial terrain risk weight, and the initial line risk weight, and the sum of the three is 1;

[0052] The ratio of the first seasonal obstacle influence coefficient to the preset standard seasonal obstacle influence coefficient is set as the terrain risk weight adjustment coefficient, and the initial terrain risk weight is compensated to obtain the adapted terrain risk weight.

[0053] The ratio of the first line status influence coefficient to the preset standard line status influence coefficient is set as the line risk weight adjustment coefficient, and the initial line risk weight is compensated to obtain the adapted line risk weight.

[0054] The adapted terrain risk weight is obtained by subtracting the adapted terrain risk weight and the adapted line risk weight from 1, and a preset weight ratio is generated based on the adapted terrain risk weight, the adapted line risk weight and the adapted electricity stealing risk weight.

[0055] Firstly, an initial weight ratio is obtained, wherein the initial weight ratio includes an initial electricity stealing risk weight, an initial terrain risk weight and an initial line risk weight, and the sum of the three is 1. The initial weight ratio is directly selected by preset initial electricity stealing risk weight, initial terrain risk weight and initial line risk weight, which needs to meet the requirement that the sum of the three risk weights is 1. For example, the initial electricity stealing risk weight is set to 0.7, the initial terrain risk weight is set to 0.2, and the initial line risk weight is set to 0.1, and the sum of the three is 1, which meets the ratio requirement.

[0056] Secondly, the ratio of the first seasonal obstacle influence coefficient to the preset standard seasonal obstacle influence coefficient is set as a terrain risk weight adjustment coefficient, the initial terrain risk weight is compensated to obtain an adapted terrain risk weight. The preset standard seasonal obstacle influence coefficient is a standard value of the seasonal obstacle influence coefficient preset for benchmark comparison. The terrain risk weight adjustment coefficient is calculated, that is, the ratio of the first seasonal obstacle influence coefficient to the preset standard seasonal obstacle influence coefficient, and the compensation of the initial terrain risk weight is completed by multiplying the adjustment coefficient by the initial terrain risk weight to obtain the adapted terrain risk weight. The calculation formula is: terrain risk weight adjustment coefficient = first seasonal obstacle influence coefficient / preset standard seasonal obstacle influence coefficient, and adapted terrain risk weight = initial terrain risk weight × terrain risk weight adjustment coefficient. For example, the first seasonal obstacle influence coefficient is 0.7, the preset standard seasonal obstacle influence coefficient is 0.5, the terrain risk weight adjustment coefficient = 0.7 / 0.5 = 1.4, the initial terrain risk weight is 0.2, and the adapted terrain risk weight = 0.2 × 1.4 = 0.28.

[0057] Further, a ratio of the first line state influence coefficient to a preset standard line state influence coefficient is set as a line risk weight adjustment coefficient, and the initial line risk weight is compensated to obtain an adapted line risk weight. The preset standard line state influence coefficient is a line state influence coefficient standard value preset for benchmark comparison. The line risk weight adjustment coefficient, i.e. the ratio of the first line state influence coefficient to the preset standard line state influence coefficient, is calculated, and the adjustment coefficient is multiplied by the initial line risk weight to complete the compensation of the initial line risk weight, and to obtain the adapted line risk weight. The calculation formula is: line risk weight adjustment coefficient = first line state influence coefficient / preset standard line state influence coefficient, and adapted line risk weight = initial line risk weight x line risk weight adjustment coefficient. For example, given that the first line state influence coefficient is 0.5, the preset standard line state influence coefficient is 0.4, the line risk weight adjustment coefficient = 0.5 / 0.4 = 1.25; the initial line risk weight is 0.1, and the adapted line risk weight = 0.1 x 1.25 = 0.125.

[0058] Further, 1 is subtracted from the adapted terrain risk weight and the adapted line risk weight to obtain an adapted electricity stealing risk weight, and a preset weight ratio is generated based on the adapted terrain risk weight, the adapted line risk weight and the adapted electricity stealing risk weight. The adapted terrain risk weight and the adapted line risk weight are subtracted from 1 to obtain the adapted electricity stealing risk weight, and the three together constitute the preset weight ratio conforming to the current scene. The calculation formula is: adapted electricity stealing risk weight = 1 - (adapted terrain risk weight + adapted line risk weight). The calculation result is that the adapted electricity stealing risk weight = 1 - (0.28 + 0.125) = 0.595, and the final preset weight ratio is: electricity stealing risk weight 0.595, terrain risk weight 0.28, and line risk weight 0.125.

[0059] Finally, after the first task block of the first electricity stealing risk coefficient, the first terrain risk coefficient and the first line risk coefficient are dimensionless, the preset weight ratio is weighted and calculated, the first task priority factor is output, and the M task priority factors of the M task blocks are calculated in turn. Dimensionless processing is a standardization processing method that eliminates the dimensional differences between different types of risk coefficients and makes each coefficient have additivity and comparability. The task priority factor is a parameter that quantifies the priority of the task block inspection. The value range is 0-1, and the larger the value, the higher the inspection priority. For each task block, extract its corresponding electricity stealing risk coefficient, terrain risk coefficient and line risk coefficient, and use the extreme value method for dimensionless processing. The coefficients of various types are uniformly mapped to the 0-1 interval to ensure that the coefficients can be directly involved in weighted calculation. Starting with the first task block, multiply the three types of risk coefficients after dimensionless processing by the corresponding weights in the preset weight ratio, and then sum the product results to obtain the first task priority factor. According to the same processing method and calculation logic, dimensionless processing and weighted calculation are performed on the remaining M-1 task blocks in turn, and finally M task priority factors corresponding to M task blocks are obtained.

[0060] For example, the three types of risk coefficients of the eight task blocks are all in the 0-1 interval and do not require additional dimensionless processing. The electricity stealing risk coefficient of task block 3 is 0.82, the terrain risk coefficient is 0.75, and the line risk coefficient is 0.68. The preset weight ratio is: electricity stealing risk weight 0.595, terrain risk weight 0.28, and line risk weight 0.125. The first task priority factor = 0.82*0.595+0.75*0.28+0.68*0.125≈0.488+0.21+0.085=0.783. The remaining seven task blocks are calculated according to the same formula, and finally eight task priority factors are obtained: 0.311, 0.364, 0.783, 0.452, 0.563, 0.632, 0.421, 0.487.

[0061] In the embodiments of the present application, by randomly selecting representative task blocks, combining historical seasonal data and real-time weather data to optimize the weight ratio, the dynamic adaptation of the weight to the actual inspection scene is realized. The dimensionless processing ensures the objectivity and comparability of the task priority factor, and accurately quantifies the inspection priority of each task block, providing a decision basis that meets the actual needs for subsequent unmanned aerial vehicle three-dimensional path optimization.

[0062] S300: In the pre-constructed three-dimensional flight task planning space, based on the attribute information of the N unmanned aerial vehicles to be executed and the M task priority factors, the unmanned aerial vehicle flight path optimization is performed to minimize the total task cost, and the optimal path planning scheme is output.

[0063] In the embodiments of the present application, in the pre-constructed three-dimensional flight task planning space, based on the attribute information of the N unmanned aerial vehicles to be executed and the M task priority factors, the unmanned aerial vehicle flight path optimization is performed with the objective of minimizing the total task cost, and an optimal path planning scheme is output. The rural power grid region has complex terrain, including poles, lines, vegetation and other three-dimensional obstacles, and two-dimensional path planning cannot avoid spatial conflicts, so a three-dimensional space needs to be constructed to ensure the safety of unmanned aerial vehicle flight. Meanwhile, different unmanned aerial vehicles have different performances, and the priority of the task block is also different. If the tasks or paths are blindly allocated, the total task cost is likely to be too high. Therefore, in the three-dimensional space, the unmanned aerial vehicle attributes and the task priority factors are combined to optimize the total task cost, balance the task adaptability and path efficiency, and ensure the feasibility and efficiency of the inspection.

[0064] The step S300 in the method provided by the embodiments of the present application includes:

[0065] The construction method of the three-dimensional flight task planning space includes:

[0066] An initial reference three-dimensional scene including lines, towers and terrain is constructed by integrating power grid GIS, laser radar and terrain data;

[0067] According to the safety regulations of lines of different voltage levels, exclusive flight layers with dynamic safety heights are respectively defined in the three-dimensional scene, and a plurality of exclusive flight layers are obtained.

[0068] The power facilities are abstracted as nodes and edges, and a three-dimensional flight task planning space is built based on the preset unmanned aerial vehicle flight rules, the initial reference three-dimensional scene and the plurality of exclusive flight layers.

[0069] First, an initial reference three-dimensional scene including lines, towers and terrain is constructed by integrating power grid GIS, laser radar and terrain data. The power grid GIS is a system integrating geographic information of power grid facilities; the laser radar data is high-precision three-dimensional terrain and object data obtained by laser scanning; and the initial reference three-dimensional scene is a basic three-dimensional model integrating power grid facilities and terrain. The line and tower coordinate information in the power grid GIS, the terrain elevation data and surface object data obtained by the laser radar are integrated to construct an initial reference three-dimensional scene including real terrain, line direction and tower position. For example, the target region contains three types of lines, i.e., 10kV main lines, 380V branch lines and 220V service lines, the tower coordinates of eight task blocks in the power grid GIS are integrated, such as task block 3 containing 10kV main line towers T101-T105, 380V branch line towers Z201-Z203 and 220V service line towers J301-J306, the terrain slope measured by the laser radar is 35°, and the vegetation height is 2-5 meters. The initial reference three-dimensional scene is constructed to clearly show the direction of lines of different voltages and the terrain undulation.

[0070] Secondly, based on the safety regulations for lines of different voltage levels, dedicated flight layers with dynamic safety altitudes are defined for each line in a three-dimensional scene, resulting in several dedicated flight layers. The dynamic safety altitude is a height value set according to the line voltage level to ensure a safe distance between the drone and the line. A dedicated flight layer is a three-dimensional flight area designated for a specific voltage level line, such as independent flight layers for 10kV main lines, 380V branch lines, and 220V service lines, with higher voltage levels corresponding to higher flight layers. According to safety regulations, dedicated flight layers are divided according to voltage levels; higher voltage levels have higher dynamic safety altitude requirements and therefore higher flight layer heights. For example, a first flight layer is assigned to the 10kV main line, a second to the 380V branch line, and a third to the 220V service line, with the first flight layer > the second > the third, to avoid overlapping flight paths of different voltage lines. Dynamic safety heights are set according to safety regulations: ≥3 meters for 10kV main lines, ≥2 meters for 380V branch lines, and ≥1 meter for 220V service lines. Based on the actual line height, dedicated flight levels are defined: First flight level = average line height 9 meters + safety distance 3 meters = 12 meters; Second flight level = average line height 6 meters + safety distance 2 meters = 8 meters; Third flight level = average line height 4 meters + safety distance 1 meter = 5 meters. The vertical intervals between the three levels are 4 meters and 3 meters respectively to avoid path overlap.

[0071] Finally, power facilities are abstracted into nodes and edges. Based on preset UAV flight rules, the initial baseline 3D scene, and several dedicated flight layers, a 3D flight mission planning space is constructed. Nodes and edges are network elements after abstracting power facilities; the 3D flight mission planning space is a 3D planning environment that includes flight rules, dedicated flight layers, and the facility abstract network. Power poles are abstracted as nodes, and line segments between adjacent poles are abstracted as edges. Combined with preset flight rules, such as prohibiting passage through buildings and avoiding areas with high vegetation, a 3D flight mission planning space that can be used for path calculation is constructed based on the initial baseline scene and dedicated flight layers. For example, T101-T105, Z201-Z203, and J301-J306 are all set as nodes, and the corresponding line segments are set as edges. The preset flight rule is: prohibiting entry into areas with vegetation height greater than 5 meters. The final constructed 3D space can intuitively display the dedicated flight areas and path constraints of different voltage lines.

[0072] Among them, based on the attribute information of N UAVs to be performed and the M task priority factors, the flight path of the UAVs is optimized with the goal of minimizing the total task cost, and the optimal path planning scheme is output, including:

[0073] N pieces of unmanned aerial vehicle attribute information of N unmanned aerial vehicles to be executed are acquired, wherein the unmanned aerial vehicle attribute information comprises endurance distance, sensor accuracy and unit distance energy consumption;

[0074] unmanned aerial vehicle performance indexes are output according to the N pieces of unmanned aerial vehicle attribute information respectively, wherein the unmanned aerial vehicle performance index is positively correlated with the endurance distance and the sensor accuracy, and is negatively correlated with the unit distance energy consumption;

[0075] in the three-dimensional flight task planning space, the N unmanned aerial vehicles are respectively subjected to task random allocation based on the M task blocks, and a first path planning scheme is generated;

[0076] based on the M task priority factors and the N unmanned aerial vehicle performance indexes, a first scheme quality coefficient is acquired by evaluating the first path planning scheme with the target of minimizing total task cost;

[0077] the random selection iteration of the path planning scheme in the three-dimensional flight task planning space is continuously performed, and the scheme iteration evaluation is continuously performed, until a preset convergence number is reached, and the path planning scheme corresponding to the maximum scheme quality coefficient in the optimization process is taken as an optimal path planning scheme.

[0078] Firstly, N pieces of unmanned aerial vehicle attribute information of N unmanned aerial vehicles to be executed are acquired, wherein the unmanned aerial vehicle attribute information comprises endurance distance, sensor accuracy and unit distance energy consumption. The endurance distance is the maximum mileage that can be continuously flown by the unmanned aerial vehicle under the condition of single full power state; the sensor accuracy is the measurement error range of the inspection sensor carried by the unmanned aerial vehicle, and the smaller the accuracy value is, the more accurate the detection is; and the unit distance energy consumption is the energy consumption of the unmanned aerial vehicle flying a unit mileage, and the smaller the value is, the better the endurance economy is. For example, it is determined that the number of unmanned aerial vehicles to be executed is N=2, which are numbered as unmanned aerial vehicle 1 and unmanned aerial vehicle 2 respectively, and are both multi-rotor inspection unmanned aerial vehicles, which are used for target rural power grid 8 task blocks of electricity stealing detection inspection. The attribute information is collected as follows: the factory parameters and measured data of the unmanned aerial vehicle 1 show that the endurance distance is 18 kilometers, an infrared thermal imaging sensor is carried, the sensor accuracy is 0.08 meters, and the unit distance energy consumption is 0.18 kilowatt hours per kilometer; the endurance distance of the unmanned aerial vehicle 2 is 12 kilometers, a normal high-definition sensor is carried, the sensor accuracy is 0.25 meters, and the unit distance energy consumption is 0.25 kilowatt hours per kilometer. After standardization and arrangement, the attribute information table is formed: unmanned aerial vehicle 1: endurance 18 km, sensor accuracy 0.08 m, unit energy consumption 0.18 kWh / km, unmanned aerial vehicle 2: endurance 12 km, sensor accuracy 0.25 m, unit energy consumption 0.25 kWh / km.

[0079] Secondly, the unmanned aerial vehicle performance is evaluated according to the N unmanned aerial vehicle attribute information respectively, and N unmanned aerial vehicle performance indexes are output, wherein the unmanned aerial vehicle performance index is positively correlated with the endurance distance and the sensor accuracy, and is negatively correlated with the unit distance energy consumption. The unmanned aerial vehicle performance index is a key index for comprehensively quantifying the unmanned aerial vehicle patrol task execution capability, is positively correlated with the endurance distance and the sensor accuracy, and the larger the value is, the stronger the corresponding attribute support for the patrol is; and is negatively correlated with the unit distance energy consumption, and the smaller the value is, the better the energy consumption is, and the higher the overall unmanned aerial vehicle performance index is, the better the comprehensive patrol capability of the unmanned aerial vehicle is. Reasonable weights are allocated to the endurance distance, the sensor accuracy and the unit distance energy consumption, the three types of attributes are standardized to eliminate dimensional differences, and the performance index of each unmanned aerial vehicle is calculated by weighted summation. For example, the attribute information of two unmanned aerial vehicles is used, and the weight distribution is: endurance distance 0.4, sensor accuracy 0.4, and unit distance energy consumption 0.2. After standardization, the attributes of the unmanned aerial vehicle 1 are all 1, and the performance index = 1*0.4 + 1*0.4 + 1*0.2 = 1.0; after standardization, the endurance of the unmanned aerial vehicle 2 is 0.667, the accuracy is 0.32, and the energy consumption is 0.72, and the performance index is approximately 0.539. Finally, the index output is: the performance index of the unmanned aerial vehicle 1 is 1.0, and the performance index of the unmanned aerial vehicle 2 is 0.539.

[0080] Further, in the three-dimensional flight task planning space, the N unmanned aerial vehicles are respectively assigned tasks based on the M task blocks, to generate a first path planning scheme. The task random assignment is a process of assigning M task blocks to N unmanned aerial vehicles without preset preference; and the first path planning scheme is an initial scheme formed after the first random assignment, which includes the task blocks, the sequence and the path of each unmanned aerial vehicle. In the three-dimensional flight task planning space constructed, the M task blocks are randomly assigned to the N unmanned aerial vehicles based on the basic endurance capability of the unmanned aerial vehicles; the task blocks assigned to each unmanned aerial vehicle are preliminarily planned in the sequence according to the geographical proximity principle, the specific flight path is generated in combination with the exclusive flight layer and the flight rules, and the first path planning scheme is formed after integration. For example, M=8 task blocks, N=2 unmanned aerial vehicles, and random assignment: the unmanned aerial vehicle 1 is responsible for the task blocks 3, 5 and 7, and the unmanned aerial vehicle 2 is responsible for the task blocks 1, 2, 4, 6 and 8. The preliminary planning of the inspection sequence is: the unmanned aerial vehicle 1 inspects the task blocks 3, 5 and 7 from south to north according to the geography, and the unmanned aerial vehicle 2 inspects the task blocks 1, 2, 4, 6 and 8 from east to west according to the geography, and the first path planning scheme is generated in combination with the 10kV, 380V and 220V exclusive flight layer planning path.

[0081] Then, the first path planning scheme is evaluated based on the M task priority factors and the N unmanned aerial vehicle performance indexes, to obtain a first scheme quality coefficient.

[0082] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0083] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0084] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0085] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0086] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0087] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes to minimize the total task cost, and a first scheme quality coefficient is obtained, including:

[0088] For example, for the task blocks assigned to each UAV in the first path planning scheme, the task priority factor of the task block and the performance index of the corresponding UAV are extracted one by one, and the cosine similarity of the two is calculated as the inspection adaptation degree of the task block; the inspection adaptation degrees of all M task blocks are added to obtain the first total inspection adaptation degree. Follow the first path scheme: UAV 1 is responsible for task blocks 3, 5, and 7, and UAV 2 is responsible for task blocks 1, 2, 4, 6, and 8. The priority factor of task block 3 is 0.783, which is assigned to UAV 1 with a performance index of 1.0, and the cosine similarity is approximately 0.99; the priority factor of task block 5 is 0.563, and the adaptation degree with UAV 1 is approximately 0.98; the priority factor of task block 7 is 0.421, and the adaptation degree with UAV 1 is approximately 0.97, and the sum of the three is approximately 2.94. Similarly, the sum of the adaptation degrees of the five task blocks responsible for UAV 2 is approximately 4.2, so the first total inspection adaptation degree = 2.94 + 4.2 = 7.14.

[0089] Secondly, the N idle path distances of the N UAVs in the first path planning scheme are counted and added to obtain the first total idle distance. The idle path distance refers to the flight distance of the UAV during the execution of the inspection task, which includes the distance from the base to the first task block, the flight distance of the non-inspection section between task blocks, and the distance from the last task block back to the base. The first total idle distance is the sum of the idle path distances of all UAVs in the first path planning scheme, and the smaller the value represents the less invalid flight. For each UAV in the first path planning scheme, the distances of each segment of the idle path are counted: the flight distance from the departure base to the first task block responsible for, the transition flight distance between task blocks without involving the inspection line, and the flight distance from the last task block back to the base; the above distances of a single UAV are added to obtain its idle path distance, and the idle path distances of all N UAVs are collected to obtain the first total idle distance. For example, follow the first path scheme: UAV 1 is responsible for task blocks 3, 5, and 7, and UAV 2 is responsible for task blocks 1, 2, 4, 6, and 8. The idle path of UAV 1: the distance from the base to task block 3 is 2 km, the non-inspection section distance from task block 3 to 5 is 3 km, the non-inspection section distance from task block 5 to 7 is 2 km, and the distance from task block 7 back to the base is 4 km, the idle path distance = 2+3+2+4=11km. The idle path of UAV 2: the distance from the base to task block 1 is 1 km, the non-inspection section distance from task block 1 to 2 is 1 km, the non-inspection section distance from task block 2 to 4 is 2 km, the non-inspection section distance from task block 4 to 6 is 3 km, the non-inspection section distance from task block 6 to 8 is 2 km, and the distance from task block 8 back to the base is 3 km, the idle path distance = 1+1+2+3+2+3=12km. The first total idle distance = 11+12=23km.

[0090] Further, a first scheme quality coefficient is calculated according to the first total inspection adaptation degree and the first total empty running distance with a view to minimizing the total task cost. The first scheme quality coefficient is a comprehensive index for measuring the overall pros and cons of the first path planning scheme, and is targeted at minimizing the total task cost. The first scheme quality coefficient is positively correlated with the first total inspection adaptation degree: the higher the adaptation degree, the greater the coefficient; and is negatively correlated with the first total empty running distance: the shorter the empty running distance, the greater the coefficient, and the higher the numerical value represents the better the scheme. The weights of the total inspection adaptation degree and the total empty running distance are set, and the adaptation degree is directly related to the task quality, so the weight of the adaptation degree is set to be higher than that of the empty running distance; the total empty running distance is standardized, for example, divided by the benchmark empty running distance, to eliminate the dimension; the quality coefficient is calculated through a weighting formula, the calculation formula is: the first scheme quality coefficient = (the first total inspection adaptation degree x the adaptation degree weight) - (the standardized first total empty running distance x the empty running weight), to realize the quantitative evaluation of the total task cost. For example, the first path scheme is followed, the first total inspection adaptation degree is 7.14, and the first total empty running distance is 23 km. The adaptation degree weight is set to 0.6, the empty running weight is set to 0.4, and the benchmark empty running distance is set to 100 km. The standardized total empty running distance = 23 / 100 = 0.23, and the first scheme quality coefficient = 7.14x0.6-0.23x0.4≈4.284-0.092=4.192.

[0091] Finally, the random selection iteration of the path planning scheme in the three-dimensional flight task planning space is continued, and the scheme iteration evaluation is continued, until the preset convergence number is reached, and the path planning scheme corresponding to the maximum scheme quality coefficient in the optimization process is taken as the optimal path planning scheme. Iteration refers to the process of repeatedly generating and evaluating the path planning scheme. The preset convergence number is a set number threshold for terminating iteration. The optimal path planning scheme is the path scheme with the maximum scheme quality coefficient and the lowest total task cost in the optimization process. In the three-dimensional flight task planning space, the processes of task random allocation, total inspection adaptation degree calculation, total empty running distance statistics, and scheme quality coefficient evaluation are repeatedly executed, new path planning schemes are continuously generated and evaluated; when the iteration number reaches the preset convergence number, the path planning scheme with the maximum scheme quality coefficient is selected from all the schemes generated by iteration, as the optimal path planning scheme.

[0092] For example, the preset convergence number is set to 100 times, and the path scheme is iteratively generated and evaluated based on 8 task blocks and 2 unmanned aerial vehicles. When the iteration reaches the 76th time, the generated scheme is that unmanned aerial vehicle 1 is responsible for task blocks 3 and 6, unmanned aerial vehicle 2 is responsible for the remaining 6 task blocks, the total inspection adaptation degree is 8.5, the total empty running distance is 16 km, and the scheme quality coefficient ≈ 8.5x0.6-(16÷100)x0.4=5.1-0.064=5.036, which is the maximum value among all the iteration schemes. After the iteration reaches 100 times, the scheme is stopped, and the scheme is the optimal path planning scheme.

[0093] In the embodiments of the present application, a three-dimensional flight task planning space is constructed, real terrain and power grid facility information are fused, the spatial safety of the unmanned aerial vehicle flight is ensured through the exclusive flight layer and flight rules, and the path conflict problem under complex terrain is solved. The path optimization process balances the inspection adaptability and the air running distance by combining the unmanned aerial vehicle performance and the task priority through iterative optimization, and the total task cost minimization is realized. The optimal path planning scheme finally output ensures that the high-priority task is executed by the high-performance unmanned aerial vehicle, and the invalid flight distance is shortened, thereby improving the inspection efficiency and quality.

[0094] S400: controlling the N unmanned aerial vehicles to perform electricity stealing detection inspection on the target power grid area according to the optimal path planning scheme.

[0095] In the embodiments of the present application, the optimal path planning scheme is the unmanned aerial vehicle inspection task allocation and three-dimensional flight path scheme with the minimum total task cost after iterative optimization. The electricity stealing detection inspection refers to the inspection operation of the unmanned aerial vehicle carrying sensors such as infrared thermal imaging and high-definition cameras to scan the power grid lines and equipment, and identify abnormal electricity stealing, line defects and other electricity stealing signs. Based on the output optimal path planning scheme, the inspection task block, sequence and three-dimensional flight path of each unmanned aerial vehicle are determined; the path information is loaded through the ground control station or the unmanned aerial vehicle autonomous navigation system to control the unmanned aerial vehicle to execute the inspection in sequence according to the planning; and the unmanned aerial vehicle uses the sensors to detect the lines, towers and other facilities of the target power grid area in all directions, and transmits the inspection data in real time.

[0096] For example, in the optimal path planning scheme, the unmanned aerial vehicle 1 inspects the task blocks 3 and 6 in the 10kV exclusive flight layer, uses infrared thermal imaging to scan the line joints and transformers, and uses a high-definition camera to take pictures of the appearance; the unmanned aerial vehicle 2 inspects the remaining six task blocks in the 380V and 220V flight layers, and synchronously collects data. After the inspection is completed, the ground station collects the data and analyzes the electricity stealing signs.

[0097] In the embodiments of the present application, the optimal path planning scheme is implemented to ensure that the high-risk area is preferentially inspected by the adaptive unmanned aerial vehicle, thereby improving the efficiency; the high-altitude perspective of the unmanned aerial vehicle and the professional sensors realize the power grid electricity stealing detection without dead angle under the complex rural terrain, and compared with the manual inspection, the electricity stealing identification accuracy and coverage are greatly improved, thereby providing technical support for electricity stealing disposal.

[0098] The embodiments of the present application achieve the following technical effects through the specific implementation manner described above.

[0099] The application provides a UAV flight path intelligent planning method and system for rural power grid inspection, which is suitable for the dynamic influence of complex terrain and seasonal obstacles of rural power grids and real-time weather, optimizes the task weight ratio, and makes the inspection priority evaluation more in line with the actual demand. Relying on a three-dimensional flight task planning space and a dedicated flight layer, the UAV performance is combined to realize path iterative optimization, balance the inspection adaptation degree and air running distance, and reduce the total task cost. Finally, the UAV is controlled to carry out electricity stealing detection inspection through the optimal scheme, improve the inspection resource allocation efficiency and flight safety, improve the electricity stealing identification accuracy and regional coverage, effectively make up for the limitations of manual inspection in complex scenes, and guarantee the quality and effectiveness of rural power grid inspection.

[0100] In an embodiment, as shown in Figure 2 The application provides a UAV flight path intelligent planning system for rural power grid inspection, which comprises:

[0101] A risk distribution map drawing module 11 is configured to draw electricity stealing risk distribution maps, terrain risk distribution maps, and line risk distribution maps based on M task association information of M task blocks of a target power grid area;

[0102] A task priority evaluation module 12 is configured to evaluate the M task blocks based on the electricity stealing risk distribution maps, the terrain risk distribution maps, and the line risk distribution maps, and obtain M task priority factors in combination with a preset weight ratio;

[0103] An optimal path planning module 13 is configured to perform UAV flight path optimization in a pre-constructed three-dimensional flight task planning space based on attribute information of N UAVs to be executed and the M task priority factors, with the objective of minimizing the total task cost, and output an optimal path planning scheme;

[0104] A UAV inspection execution module 14 is configured to control the N UAVs to perform electricity stealing detection inspection on the target power grid area according to the optimal path planning scheme.

[0105] In one embodiment, the risk distribution map drawing module 11 is further configured to:

[0106] The target rural power grid inspection area is divided into M task blocks according to a preset line distance, wherein the preset line distance is 5 kilometers, and M is an integer greater than 1;

[0107] According to electricity stealing association information, terrain association information, and line association information of the M task blocks, electricity stealing risk evaluation, terrain risk evaluation, and line risk evaluation are respectively performed, and electricity stealing risk coefficient distribution, terrain risk coefficient distribution, and line risk coefficient distribution are output;

[0108] Draw a power stealing risk distribution map, a terrain risk distribution map and a line risk distribution map based on the power stealing risk coefficient distribution, the terrain risk coefficient distribution and the line risk coefficient distribution.

[0109] The power stealing associated information includes historical power stealing records, user power consumption abnormal data and seasonal power stealing rules, the terrain associated information includes terrain complexity parameters extracted from satellite remote sensing data, and the line associated information includes line voltage grades and types, line aging degrees and line importance degrees, wherein the terrain complexity parameters at least include slope, slope direction and vegetation coverage.

[0110] In one embodiment, the task priority evaluation module 12 is also used for:

[0111] Randomly selecting a first task block from the M task blocks, collecting a historical seasonal obstacle feature set of a region where the first task block is located in a current time zone based on historical unmanned aerial vehicle inspection records, performing seasonal obstacle influence evaluation, and obtaining a first seasonal obstacle influence coefficient;

[0112] Collecting real-time environmental meteorological data of the region where the first task block is located in the current time zone to perform line state influence evaluation, and obtaining a first line state influence coefficient;

[0113] Optimizing an initial weight ratio based on the first seasonal obstacle influence coefficient and the first line state influence coefficient to obtain a preset weight ratio;

[0114] After non-dimensional processing of the first power stealing risk coefficient, the first terrain risk coefficient and the first line risk coefficient of the first task block, performing weighted calculation according to the preset weight ratio, outputting a first task priority factor, and sequentially calculating M task priority factors of the M task blocks.

[0115] The optimization of the initial weight ratio based on the first seasonal obstacle influence coefficient and the first line state influence coefficient to obtain the preset weight ratio includes:

[0116] Obtaining an initial weight ratio, wherein the initial weight ratio includes an initial power stealing risk weight, an initial terrain risk weight and an initial line risk weight, and the sum of the three is 1;

[0117] Setting a ratio of the first seasonal obstacle influence coefficient to a preset standard seasonal obstacle influence coefficient as a terrain risk weight adjustment coefficient, compensating the initial terrain risk weight to obtain an adaptive terrain risk weight;

[0118] The ratio of the first line state influence coefficient to a preset standard line state influence coefficient is set as a line risk weight adjustment coefficient, and the initial line risk weight is compensated to obtain an adapted line risk weight;

[0119] The adapted terrain risk weight, the adapted line risk weight, and the adapted electricity stealing risk weight are used to generate a preset weight ratio.

[0120] In one embodiment, the optimal path planning module 13 is further configured to:

[0121] The method for constructing the three-dimensional flight task planning space comprises:

[0122] An initial reference three-dimensional scene including lines, towers, and terrain is constructed by integrating power grid GIS, laser radar, and terrain data.

[0123] According to the safety regulations of lines of different voltage levels, exclusive flight layers with dynamic safety heights are respectively defined in the three-dimensional scene, and a plurality of exclusive flight layers are obtained.

[0124] The power facilities are abstracted as nodes and edges, and a three-dimensional flight task planning space is built based on preset unmanned aerial vehicle flight rules, the initial reference three-dimensional scene, and the plurality of exclusive flight layers.

[0125] The attributes information of the N unmanned aerial vehicles to be executed and the M task priority factors are used to optimize the flight path of the unmanned aerial vehicles to minimize the total task cost, and an optimal path planning scheme is output, comprising:

[0126] N unmanned aerial vehicle attribute information of N unmanned aerial vehicles to be executed is obtained, wherein the unmanned aerial vehicle attribute information comprises endurance distance, sensor accuracy, and unit distance energy consumption.

[0127] The N unmanned aerial vehicle attribute information is used to respectively evaluate the performance of the unmanned aerial vehicles, and N unmanned aerial vehicle performance indexes are output, wherein the unmanned aerial vehicle performance index is positively correlated with the endurance distance and the sensor accuracy, and is negatively correlated with the unit distance energy consumption.

[0128] In the three-dimensional flight task planning space, the N unmanned aerial vehicles are respectively assigned tasks based on the M task blocks, and a first path planning scheme is generated.

[0129] The M task priority factors and the N unmanned aerial vehicle performance indexes are used to evaluate the first path planning scheme to minimize the total task cost, and a first scheme quality coefficient is obtained.

[0130] Continue to carry out the random selection iteration of the path planning scheme in the three-dimensional flight task planning space, and carry out scheme iteration evaluation until a preset convergence number is reached, and the path planning scheme corresponding to the maximum scheme quality coefficient in the optimization process is taken as the optimal path planning scheme.

[0131] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes, and a first scheme quality coefficient is obtained, including:

[0132] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes, and a first scheme quality coefficient is obtained, including:

[0133] The N empty running path distances of the N UAVs in the first path planning scheme are counted and summed to obtain a first total empty running distance.

[0134] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes, and a first scheme quality coefficient is obtained, including:

[0135] The first path planning scheme is evaluated based on the M task priority factors and the N UAV performance indexes, and a first scheme quality coefficient is obtained, including:

[0136] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0137] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0138] The present application is only an exemplary description of the present application, and is considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalents, the present application intends to include these modifications and changes.

Claims

1. A method for intelligent flight path planning of unmanned aerial vehicles (UAVs) for rural power grid inspection, characterized in that, The methods include: Based on the M task association information of M task blocks in the target power grid area, draw the electricity theft risk distribution map, terrain risk distribution map and line risk distribution map; Based on the electricity theft risk distribution map, terrain risk distribution map, and line risk distribution map, and combined with a preset weighting ratio, the M task blocks are evaluated to obtain M task priority factors; Within a pre-constructed three-dimensional flight mission planning space, based on the attribute information of N UAVs to be executed and the M mission priority factors, the flight path of the UAVs is optimized with the goal of minimizing the total mission cost, and the optimal path planning scheme is output. The N drones are controlled to conduct electricity theft detection and inspection of the target power grid area according to the optimal path planning scheme. Based on the task association information of M task blocks in the target power grid area, draw electricity theft risk distribution maps, terrain risk distribution maps, and line risk distribution maps, including: According to the preset line distance, the target rural power grid inspection area is divided into M task blocks, where the preset line distance is 5 kilometers and M is an integer greater than 1; Based on the electricity theft association information, terrain association information and line association information of the M task blocks, electricity theft risk assessment, terrain risk assessment and line risk assessment are performed respectively, and the distribution of electricity theft risk coefficient, terrain risk coefficient and line risk coefficient are output. Based on the distribution of electricity theft risk coefficients, terrain risk coefficients, and line risk coefficients, electricity theft risk distribution maps, terrain risk maps, and line risk maps are drawn respectively. Based on the aforementioned electricity theft risk distribution map, terrain risk distribution map, and line risk distribution map, and combined with a preset weighting ratio, the M task blocks are evaluated to obtain M task priority factors, including: Randomly select the first task block from the M task blocks, collect the historical seasonal obstacle feature set of the area where the first task block is located in the current time zone based on historical UAV inspection records, conduct seasonal obstacle impact assessment, and obtain the first seasonal obstacle impact coefficient. Real-time environmental meteorological data of the area where the first task block is located in the current time zone are collected to assess the impact of the line status and obtain the first line status impact coefficient. The initial weight allocation is optimized based on the first seasonal obstacle influence coefficient and the first line status influence coefficient to obtain the preset weight allocation; After dimensionless processing of the first electricity theft risk coefficient, the first terrain risk coefficient, and the first line risk coefficient of the first task block, a weighted calculation is performed according to the preset weight ratio to output the first task priority factor, and the M task priority factors of the M task blocks are calculated in sequence.

2. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 1, characterized in that, The electricity theft association information includes historical electricity theft records, abnormal user electricity consumption data, and seasonal electricity theft patterns. The terrain association information includes terrain complexity parameters extracted from satellite remote sensing data. The line association information includes line voltage level and type, line aging degree, and line importance. The terrain complexity parameters include at least slope, aspect, and vegetation coverage.

3. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 1, characterized in that, The initial weight allocation is optimized based on the first seasonal obstacle influence coefficient and the first line condition influence coefficient to obtain a preset weight allocation, including: Obtain the initial weight ratio, wherein the initial weight ratio includes the initial electricity theft risk weight, the initial terrain risk weight, and the initial line risk weight, and the sum of the three is 1; The ratio of the first seasonal obstacle influence coefficient to the preset standard seasonal obstacle influence coefficient is set as the terrain risk weight adjustment coefficient, and the initial terrain risk weight is compensated to obtain the adapted terrain risk weight. The ratio of the first line status influence coefficient to the preset standard line status influence coefficient is set as the line risk weight adjustment coefficient, and the initial line risk weight is compensated to obtain the adapted line risk weight. The adaptive terrain risk weight and the adaptive line risk weight are subtracted from 1 to obtain the adaptive electricity theft risk weight, and a preset weight ratio is generated based on the adaptive terrain risk weight, the adaptive line risk weight and the adaptive electricity theft risk weight.

4. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 1, characterized in that, The method for constructing the three-dimensional flight mission planning space includes: By integrating power grid GIS, lidar, and terrain data, an initial benchmark 3D scene containing power lines, towers, and topography is constructed. Based on the safety regulations for lines of different voltage levels, a dedicated flight layer with dynamic safety height is defined for each of them in the three-dimensional scene, resulting in several dedicated flight layers. Power facilities are abstracted into nodes and edges, and a three-dimensional flight mission planning space is built based on preset UAV flight rules, the initial baseline three-dimensional scene, and several dedicated flight layers.

5. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 1, characterized in that, Based on the attribute information of N UAVs to be performed and the M task priority factors, the flight path of the UAVs is optimized with the goal of minimizing the total task cost, and the optimal path planning scheme is output, including: Obtain N drone attribute information of N drones to be executed, wherein the drone attribute information includes flight range, sensor accuracy and energy consumption per unit distance; Based on the N drone attribute information, the drone performance is evaluated and N drone performance indices are output. The drone performance index is positively correlated with the flight range and sensor accuracy, and negatively correlated with the energy consumption per unit distance. Within the three-dimensional flight mission planning space, based on the M mission blocks, N UAVs are randomly assigned missions to generate a first path planning scheme. Based on the M task priority factors and N UAV performance indices, with the goal of minimizing the total task cost, the first path planning scheme is evaluated to obtain the quality coefficient of the first scheme. Continue to randomly select and iterate path planning schemes within the three-dimensional flight mission planning space, and evaluate the schemes iteratively until the preset number of convergences is reached. The path planning scheme corresponding to the scheme with the highest quality coefficient during the optimization process is taken as the optimal path planning scheme.

6. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 5, characterized in that, Based on the M task priority factors and N UAV performance indices, with the objective of minimizing the total task cost, the first path planning scheme is evaluated to obtain the quality coefficient of the first scheme, including: Based on the M task priority factors and N UAV performance indices, the inspection fit degree of the M task blocks in the first path planning scheme is calculated, and the M inspection fit degrees are added together to obtain the first total inspection fit degree, wherein the inspection fit degree is the cosine similarity between the task priority factor of the task block and the UAV performance index of the assigned UAV. Calculate the distances of the N empty run paths of the N drones in the first path planning scheme, and sum them to obtain the first total empty run distance; With the goal of minimizing the total task cost, the first scheme quality coefficient is calculated by weighting the first total inspection adaptability and the first total empty running distance.

7. The intelligent flight path planning method for UAVs used in rural power grid inspection according to claim 6, characterized in that, The quality coefficient of the first scheme is positively correlated with the adaptability of the first total inspection, and negatively correlated with the first total empty running distance.

8. An intelligent flight path planning system for unmanned aerial vehicles (UAVs) used for rural power grid inspection, characterized in that: The system is used to implement the intelligent flight path planning method for unmanned aerial vehicles (UAVs) for rural power grid inspection as described in any one of claims 1-7, the system comprising: The risk distribution map drawing module is used to draw electricity theft risk distribution maps, terrain risk distribution maps, and line risk distribution maps based on the M task association information of M task blocks in the target power grid area. The task priority evaluation module is used to evaluate the M task blocks based on the electricity theft risk distribution map, terrain risk distribution map and line risk distribution map, and to obtain M task priority factors by combining the preset weight ratio. The optimal path planning module is used to optimize the flight path of the UAVs within a pre-constructed three-dimensional flight mission planning space, based on the attribute information of the N UAVs to be executed and the M mission priority factors, with the goal of minimizing the total mission cost, and output the optimal path planning scheme. The UAV inspection execution module is used to control the N UAVs to conduct electricity theft detection and inspection of the target power grid area according to the optimal path planning scheme.

Images