Line patrol method of charging type power line monitoring unmanned aerial vehicle

Abstract

The invention relates to a line patrol method of a rechargeable power line monitoring unmanned aerial vehicle, which solves a power line patrol problem by using a rechargeable power monitoring unmanned aerial vehicle and a plant community algorithm, and comprises the following steps: step 1, global line patrol optimization of the rechargeable power line monitoring unmanned aerial vehicle; 2, optimizing a local detailed survey route of the charging type power line monitoring unmanned aerial vehicle; 3, optimizing a local obstacle avoidance route of the charging type power line monitoring unmanned aerial vehicle; and step 4, optimizing a local charging route of the charging type power line monitoring unmanned aerial vehicle. According to the method, line patrol optimization of the charging type power line monitoring unmanned aerial vehicle can be efficiently completed.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and in particular relates to power line inspection and artificial intelligence technology, and specifically relates to a power line inspection method of a rechargeable power line monitoring drone. Background Art

[0002] The electricity generated by the power plant needs to be sent to all parts of the country through power lines. Therefore, the terrain where the power lines are located is all-encompassing, running through the north and south, with hilly and mountainous areas, as well as river and lake areas. The climatic conditions faced by the power lines are also very complex, including scorching sun, strong winds, rain, snow, hail, etc., which will have an impact on the power lines. In many cases, birds and animals will climb onto or hit the power lines, causing damage to the power lines. Even if there are no external factors, the strong gravity of the power lines themselves will stretch the power lines or high-voltage towers, and the high-voltage and high-current flowing through them will also cause the power lines to heat up or discharge. The combined effect of one or more of the above factors can easily cause abnormalities in the power lines during long-term work, including deformation of the power lines, deformation of the high-voltage towers, damage to the insulation of the wires, damage to the insulators, damage or loss of the shock-absorbing hammers, wire clamps, tower bodies, and hardware. In order to ensure the stability of the power supply of the power lines, early detection and elimination of power line faults, monitoring and inspection of the power lines has become an important task.

[0003] The power line monitoring and inspection methods have mainly gone through three stages: manual inspection stage, helicopter inspection stage, and drone inspection stage.

[0004] Manual inspection is the earliest form of power inspection, that is, inspectors personally inspect around power lines. This traditional inspection method is slow and easily interfered by human and weather factors. In addition, the distance between power lines and the ground is very high, making it difficult to ensure the quality of inspection. The terrain through which power lines pass is often very complex, including forests, reservoirs, lakes and mountains. In many places, there is a lack of inspection roads, which can easily affect the smooth progress of inspection tasks, increase inspection costs, and fail to ensure the safety of inspectors.

[0005] With the invention and promotion of helicopters, manned helicopter power line inspection methods began to appear. This new inspection method greatly improves the inspection speed and efficiency, and can inspect power lines at high altitudes, avoiding the interference of ground terrain and objects, and greatly improving the safety of inspections. However, the purchase, operation, and maintenance of manned helicopters require a lot of costs, plus the wages of pilots, and the need to apply for airspace use rights for helicopter flights, resulting in helicopter inspections lacking economy and flexibility, making them difficult to use on a large scale. During helicopter inspections, the distance between the helicopter and power lines or high-voltage towers needs to be maintained, such as between 15 meters and 25 meters. However, if the distance is too close due to terrain or climate, it will also endanger the safety of the pilot and the helicopter.

[0006] In recent years, new drone inspections have emerged with the emergence of drones, and have been widely popular since their inception. Drones are very low-cost, do not require pilots, and do not have low-altitude airspace use rights issues. They are almost unaffected by topography and landforms, bringing unprecedented convenience and safety to power line inspections. Small drones used for power line inspections are generally considered to be aviation models. When drones are inspecting, they are often tens of meters away from power lines, and there are almost no issues with low-altitude airspace use rights. However, there are almost no restrictions on drones for power line corridors that aircraft need to avoid. With the development of the economy and the growth of industrial electricity consumption, the scale of power lines is expanding, and drone power line inspections will become the main method of power line inspections in the future.

[0007] UAV power inspection has now been expanded in many ways, such as being equipped with GPS positioning equipment, optical cameras, laser scanners, infrared thermal imagers, remote communication modules and other advanced equipment and sensors. During the UAV power inspection, with the help of the control of the UAV operator, multiple images can be taken from different angles at close range and transmitted to the ground inspection workstation in real time. The inspection images are mined and processed by the image processing algorithm to find the suspected fault points of the power lines and power equipment, ensuring the early discovery and timely handling of power line faults and hidden dangers. Especially in complex terrain and areas that are difficult for personnel to reach, the advantages of UAV power inspection are very large, greatly improving the inspection speed and efficiency, and the inspection cost is very low, and the information is transmitted in a timely manner, which helps to analyze and diagnose suspected power line faults and ensure the normal operation of the power lines.

[0008] However, the current drone power line inspection method has the following disadvantages:

[0009] 1: The endurance of drones is insufficient. The existing drones are small in size, which also limits the maximum capacity of fuel or batteries, and thus limits the endurance of drones. The endurance of traditional drones is estimated by the operator's experience. The drone is operated to return home as soon as possible before the drone's battery is about to run out. This makes it difficult to take into account the travel methods and energy replenishment methods of different types of drones. Traditional drones have two ways of using fuel and electricity. The locations of fuel replenishment points and charging stations or battery replacement points are fixed or determined by drone operators. Their layout models and route algorithms rarely analyze the energy consumption characteristics and flight behaviors of different drones and detailed data on energy consumption, and cannot provide reference or help for the calculation of drone endurance and charging load. If the endurance of drones needs to be improved, it is necessary to obtain and analyze the flight behaviors and endurance laws of different types of drones in real time and calculate the energy consumption load curve of drones. Scientific methods are used to model and statistically analyze the endurance of drones, and to construct energy consumption load curves. This is very difficult in engineering practice. When the drone's flight speed continues to increase, resulting in a large increase in the drone's load, the drone's energy consumption will be too large and the endurance will be insufficient. In addition, the temperature environment in winter and summer causes changes in the operation of drone batteries, which has seasonal load characteristics.

[0010] 2: The drone’s detailed inspection capability is insufficient. During the line inspection process, the drone may find some suspicious fault points or targets. At this time, it is necessary to observe closely and shoot at close range. On the one hand, it will consume more power, and on the other hand, getting too close will also bring the risk of collision to the drone. The current drone inspection methods are relatively simple. The drone operator can only use the naked eye to identify, and it is impossible to observe the fault points of power lines or high-voltage towers in detail. If a power line fault is found during the drone inspection, it is completely based on the operator’s experience to make a judgment, and then control the drone to climb to the fault site for a detailed inspection. Sometimes the operator even accidentally fails to notice the suspicious fault point, which leads to the omission of the suspicious obstacle. When the power line or power equipment fails, if it cannot be detected and repaired in time, it may cause the fault to cascade and cause a larger range of power line faults, thus failing to ensure the stability of power supply.

[0011] 3: Drones have insufficient obstacle avoidance capabilities. Although drones are not easily affected by terrain during inspections, inspections still require drones to face various obstacles, including power lines, high-voltage towers, tall buildings, hillsides, large trees, signal towers and other tall and hard objects. In addition, flying birds may also cause damage to drones. Drones are often made of lightweight materials, and they are prone to crashing and being damaged, or even beyond repair, when they encounter obstacles. Many power lines are deployed in mountainous and hilly areas, and the inspection terrain is complex, which greatly increases the probability of drones encountering obstacles. Due to the influence of vegetation and climate, the obstacles faced by drones will also change with the seasons.

[0012] 4: Drones are not intelligent enough. During inspections, drones often face various complex situations, including detailed shots of suspected fault points, avoidance of obstacles, insufficient endurance, and interference from crosswinds or birds. These require drones to have a certain degree of intelligence to cope with them well. Fourth, there are great safety hazards. Since power lines are often located in mountains and hills, when personnel go to the fault area or operators go to the site for work, the safety of personnel cannot be effectively guaranteed. Traditional drone line inspection models and algorithms are usually static analysis, which makes it difficult to consider the impact of terrain and the random behavior of drone flight. In fact, drone flight behavior has a great impact on the drone energy consumption load curve, especially the sharp increase in energy consumption after the rapid increase in flight speed. In addition, the rate of increase in the power load of drones working during the day or at night is also different, and the power consumption in winter also increases faster than in summer, which puts higher requirements on the intelligence of drones. Summary of the invention

[0013] In view of the above-mentioned problems existing in the prior art, the present invention proposes a power line inspection method of a rechargeable power line monitoring drone. By using a rechargeable power monitoring drone and a plant community algorithm to solve the power line inspection problem, the design is reasonable, which solves the shortcomings of the prior art and greatly improves the power line monitoring and inspection efficiency while improving the algorithm efficiency.

[0014] A method for patrolling a power line by a rechargeable power line monitoring drone comprises the following steps:

[0015] Step 1: Use a rechargeable power line monitoring drone to perform global line inspection optimization;

[0016] Step 2: Use a rechargeable power line monitoring drone to optimize the local detailed inspection route;

[0017] Step 3: Use a rechargeable power line monitoring drone to optimize the local obstacle avoidance route;

[0018] Step 4: Use a rechargeable power line monitoring drone to optimize local charging routes.

[0019] In step 1, the following sub-steps are included:

[0020] Sub-step 1-1, obtain the global patrol task through the rechargeable power line monitoring drone, input the power line coordinates, obstacle coordinates, and charging station coordinates; the patrol optimization problem of the rechargeable power line monitoring drone includes multiple power lines, multiple obstacles, and multiple charging stations. A charging station may contain multiple charging piles, multiple mileage-restricted, pure electric drones or hybrid drones, and drone patrol charging scheduling optimization problems with time window constraints. In the patrol task, the rechargeable power line monitoring drone needs to access all power line coordinates, avoid obstacle coordinates, and access the charging station coordinates before the battery is exhausted to ensure that the patrol task is completed and returned;

[0021] Sub-step 1-2, the rechargeable power line monitoring drone runs the plant community algorithm to perform global patrol optimization on the power line coordinates, obstacle coordinates, and charging station coordinates;

[0022] Sub-step 1-2a, the plant community algorithm initializes the parameters of the patrol optimization problem of the rechargeable power line monitoring drone; the plant community algorithm defines the parameters of the patrol optimization problem of the rechargeable power line monitoring drone, including solving the target set, power line parameters, obstacle parameters, charging station parameters, and drone parameters;

[0023] In sub-step 1-2b, the plant community algorithm iteratively optimizes the patrol optimization problem of the rechargeable power line monitoring drone; the plant individuals are encoded according to the patrol optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of the corresponding power line, obstacle, and charging station; each bit of the plant individual encoding uses 0 to indicate that the corresponding power line, obstacle, and charging station are not selected;

[0024] In sub-steps 1-3, the plant community algorithm outputs the optimal global patrol route, and the rechargeable power line monitoring drone patrols the line according to the route; the plant community algorithm continuously repeats sub-steps 1-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal global patrol optimization route; the rechargeable power line monitoring drone patrols the line globally according to the route. If a suspicious power line target or suspicious obstacle is encountered during the flight, or the remaining power is insufficient, the plant community algorithm needs to be used to locally adjust the patrol route, and then go to step 2, step 3, or step 4.

[0025] In sub-step 1-1, it is assumed that the location and number of power lines are known, the location and number of obstacles are known, the location of charging stations and charging piles are known, the number of charging piles at each charging station is known, and the maximum charging capacity of each type of drone and the total weight of the drone are known; it is assumed that a charging pile can only charge one drone, the charging pile has a queuing time limit, the drone cannot be overcharged, cannot be charged beyond capacity, the drone flies at a constant speed between nodes, all drones are fully charged each time, the drones are in good condition, the air flight channel is unobstructed, and the sudden appearance of birds, drone failures, and interference or intervention by other drones not for line patrol tasks are not considered.

[0026] In sub-step 1-2a, the target set is solved, including the power line set, the obstacle set, the charging station set, the charging pile set of the charging station, the drone set, and the drone type set; the power line parameters, including the position and number of power lines; the obstacle parameters, including the position and number of obstacles; the charging station parameters, including the position and number of the charging station and each charging pile, the maximum charging load of the charging station, the impact coefficient of drone charging on the grid load, the scalable charging load peak of the charging station, the flight distance between charging stations, the charging service time of each charging station, the charging time window of the charging station, the unit fixed cost, and the unit time charging cost; the drone parameters, including the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance that the drone can fly when fully charged, the current remaining power of the drone, the charging capacity of the drone at the charging station, the time when the drone arrives at the charging station, the flight time of the drone between charging stations, the early arrival penalty cost coefficient of the drone, the late arrival penalty cost coefficient of the drone, the unit flight cost of the drone, and the total flight cost of the drone.

[0027] In sub-step 1-2b, the plant community algorithm uses the shortest total flight time as the objective function of the patrol optimization problem of the rechargeable power line monitoring drone to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 1-2a to set constraints, and the patrol optimization problem of the rechargeable power line monitoring drone should not exceed the total number of power lines, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the patrol optimization problem of the rechargeable power line monitoring drone.

[0028] In step 2, the following sub-steps are included:

[0029] Sub-step 2-1, the rechargeable power line monitoring drone continuously extracts power line features in the monitoring image during the global line patrol process, and mines suspicious power line targets; after the rechargeable power line monitoring drone starts the global line patrol, it continuously shoots monitoring images of the power lines and stores them in the drone memory; further, the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm, including filtering, noise reduction, image segmentation, edge detection, target classification, image enhancement, and extraction of power line features; further, the drone stores a suspicious power line target feature library, including power line deformation, power line insulation damage, high-voltage insulator damage, high-voltage tower deformation, transformer damage, etc., suspicious power line target features that may endanger the normal operation of the power line; further, the power line features in the monitoring image are compared with the suspicious power line target feature library through the image processing algorithm to mine whether there are power line features that meet the suspicious power line target feature library. If not, the drone continues to patrol according to the optimal global line patrol route in sub-steps 1-3; otherwise, the drone conducts a local detailed inspection of the suspicious power line target according to sub-step 2-2;

[0030] Sub-step 2-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local detailed inspection route of the suspicious power line target;

[0031] In sub-step 2-2a, the plant community algorithm initializes the parameters of the local detailed inspection route optimization problem of suspicious power line targets; the plant community algorithm defines the parameters of the local detailed inspection route optimization problem of suspicious power line targets of the rechargeable power line monitoring drone, including solving the target set, suspicious power line target parameters, obstacle parameters, charging station parameters, and drone parameters; further, solving the target set includes a suspicious power line target set, an obstacle set, a charging station set, a charging pile set of charging stations, a drone set, and a drone type set; suspicious power line target parameters include the location and number of suspicious power line targets; obstacle parameters include the location and number of obstacles; charging station parameters include the location and number of charging stations and each charging pile; drone parameters include the maximum flight speed of the drone, the maximum charging capacity of the drone, the longest distance the drone can fly on a full charge, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone;

[0032] In sub-step 2-2b, the plant community algorithm iteratively optimizes the local detailed inspection route optimization problem of the suspicious power line target of the rechargeable power line monitoring drone; the plant individual is encoded according to the local detailed inspection route optimization problem of the suspicious power line target of the rechargeable power line monitoring drone, and one plant individual corresponds to a feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of several shooting positions around the corresponding suspicious power line target; each bit of the plant individual encoding uses 0 to indicate that the corresponding several shooting positions around the suspicious power line target are not selected; further, the plant community algorithm uses the shortest total flight time as the suspicious power line target of the rechargeable power line monitoring drone. The objective function of the local detailed inspection route optimization problem of power line targets is used to evaluate the fitness of individual plants; further, the plant community algorithm uses the parameters in sub-step 2-2a to set constraints, and the local detailed inspection route optimization problem of suspicious power line targets of the rechargeable power line monitoring drone should not exceed the total number of suspicious targets, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; further, the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness, and use it as the optimal solution to the local detailed inspection route optimization problem of suspicious power line targets of the rechargeable power line monitoring drone;

[0033] In sub-step 2-3, the plant community algorithm outputs the optimal local detailed inspection route, and the rechargeable power line monitoring drone photographs the suspicious power line target from different angles along the route; the plant community algorithm continuously repeats sub-step 2-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; further, starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal local detailed inspection route for the suspicious power line target; further, the rechargeable power line monitoring drone conducts a local detailed inspection and patrol along the route, and photographs the suspicious power line target along the surrounding upper, lower, left and right shooting positions determined by the optimal solution until all the shooting tasks of the local detailed inspection of the suspicious power line target are completed, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to conduct a global patrol along the route; further, if a new suspicious power line target, or a suspicious obstacle, or insufficient remaining power is encountered in the local detailed inspection route, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 2, or step 3, or step 4;

[0034] In step 3, the following sub-steps are included:

[0035] Sub-step 3-1, the rechargeable power line monitoring drone continuously extracts obstacle features in the monitoring image during the global line patrol process and mines suspicious obstacles; the rechargeable power line monitoring drone starts the global line patrol or after the local detailed line patrol, continuously shoots the monitoring images of the power lines and stores them in the drone memory; further, the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm, including filtering, noise reduction, image segmentation, edge detection, target classification, image enhancement, and extraction of obstacle features; further, the drone stores a suspicious obstacle feature library, including power lines, high-voltage insulators, high-voltage towers, transformers, hillsides, trees Wood, stone, high-rise building, bridge, signal transmission tower, etc., are suspicious obstacle features that may hinder the normal flight of the drone; further, the drone determines whether it is possible to collide with the suspicious obstacle if it continues to fly. If there is no possibility of collision, it maintains the current flight attitude and continues to fly. If it is determined that continuing to fly may cause a collision, further, the obstacle features in the monitoring image are compared with the suspicious obstacle feature library through the image processing algorithm to find out whether there are image features that meet the suspicious obstacle feature library. If not, the drone continues to patrol the line according to the optimal global patrol route in sub-steps 1-3; otherwise, the drone performs local obstacle avoidance on the suspicious obstacle according to sub-step 3-2;

[0036] Sub-step 3-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local obstacle avoidance route for suspicious obstacles;

[0037] In sub-step 3-2a, the plant community algorithm initializes the parameters of the local obstacle avoidance route optimization problem of suspicious obstacles; the plant community algorithm defines the parameters of the local obstacle avoidance route optimization problem of suspicious obstacles of the rechargeable power line monitoring drone, including solving the target set, suspicious obstacle parameters, charging station parameters, and drone parameters; further, solving the target set includes a suspicious obstacle set, a charging station set, a charging pile set of charging stations, a drone set, and a drone type set; the suspicious obstacle parameters include the location and number of suspicious obstacles; the charging station parameters include the location and number of charging stations and each charging pile; the drone parameters include the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly when fully charged, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone;

[0038] In sub-step 3-2b, the plant community algorithm iteratively optimizes the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone for suspected obstacles; the plant individuals are encoded according to the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone for suspected obstacles, and one plant individual corresponds to one feasible decoding; each bit of the plant individual encoding uses 1 to indicate that several obstacle avoidance positions around the corresponding suspected obstacle are selected; each bit of the plant individual encoding uses 0 to indicate that several obstacle avoidance positions around the corresponding suspected obstacle are not selected; further, the plant community algorithm uses the shortest total flight time as the suspected obstacle of the rechargeable power line monitoring drone. The objective function of the local obstacle avoidance route optimization problem is used to evaluate the fitness of individual plants; further, the plant community algorithm uses the parameters in sub-step 3-2a to set constraints, and the local obstacle avoidance route optimization problem of suspicious obstacles for rechargeable power line monitoring drones should not exceed the total number of suspicious obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of drones, the maximum charging capacity of drones, and the maximum distance that drones can fly on a full charge; further, the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness, and use it as the optimal solution to the local obstacle avoidance route optimization problem of suspicious obstacles for rechargeable power line monitoring drones;

[0039] In sub-step 3-3, the plant community algorithm outputs the optimal local obstacle avoidance route, and the rechargeable power line monitoring drone avoids suspicious obstacles along the route; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; further, starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal local obstacle avoidance route for the suspicious obstacle; further, the rechargeable power line monitoring drone performs local obstacle avoidance along the route, flies around and avoids obstacles along several obstacle avoidance positions around the suspicious obstacle determined by the optimal solution, until the suspicious obstacle is completely avoided, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to perform global patrol along the route; further, if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4.

[0040] In step 4, the following sub-steps are included:

[0041] Sub-step 4-1, the rechargeable power line monitoring drone continuously detects the remaining power during the global line patrol process; the rechargeable power line monitoring drone has a power detection sensor, which can obtain the latest value of the power and compare it with the remaining power warning threshold; further, the remaining power warning threshold is not fixed, but is estimated based on the current flight speed and power consumption speed of the drone, and an estimated safety capacity is added; further, the remaining power warning threshold includes the power warning threshold for the drone to return, that is, whether the remaining power can support the drone to return smoothly; further, the remaining power warning threshold includes the power warning threshold for the drone to charge, that is, whether the remaining power can support the drone to fly smoothly to the nearest charging station;

[0042] Sub-step 4-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local charging route based on the power line coordinates, obstacle coordinates, and charging station coordinates according to the remaining power;

[0043] Sub-step 4-2a, the plant community algorithm initializes the parameters of the local charging route optimization problem; the plant community algorithm defines the parameters of the local charging route optimization problem of the rechargeable power line monitoring drone, including the solution target set, suspicious obstacle parameters, charging station parameters, and drone parameters; further, the solution target set includes a suspicious obstacle set, a charging station set, a charging pile set of the charging station, a drone set, and a drone type set; the suspicious obstacle parameters include the location and number of suspicious obstacles; the charging station parameters include the location and number of the charging station and each charging pile; the drone parameters include the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly when fully charged, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone;

[0044] In sub-step 4-2b, the plant community algorithm iteratively optimizes the local charging route optimization problem of the rechargeable power line monitoring drone; the plant community algorithm selects the nearest available charging station as the target node of the local charging route, and encodes the plant individuals according to the local charging route optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible decompression code; each bit of the plant individual code uses 1 to indicate the spatial position passed during the flight; each bit of the plant individual code uses 0 to indicate that the corresponding spatial position is not selected during the flight; further, the plant community algorithm uses the shortest total flight time as the rechargeable power line monitoring drone. The objective function of the local charging route optimization problem of the drone is used to evaluate the fitness of individual plants; further, the plant community algorithm uses the parameters in sub-step 4-2a to set constraints, and the local charging route optimization problem of the rechargeable power line monitoring drone should not exceed the total number of suspicious obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; further, the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the local charging route optimization problem of the rechargeable power line monitoring drone;

[0045] In sub-step 4-3, the plant community algorithm outputs the optimal local charging route, and the rechargeable power line monitoring drone goes to the charging station along the route to complete charging; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; further, starting from the starting condition until the termination condition is met, the plant community algorithm uses the obtained optimal solution as the optimal local charging route; further, the rechargeable power line monitoring drone goes to the nearest available charging station along the route to charge until charging is completed, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to patrol the line globally along the route;

[0046] Furthermore, if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, and the plant community algorithm needs to be used to locally adjust the patrol route, then proceed to step 3 or step 4;

[0047] Furthermore, if the drone cannot reach the nearest available charging station before the battery runs out, the drone should use the GPS module it carries to report its latest location to the drone operator and send a battery depletion warning before the battery runs out;

[0048] Furthermore, if the drone cannot reach the end point of the global power line patrol after being fully charged, it should use the carried GPS module to report the latest position to the drone operator and send a warning that the remaining power cannot reach the end point; further, the drone should continuously search for the nearest available charging station, and if the nearest available charging station can be found, return to sub-step 4-2; further, if the drone cannot reach the end point of the global power line patrol after being fully charged, and cannot search for the nearest available charging station, it should use the carried GPS module to report the latest position to the drone operator and send a warning that there is no nearby charging station.

[0049] Compared with the prior art, the present invention has the following technical effects:

[0050] 1. The invention improves the endurance of drones. The present invention uses a rechargeable power monitoring drone, and arranges charging stations and charging piles on the monitored power line, which can provide power supplement for the drone's endurance. This method can take into account the flight behavior and charging methods of different drones. This method deploys multiple drone charging piles on the power line, collects and analyzes the power consumption, remaining power and charging data of each drone in real time, and searches for the nearest charging station and charging pile by the plant community algorithm to charge the drone. By analyzing the global patrol route of each drone and calculating the remaining power and flight route of the drone, whether the drone can reach the nearest charging pile and the end point of the patrol line are calculated and analyzed respectively. If the drone can reach the nearest charging station, the flight route from the global patrol line to the charging station is optimized, as well as the optimal route back to the global patrol line, to improve the endurance of the drone's remaining power. Furthermore, the calculation of this method takes into account the impact of the charging station location on the drone's endurance, and optimizes the cruising capability while optimizing the global patrol line.

[0051] 2. The invention improves the detailed inspection capability of drones. The present invention uses a rechargeable power line monitoring drone to perform global line inspection optimization. When a suspicious power line target is found, the plant community algorithm is used to select several shooting positions around the corresponding suspicious power line target up, down, left and right, so as to complete the detailed inspection of the suspicious power line target by itself without human intervention. The drone line inspection model of this method takes into account the tasks of global line inspection and local detailed line inspection. Under the overall goal of global line inspection, the flight route can be adjusted according to the suspicious power line targets found to conduct local detailed line inspections, and then return to continue global line inspections after completing the local detailed inspection tasks. Furthermore, if a new suspicious power line target, suspicious obstacle, or insufficient remaining power is encountered during the local detailed inspection route, it is necessary to use the plant community algorithm to locally adjust the line inspection route, then go to step 2, or step 3, or step 4, that is, the plant community algorithm is always online to guide the drone's detailed inspection process and return route.

[0052] 3. The invention improves the obstacle avoidance capability of the UAV. When a suspicious obstacle is found, the present invention uses a plant community algorithm to search for several obstacle avoidance positions around the suspicious obstacle, build an optimal local obstacle avoidance route for the suspicious obstacle, and automatically complete the obstacle avoidance of the UAV without human intervention. The obstacle avoidance model of the UAV in this method is dependent on the plant community and is a dynamic analysis that can take into account the randomness of obstacles and the random behavior of the obstacle avoidance process of the UAV. Furthermore, if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4, that is, the plant community algorithm is always online to guide the obstacle avoidance process and return route of the UAV.

[0053] 4. The invention improves the intelligence of drones. The present invention uses a rechargeable power line monitoring drone to perform global patrol optimization, local detailed inspection route optimization, local obstacle avoidance route optimization, and local charging route optimization, all of which use a plant community algorithm for heuristic search, which greatly reduces manual intervention and improves the level of intelligence. The rechargeable power line monitoring drone of the present invention makes full use of the flight control technology, wireless image return technology, communication remote control technology, track planning and artificial plant community algorithm of existing drones. The drone patrol route planning of the present invention is achieved by designing an artificial plant community algorithm for random search. Since the drone flies in the air, the search space range of the artificial plant community algorithm is not large, and the optimal solution to the patrol route optimization problem can be obtained in a very short time, while also reducing the consumption of drone computing power and energy. While optimizing the global patrol, the present invention also takes into account the local detailed inspection task, avoids obstacles in the flight environment, and optimizes charging, thereby generating an optimal route that meets the needs between the starting point and the target point. The patrol optimization method of the present invention enables drones to promptly discover and eliminate safety hazards of power lines during power patrol inspections, improve patrol efficiency, and achieve early detection and early elimination of power line faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0055] Figure 1 It is a flow chart of the working method of the present invention. DETAILED DESCRIPTION

[0056] A method for patrolling a power line by a rechargeable power line monitoring drone comprises the following steps:

[0057] Step 1: Use a rechargeable power line monitoring drone to perform global line inspection optimization;

[0058] Step 2: Use a rechargeable power line monitoring drone to optimize the local detailed inspection route;

[0059] Step 3: Use a rechargeable power line monitoring drone to optimize the local obstacle avoidance route;

[0060] Step 4: Use a rechargeable power line monitoring drone to optimize local charging routes.

[0061] In step 1, the following sub-steps are included:

[0062] Sub-step 1-1, obtain the global patrol task through the rechargeable power line monitoring drone, input the power line coordinates, obstacle coordinates, and charging station coordinates; the patrol optimization problem of the rechargeable power line monitoring drone includes multiple power lines, multiple obstacles, and multiple charging stations. A charging station may contain multiple charging piles, multiple mileage-restricted, pure electric drones or hybrid drones, and drone patrol charging scheduling optimization problems with time window constraints. In the patrol task, the rechargeable power line monitoring drone needs to access all power line coordinates, avoid obstacle coordinates, and access the charging station coordinates before the battery is exhausted to ensure that the patrol task is completed and returned;

[0063] Sub-step 1-2, the rechargeable power line monitoring drone runs the plant community algorithm to perform global patrol optimization on the power line coordinates, obstacle coordinates, and charging station coordinates;

[0064] Sub-step 1-2a, the plant community algorithm initializes the parameters of the patrol optimization problem of the rechargeable power line monitoring drone; the plant community algorithm defines the parameters of the patrol optimization problem of the rechargeable power line monitoring drone, including solving the target set, power line parameters, obstacle parameters, charging station parameters, and drone parameters;

[0065] In sub-step 1-2b, the plant community algorithm iteratively optimizes the patrol optimization problem of the rechargeable power line monitoring drone; the plant individuals are encoded according to the patrol optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of the corresponding power line, obstacle, and charging station; each bit of the plant individual encoding uses 0 to indicate that the corresponding power line, obstacle, and charging station are not selected;

[0066] In sub-steps 1-3, the plant community algorithm outputs the optimal global patrol route, and the rechargeable power line monitoring drone patrols the line according to the route; the plant community algorithm continuously repeats sub-steps 1-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal global patrol optimization route; the rechargeable power line monitoring drone patrols the line globally according to the route. If a suspicious power line target or suspicious obstacle is encountered during the flight, or the remaining power is insufficient, the plant community algorithm needs to be used to locally adjust the patrol route, and then go to step 2, step 3, or step 4.

[0067] In sub-step 1-1, it is assumed that the location and number of power lines are known, the location and number of obstacles are known, the location of charging stations and charging piles are known, the number of charging piles at each charging station is known, and the maximum charging capacity of each type of drone and the total weight of the drone are known; it is assumed that a charging pile can only charge one drone, the charging pile has a queuing time limit, the drone cannot be overcharged, cannot be charged beyond capacity, the drone flies at a constant speed between nodes, all drones are fully charged each time, the drones are in good condition, the air flight channel is unobstructed, and the sudden appearance of birds, drone failures, and interference or intervention by other drones not for line patrol tasks are not considered.

[0068] In sub-step 1-2a, the target set is solved, including the power line set, the obstacle set, the charging station set, the charging pile set of the charging station, the drone set, and the drone type set; the power line parameters, including the position and number of power lines; the obstacle parameters, including the position and number of obstacles; the charging station parameters, including the position and number of the charging station and each charging pile, the maximum charging load of the charging station, the impact coefficient of drone charging on the grid load, the scalable charging load peak of the charging station, the flight distance between charging stations, the charging service time of each charging station, the charging time window of the charging station, the unit fixed cost, and the unit time charging cost; the drone parameters, including the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance that the drone can fly when fully charged, the current remaining power of the drone, the charging capacity of the drone at the charging station, the time when the drone arrives at the charging station, the flight time of the drone between charging stations, the early arrival penalty cost coefficient of the drone, the late arrival penalty cost coefficient of the drone, the unit flight cost of the drone, and the total flight cost of the drone.

[0069] In sub-step 1-2b, the plant community algorithm uses the shortest total flight time as the objective function of the patrol optimization problem of the rechargeable power line monitoring drone to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 1-2a to set constraints, and the patrol optimization problem of the rechargeable power line monitoring drone should not exceed the total number of power lines, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the patrol optimization problem of the rechargeable power line monitoring drone.

[0070] In step 2, the following sub-steps are included:

[0071] Sub-step 2-1, the rechargeable power line monitoring drone continuously extracts power line features in the monitoring image during the global line patrol process, and mines suspicious power line targets; after the rechargeable power line monitoring drone starts the global line patrol, it continuously takes monitoring images of the power lines and stores them in the drone memory; the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm to extract the features of the power lines; the power line features in the monitoring image are compared with the feature library of suspicious power line targets through the image processing algorithm to mine whether there are power line features that meet the feature library of suspicious power line targets, if not, the drone continues to patrol according to the optimal global line patrol route in sub-steps 1-3; otherwise, the drone conducts a local detailed inspection of the suspicious power line targets according to sub-step 2-2;

[0072] Sub-step 2-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local detailed inspection route of the suspicious power line target;

[0073] Sub-step 2-2a, the plant community algorithm initializes the parameters of the local detailed inspection route optimization problem of the suspicious power line target; the plant community algorithm defines the parameters of the local detailed inspection route optimization problem of the suspicious power line target of the rechargeable power line monitoring drone, including solving the target set, suspicious power line target parameters, obstacle parameters, charging station parameters, and drone parameters;

[0074] In sub-step 2-2b, the plant community algorithm iteratively optimizes the local detailed inspection route optimization problem of the rechargeable power line monitoring drone for suspicious power line targets; the plant individuals are encoded according to the local detailed inspection route optimization problem of the rechargeable power line monitoring drone for suspicious power line targets, and one plant individual corresponds to a feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of several shooting positions around the corresponding suspicious power line target up, down, left, and right; each bit of the plant individual encoding uses 0 to indicate that several shooting positions around the corresponding suspicious power line target are not selected; the plant community algorithm uses the shortest total flight time as the suspicious power line target of the rechargeable power line monitoring drone. The objective function of the power line target local detailed inspection route optimization problem is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 2-2a to set the constraints. The suspicious power line target local detailed inspection route optimization problem of the rechargeable power line monitoring drone should not exceed the total number of suspicious targets, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the suspicious power line target local detailed inspection route optimization problem of the rechargeable power line monitoring drone;

[0075] In sub-step 2-3, the plant community algorithm outputs the optimal local detailed inspection route, and the rechargeable power line monitoring drone photographs the suspicious power line target from different angles along the route; the plant community algorithm continuously repeats sub-step 2-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal local detailed inspection route for the suspicious power line target; the rechargeable power line monitoring drone conducts local detailed inspection and patrol along the route, and photographs the suspicious power line target along the surrounding upper, lower, left and right shooting positions determined by the optimal solution until all the shooting tasks of the local detailed inspection of the suspicious power line target are completed, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to conduct global patrol along the route; if a new suspicious power line target, or a suspicious obstacle, or insufficient remaining power is encountered in the local detailed inspection route, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 2, step 3, or step 4;

[0076] In step 3, the following sub-steps are included:

[0077] Sub-step 3-1, the rechargeable power line monitoring drone continuously extracts obstacle features in the monitoring image during the global line patrol process and mines suspicious obstacles; the rechargeable power line monitoring drone starts the global line patrol or after the local detailed line patrol, continuously shoots the monitoring images of the power lines and stores them in the drone memory; the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm; the drone stores a suspicious obstacle feature library; the drone determines whether it is possible to collide with the suspicious obstacle if it continues to fly. If there is no possibility of collision, it maintains the current flight attitude and continues to fly. If it is determined that continuing to fly may cause a collision; the obstacle features in the monitoring image are compared with the suspicious obstacle feature library through the image processing algorithm to mine whether there are image features that meet the suspicious obstacle feature library. If not, the drone continues to patrol according to the optimal global line patrol route in sub-steps 1-3; otherwise, the drone performs local obstacle avoidance on the suspicious obstacle according to sub-step 3-2;

[0078] Sub-step 3-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local obstacle avoidance route for suspicious obstacles;

[0079] Sub-step 3-2a, the plant community algorithm initializes the parameters of the local obstacle avoidance route optimization problem of suspicious obstacles; the plant community algorithm defines the parameters of the local obstacle avoidance route optimization problem of suspicious obstacles of the rechargeable power line monitoring drone, including the solution target set, suspicious obstacle parameters, charging station parameters, and drone parameters; the solution target set includes a suspicious obstacle set, a charging station set, a charging pile set of charging stations, a drone set, and a drone type set; the suspicious obstacle parameters include the location and number of suspicious obstacles; the charging station parameters include the location and number of charging stations and each charging pile; the drone parameters include the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly when fully charged, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone;

[0080] In sub-step 3-2b, the plant community algorithm iteratively optimizes the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone; the plant individuals are encoded according to the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible solution code; each bit of the plant individual code uses 1 to indicate that several obstacle avoidance positions around the corresponding suspicious obstacle are selected; each bit of the plant individual code uses 0 to indicate that several obstacle avoidance positions around the corresponding suspicious obstacle are not selected; the plant community algorithm uses the shortest total flight time as the feasible solution code of the rechargeable power line monitoring drone. The objective function of the local obstacle avoidance route optimization problem of suspected obstacles is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 3-2a to set the constraints. The local obstacle avoidance route optimization problem of suspected obstacles for rechargeable power line monitoring drones should not exceed the total number of suspected obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of drones, the maximum charging capacity of drones, and the maximum distance that drones can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the local obstacle avoidance route optimization problem of suspected obstacles for rechargeable power line monitoring drones;

[0081] In sub-step 3-3, the plant community algorithm outputs the optimal local obstacle avoidance route, and the rechargeable power line monitoring drone avoids suspicious obstacles along the route; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal local obstacle avoidance route for the suspicious obstacle; the rechargeable power line monitoring drone performs local obstacle avoidance along the route, flies around and avoids obstacles along several obstacle avoidance positions around the suspicious obstacle determined by the optimal solution, until the suspicious obstacle is completely avoided, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to perform global patrol along the route; if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4.

[0082] In step 4, the following sub-steps are included:

[0083] Sub-step 4-1, the rechargeable power line monitoring drone continuously detects the remaining power during the global line patrol process; the rechargeable power line monitoring drone has a power detection sensor, which can obtain the latest value of the power and compare it with the remaining power warning threshold; the remaining power warning threshold is not fixed, but is estimated based on the current flight speed and power consumption speed of the drone, and an estimated safety capacity is added; the remaining power warning threshold includes the power warning threshold for the drone to return, that is, whether the remaining power can support the drone to return smoothly; the remaining power warning threshold includes the power warning threshold for the drone to charge, that is, whether the remaining power can support the drone to fly smoothly to the nearest charging station;

[0084] Sub-step 4-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local charging route based on the power line coordinates, obstacle coordinates, and charging station coordinates according to the remaining power;

[0085] Sub-step 4-2a, the plant community algorithm initializes the parameters of the local charging route optimization problem; the plant community algorithm defines the parameters of the local charging route optimization problem of the rechargeable power line monitoring drone, including the solution target set, suspicious obstacle parameters, charging station parameters, and drone parameters; the solution target set includes a suspicious obstacle set, a charging station set, a charging pile set of charging stations, a drone set, and a drone type set; suspicious obstacle parameters include the location and number of suspicious obstacles; charging station parameters include the location and number of charging stations and each charging pile; drone parameters include the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly when fully charged, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone;

[0086] In sub-step 4-2b, the plant community algorithm iteratively optimizes the local charging route optimization problem of the rechargeable power line monitoring drone; the plant community algorithm selects the nearest available charging station as the target node of the local charging route, and encodes the plant individuals according to the local charging route optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible solution code; each bit of the plant individual code uses 1 to indicate the spatial position passed during the flight; each bit of the plant individual code uses 0 to indicate that the corresponding spatial position is not selected during the flight; the plant community algorithm uses the shortest total flight time as the rechargeable power line The objective function of the local charging route optimization problem of the monitoring drone is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 4-2a to set the constraints. The local charging route optimization problem of the rechargeable power line monitoring drone should not exceed the total number of suspicious obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses the seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the local charging route optimization problem of the rechargeable power line monitoring drone;

[0087] In sub-step 4-3, the plant community algorithm outputs the optimal local charging route, and the rechargeable power line monitoring drone goes to the charging station along the route to complete charging; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the obtained optimal solution as the optimal local charging route; the rechargeable power line monitoring drone goes to the nearest available charging station along the route to charge until charging is completed, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to patrol the line globally along the route;

[0088] If a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4;

[0089] If the drone cannot reach the nearest available charging station before the battery runs out, the drone should use the GPS module it carries to report its latest location to the drone operator and send a battery depletion warning before the battery runs out;

[0090] If the drone cannot reach the end point of the global power line patrol after being fully charged, it should use the GPS module it carries to report its latest position to the drone operator and send a warning that the remaining power cannot reach the end point; the drone should continuously search for the nearest available charging station, and if the nearest available charging station can be found, it returns to sub-step 4-2; if the drone cannot reach the end point of the global power line patrol after being fully charged, and cannot search for the nearest available charging station, it should use the GPS module it carries to report its latest position to the drone operator and send a warning that there is no nearby charging station.

[0091] Example:

[0092] like Figure 1 As shown, a patrol method for a rechargeable power line monitoring drone is used to optimize the patrol optimization problem of the rechargeable power line monitoring drone by simulating the seeding, growth, and fruiting behaviors of plant communities.

[0093] The rechargeable power line monitoring drone is a drone with a camera and network communication equipment that can implement power line monitoring. In terms of power, it includes pure electric drones and hybrid drones that use electricity and other energy sources. Furthermore, the rechargeable power line monitoring drone is equipped with a wireless charging module, which can be modified by adding a mobile phone wireless charging module to a publicly available drone. It is small in size and light in weight. Preferably, the drone uses an electromagnetic induction wireless fast charging module. When the drone stops above the charging pile, it can undergo electromagnetic induction with the charging pile to complete the charging process. There is no need to add an additional charging interface to the drone, nor is it necessary to fully contact the charging pile.

[0094] The charging station includes several wireless charging piles, which are deployed on the power line and use the power line to obtain the electricity required for charging without the need to add power lines. Preferably, a charging station often includes multiple charging piles, and a drone helipad is placed above each charging pile to provide rain and sun protection for the charging pile below. One charging pile can only wirelessly charge several drones at a time, including pure electric drones and hybrid drones. Preferably, the charging pile uses an air-spaced 15-28MM wireless fast charging module with a product size of no more than φ80*30MM. One charging pile can use multiple wireless fast charging modules, compatible with 10W / 20W / 30W / 40W / 50W charging power of different drones, with a charging current of 3A. The drone helipad is set at a distance of no more than 15-28MM from the charging pile to ensure that several drones can share charging when parked above the charging pile helipad. Preferably, an electromagnetic induction wireless fast charging module is used, with an operating temperature of -10 to 40°C, a storage temperature of -40 to 70°C, a maximum altitude of 5000 meters, a power supply DC voltage range of 12V DC to 24V DC, a voltage ripple ≤200mV, a rated input current of 6A, a standby power consumption of 0.5W, and overvoltage protection, undervoltage protection, and over-temperature protection.

[0095] The patrol optimization problem of the rechargeable power line monitoring drone refers to making full use of the drone's battery power and charging stations, reasonably optimizing all local detailed inspection routes, obstacle avoidance routes and charging routes, and completing the global patrol process; the patrol optimization problem of the rechargeable power line monitoring drone is a drone charging scheduling optimization problem with multiple charging stations, mileage restrictions, multiple types of pure electric drones or hybrid electric drones, and time window constraints. The specific description is: within a certain geographical range, there are multiple power lines, multiple charging stations, multiple obstacles and multiple drones with charging needs, including pure electric drones or hybrid electric drones; the charging station is provided with a wireless charging interface and can have multiple pure drones or hybrid drones with different specifications that can be wirelessly charged. Each drone has its own charging needs and has a specific service time window. Each charging pile can provide charging services to multiple drones at the same time; electric drones cannot be charged during flight. After the flight is completed, all drones need to return to the charging station. By reasonably arranging the charging time, flight route and flight time, under the conditions of completing the patrol task and meeting various constraints, the goal of minimizing the sum of fixed cost, flight cost, penalty cost and charging cost is finally achieved.

[0096] The multi-objective function of the UAV patrol line optimization problem is used to evaluate the quality or performance of feasible solutions to the patrol line optimization problem of a rechargeable power line monitoring UAV; further, different performance indicators require different multi-objective functions of the patrol line optimization of a rechargeable power line monitoring UAV to measure the effectiveness of UAV patrol line from different perspectives.

[0097] The plant community algorithm is an artificial intelligence algorithm deployed on a drone to solve the patrol route optimization problem;

[0098] The plant community is used to simulate the solution space of the patrol optimization problem of the rechargeable power line monitoring drone; further, it includes a seeding operation, a growth operation, and a result operation; through the continuous cycle and iteration of the three operations, the optimal solution of the patrol optimization problem of the rechargeable power line monitoring drone is found.

[0099] The plant individuals are used to simulate a feasible solution to the patrol optimization problem of the rechargeable power line monitoring drone; the minimum value of the plant individual is 0, and the maximum value is an integer with all 1s in each digit, and the number of digits depends on the number of power lines, the number of charging stations, the number of obstacles, and the size of the rechargeable power line monitoring drone;

[0100] The plant community population size, i.e., the number of plant individuals in the plant community, is used to simulate the number of feasible solutions to the patrol optimization problem of a rechargeable power line monitoring drone; preferably, the number of feasible solutions may be consistent with or inconsistent with the scale of the problem to be solved;

[0101] The fitness function of the plant planting is used to simulate and evaluate the performance of solving the patrol optimization problem of the rechargeable power line monitoring drone. The user can select different drone patrol optimization problem performance functions as the fitness function according to the usage needs and different application scenarios; further, the user can use all or part of the multi-objective function of the drone patrol optimization problem as the fitness function of the plant planting;

[0102] The sowing operation of the plant community simulates the seed sowing process of a real plant community in nature, and is used to generate the initial feasible solution value required for each round of iterative calculation. The feasible solution value can be used to complete the subsequent growth operation and result operation; the sowing operation is both random and limited to the vicinity of a certain plant;

[0103] The growth operation of the plant community simulates the growth and mutation process of real plant communities in nature. The sowing operation of plants is random, but only seeds near water sources can continue to grow, and plant individuals that cannot obtain water sources can only die; during the growth process of plants, they may also be affected by water sources or the environment and mutate to become new plant individuals; the growth operation is used to generate new feasible solutions, retaining the main information of the original feasible solutions, but also expanding the search range; the growth mutation probability of the plant community simulates the probability of mutation of real plants in nature during the growth operation, that is, the feasible solution changes in a small range with the mutation probability during the growth process; further, a larger growth mutation probability facilitates the plant community to expand the search space, but it will also lead to a slower convergence speed; a smaller mutation probability is easy to increase the convergence speed of the plant community, but it is also easy to cause the plant community to converge to the local optimal solution earlier;

[0104] The result operation of the plant community is used to simulate the learning function of individual plants in a real plant community in nature, that is, the process in which different plant individuals exchange genetic material through pollen and perform result operation on the optimal solution; the result operation can realize the exchange of data of feasible solutions of different plant individuals, generate new feasible solutions, and simulate the social learning function of the plant community; the result probability of the plant community refers to the probability that different plant individuals learn from each other and perform result operation; a smaller result probability facilitates the plant community to expand the search space, so that the optimal solution retains less, but it will also lead to a slower convergence speed; a larger result probability is easy to increase the convergence speed of the plant community, and it is easy to retain more optimal solutions of the plant community, but it is also easy to cause the plant community to converge to the local optimal solution earlier;

[0105] A method for patrolling a power line by a rechargeable power line monitoring drone comprises the following main steps:

[0106] Step 1: Global patrol optimization of the rechargeable power line monitoring drone;

[0107] Step 2, local detailed inspection route optimization of the rechargeable power line monitoring drone;

[0108] Step 3, local obstacle avoidance route optimization of the rechargeable power line monitoring drone;

[0109] Step 4: Optimize the local charging route of the rechargeable power line monitoring drone.

[0110] In step 1, the following sub-steps are included:

[0111] Sub-step 1-1, the rechargeable power line monitoring drone obtains a global patrol task and inputs the power line coordinates, obstacle coordinates, and charging station coordinates;

[0112] Since the actual flight and charging process of the drone is affected by many external factors, some of which cannot be quantitatively expressed in the model, the following assumptions are made:

[0113] 1) The locations of charging stations and charging piles are known;

[0114] 2) The number of charging piles at each charging station is known;

[0115] 3) The maximum charging capacity and total weight of each drone are known and fixed, and the change of charging capacity with battery life and the change of total weight with flight mileage are not considered;

[0116] 4) There are multiple charging stations, all deployed on power lines;

[0117] 5) The charging station at the starting point of the line patrol fully charges the drone. The starting point is the starting and ending point of all line patrol flight missions;

[0118] 6) Charging piles have a queue time limit, and if the charging power line monitoring drone does not arrive at the queue within the specified time, a penalty cost will be incurred;

[0119] 7) There is no waiting time after the rechargeable power line monitoring drone starts charging, and the charging start and stop time of the charging pile is not considered, nor is the failure of the charging pile and the apron congestion problem. It is assumed that all drones have the same service time at the charging pile;

[0120] 8) Each charging station can be serviced at least once;

[0121] 9) The drone cannot be overcharged or overcharged;

[0122] 10) The UAV flies at a constant speed between nodes on each patrol route, without considering factors such as acceleration, deceleration, waiting, and accidents during the flight;

[0123] 11) All drones are fully charged every time and are fully charged when departing from the charging station;

[0124] 12) The drone is in good condition and the air flight path is unobstructed. Sudden drone failures, road obstacles, drone operator errors, and collisions caused by other drones are not considered;

[0125] The parameters used in the patrol optimization problem of a rechargeable power line monitoring drone are defined as follows:

[0126] O, represents the charging station collection

[0127] C, represents the charging pile set {1,2,3…n}

[0128] D, D = O∪C, represents the charging pile set of the charging station

[0129] N, represents the set of natural numbers

[0130] K, represents the drone set 1, 2, 3…n

[0131] M, represents the set of drone types 1, 2, 3…m

[0132] f m , represents the impact coefficient of the mth UAV on the grid load, which is usually less than 1

[0133] g m , represents the charging capacity factor of the mth drone, usually less than 1

[0134] p im , represents the charging load peak value that can be expanded at the i-th charging station

[0135] represents the charging capacity of the kth UAV at the i-th charging station

[0136] Indicates the time when the kth drone arrives at point i

[0137] d ij , represents the flight distance between charging stations i and j, in km

[0138] t s , represents the service time of each charging station i, unit min

[0139] t ij , represents the flight time of the drone from charging station i to charging station j, in min

[0140] [a i ,b j ], represents the charging time window of charging station i

[0141] h, represents the early arrival penalty cost coefficient, unit: yuan / min

[0142] e, represents the penalty cost coefficient for lateness, unit: RMB / min

[0143] v, represents the flight speed of the drone, unit Km / min

[0144] V, represents the maximum charging load of the charging station, that is, the maximum charging capacity of the charging station, unit KWH

[0145] W, represents the maximum charging capacity of the drone, unit KWH

[0146] r, indicates the maximum distance the drone can fly after fully charging, in kilometers

[0147] Z, represents the total cost of UAV flight, unit yuan

[0148] c 1 , represents unit fixed cost, unit yuan / station

[0149] c 2 , represents the unit flight cost, unit yuan / km

[0150] c b , represents the charging cost per unit time, RMB / time

[0151] 0-1 variable, charging station construction selection bit, indicating that node i is selected to build a charging station, and electric drone k can fly from charging station i to point j, then the selection bit of charging station i takes the value of 1, otherwise it takes the value of 0

[0152] 0-1 variable, the selection bit of the charging station selected by multiple types of drones, indicating that when drone k uses the m-type drone of charging station i, the value is 1, otherwise the value is 0

[0153] 0-1 variable, the selection bit of the electric drone to select the charging station, indicating that when there is drone k using the drone of charging station i, the value is 1, otherwise the value is 0

[0154] Sub-step 1-2, the rechargeable power line monitoring drone runs the plant community algorithm to perform global patrol optimization on the power line coordinates, obstacle coordinates, and charging station coordinates;

[0155] Sub-step 1-2a, the plant community algorithm initializes the parameters of the patrol optimization problem of the rechargeable power line monitoring drone; further, initialize the plant community parameters, including the size of the plant population in the plant community, the numerical type and numerical size of the individual plant, the multi-objective function of the patrol optimization of the rechargeable power line monitoring drone of the individual plant is cleared, and the growth variation probability and result probability of the plant community are set; the charging station and drone set of the individual plant are cleared, that is, no charging station and drone are selected into the optimal solution, the plant plant sowing set, the plant community result pairing set, and the plant plant individual with the highest sorting priority are cleared; further, the plant plant individual is encoded according to the selection bit of the power line, charging station, obstacle and drone, the encoding length depends on the number of power lines, charging stations, obstacles and drones, and the encoded value depends on the power line, charging station, obstacle selection bit and whether the drone selects the charging station; preferably, 1 indicates that the node is selected as the corresponding power line, charging station, obstacle, otherwise, 0 indicates that the node does not select the corresponding power line, charging station, obstacle;

[0156] Sub-step 1-2a1, further, from the perspective of global patrol, the plant community takes the lowest total cost as the model decision objective and fitness function. The total cost includes fixed cost, flight cost, time window penalty cost and charging cost.

[0157] Sub-step 1-2a2, further, calculates fixed costs, which refers to the relatively fixed costs in the construction of charging piles and the use of electric drones. It is an important part of the total cost, mainly including the purchase cost of charging piles, the construction cost of charging stations, the operation and maintenance costs of charging stations, material costs, the purchase cost of drones, the wages of operators, and the loss of drones. Assume that the fixed cost per unit of the charging station is c 1 , then the fixed cost can be calculated as:

[0158]

[0159] Sub-step 1-2a3, further, calculates the flight cost, which is the cost of the electric energy consumed by the pure electric drone or hybrid drone during the flight. Compared with the pure fuel drone, which mainly uses fuel consumption as the flight cost, the pure electric drone or hybrid drone uses electricity as the power source, and its energy consumption is closely related to its flight mileage. Assume that the unit distance flight cost of the rechargeable power line monitoring drone is c 2 , then the total flight cost of the drone is:

[0160]

[0161] Sub-step 1-2a4, further, calculate the time window penalty cost. In the process of using the rechargeable power line monitoring drone, due to the influence of various force majeure, the drone misses the charging time in the hard time window, so a soft time window is adopted. The soft time window will not refuse to serve the rechargeable power line monitoring drone that arrives early or late, but a certain penalty cost will be charged, and the penalty fee will be given according to the difference from the specified time window. Assume that the time for the rechargeable power line monitoring drone k to arrive at the charging pile i is According to the time window requirements, the penalty cost function generated at charging pile i is:

[0162]

[0163] Therefore, the total time window penalty cost of the rechargeable power line monitoring drone is:

[0164]

[0165] Sub-step 1-2a5, further, calculate the charging cost. During the use of the rechargeable power line monitoring drone, each charging at the charging pile will increase the charging cost. To prevent overcharging and increase the total cost, the number of charging times and charging time are constrained by constraining the charging cost. Assume that the charging cost per unit time is c b , then the charging cost during the entire drone use process is:

[0166]

[0167] Sub-step 1-2a6, further, calculates the total cost of the rechargeable power line monitoring drone. The objective function is obtained by the above formulas (1), (2), (3), (4), and (5) to obtain the sum of the four costs:

[0168]

[0169] Sub-step 1-2a7, further, calculates the charging station constraints for the patrol optimization of the rechargeable power line monitoring drone; preferably, five constraints are considered: charging station constraint, charging time constraint, charging capacity constraint, drone load constraint, and drone mileage constraint.

[0170] Sub-step 1-2a8, further, calculate the charging station constraints. The charging demand of drones during the flight process is splittable. Therefore, one charging station can provide charging services for multiple drones, as shown in formula (7); electric drones can enjoy charging services at different charging stations, but the number of times a drone visits any charging station is the same as the number of times it leaves any charging station, as shown in formula (8);

[0171]

[0172] Furthermore, after completing the inspection mission, the drone must return to the charging station, as shown in equations (9) and (10).

[0173]

[0174] Sub-step 1-2a9, calculates the charging time constraint for patrol optimization of the rechargeable power line monitoring drone; the charging time constraint, that is, the time when the electric drone departs from charging station i and arrives at charging station j, is equal to the sum of the time when the drone arrives at the previous node i, the service time at node i and the flight time at nodes i, j, as shown in formula (11).

[0175]

[0176] Sub-step 1-2a10, calculate the charging capacity constraint of the patrol optimization of the rechargeable power line monitoring drone; the drone charging capacity constraint refers to the limited charging capacity of pure electric drones or hybrid electric drones, and the total charging capacity of the charging station is also limited, so the charging capacity of a drone shall not exceed the maximum charging capacity limit of the drone, and the total charging capacity of the drones served by a charging station shall not exceed the grid load of the charging station. Formula (12) indicates that the charging capacity of a drone shall not be greater than the maximum charging capacity limit of the drone. Formula (13) indicates that the total charging capacity of all types of drones served by a charging station shall not exceed the grid load of the charging station.

[0177]

[0178] Sub-step 1-2a11, calculate the extended grid capacity constraint for patrol optimization of rechargeable power line monitoring drones; the extended grid capacity constraint, during the long-term use of the charging station, the charging station may add new charging piles, but it cannot exceed the site area constraint, and the power capacity of the expanded power line equipment, lines and charging interfaces should not exceed the maximum expandable charging load constraint of the charging station. The sum of the charging loads of all types of drones at each charging station should not exceed the expandable charging load peak of the charging station, as shown in formula (14); the charging load of each electric drone should not exceed the expandable charging load peak of the charging station, as shown in formula (15).

[0179]

[0180] Sub-step 1-2a12, calculate the drone mileage constraint for patrol optimization of the rechargeable power line monitoring drone; the drone mileage constraint, due to the limitation of the battery capacity of pure electric drones or hybrid drones, and the fact that drone operators cannot operate the drone for a long time, a rated drone maximum flight mileage must be set. Considering factors such as the complexity of the task during the patrol flight, the actual maximum flight mileage of the drone should be less than the rated drone maximum flight mileage. It is stipulated that the total flight mileage of all drones is not allowed to exceed 80% of the rated drone maximum flight mileage, as shown in formula (16).

[0181]

[0182] Sub-step 1-2a13, calculating the multi-objective function and constraint conditions of the patrol optimization of the rechargeable power line monitoring drone; In summary, the multi-objective function and constraint conditions of the patrol optimization of the rechargeable power line monitoring drone are calculated as follows:

[0183]

[0184]

[0185] In sub-step 1-2b, the plant community algorithm iteratively optimizes the patrol optimization problem of the rechargeable power line monitoring drone; further, the plant community algorithm uses the parameters in sub-step 1-2a to set the constraints; further, in the growth operation, the plant community compares the fitness of individual plants, and the plant individuals with small fitness will die, and only the plant individuals with large fitness will survive; further, in the fruit operation, the plant individuals with large fitness will learn from each other and exchange some feasible solution information with each other, thereby producing more fruits, and the plant community population size increases; further, in the sowing operation, the fitness of individual plants is compared again, and the plant individuals with small fitness, that is, the fruits, will die during the sowing process, and the plants will produce random mutations to form a part of new seeds, and the fruits of the plant community with large fitness will become seeds and survive, and enter the next round of production operations; after continuous cycles of the three operations of sowing, production, and fruiting, the optimal solution to the patrol optimization problem is found;

[0186] In sub-steps 1-3, the plant community algorithm outputs the optimal global patrol route, and the rechargeable power line monitoring drone patrols the line according to the route; the plant community algorithm continuously repeats sub-steps 1-2 until the termination condition is met, and the plant community algorithm uses the obtained optimal solution as the optimal global patrol optimization route; further, the rechargeable power line monitoring drone patrols the line globally according to the route. If it encounters a suspicious power line target or a suspicious obstacle during the flight, or the remaining power is insufficient, it is necessary to use the plant community algorithm to partially adjust the patrol route, and then go to step 2, step 3, or step 4;

[0187] In step 2, the following sub-steps are included:

[0188] Sub-step 2-1, the rechargeable power line monitoring drone continuously extracts power line features in the monitoring image during the global line patrol process, and mines suspicious power line targets; further, it mines whether there are power line features that meet the suspicious power line target feature library. If not, the drone continues to patrol the line according to the optimal global line patrol route in sub-step 1-3; otherwise, the drone conducts a local detailed inspection of the suspicious power line target according to sub-step 2-2;

[0189] Sub-step 2-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local detailed inspection route of the suspicious power line target;

[0190] In sub-step 2-2a, the plant community algorithm initializes parameters for the optimization problem of the local detailed inspection route of the suspicious power line target; the plant community selects and designs the fitness function with reference to sub-steps 1-2a1 to 1-2a13, and completes the local detailed inspection of the suspicious power line target with the optimal route;

[0191] In sub-step 2-2b, the plant community algorithm iteratively optimizes the local detailed inspection route optimization problem of the suspicious power line target of the rechargeable power line monitoring drone; further, the plant community algorithm uses the parameters in sub-step 2-2a to set the constraints; further, in the growth operation, the plant community compares the fitness of individual plants, and the plant individuals with small fitness will die, and only the plant individuals with large fitness will survive; further, in the fruit operation, the plant individuals with large fitness will learn from each other and exchange some feasible solution information with each other, thereby producing more fruits, and the plant community population size increases; further, in the sowing operation, the fitness of individual plants is compared again, and the plant individuals with small fitness, that is, the fruits, will die during the sowing process, and the plant plants will produce random mutations to form a part of new seeds, and the fruits of the plant community with large fitness will become seeds and survive, and enter the next round of production operations; after continuous cycles of the three operations of sowing, production, and fruiting, the optimal solution to the line inspection optimization problem is found;

[0192] In sub-step 2-3, the plant community algorithm outputs the optimal local detailed inspection route, and the rechargeable power line monitoring drone takes photos of the suspicious power line targets from different angles along the route; the plant community algorithm continuously repeats sub-step 2-2 until the termination condition is met, and the plant community algorithm uses the optimal solution obtained as the optimal local detailed inspection route for the suspicious power line targets; further, until all the shooting tasks of the local detailed inspection of the suspicious power line targets are completed, the optimal global patrol route in sub-step 1-3 is returned; further, if a new suspicious power line target, a suspicious obstacle, or insufficient remaining power is encountered in the local detailed inspection route, the plant community algorithm needs to be used to locally adjust the patrol route, and then go to step 2, step 3, or step 4;

[0193] In step 3, the following sub-steps are included:

[0194] Sub-step 3-1, the rechargeable power line monitoring drone continuously extracts obstacle features in the monitoring image during the global line patrol process, and mines suspicious obstacles; further, it mines whether there are image features that meet the suspicious obstacle feature library. If not, the drone continues to patrol the line according to the optimal global line patrol route in sub-step 1-3; otherwise, the drone performs local obstacle avoidance for the suspicious obstacle according to sub-step 3-2;

[0195] Sub-step 3-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local obstacle avoidance route for suspicious obstacles;

[0196] In sub-step 3-2a, the plant community algorithm initializes parameters for the optimization problem of the local obstacle avoidance route of the suspicious obstacle; the plant community selects and designs the fitness function with reference to sub-steps 1-2a1 to 1-2a13, and completes the local obstacle avoidance of the suspicious obstacle with the optimal route;

[0197] In sub-step 3-2b, the plant community algorithm iteratively optimizes the optimization problem of the local obstacle avoidance route of the rechargeable power line monitoring drone for suspicious obstacles; further, the plant community algorithm uses the parameters in sub-step 3-2a to set the constraints; further, in the growth operation, the plant community compares the fitness of individual plants, and the plant individuals with small fitness will die, and only the plant individuals with large fitness will survive; further, in the fruit operation, the plant individuals with large fitness will learn from each other and exchange some feasible solution information with each other, thereby producing more fruits, and the plant community population size increases; further, in the sowing operation, the fitness of individual plants is compared again, and the plant individuals with small fitness, that is, the fruits, will die during the sowing process, and at the same time, the plants will produce random mutations to form a part of new seeds, and the fruits of the plant community with large fitness will become seeds and survive, and enter the next round of production operations; after continuous cycles of the three operations of sowing, production, and fruiting, the optimal solution to the line patrol optimization problem is found;

[0198] In sub-step 3-3, the plant community algorithm outputs the optimal local obstacle avoidance route, and the rechargeable power line monitoring drone avoids suspicious obstacles along the route; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the plant community algorithm uses the optimal solution obtained as the optimal local obstacle avoidance route for suspicious obstacles; further, until the suspicious obstacles are completely avoided, the optimal global patrol route in sub-step 1-3 is returned; further, if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, the plant community algorithm needs to be used to locally adjust the patrol route, then go to step 3 or step 4.

[0199] In step 4, the following sub-steps are included:

[0200] Sub-step 4-1, the rechargeable power line monitoring drone continuously detects the remaining power during the global line patrol process;

[0201] Sub-step 4-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local charging route based on the power line coordinates, obstacle coordinates, and charging station coordinates according to the remaining power;

[0202] Sub-step 4-2a, the plant community algorithm initializes parameters for the local charging route optimization problem; the plant community refers to sub-steps 1-2a1 to 1-2a13 to select and design the fitness function, and completes the nearest charging using the optimal route;

[0203] In sub-step 4-2b, the plant community algorithm iteratively optimizes the local charging route optimization problem of the rechargeable power line monitoring drone; further, the plant community algorithm uses the parameters in sub-step 4-2a to set constraints; further, in the growth operation, the plant community compares the fitness of individual plants, and the plant individuals with small fitness will die, and only the plant individuals with large fitness will survive; further, in the fruit operation, the plant individuals with large fitness will learn from each other and exchange some feasible solution information with each other, thereby producing more fruits, and the plant community population size increases; further, in the sowing operation, the fitness of individual plants is compared again, and the plant individuals with small fitness, that is, the fruits, will die during the sowing process, and at the same time, the plants will produce random mutations to form a part of new seeds, and the fruits of the plant community with large fitness will become seeds and survive, and enter the next round of production operations; after continuous cycles of the three operations of sowing, production, and fruiting, the optimal solution to the line patrol optimization problem is found;

[0204] In sub-step 4-3, the plant community algorithm outputs the optimal local charging route, and the rechargeable power line monitoring drone goes to the charging station along the route to complete charging; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the plant community algorithm uses the optimal solution obtained as the optimal local charging route; further, until charging is completed, it returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to patrol the line globally along the route; further, if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4.

Claims

1. A method for patrolling power lines using a rechargeable power line monitoring drone, characterized in that: It includes the following steps: Step 1: Use a rechargeable power line monitoring drone to perform global line inspection optimization; Step 2: Use a rechargeable power line monitoring drone to optimize the local detailed inspection route; Step 3: Use a rechargeable power line monitoring drone to optimize the local obstacle avoidance route; Step 4: Use a rechargeable power line monitoring drone to optimize local charging routes.

2. The method according to claim 1, characterized in that In step 1, the following sub-steps are included: Sub-step 1-1, obtain the global patrol task through the rechargeable power line monitoring drone, input the power line coordinates, obstacle coordinates, and charging station coordinates; the patrol optimization problem of the rechargeable power line monitoring drone includes multiple power lines, multiple obstacles, and multiple charging stations. A charging station may contain multiple charging piles, multiple mileage-restricted, pure electric drones or hybrid drones, and drone patrol charging scheduling optimization problems with time window constraints. In the patrol task, the rechargeable power line monitoring drone needs to access all power line coordinates, avoid obstacle coordinates, and access the charging station coordinates before the battery is exhausted to ensure that the patrol task is completed and returned; Sub-step 1-2, the rechargeable power line monitoring drone runs the plant community algorithm to perform global patrol optimization on the power line coordinates, obstacle coordinates, and charging station coordinates; Sub-step 1-2a, the plant community algorithm initializes the parameters of the patrol optimization problem of the rechargeable power line monitoring drone; the plant community algorithm defines the parameters of the patrol optimization problem of the rechargeable power line monitoring drone, including solving the target set, power line parameters, obstacle parameters, charging station parameters, and drone parameters; In sub-step 1-2b, the plant community algorithm iteratively optimizes the patrol optimization problem of the rechargeable power line monitoring drone; the plant individuals are encoded according to the patrol optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of the corresponding power line, obstacle, and charging station; each bit of the plant individual encoding uses 0 to indicate that the corresponding power line, obstacle, and charging station are not selected; In sub-steps 1-3, the plant community algorithm outputs the optimal global patrol route, and the rechargeable power line monitoring drone patrols the line according to the route; the plant community algorithm continuously repeats sub-steps 1-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal global patrol optimization route; The rechargeable power line monitoring drone conducts a global line patrol along the route. If it encounters a suspicious power line target, a suspicious obstacle, or insufficient remaining power during the flight, it is necessary to use the plant community algorithm to locally adjust the line patrol route, and then proceed to step 2, step 3, or step 4.

3. The method according to claim 2, characterized in that In sub-step 1-1, it is assumed that the location and number of power lines are known, the location and number of obstacles are known, the location of charging stations and charging piles are known, the number of charging piles at each charging station is known, and the maximum charging capacity of each type of drone and the total weight of the drone are known; it is assumed that a charging pile can only charge one drone, the charging pile has a queuing time limit, the drone cannot be overcharged, cannot be charged beyond capacity, the drone flies at a constant speed between nodes, all drones are fully charged each time, the drones are in good condition, the air flight channel is unobstructed, and the sudden appearance of birds, drone failures, and interference or intervention by other drones not for line patrol tasks are not considered.

4. The method according to claim 2, characterized in that: In sub-step 1-2a, solve the target set, including the power line set, obstacle set, charging station set, charging pile set of charging station, drone set, and drone type set; power line parameters, including the position and number of power lines; obstacle parameters, including the position and number of obstacles; charging station parameters, including the position and number of charging stations and charging piles, the maximum charging load of the charging station, the impact coefficient of drone charging on the grid load, the scalable charging load peak of the charging station, the flight distance between charging stations, the charging service time of each charging station, the charging time window of the charging station, the unit fixed cost, and the unit time charging cost; Drone parameters, including the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly on a full charge, the current remaining power of the drone, the charging capacity of the drone at the charging station, the time when the drone arrives at the charging station, the flight time of the drone between charging stations, the penalty cost coefficient for early arrival of the drone, the penalty cost coefficient for late arrival of the drone, the unit flight cost of the drone, and the total flight cost of the drone.

5. The method according to claim 3 or 4, characterized in that: In sub-step 1-2b, the plant community algorithm uses the shortest total flight time as the objective function of the patrol optimization problem of the rechargeable power line monitoring drone to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 1-2a to set constraints, and the patrol optimization problem of the rechargeable power line monitoring drone should not exceed the total number of power lines, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the patrol optimization problem of the rechargeable power line monitoring drone.

6. The method according to claim 1, characterized in that In step 2, the following sub-steps are included: Sub-step 2-1, the rechargeable power line monitoring drone continuously extracts power line features in the monitoring image during the global line patrol process, and mines suspicious power line targets; after the rechargeable power line monitoring drone starts the global line patrol, it continuously takes monitoring images of the power lines and stores them in the drone memory; the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm to extract the features of the power lines; The power line features in the monitoring image are compared with the suspicious power line target feature library through the image processing algorithm to find out whether there are power line features that meet the suspicious power line target feature library. If not, the UAV continues to patrol the line according to the optimal global patrol route in sub-steps 1-3; otherwise, the UAV conducts a local detailed inspection of the suspicious power line target according to sub-step 2-2; Sub-step 2-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local detailed inspection route of the suspicious power line target; Sub-step 2-2a, the plant community algorithm initializes the parameters of the local detailed inspection route optimization problem of the suspicious power line target; the plant community algorithm defines the parameters of the local detailed inspection route optimization problem of the suspicious power line target of the rechargeable power line monitoring drone, including solving the target set, suspicious power line target parameters, obstacle parameters, charging station parameters, and drone parameters; In sub-step 2-2b, the plant community algorithm iteratively optimizes the local detailed inspection route optimization problem of the rechargeable power line monitoring drone for suspicious power line targets; the plant individuals are encoded according to the local detailed inspection route optimization problem of the rechargeable power line monitoring drone for suspicious power line targets, and one plant individual corresponds to a feasible decoding; each bit of the plant individual encoding uses 1 to indicate the selection of several shooting positions around the corresponding suspicious power line target up, down, left, and right; each bit of the plant individual encoding uses 0 to indicate that several shooting positions around the corresponding suspicious power line target are not selected; the plant community algorithm uses the shortest total flight time as the suspicious power line target of the rechargeable power line monitoring drone. The objective function of the power line target local detailed inspection route optimization problem is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 2-2a to set the constraints. The suspicious power line target local detailed inspection route optimization problem of the rechargeable power line monitoring drone should not exceed the total number of suspicious targets, the total number of obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the suspicious power line target local detailed inspection route optimization problem of the rechargeable power line monitoring drone; In sub-step 2-3, the plant community algorithm outputs the optimal local detailed inspection route, and the rechargeable power line monitoring drone takes photos of suspicious power line targets from different angles along the route; the plant community algorithm continuously repeats sub-step 2-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the obtained optimal solution as the optimal local detailed inspection route for suspicious power line targets; The rechargeable power line monitoring drone conducts local detailed inspection and patrol along the route, and shoots suspicious power line targets along several shooting positions around the optimal solution until all shooting tasks of local detailed inspection of suspicious power line targets are completed, and then returns to the optimal global patrol route in sub-steps 1-3, and the rechargeable power line monitoring drone continues to conduct global patrol along this route; if a new suspicious power line target, or a suspicious obstacle, or insufficient remaining power is encountered during the local detailed inspection route, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 2, step 3, or step 4.

7. The method according to claim 1, characterized in that In step 3, the following sub-steps are included: Sub-step 3-1, the rechargeable power line monitoring drone continuously extracts obstacle features in the monitoring image during the global line patrol process and mines suspicious obstacles; the rechargeable power line monitoring drone starts the global line patrol or after the local detailed line patrol, continuously shoots the monitoring images of the power lines and stores them in the drone memory; the stored power line monitoring images are analyzed and processed in real time by the image processing algorithm; the drone stores a suspicious obstacle feature library; the drone determines whether it is possible to collide with the suspicious obstacle if it continues to fly. If there is no possibility of collision, it maintains the current flight attitude and continues to fly. If it is determined that continuing to fly may cause a collision; the obstacle features in the monitoring image are compared with the suspicious obstacle feature library through the image processing algorithm to mine whether there are image features that meet the suspicious obstacle feature library. If not, the drone continues to patrol according to the optimal global line patrol route in sub-steps 1-3; otherwise, the drone performs local obstacle avoidance on the suspicious obstacle according to sub-step 3-2; Sub-step 3-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local obstacle avoidance route for suspicious obstacles; In sub-step 3-2a, the plant community algorithm initializes parameters for the optimization problem of the local obstacle avoidance route of the suspicious obstacle; the plant community algorithm defines the parameters of the optimization problem of the local obstacle avoidance route of the suspicious obstacle of the rechargeable power line monitoring drone, including solving the target set, the suspicious obstacle parameters, the charging station parameters, and the drone parameters; the solving target set includes the suspicious obstacle set, the charging station set, the charging pile set of the charging station, the drone set, and the drone type set; Suspicious obstacle parameters, including the location and number of suspicious obstacles; charging station parameters, including the location and number of charging stations and charging piles; drone parameters, including the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly on a full charge, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone; In sub-step 3-2b, the plant community algorithm iteratively optimizes the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone; the plant individuals are encoded according to the local obstacle avoidance route optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible solution code; each bit of the plant individual code uses 1 to indicate that several obstacle avoidance positions around the corresponding suspicious obstacle are selected; each bit of the plant individual code uses 0 to indicate that several obstacle avoidance positions around the corresponding suspicious obstacle are not selected; the plant community algorithm uses the shortest total flight time as the feasible solution code of the rechargeable power line monitoring drone. The objective function of the local obstacle avoidance route optimization problem of suspected obstacles is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 3-2a to set the constraints. The local obstacle avoidance route optimization problem of suspected obstacles for rechargeable power line monitoring drones should not exceed the total number of suspected obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of drones, the maximum charging capacity of drones, and the maximum distance that drones can fly on a full charge; the plant community algorithm uses seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the local obstacle avoidance route optimization problem of suspected obstacles for rechargeable power line monitoring drones; In sub-step 3-3, the plant community algorithm outputs the optimal local obstacle avoidance route, and the rechargeable power line monitoring drone follows the route to avoid suspicious obstacles; The plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, which includes the calculation end time or the number of iteration calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the optimal solution obtained as the optimal local obstacle avoidance route for suspicious obstacles; the rechargeable power line monitoring drone performs local obstacle avoidance along the route, flies around and avoids obstacles along several obstacle avoidance positions around the suspicious obstacles determined by the optimal solution, until the suspicious obstacles are completely avoided, and then returns to the optimal global patrol route in sub-steps 1-3, and the rechargeable power line monitoring drone continues to perform global patrol along the route; if a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4.

8. The method according to claim 1, characterized in that: In step 4, the following sub-steps are included: Sub-step 4-1, the rechargeable power line monitoring drone continuously detects the remaining power during the global line patrol process; the rechargeable power line monitoring drone has a power detection sensor, which can obtain the latest value of the power and compare it with the remaining power warning threshold; The remaining power warning threshold is not fixed, but is estimated based on the current flight speed and power consumption rate of the drone, plus an estimated safety capacity; the remaining power warning threshold includes the power warning threshold for the drone to return home, that is, whether the remaining power can support the drone to return home smoothly; the remaining power warning threshold includes the power warning threshold for the drone to charge, that is, whether the remaining power can support the drone to fly smoothly to the nearest charging station; Sub-step 4-2, the rechargeable power line monitoring drone runs a plant community algorithm to optimize the local charging route based on the power line coordinates, obstacle coordinates, and charging station coordinates according to the remaining power; Sub-step 4-2a, the plant community algorithm initializes the parameters of the local charging route optimization problem; the plant community algorithm defines the parameters of the local charging route optimization problem of the rechargeable power line monitoring drone, including the solution target set, suspicious obstacle parameters, charging station parameters, and drone parameters; the solution target set includes a suspicious obstacle set, a charging station set, a charging pile set of the charging station, a drone set, and a drone type set; Suspicious obstacle parameters, including the location and number of suspicious obstacles; charging station parameters, including the location and number of charging stations and charging piles; drone parameters, including the maximum flight speed of the drone, the maximum charging capacity of the drone, the maximum distance the drone can fly on a full charge, the current remaining power of the drone, the unit flight cost of the drone, and the total flight cost of the drone; In sub-step 4-2b, the plant community algorithm iteratively optimizes the local charging route optimization problem of the rechargeable power line monitoring drone; the plant community algorithm selects the nearest available charging station as the target node of the local charging route, and encodes the plant individuals according to the local charging route optimization problem of the rechargeable power line monitoring drone, and one plant individual corresponds to one feasible solution code; each bit of the plant individual code uses 1 to indicate the spatial position passed during the flight; each bit of the plant individual code uses 0 to indicate that the corresponding spatial position is not selected during the flight; the plant community algorithm uses the shortest total flight time as the rechargeable power line The objective function of the local charging route optimization problem of the monitoring drone is used to evaluate the fitness of individual plants; the plant community algorithm uses the parameters in sub-step 4-2a to set the constraints. The local charging route optimization problem of the rechargeable power line monitoring drone should not exceed the total number of suspicious obstacles, the total number of charging stations, the total number of drones, the maximum flight speed of the drone, the maximum charging capacity of the drone, and the maximum distance that the drone can fly on a full charge; the plant community algorithm uses the seeding, growth, and result operations to select the plant individual with the highest fitness as the optimal solution to the local charging route optimization problem of the rechargeable power line monitoring drone; In sub-step 4-3, the plant community algorithm outputs the optimal local charging route, and the rechargeable power line monitoring drone goes to the charging station along the route to complete charging; the plant community algorithm continuously repeats sub-step 3-2 until the termination condition is met, and the termination condition includes the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; starting from the starting condition until the termination condition is met, the plant community algorithm uses the obtained optimal solution as the optimal local charging route; the rechargeable power line monitoring drone goes to the nearest available charging station along the route to charge until charging is completed, and then returns to the optimal global patrol route in sub-step 1-3, and the rechargeable power line monitoring drone continues to patrol the line globally along the route; If a new suspicious obstacle is encountered in the local obstacle avoidance route, or the remaining power is insufficient, it is necessary to use the plant community algorithm to locally adjust the patrol route, then go to step 3 or step 4; If the drone cannot reach the nearest available charging station before the battery runs out, the drone should use the GPS module it carries to report its latest location to the drone operator and send a battery depletion warning before the battery runs out; If the drone cannot reach the end point of the global power line patrol after being fully charged, it should use the GPS module it carries to report its latest position to the drone operator and send a warning that the remaining power cannot reach the end point; the drone should continuously search for the nearest available charging station, and if the nearest available charging station can be found, it returns to sub-step 4-2; if the drone cannot reach the end point of the global power line patrol after being fully charged, and cannot search for the nearest available charging station, it should use the GPS module it carries to report its latest position to the drone operator and send a warning that there is no nearby charging station.

Images