Substation unmanned aerial vehicle inspection route planning processing method and system based on laser point cloud

Abstract

The invention provides a transformer substation unmanned aerial vehicle inspection route planning processing method and system based on laser point cloud. The method comprises the steps of collecting point cloud data of a target transformer substation; according to the issued inspection task, selecting and matching corresponding to-be-inspected waypoints in historical existing routes, and generating a to-be-inspected waypoint list; on the 3D grid map, gradually searching the nearest grid point from the starting waypoint, planning a safe path between waypoints to be inspected of all the grid points by adopting an A * algorithm, and calculating a path weight coefficient between the selected waypoints; setting a starting waypoint and an ending waypoint based on the first safe path, optimizing the first safe path by adopting an improved traveling salesman problem algorithm, and calculating a shortest path; and inserting the route planned by the A * algorithm into the shortest path to generate a final path. According to the scheme, the problem of low efficiency of manual route planning is effectively solved, the repeated utilization rate of the route is improved, and the automation level of route planning is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) inspection technology, and in particular to a method and system for planning and processing flight routes for UAV inspections of substations based on laser point clouds. Background Technology

[0002] With rapid economic development, the demand for reliable power supply has increased significantly. As the core hub of power transmission, ensuring the stability and reliability of substation system equipment is crucial. Therefore, inspection tasks are not only indispensable but also increasingly important. Currently, drones, with their efficient inspection capabilities and comprehensive data collection advantages, have become a mainstream inspection method. However, due to the large number of devices in substations and the existence of many blind spots, manually operating drones can easily lead to collisions. Therefore, ensuring flight path safety is paramount.

[0003] Currently, substations commonly use laser point cloud technology. By manually selecting the points to be inspected and setting the shooting angle and distance, waypoints are generated to construct a safe inspection route. However, a significant amount of time is required to adjust the route before each inspection to ensure its safety, making the route planning work extremely demanding.

[0004] For example, consider two existing technical solutions: Chinese patent invention CN120255543A - A Dynamic Route Planning Method and System for Substation UAVs. The specific solution involves converting inspection requirements into a detailed set of point data. Through coordinate transformation, safety radius calculation, and shooting parameter settings, a unified set of safe inspection points is generated. Based on an improved ant colony algorithm, a path selection probability formula with safety constraints is introduced, combined with multi-ant colony parallel search and an adaptive pheromone update mechanism to generate an optimized shortest route. Subsequently, the optimized path point set is smoothed by using a third-order Bézier curve to fit the starting point, ending point, and intermediate inspection points, achieving smooth route connections and satisfying constraints on safe turning angle, path smoothness, and safety radius. This ensures a stable and smooth UAV flight path, ultimately outputting a temporary optimal route. This method also adjusts the waypoint order using an ant colony algorithm to achieve global path shortestization. However, when safety constraints are not met between any two waypoints, this method cannot plan a safe path.

[0005] Chinese patent invention CN120355049A - Method and Equipment for Generating UAV Inspection Routes. The method for generating UAV inspection routes includes the following steps: First, identifying the three-dimensional spatial model of the target substation and obtaining a list of key nodes; this list contains N key inspection points, representing equipment points in the target substation that need to be inspected. Second, obtaining a list of obstacles. Next, using the A* algorithm and a preset evaluation function, a first route is generated based on the UAV's start point, end point, the list of key nodes, and the list of obstacles; the preset evaluation function describes the total cost of each waypoint in the first route, including a dynamic weighting factor based on the actual distance between the key inspection point and the UAV and the corresponding minimum safe distance, to balance the total cost of the waypoints. Finally, the first route is optimized to obtain the target route. This method first uses the A* algorithm to generate the first route, then uses an evaluation function to describe the priority of each waypoint, determines the initial route, and optimizes it. Summary of the Invention

[0006] This invention proposes a method and system for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds. It can reduce the work of repeatedly marking inspection points by utilizing existing routes, ensure the shortest route path, and guarantee route safety, thereby effectively improving the efficiency and safety of substation route planning.

[0007] This invention provides a method and system for UAV flight path planning in substations based on 3D laser technology. This method generates a global flight path by collecting substation laser point cloud data in a single step and associating it with substation ledger information and safety waypoints, thereby reducing the tedious work of laser data acquisition and repetitive waypoint generation.

[0008] In a first aspect, the present invention provides a method for planning the flight path of a substation UAV based on three-dimensional laser, comprising:

[0009] S101. Scan the target substation using a 3D laser scanning device to collect point cloud data of the target substation;

[0010] S102. Based on the issued inspection task, select the corresponding waypoints to be inspected from the existing historical routes and generate a list of waypoints to be inspected.

[0011] S103. A 3D grid map is pre-constructed. On the 3D grid map, starting from the starting waypoint, the nearest grid point is searched step by step. The A* algorithm is used to plan the safe path between all grid points to be inspected waypoints, and the path weight coefficient (i.e., safe path length) between the selected waypoints is calculated. The total safe path distance formed by all the waypoints to be inspected is calculated by cumulative integration and recorded as the total path weight coefficient of the route. This is recorded as the first safe path.

[0012] S104. Based on the first safe path, set the start and end waypoints, optimize the first safe path using an improved Traveling Salesman Problem (TSP) algorithm, and calculate the shortest path;

[0013] S105. Insert the route planned by the A* algorithm into the shortest path to generate the final path.

[0014] Preferably, as a specific feasible implementation; the step of scanning the target substation using a three-dimensional laser scanning device to collect point cloud data of the target substation includes:

[0015] S1011. Scan the surrounding environment of the target area using a three-dimensional lidar to collect laser point cloud data around the target substation;

[0016] S102. Construct a KD tree point cloud from the acquired laser point cloud data, set a corresponding first safety distance, select and generate corresponding waypoints on the laser point cloud in the KD tree point cloud, to ensure that each selected waypoint meets the condition that the shortest distance from the waypoint to the laser point cloud is greater than the set first safety distance, and save the current route to form a ledger to be inspected.

[0017] Preferably, as a specific feasible implementation, after saving the current route, it also includes:

[0018] Select the appropriate parameter value for the first safe distance based on the different mission types, and then generate the flight path corresponding to the specific mission type based on the first safe distance.

[0019] Preferably, as a specific feasible implementation; the step of selecting and matching corresponding waypoints to be inspected from existing historical routes based on the issued inspection task, and generating a list of waypoints to be inspected, includes:

[0020] S1021. First, obtain the issued inspection tasks in real time, determine the task type according to the issued inspection tasks, and select the ledger to be inspected according to the task type.

[0021] S1022. Search for the waypoints to be inspected in the existing historical routes that correspond to the ledger to be inspected, and generate a list of waypoints to be inspected.

[0022] Preferably, as a specific feasible implementation; the step of selecting and matching corresponding waypoints to be inspected from existing historical routes based on the issued inspection task, and generating a list of waypoints to be inspected, includes:

[0023] The A* algorithm is used to plan safe paths between waypoints to be inspected, and the path weight coefficients (i.e., safe path lengths) between selected waypoints are calculated.

[0024] Preferably, as a specific feasible implementation; the step of setting start and end waypoints based on the first safe path, optimizing the first safe path using an improved Traveling Salesman Problem (TSP) algorithm, and calculating the shortest path includes;

[0025] S1041. First, set the start and end waypoints to determine the waypoints to be inspected;

[0026] S1042. Then, the TSP algorithm is used to optimize and adjust the order of each waypoint to be inspected to form the shortest path.

[0027] Preferably, as a specific feasible implementation, the TSP algorithm includes the ant colony algorithm.

[0028] The calculation formula for the ant colony algorithm is as follows:

[0029] ;

[0030] This formula is the probability transition formula between two waypoints, representing the probability that ant k moves from waypoint i to waypoint j at time t; where, It is the reciprocal of the path weight coefficient in S103, which means that the smaller the distance, the greater the transition probability; and These represent the relative importance weights of pheromones and expected heuristic factors, respectively. Representing a path The pheromone values ​​on each path are initially set to 0. After one traversal is completed, the pheromone values ​​of each path will be updated.

[0031] Waypoints i and j are both waypoints; This represents the set of possible cities that ant k can visit at time t. Ants that have already visited a city will leave this set.

[0032] Preferably, as a specific feasible implementation; Representing a path The pheromone values ​​on each path are updated using the following formula:

[0033] ;

[0034] In the formula: The pheromone evaporation rate along the path, Indicates the path in this iteration The pheromone increment, i.e.:

[0035] ;

[0036] In the formula: The k-th ant left on the path in this iteration The amount of information on the data is 0 if it has not been accessed.

[0037] ;

[0038] In the formula: Q is a constant, Let be the length of the path traversed by the k-th ant in this operation.

[0039] Preferably, as a specific feasible implementation, before constructing the KD-tree point cloud, the method further includes: downsampling the KD-tree point cloud using voxel filtering, wherein the downsampling size is 0.03 meters.

[0040] This invention provides a substation UAV inspection route planning and processing system based on laser point clouds, including a data acquisition module, a matching module, an initial path processing module, a shortest path processing module, and an insertion module.

[0041] The acquisition module is used to scan the target substation using a 3D laser scanning device and acquire point cloud data of the target substation.

[0042] The matching module is used to select and match the corresponding waypoints to be inspected from the existing historical routes based on the issued inspection tasks, and generate a list of waypoints to be inspected.

[0043] The initial path processing module is used to pre-build a 3D grid map. On the 3D grid map, starting from the starting waypoint, it searches for the nearest grid point step by step. The A* algorithm is used to plan the safe path between all grid points to be inspected waypoints and calculate the path weight coefficient (i.e., the safe path length) between the selected waypoints. The total distance of the safe path formed by all the waypoints to be inspected is calculated by cumulative integration and recorded as the total path weight coefficient of the route. This is recorded as the first safe path.

[0044] The shortest path processing module is used to set the start and end waypoints based on the first safe path, optimize the first safe path using an improved Traveling Salesman Problem (TSP) algorithm, and calculate the shortest path.

[0045] The insertion module is used to insert the route planned by the A* algorithm into the shortest path to generate the final path.

[0046] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, enables a method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds.

[0047] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0048] This invention provides a method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds. The method includes: scanning the target substation using a 3D laser scanning device to collect point cloud data; selecting matching waypoints from existing historical routes according to the issued inspection task and generating a list of waypoints to be inspected; pre-constructing a 3D grid map; progressively searching for the nearest grid points starting from the initial waypoint on the 3D grid map; using the A* algorithm to plan safe paths between all the waypoints to be inspected and calculating the path weight coefficients (i.e., safe path lengths) between the selected waypoints; calculating the total safe path distance formed by all the waypoints to be inspected using cumulative integration, recording this as the total path weight coefficient of the route, and recording this as the first safe path; setting the starting and ending waypoints based on the first safe path; optimizing the first safe path using an improved Traveling Salesman Problem (TSP) algorithm to calculate the shortest path; and inserting the route planned by the A* algorithm into the shortest path to generate the final path.

[0049] The above method generates a global ledger route by collecting substation laser point cloud data in a single step and associating it with substation ledger information and safety waypoints, thereby reducing the tedious work of laser data collection and repeated waypoint generation. Simultaneously, it acquires issued inspection tasks, determines the task type based on the task, and selects the ledger to be inspected according to the task type. Subsequently, the system automatically searches for the corresponding waypoints in the existing routes that match the ledgers to be inspected, generating a list of waypoints to be inspected. This allows the system to automatically generate corresponding inspection routes when subsequent inspection tasks are issued, based on the ledger information. This invention reuses existing routes, avoiding the preliminary work of regenerating waypoints for multiple inspections. It also uses an ant colony algorithm to automatically adjust the order of waypoints in the route, ensuring the shortest route and guaranteeing route safety, without requiring post-processing adjustments to the generated route. The above technical solution effectively solves the problem of low efficiency in manual route planning, improves the reuse rate of routes, and significantly enhances the automation level of route planning. Attached Figure Description

[0050] Figure 1 A main flowchart of a substation UAV inspection route planning and processing method based on laser point cloud provided in an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of a specific implementation process of a substation UAV inspection route planning and processing method based on laser point cloud provided by an embodiment of the present invention;

[0052] Figure 3This is a schematic diagram of another specific implementation process of a substation UAV inspection route planning and processing method based on laser point cloud provided by an embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram of a substation UAV inspection route planning and processing system based on laser point cloud, provided as an embodiment of the present invention.

[0054] Labels: Acquisition module 10, Matching module 20, Initial path processing module 30, Shortest path processing module 40, Insertion module 50. Detailed Implementation

[0055] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the relevant invention and are not intended to limit the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0056] like Figure 1 As shown in the figure, this embodiment presents a method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds, which specifically includes the following steps:

[0057] S101. Scan the target substation using a 3D laser scanning device to collect point cloud data of the target substation.

[0058] Preferred options, see details. Figure 2 The above step S101 further includes:

[0059] S1011. Scan the surrounding environment of the target area using a three-dimensional lidar to collect laser point cloud data around the target substation;

[0060] S1012. Construct a KD tree from the acquired laser point cloud data to generate a safety detection module for analyzing the safe distance to waypoints.

[0061] S102. Based on the issued inspection task, select and match the corresponding waypoints to be inspected from the existing historical routes. When an inspection task is issued, the system can select the corresponding waypoints based on the ledger information and automatically generate the corresponding inspection route. This invention can reuse existing routes, avoiding the preliminary preparation work of regenerating waypoints during multiple inspections. It can also automatically adjust the order of waypoints in the route to ensure the shortest route path and guarantee route safety, without requiring post-processing adjustments to the generated route. This innovation effectively solves the problem of low efficiency in manual route planning, improves the reuse rate of routes, and significantly enhances the automation level of route planning.

[0062] The explanation explains that, based on the substation ledger or photo point name, a corresponding first safety distance is set. Corresponding waypoints are then selected and generated on the laser point cloud, ensuring that each waypoint meets safety standards (i.e., the nearest distance to the laser point cloud is greater than the set first safety distance), and the current flight path is saved. In subsequent operations, appropriate waypoints can be selected according to different task types, and the system will automatically generate specific flight paths based on the selected options. It is important to note that a KD-tree is a tree-shaped data structure used to organize multidimensional spatial data, widely applied for range queries and nearest neighbor searches in spatial data.

[0063] It should be noted that before constructing the KD tree, voxel filtering is used for downsampling to reduce the number of point clouds and avoid memory overflow. The downsampling size is 0.03 meters. The actual safe distance is set as an input field; when detection is needed, the flight path or waypoint and the safe distance can be used to calculate whether it is safe. Since the KD tree point cloud has already been thinned, the first safe distance can be set slightly larger (greater than the actual safe distance by 0.03 meters) to ensure absolute safety.

[0064] Specifically, first, the system obtains the issued inspection task, determines the task type based on the issued inspection task, and selects the ledger to be inspected based on the task type; then, the system will automatically search for the waypoints to be inspected corresponding to the ledgers to be inspected in the existing routes, and generate a list of waypoints to be inspected.

[0065] S103. Use the A* algorithm to plan safe paths between all waypoints (i.e., waypoints to be inspected) and calculate the path weight coefficients (i.e., safe path lengths) between the selected waypoints.

[0066] Specifically, the path weight coefficient between any two waypoints is determined by the safe path distance. If the distance between the line connecting the two waypoints and the nearest obstacle point cloud is greater than a set safe distance, the route is considered safe, and the weight coefficient is the distance between the two waypoints. Otherwise, it is considered unsafe, and the A* algorithm is used to calculate the safe path between the two points, simplifying the path nodes. The total distance of the generated safe path is the path weight coefficient for this route, and this safe path is recorded for generating the final route. This method ensures that a safe route exists between any two waypoints.

[0067] In detail, the A* algorithm constructs a 3D mesh map using laser point clouds, starting from the initial waypoint and using a heuristic function. Calculate the cost of reaching surrounding 3D grid points. Based on the principle of cost minimization, progressively search for the nearest grid points until the final waypoint is reached. During this process, Let n be the total cost of node n. It is the actual cost from node n to the starting waypoint. It is the estimated cost from node n to the final waypoint.

[0068] It should be noted that an improved TSP algorithm will be used to adjust the waypoint order, which requires calculating path weight coefficients between waypoints. Simultaneously, the A* algorithm will be used to calculate safe paths, ensuring that a safe path exists between every two waypoints. Finally, when connecting all waypoints, the safety of the final path will be guaranteed.

[0069] S104. Set the starting and ending waypoints, and use the improved Traveling Salesman Problem (TSP) algorithm to calculate the shortest path.

[0070] Specifically, the process begins by setting start and end waypoints to determine the waypoints to be inspected. Then, the TSP algorithm is used to optimize the order of these waypoints to form the shortest path. The core of the TSP algorithm is to solve the problem of finding the shortest path from the start point, passing through all other waypoints, and finally reaching the end point. Ant colony optimization, as an intelligent optimization algorithm, can be effectively applied to solving the TSP problem.

[0071] The core formula of the ant colony algorithm:

[0072] ;

[0073] This formula represents the probability transition between two waypoints, indicating the probability that ant k will move from waypoint i to waypoint j at time t. Wherein, It is the reciprocal of the path weight coefficient in S103, which means that the smaller the distance, the greater the transition probability. and These represent the relative importance of pheromones and the expected heuristic factor, respectively. Representing a path The pheromone values ​​on each path are initially set to 0. After one traversal is completed, the pheromone values ​​of each path will be updated.

[0074] ;

[0075] In the formula: The pheromone evaporation rate along the path, Indicates the path in this iteration The pheromone increment, i.e.:

[0076] ;

[0077] In the formula: The k-th ant left on the path in this iteration The amount of information on the data is 0 if it has not been accessed.

[0078] ;

[0079] In the formula: Q is a constant, Let be the length of the path traversed by the k-th ant in this operation.

[0080] For detailed implementation steps of the ant colony algorithm, please refer to [link / reference]. Figure 3 ;

[0081] about Figure 3 The specific steps for implementing the ant colony algorithm are as follows:

[0082] Step 1: Implement initialization parameters;

[0083] Step 2: Place ants at the initial waypoint;

[0084] Step 3: Calculate the heuristic value based on the path weight coefficient;

[0085] Step 3: Path information concentration probability;

[0086] Step 4: Select the next waypoint according to the roulette method;

[0087] Step 5: Determine if all waypoints have been traversed;

[0088] If yes, then the 2-opt method performs local path optimization; otherwise, return to step 2.

[0089] Step 6: Determine if the ant colony traversal is complete;

[0090] If yes, proceed to step 7 to update the pheromone concentration for each path; otherwise, proceed to step 8.

[0091] Step 8: Determine if the maximum number of iterations has been reached.

[0092] If so, output the optimal path.

[0093] It should be noted that this embodiment of the invention combines the ant colony algorithm and the 2-opt algorithm to solve the TSP problem. First, key parameters such as the number of ants, pheromone evaporation rate, and initial pheromone intensity are set. Each ant starts from the initial waypoint and calculates the transition probability density function of the remaining waypoints. Using a roulette wheel algorithm, a random number between 0 and 1 is generated. When the probability density of a remaining waypoint is greater than this random number, that waypoint is selected as the next target, until the last waypoint is reached. After traversing all waypoints, the 2-opt method is used to locally optimize the traversal order. Then, the pheromone value of the current path is updated, and this process is repeated until all ants have completed the calculation and the maximum number of traversals has been reached, thus generating the optimal path. The ant colony algorithm ensures the optimization of the global path, while the 2-opt algorithm safeguards the optimization of the local path, ultimately achieving overall path optimization.

[0094] Regarding the number of iterations: the number of iterations and the number of ant colonies are determined by the number of waypoints. When the number of waypoints is less than 100, the number of iterations is fixed at 250; when it exceeds 100, the formula iter=(wayPtNum-100)*0.5 is used for calculation, but the minimum number of iterations must not be less than 10. The number of ant colonies is set to 1.5 times the number of waypoints and needs to be controlled between 30 and 500. This configuration ensures both computational speed and optimizes algorithm results.

[0095] It should be noted that the 2-opt algorithm sequentially swaps every two waypoints between the start and end waypoints. If the distance of the new path is less than the original path, the new path is updated. Then, it starts swapping waypoints again from the start waypoint until no better path can be found. Verification has shown that the 2-opt algorithm can effectively improve the rationality of planned routes.

[0096] S105. Insert the route planned by the A* algorithm into the shortest path to generate the final path.

[0097] Specifically, after generating the optimal path between existing waypoints, if there is a route planned by the A* algorithm between the current waypoints, it needs to be inserted into the current shortest route path.

[0098] This invention provides a method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds. The method includes: scanning the target substation using a 3D laser scanning device to collect point cloud data; selecting matching waypoints from existing historical routes according to the issued inspection task and generating a list of waypoints to be inspected; pre-constructing a 3D grid map; progressively searching for the nearest grid points starting from the initial waypoint on the 3D grid map; using the A* algorithm to plan safe paths between all the waypoints to be inspected and calculating the path weight coefficients (i.e., safe path lengths) between the selected waypoints; calculating the total safe path distance formed by all the waypoints to be inspected using cumulative integration, recording this as the total path weight coefficient of the route, and recording this as the first safe path; setting the starting and ending waypoints based on the first safe path; optimizing the first safe path using an improved Traveling Salesman Problem (TSP) algorithm to calculate the shortest path; and inserting the route planned by the A* algorithm into the shortest path to generate the final path.

[0099] The above method generates a global ledger route by collecting substation laser point cloud data in a single step and associating it with substation ledger information and safety waypoints, thereby reducing the tedious work of laser data collection and repeated waypoint generation. Simultaneously, it acquires issued inspection tasks, determines the task type based on the task, and selects the ledger to be inspected according to the task type. Subsequently, the system automatically searches for the corresponding waypoints in the existing routes that match the ledgers to be inspected, generating a list of waypoints to be inspected. This allows the system to automatically generate corresponding inspection routes when subsequent inspection tasks are issued, based on the ledger information. This invention reuses existing routes, avoiding the preliminary work of regenerating waypoints for multiple inspections. It also uses an ant colony algorithm to automatically adjust the order of waypoints in the route, ensuring the shortest route and guaranteeing route safety, without requiring post-processing adjustments to the generated route. The above technical solution effectively solves the problem of low efficiency in manual route planning, improves the reuse rate of routes, and significantly enhances the automation level of route planning.

[0100] Compared to existing technologies, this invention ensures the generation of safe paths by pre-planning safe routes between every two waypoints. Furthermore, it employs the 2-opt method to optimize local paths, resulting in a more reasonable final route. The route priority problem in this invention is solved using an ant colony algorithm, which differs from the aforementioned methods.

[0101] Example 2

[0102] like Figure 4 As shown, this invention provides a substation UAV inspection route planning and processing system based on laser point clouds, including a data acquisition module 10, a matching module 20, an initial path processing module 30, a shortest path processing module 40, and an insertion module 50.

[0103] The acquisition module 10 is used to scan the target substation using a three-dimensional laser scanning device and acquire point cloud data of the target substation.

[0104] The matching module 20 is used to select and match the corresponding waypoints to be inspected from the existing historical routes according to the issued inspection task, and generate a list of waypoints to be inspected.

[0105] The initial path processing module 30 is used to pre-build a 3D grid map. On the 3D grid map, starting from the starting waypoint, it searches for the nearest grid point step by step. It uses the A* algorithm to plan the safe path between all grid points to be inspected waypoints and calculates the path weight coefficient (i.e., the safe path length) between the selected waypoints. It calculates the total safe path distance formed by all the waypoints to be inspected by the cumulative integration method and records it as the total path weight coefficient of the route. This is recorded as the first safe path.

[0106] The shortest path processing module 40 is used to set the start and end waypoints based on the first safe path, optimize the first safe path using an improved Traveling Salesman Problem (TSP) algorithm, and calculate the shortest path.

[0107] The insertion module 50 is used to insert the route planned by the A* algorithm into the shortest path to generate the final path.

[0108] Embodiment 2 of the present invention provides a substation UAV inspection route planning and processing system based on laser point cloud; the system includes: a data acquisition module 10, a matching module 20, an initial path processing module 30, a shortest path processing module 40, and an insertion module 50.

[0109] The aforementioned substation UAV inspection route planning and processing system based on laser point clouds collects substation laser point cloud data in a single step and associates it with substation ledger information and safety waypoints to generate a global ledger route, thereby reducing the tedious work of laser data collection and repetitive waypoint generation. Simultaneously, it acquires issued inspection tasks, determines the task type based on the task, and selects the ledger to be inspected according to the task type. Subsequently, the system automatically searches the existing routes for the corresponding waypoints to be inspected for each ledger, generating a list of waypoints to be inspected. Thus, when subsequent inspection tasks are issued, the system can select the corresponding waypoints based on the ledger information and automatically generate the corresponding inspection route.

[0110] This invention enables the reuse of existing routes, avoiding the preliminary preparation work of regenerating waypoints during multiple inspections. Simultaneously, it employs an ant colony algorithm to automatically adjust the order of waypoints within the route, ensuring the shortest path and guaranteeing route safety, without requiring post-processing adjustments to the generated route. The above technical solution effectively solves the problem of low efficiency in manual route planning, improves the reuse rate of routes, and significantly enhances the automation level of route planning.

[0111] Example 3

[0112] Embodiment 3 of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can realize a method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds.

[0113] This invention provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps of any of the point cloud-based pole-shaped tower deflection calculation methods described above. In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0114] The present invention provides a method and system for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds. This system collects laser point cloud data from the substation in a single operation and marks it using waypoints from a logbook, generating route information. In subsequent inspection tasks, the generated route can be reused without requiring data collection or repeated waypoint marking. Furthermore, by integrating multiple algorithms, the system ensures that the generated final route is not only the shortest but also possesses extremely high safety, thereby reducing the workload of manually adjusting routes to ensure safety.

[0115] The foregoing has provided a detailed description of the method, system, and storage medium for substation UAV inspection route planning and processing based on laser point clouds provided by this invention. Specific examples have been used in the embodiments of this invention to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and core ideas of this invention. It should be noted that those skilled in the art can make various improvements and modifications to this invention without departing from the principles of this invention, and these improvements and modifications also fall within the protection scope of this invention.

Claims

1. A method for planning and processing unmanned aerial vehicle (UAV) inspection routes for substations based on laser point clouds, characterized in that, The method includes: S101. Scan the target substation using a 3D laser scanning device to collect point cloud data of the target substation; S102. Based on the issued inspection task, select the corresponding waypoints to be inspected from the existing historical routes and generate a list of waypoints to be inspected. S103. A 3D grid map is pre-constructed. On the 3D grid map, starting from the starting waypoint, the nearest grid point is searched step by step. The A* algorithm is used to plan the safe path between all grid points to be inspected waypoints, and the path weight coefficient between the selected waypoints is calculated. The total distance of the safe path formed by all the waypoints to be inspected is calculated by cumulative integration and recorded as the total path weight coefficient of the route. This is recorded as the first safe path. S104. Based on the first safe path, set the start and end waypoints, optimize the first safe path using an improved traveling salesman problem algorithm, and calculate the shortest path; S105. Insert the route planned by the A* algorithm into the shortest path to generate the final path.

2. The method according to claim 1, characterized in that, The process of scanning the target substation using a 3D laser scanning device to collect point cloud data of the target substation includes: S1011. Scan the surrounding environment of the target area using a three-dimensional lidar to collect laser point cloud data around the target substation; S1012. Construct a KD tree point cloud from the acquired laser point cloud data, set a corresponding first safety distance, select and generate corresponding waypoints on the laser point cloud in the KD tree point cloud, to ensure that each selected waypoint meets the condition that the shortest distance from the waypoint to the laser point cloud is greater than the set first safety distance, and save the current route to form a ledger to be inspected.

3. The method according to claim 2, characterized in that, After saving the current route, it also includes: Select the appropriate parameter value for the first safe distance based on the different mission types, and then generate the flight path corresponding to the specific mission type based on the first safe distance.

4. The method according to claim 1, characterized in that, The process involves selecting and matching corresponding waypoints to be inspected from existing historical routes based on the issued inspection tasks, and generating a list of waypoints to be inspected, including: S1021. First, obtain the issued inspection tasks in real time, determine the task type according to the issued inspection tasks, and select the ledger to be inspected according to the task type. S1022. Search for the waypoints to be inspected in the existing historical routes that correspond to the ledger to be inspected, and generate a list of waypoints to be inspected from the waypoints to be inspected.

5. The method according to claim 1, characterized in that, The step of setting start and end waypoints based on the first safe path, optimizing the first safe path using an improved traveling salesman problem algorithm, and calculating the shortest path includes: S1041. First, set the start and end waypoints to determine the waypoints to be inspected; S1042. Then, the TSP algorithm is used to optimize and adjust the order of each waypoint to be inspected to form the shortest path.

6. The method according to claim 5, characterized in that, The TSP algorithm includes the ant colony algorithm; The calculation formula for the ant colony algorithm is as follows: ； This formula is the probability transition formula between two waypoints, representing the probability that ant k moves from waypoint i to waypoint j at time t; where, It is the reciprocal of the path weight coefficient in S103, which means that the smaller the distance, the greater the transition probability; and These represent the relative importance weights of pheromones and expected heuristic factors, respectively. Representing a path The pheromone values ​​on each path are initially 0. After one traversal is completed, the pheromone values ​​of each path will be updated. Waypoints i and j are both waypoints; It represents the set of possible destinations for ant k at time t.

7. The method according to claim 6, characterized in that, Representing a path The pheromone values ​​on each path are updated using the following formula: ； In the formula: The pheromone evaporation rate along the path, Indicates the path in this iteration The pheromone increment, i.e.: ； In the formula: The k-th ant left on the path in this iteration The amount of information on the data is 0 if it has not been accessed. ； In the formula: Q is a constant, Let be the length of the path traversed by the k-th ant in this operation.

8. The method according to claim 1, characterized in that, Before constructing the KD tree point cloud, the method further includes downsampling the KD tree point cloud using voxel filtering, wherein the downsampling size is 0.03 meters.

9. A substation UAV inspection route planning and processing system based on laser point clouds, characterized in that, It includes a data acquisition module, a matching module, an initial path processing module, a shortest path processing module, and an insertion module. The acquisition module is used to scan the target substation using a 3D laser scanning device and acquire point cloud data of the target substation. The matching module is used to select and match the corresponding waypoints to be inspected from the existing historical routes based on the issued inspection tasks, and generate a list of waypoints to be inspected. The initial path processing module is used to pre-build a 3D grid map. On the 3D grid map, starting from the starting waypoint, it searches for the nearest grid point step by step. The A* algorithm is used to plan the safe path between all grid points to be inspected waypoints and calculate the path weight coefficient between the selected waypoints. The total distance of the safe path formed by all the waypoints to be inspected is calculated by cumulative integration and recorded as the total path weight coefficient of the route. This is recorded as the first safe path. The shortest path processing module is used to set the start and end waypoints based on the first safe path, optimize the first safe path using an improved Traveling Salesman Problem (TSP) algorithm, and calculate the shortest path. The insertion module is used to insert the route planned by the A* algorithm into the shortest path to generate the final path.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the substation UAV inspection route planning and processing method based on laser point clouds as described in any one of claims 1 to 8.

Images