Unmanned aerial vehicle lightweight map representation method and system based on spherical convex hull

Abstract

The invention discloses an unmanned aerial vehicle lightweight map characterization method and system based on a spherical convex hull. The method comprises the steps of performing spatial sampling on a point cloud map; searching and calculating the nearest point cloud distance of the sampling points, if the nearest point cloud distance of the sampling points is smaller than a distance set value, deleting the sampling points, and repeatedly sampling until a preset specified number of sampling points are obtained; traversing all sampling points to calculate the minimum distance to the nearest point cloud; generating a spherical convex hull by taking the sampling point with the maximum minimum distance as a sphere center and the minimum distance as a radius after sequencing; deleting sampling points in the spherical convex hull; and the steps are repeated until the filling rates of all the spherical convex hulls are larger than a preset threshold value. The method aims at achieving light-weight map representation, optimizing calculation efficiency, enhancing dynamic adaptability and achieving real-time updating while ensuring data precision so as to improve the navigation capability and safety of the unmanned aerial vehicle in the power industry in complex environments such as cable tunnels, and the method is particularly suitable for map representation in a large space.

Description

Technical Field

[0001] The present invention relates to a complex environment inspection technology for a micro-UAV in the electric power field, and in particular to a lightweight map representation method and system for an UAV based on a spherical convex hull. Background Art

[0002] The use of drones is rapidly growing in the power industry, becoming a crucial tool for improving the management and maintenance efficiency of power facilities. Due to their efficiency, flexibility, and cost-effectiveness, drones are widely used in power line inspections, equipment monitoring, and troubleshooting. Equipped with high-precision cameras and sensors, drones can quickly capture real-time data over a wide area, helping operators identify problems and optimize maintenance plans, thereby improving the safety and reliability of power systems.

[0003] Drone navigation in denied environments, such as cable tunnels, presents unique challenges. These environments typically feature complex spatial structures and a variety of potential obstacles, requiring drones to possess efficient obstacle avoidance capabilities and accurate path planning. The complexity and uncertainty of denied environments necessitate drones to identify and adapt to these changes in real time to ensure safe flight. Efficient path planning and real-time data processing are key to successful drone operation in these environments. Accurate map representation enables drones to adjust their paths in dynamic environments, ensuring the safe and smooth completion of inspection missions. Therefore, the value of drone applications in the power industry lies in improving inspection efficiency, reducing labor costs, and enhancing safety. Precise environmental navigation capabilities provide a crucial guarantee for their application in complex scenarios.

[0004] Environmental map representation is a prerequisite for UAV navigation. It determines traversable and inaccessible areas. Only with accurate map representation can a UAV perform obstacle avoidance. Current UAV map representation methods typically use sensors to generate raw point cloud maps, which are then converted into occupancy grid maps or octree maps. While environmental map representation is essential for successful mission execution in UAV navigation, it currently faces several challenges. First, the point cloud data generated by sensors can contain noise and errors due to environmental factors and sensor accuracy, which can affect the determination of traversable areas. Second, the process of converting point cloud maps into occupancy grid maps or octree maps is computationally complex, especially at high resolutions, which can affect real-time performance. Furthermore, traditional static map representations lack adaptability in dynamic environments and struggle to reflect changes in obstacles, thereby reducing the UAV's obstacle avoidance capabilities. Furthermore, while increasing map resolution improves accuracy, it also leads to a significant increase in data volume, limiting storage and processing capabilities. Finally, existing methods generally lack effective real-time update mechanisms, making them unable to respond promptly to environmental changes. To solve these problems, in-depth research is needed to improve data accuracy, optimize computing efficiency, enhance dynamic adaptability, and achieve real-time updates, so as to enhance the navigation capabilities and safety of drones in complex environments. In drone navigation, traditional occupancy grid maps identify obstacles by discretizing space into several grids. The principle of occupancy grid maps is to discretize space into several grids according to resolution, setting the grid that is not an obstacle to "0" and the grid that is an obstacle to "1". The amount of data in the occupancy grid map increases with the inverse of the resolution. In a cubic space, let the data amount be m and the resolution be r. Their relationship is: , As can be seen, as the resolution increases, the amount of data increases at a cubic rate, consuming a large amount of memory in larger scenes. On small robots like drones, there isn't much memory left to allocate to map representation, so this method can only be used to represent maps in small spaces. Summary of the Invention

[0005] Technical problem to be solved by the present invention: In response to the above-mentioned problems of the prior art, a lightweight map representation method and system for UAVs based on spherical convex hull are provided. The present invention aims to achieve lightweight map representation, optimize computing efficiency, enhance dynamic adaptability and realize real-time updates while ensuring data accuracy, so as to improve the navigation capability and safety of UAVs in complex environments such as cable tunnels in the power industry, and is particularly suitable for map representation in large spaces.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: A lightweight map representation method for unmanned aerial vehicles based on spherical convex hull includes the following steps: S1, obtain the point cloud map collected by the drone; S2, uniform spatial sampling in the space of the point cloud map; S3, performing a nearest neighbor search on the sampled sampling points to find the nearest point cloud distance of the sampling points. If the nearest point cloud distance of the sampling point is less than the set distance value, the sampling point is deleted and the process goes to step S2 until a preset number of sampling points is obtained; S4, traverse all sampling points and calculate the minimum distance to the nearest point cloud respectively; S5, sorting the sampling points by minimum distance, taking the sampling point with the largest minimum distance as the center of the sphere, and generating a spherical convex hull with the minimum distance as the radius; S6, delete all sampling points within the spherical convex hull; S7, calculate the filling rate of all spherical convex hulls. If the filling rate is greater than or equal to a preset threshold, the map representation is determined to be complete, and the drone is controlled to fly within the spherical convex hull to avoid obstacles; otherwise, jump to step S4.

[0007] Optionally, before step S3, the method further includes determining the size of a preset specified number according to the size of the scene inspected by the drone, and the size of the preset specified number is positively correlated with the size of the scene inspected by the drone.

[0008] Optionally, before step S3, the method further includes determining the size of the distance setting value according to the size of the scene inspected by the drone, and the size of the distance setting value is positively correlated with the size of the scene inspected by the drone.

[0009] Optionally, performing a nearest neighbor search on the sampled sampling points in step S3 to find the nearest point cloud distance of the sampling points includes: performing a nearest neighbor search on the sampled sampling points in the point cloud map to find a neighbor point set consisting of all neighbor points within a preset distance range, calculating the distance between each neighbor point in the neighbor point set and the sampled sampling point, and taking the distance with the smallest distance to the sampled sampling point as the nearest point cloud distance of the sampling point.

[0010] Optionally, calculating the minimum distance to the nearest point cloud in step S4 includes: performing a nearest neighbor search on each of the traversed sampling points to find a neighbor sampling point set consisting of all neighbor sampling points within a preset distance range, calculating the distance between each neighbor sampling point in the neighbor sampling point set and the traversed sampling point, and taking the distance with the minimum distance to the traversed sampling point as the minimum distance to the nearest point cloud of the traversed sampling point.

[0011] Optionally, in step S5, the sampling points are sorted by minimum distance, and the sampling point with the largest minimum distance is used as the center of the sphere. When a spherical convex hull is generated with the minimum distance as the radius, if the distance between the center of the sphere and the ground is greater than or equal to the radius, the generated spherical convex hull is a sphere; if the distance between the center of the sphere and the ground is less than the radius, the generated spherical convex hull is the spatial portion of the sphere located above the ground.

[0012] Optionally, calculating the filling rates of all spherical convex hulls in step S7 includes: calculating the volumes of all spherical convex hulls separately, adding the volumes of all spherical convex hulls to obtain the total volume of the spherical convex hulls, and then dividing the total volume of the convex hulls by the size of the obstacle-free space detected by the drone, thereby obtaining the filling rates of all spherical convex hulls.

[0013] In addition, the present invention also provides a UAV lightweight map representation system based on spherical convex hull, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the UAV lightweight map representation method based on spherical convex hull.

[0014] In addition, the present invention also provides a computer-readable storage medium, which stores a computer program or instruction, and the computer program or instruction is programmed or configured to execute the spherical convex hull-based UAV lightweight map representation method through a processor.

[0015] In addition, the present invention also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the spherical convex hull-based UAV lightweight map representation method through a processor.

[0016] Compared with the existing technology, the present invention can mainly achieve the following beneficial effects: The core of the present invention is to use multiple spherical convex hulls to fill the obstacle-free space to ensure the safety of the drone when flying within these spheres. The innovative points of the present invention's lightweight map representation method for drones based on spherical convex hulls include: Uniform sampling: Uniformly sampling the candidate points for the center of the spherical convex hull to generate the spherical convex hull within the entire map range, regardless of whether these points are within obstacles. Sorting processing: The center candidate points are sorted from large to small according to the distance to the nearest point cloud to select the most suitable center. Deleting redundant candidate points: When generating the spherical convex hull, all remaining center candidate points within it are deleted to optimize the map representation. Valid distance restriction: The distance between the valid sampling point and the nearest point cloud must be greater than a certain threshold, otherwise re-sampling is required to ensure the safe flight of the drone. Through the combination of these methods, the present invention achieves efficient space filling and collision avoidance, providing reliable technical support for the safe navigation of drones in complex environments, and is particularly suitable for map representation in large spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 Schematic diagram of the basic process of the method of the embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be further described in detail below with reference to the accompanying drawings in the embodiments of the present invention.

[0019] like Figure 1 As shown, the lightweight map representation method for UAV based on spherical convex hull in this embodiment includes the following steps: S1, obtain the point cloud map collected by the drone; S2, uniform spatial sampling in the space of the point cloud map; S3, performing a nearest neighbor search on the sampled sampling points to find the nearest point cloud distance of the sampling points. If the nearest point cloud distance of the sampling point is less than the set distance value, the sampling point is deleted and the process goes to step S2 until a preset number of sampling points is obtained; S4, traverse all sampling points and calculate the minimum distance to the nearest point cloud respectively; S5, sorting the sampling points by minimum distance, taking the sampling point with the largest minimum distance as the center of the sphere, and generating a spherical convex hull with the minimum distance as the radius; S6, delete all sampling points within the spherical convex hull; S7, calculate the filling rate of all spherical convex hulls. If the filling rate is greater than or equal to a preset threshold, the map representation is determined to be complete, and the drone is controlled to fly within the spherical convex hull to avoid obstacles; otherwise, jump to step S4.

[0020] When step S1 obtains the point cloud map collected by the drone, the point cloud map is collected in the area where the drone needs to fly, with the purpose of obtaining complete obstacle information and describing the infeasible area of ​​the entire space.

[0021] Step S2 performs uniform spatial sampling in the space of the point cloud map. The uniform spatial sampling can improve the accuracy of obstacle information collection.

[0022] Before step S3, the method also includes determining the size of a preset specified number according to the size of the scene inspected by the drone, and the size of the preset specified number is positively correlated with the size of the scene inspected by the drone, and the number of samples changes with the size of the scene.

[0023] Prior to step S3, the process also includes determining a distance setting value based on the size of the scene being inspected by the drone, and the distance setting value is positively correlated with the size of the scene being inspected. As an optional implementation, in this embodiment, the distance setting value is 0.2m. A nearest neighbor search is performed on each completed sampling point to find the minimum distance to the nearest point cloud. If this minimum distance is less than 0.2m, the sampled point is removed, and spatial sampling is resumed, with a new point sampled.

[0024] In step S3 of this embodiment, performing a nearest neighbor search on the sampled sampling points to find the nearest point cloud distance of the sampling points includes: performing a nearest neighbor search on each of the sampled sampling points in the point cloud map to find a neighbor point set consisting of all neighbor points within a preset distance range, calculating the distance between each neighbor point in the neighbor point set and the sampled sampling point, and taking the distance with the smallest distance to the sampled sampling point as the nearest point cloud distance of the sampling point.

[0025] In step S4 of this embodiment, calculating the minimum distance to the nearest point cloud includes: performing a nearest neighbor search on each of the traversed sampling points to find a set of neighbor sampling points consisting of all neighbor sampling points within a preset distance range, calculating the distance between each neighbor sampling point in the set of neighbor sampling points and the traversed sampling point, and using the distance with the smallest distance to the traversed sampling point as the minimum distance to the nearest point cloud of the traversed sampling point.

[0026] In step S5 of this embodiment, the sampling points are sorted by minimum distance, and the sampling point with the largest minimum distance is used as the center. When generating a spherical convex hull with this minimum distance as the radius, if the distance between the center and the ground is greater than or equal to the radius, the generated spherical convex hull is a sphere; if the distance between the center and the ground is less than the radius, the generated spherical convex hull is the spatial portion of the sphere located above the ground. In this embodiment, each spherical convex hull is represented by a four-dimensional vector, including the Cartesian coordinates (x, y, z) of its center and the radius r. By filling areas without obstacles with overlapping spheres, the environmental information can be simplified, retaining only the overlapping spheres. The ground can be determined based on the flight altitude of the drone.

[0027] Calculating the fill rate of all spherical convex hulls in step S7 of this embodiment includes calculating the volumes of all spherical convex hulls individually, adding the volumes of all spherical convex hulls to obtain a total volume, and then dividing the total volume by the size of the obstacle-free space detected by the drone to obtain the fill rate of all spherical convex hulls. As an optional embodiment, in this embodiment, when the fill rate is greater than 90%, the program exits, completing the map representation.

[0028] In summary, the drone lightweight map representation method based on spherical convex hull in this embodiment can extract useful feasible areas for drones in complex point cloud maps, uses a combination of multiple spherical convex hulls to fill the obstacle-free areas in the point cloud map, and finally determines the end of the algorithm based on the filling rate judgment standard. Compared with the above grid map, the memory usage is lower, but the search complexity increases from O(1) of the grid map to O(logn). The drone lightweight map representation method based on spherical convex hull in this embodiment can achieve efficient space filling and collision avoidance, providing reliable technical support for the safe navigation of drones in complex environments, and is particularly suitable for map representation in large spaces. The drone lightweight map representation method based on spherical convex hull in this embodiment can significantly improve inspection efficiency, and the inspection robot can autonomously navigate and inspect, reducing the labor intensity of inspection personnel and improving inspection safety. Through real-time monitoring and alarm, safety hazards in cable tunnels can be discovered in a timely manner, thereby improving the level of power safety production.

[0029] In addition, this embodiment also provides a UAV lightweight map representation system based on spherical convex hull, including a microprocessor and a memory connected to each other, and the microprocessor is programmed or configured to execute the UAV lightweight map representation method based on spherical convex hull.

[0030] In addition, this embodiment also provides a computer-readable storage medium, which stores a computer program or instruction, and the computer program or instruction is programmed or configured to execute the spherical convex hull-based drone lightweight map representation method through a processor.

[0031] In addition, this embodiment also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the spherical convex hull-based drone lightweight map representation method through a processor.

[0032] Those skilled in the art should understand that the technical solutions that can be provided by the embodiments of the present invention may be in the form of methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment that combines software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the functions in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including the instruction device, which implements the function specified in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0033] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A lightweight map representation method for UAV based on spherical convex hull, characterized by: The steps include: S1, obtain the point cloud map collected by the drone; S2, uniform spatial sampling in the space of the point cloud map; S3, performing a nearest neighbor search on the sampled sampling points to find the nearest point cloud distance of the sampling points. If the nearest point cloud distance of the sampling point is less than the set distance value, the sampling point is deleted and the process goes to step S2 until a preset number of sampling points is obtained; S4, traverse all sampling points and calculate the minimum distance to the nearest point cloud respectively; S5, sorting the sampling points by minimum distance, taking the sampling point with the largest minimum distance as the center of the sphere, and generating a spherical convex hull with the minimum distance as the radius; S6, delete all sampling points within the spherical convex hull; S7, calculate the filling rate of all spherical convex hulls. If the filling rate is greater than or equal to a preset threshold, the map representation is determined to be complete, and the drone is controlled to fly within the spherical convex hull to avoid obstacles; otherwise, jump to step S4.

2. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: Before step S3, the method further includes determining the size of the preset specified number according to the size of the scene inspected by the drone, and the size of the preset specified number is positively correlated with the size of the scene inspected by the drone.

3. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: Before step S3, the process also includes determining the size of the distance setting value according to the size of the scene inspected by the drone, and the size of the distance setting value is positively correlated with the size of the scene inspected by the drone.

4. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: In step S3, performing a nearest neighbor search on the sampled sampling points to find the nearest point cloud distance of the sampling points includes: performing a nearest neighbor search on the sampled sampling points in the point cloud map to find a neighbor point set consisting of all neighbor points within a preset distance range, calculating the distance between each neighbor point in the neighbor point set and the sampled sampling point, and taking the distance with the smallest distance to the sampled sampling point as the nearest point cloud distance of the sampling point.

5. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: Calculating the minimum distance to the nearest point cloud in step S4 includes: performing a nearest neighbor search on each of the traversed sampling points to find a set of neighbor sampling points consisting of all neighbor sampling points within a preset distance range, calculating the distance between each neighbor sampling point in the set of neighbor sampling points and the traversed sampling point, and taking the distance with the minimum distance to the traversed sampling point as the minimum distance to the nearest point cloud of the traversed sampling point.

6. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: In step S5, the sampling points are sorted by minimum distance, and the sampling point with the largest minimum distance is used as the center of the sphere. When a spherical convex hull is generated with the minimum distance as the radius, if the distance between the center of the sphere and the ground is greater than or equal to the radius, the generated spherical convex hull is a sphere; if the distance between the center of the sphere and the ground is less than the radius, the generated spherical convex hull is the spatial portion of the sphere located above the ground.

7. The UAV lightweight map representation method based on spherical convex hull according to claim 1 is characterized in that: Calculating the filling rates of all spherical convex hulls in step S7 includes: calculating the volumes of all spherical convex hulls respectively, adding the volumes of all spherical convex hulls to obtain the total volume of the spherical convex hulls, and then dividing the total volume of the convex hulls by the size of the obstacle-free space detected by the drone to obtain the filling rates of all spherical convex hulls.

8. A lightweight map representation system for unmanned aerial vehicles based on spherical convex hull, comprising a microprocessor and a memory connected to each other, characterized in that: The microprocessor is programmed or configured to execute the spherical convex hull-based drone lightweight map representation method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program or instruction stored therein, characterized in that: The computer program or instruction is programmed or configured to execute the spherical convex hull-based drone lightweight map representation method according to any one of claims 1 to 7 through a processor.

10. A computer program product comprising a computer program or instructions, characterized in that The computer program or instruction is programmed or configured to execute the spherical convex hull-based drone lightweight map representation method according to any one of claims 1 to 7 through a processor.

Images