A lightning-rod-avoiding flight inspection method and system based on a laser-calibrated unmanned aerial vehicle

Abstract

This application discloses a method and system for aerial inspection and testing of lightning rods based on a laser-calibrated UAV, relating to the field of lightning rod inspection. The method includes: taking photos of the lightning rod target inspection point via aerial flight using a laser-calibrated UAV, and obtaining the optimal safety viewpoint using multi-target optimization; first emitting a point laser, then emitting a ring laser centered on the laser point to define the target outline; capturing inspection images under laser illumination, and cropping the images based on the ring laser range using edge detection; using the laser point as a positioning reference point, performing fault detection analysis on the cropped image to obtain the analysis results. This application achieves accurate calibration of the target inspection point by setting an adjustable ring laser module on the UAV, enhancing the inspection capability of the target inspection point in complex inspection environments with obstructions; and achieves rapid cropping of the inspection image through ring laser outline selection, reducing the difficulty and complexity of fault analysis and lowering computational costs.

Description

Technical Field

[0001] This application relates to the field of lightning rod inspection, and in particular to a method and system for inspection and testing of lightning rods by flying around a laser-calibrated UAV. Background Technology

[0002] To strengthen lightning protection at substations, lightning rods, typically over 30 meters high, are installed around each substation. However, manual inspections have blind spots, making it difficult to monitor the operational status of the tops of these rods. Most lightning rods were constructed concurrently with the substations and have been in operation for many years, enduring sun, rain, and weathering. This can lead to severe corrosion at connection points, potentially causing breakage, collapse, or even disintegration, posing safety hazards to substation equipment and personnel, and easily triggering accidents. Traditional lightning rod inspections rely on manual climbing, ground-based telescope observation, or manned helicopter operations, which are inefficient, risky, and costly. In recent years, the rapid development of aerial inspection technology (i.e., drones orbiting a lightning rod to inspect designated points) has provided a new solution for power line inspection. By equipping sensors such as visible light cameras, it enables non-contact inspection of equipment at heights. However, lightning rods are typically installed at the top of towers or high up in buildings, in complex environments often obstructed by hardware, guy wires, and other structures. This makes it difficult for drones to obtain clear and complete target images during circling flights, often resulting in the inability to locate the target during subsequent fault analysis. Furthermore, existing drone inspection methods largely rely on GPS positioning and preset flight paths, making them susceptible to airflow disturbances and electromagnetic interference during close-range inspections. Precise control of the observation angle and distance is difficult, resulting in small target areas and cluttered backgrounds in the captured images, making subsequent fault identification algorithms difficult to process and computationally expensive.

[0003] In summary, existing UAV lightning rod inspection methods are prone to poor image acquisition quality due to inaccurate target positioning in complex and obstructed environments, resulting in time-consuming and labor-intensive subsequent identification; the background of the captured inspection images is cluttered, and cropping and preprocessing rely on manual screening and intervention, leading to high computational costs for fault analysis; moreover, traditional detection algorithms are insufficient in identifying minor or typical defects, making it difficult to achieve accurate fault diagnosis. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for aerial inspection and testing around lightning rods based on laser-calibrated UAVs. This method can accurately calibrate target inspection points, enhance the inspection capability of target inspection points in complex inspection environments with obstructions, and enable rapid cropping of inspection images through ring-shaped laser contour selection, thereby reducing the difficulty and complexity of fault analysis and lowering computational costs.

[0005] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for aerial inspection of lightning rods based on a laser calibration drone. The laser calibration drone is a drone equipped with a ring-shaped adjustable laser module and an industrial camera. The method includes: guiding the laser calibration drone to fly around a target inspection point of a lightning rod, acquiring 360° ring images of the environment around the target inspection point; based on the 360° ring images, using a multi-objective constraint optimization algorithm to analyze the optimal observation angle and optimal observation distance of the laser calibration drone under safe conditions, obtaining the optimal safe viewpoint; controlling the ring-shaped adjustable laser module to first emit a point laser towards the target inspection point, and then using a laser beam illuminating the target inspection point... A circular laser is emitted and adjusted with the spot as the center. The diameter of the circular adjustable laser is adjusted by controlling the circular adjustable laser module until the outline of the target inspection point is completely defined. Under the simultaneous illumination of the spot laser and the circular laser, an inspection image of the target inspection point is captured by the industrial camera. Based on the range of the circular laser in the inspection image, the inspection image is cropped using an edge detection method to obtain a cropped inspection image. Based on the cropped inspection image, the laser point is used as the positioning reference point, and an improved YOLOv8 algorithm is used to perform fault detection analysis on the target inspection point to obtain the analysis results. The improved YOLOv8 algorithm incorporates a spatial attention mechanism into the YOLOv8 algorithm.

[0006] Furthermore, the laser calibration drone flies around the target inspection point of the lightning rod while maintaining a dynamic safe distance; the dynamic safe distance is obtained by the laser calibration drone dynamically adjusting its posture after real-time ranging of the target inspection point by the ring adjustable laser module.

[0007] Furthermore, the constraints of the multi-objective constrained optimization algorithm include at least geometric constraints and observation constraints.

[0008] Furthermore, controlling the ring-shaped adjustable laser module to first emit a point laser towards the target inspection point, and then emitting and adjusting a ring-shaped laser with the laser point illuminating the target inspection point as the center, specifically includes: controlling the ring-shaped adjustable laser module to first emit a point laser, using the industrial camera to locate the center of the laser spot in real time; adjusting the laser point to the target inspection point by servo fine-tuning the pose of the laser calibration drone; after the laser spot is locked, controlling the ring-shaped adjustable laser module to activate the ring laser; and aligning the center of the ring laser with the center of the laser spot.

[0009] Furthermore, controlling the ring-shaped adjustable laser module to adjust the diameter of the ring laser until it completely defines the contour range of the target inspection point specifically includes: acquiring images with the ring laser in real time using an industrial camera; extracting the inner and outer edges of the ring laser and the actual contour edge of the target inspection point in the image with the ring laser using an edge detection algorithm; calculating the fitting degree between the inner and outer edges of the ring laser and the actual contour edge of the target inspection point in real time; and adjusting the diameter of the ring laser using a PID control zoom lens group based on the fitting degree until the inner diameter of the ring laser completely defines the contour range of the target inspection point.

[0010] Furthermore, based on the range of the ring laser in the inspection image, the inspection image is cropped using an edge detection method to obtain the cropped inspection image. Specifically, this includes: detecting the inner and outer boundaries of the ring laser in the inspection image, fitting the ring boundary using an active contour model to generate a mask; cropping the inspection image with the minimum bounding rectangle of the ring laser, and removing the background outside the mask to obtain the cropped inspection image.

[0011] Furthermore, before detecting the inner and outer boundaries of the ring laser in the inspection image and fitting the ring boundary using an active contour model to generate a mask, the process also includes: if the ring laser has partial boundary blurring or partial line missing due to the surface curvature of the target detection point, the center of the laser spot is used as a geometric constraint, and the missing part of the ring laser is fitted using the RANSAC algorithm.

[0012] Furthermore, based on the cropped inspection image, using the laser point as the positioning reference point, an improved YOLOv8 algorithm is used to perform fault detection analysis on the target inspection point, and the analysis results are obtained. Specifically, the center coordinates of the laser point are used as the positioning reference point of the cropped inspection image; the position of the positioning reference point is used as the center prior of the spatial attention map through the improved YOLOv8 algorithm, and the fault detection analysis on the target inspection point is performed based on the cropped inspection image, and the analysis results are obtained.

[0013] Furthermore, the fault types corresponding to the fault detection and analysis process include: corrosion, cracks, deformation, arc burns, loose connections, and foreign matter adhesion.

[0014] Secondly, this application also provides a computer system, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the laser calibration UAV-based lightning rod-avoiding flight inspection and detection method described in the first aspect.

[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application utilizes a ring-shaped adjustable laser module and an industrial camera mounted on a drone to guide the laser calibration drone in a 360° circular flight around a lightning rod inspection point, acquiring surrounding environmental information. Based on this 360° circular image, a multi-objective constraint optimization algorithm is employed to analyze the optimal observation angle and distance under safe conditions, obtaining the optimal safe viewpoint and ensuring inspection accuracy and flight safety. The ring-shaped adjustable laser module first emits a point laser, then emits a ring laser centered on that point, dynamically adjusting the diameter of the ring laser until it completely outlines the target inspection point. During this process, active laser calibration overcomes interference from obstructions in complex environments, enhancing the location and inspection capabilities of the inspection point. Under simultaneous illumination by the point and ring lasers, the industrial camera captures inspection images. Subsequently, based on the ring laser range, edge detection is used to quickly crop the image, retaining only the defined area, significantly reducing the difficulty and computational cost of subsequent fault analysis. Finally, using the laser point as the positioning reference point, the improved YOLOv8 algorithm with the introduction of spatial attention mechanism is used to perform accurate fault detection analysis on the cropped image, which improves the ability to identify small or typical defects. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for aerial inspection and detection around lightning rods based on a laser-calibrated UAV, as provided in an embodiment of this application.

[0018] Figure 2 A diagram showing the positional relationship between a laser calibration drone and a lightning rod, provided for an embodiment of this application.

[0019] Figure 3 This is a schematic diagram of the structure of a laser calibration drone provided in an embodiment of this application.

[0020] Figure 4 This is a schematic diagram of the environment near the lightning rod target checkpoint provided in an embodiment of this application.

[0021] Figure 5 A schematic diagram of the network structure of the improved YOLOv8 algorithm provided in the embodiments of this application.

[0022] Figure 6 This is an internal structure diagram of a computer system provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] Example 1, such as Figures 1-4 As shown, this embodiment provides a method for aerial inspection and testing around lightning rods based on a laser calibration drone. The laser calibration drone is a drone equipped with a ring-shaped adjustable laser module and an industrial camera. The method includes: S1. As Figure 2 As shown, the laser calibration UAV is guided to fly around the target checkpoint of the lightning rod, collecting 360° circular images of the environment around the target checkpoint.

[0026] Furthermore, the laser calibration drone flies around the target inspection point of the lightning rod while maintaining a dynamic safe distance; the dynamic safe distance is obtained by the laser calibration drone dynamically adjusting its posture after real-time ranging of the target inspection point by the ring adjustable laser module.

[0027] In practical applications, operators of the laser calibration drone select the target inspection point of the lightning rod via a ground station. The drone then automatically flies to the vicinity of the lightning rod. Centering on the target point, and considering obstructions such as hardware, guy wires, and structures near the lightning rod, the drone maintains a dynamic safe distance and circles the target point. During this flight, industrial cameras continuously capture images from different angles, which are then stitched together in real time to generate a 360° ring image covering the surrounding environment. This allows for rapid acquisition of panoramic environmental information around the target inspection point without blind spots, overcoming the obstruction and field-of-view limitations caused by a single perspective.

[0028] S2. Based on 360° ring imagery, a multi-objective constraint optimization algorithm is used to analyze the optimal observation angle and optimal observation distance of the laser calibration UAV under safe conditions, and the optimal safe viewpoint is obtained.

[0029] Furthermore, the constraints of the multi-objective constrained optimization algorithm include at least geometric constraints and observation constraints.

[0030] Optionally, the expression for the geometric constraint is: .

[0031] in, The spatial coordinates of the UAV; Let be the coordinates of any point on the surface of the target checkpoint; It is the set of all points on the surface of the obstacle; Euclidean distance; This is the safety threshold.

[0032] Optionally, the expression for the observation constraint is: .

[0033] .

[0034] in, The spatial coordinates of the target checkpoint; The coordinates of the optical center position of the industrial camera; The unit direction is the direction of the camera's optical axis; The distance from the camera to the target checkpoint; To determine the maximum allowable deviation angle, it is typically taken as 1 / 2 or 1 / 3 of the field of view (e.g., 15°). This is the minimum effective observation distance between the industrial camera and the target inspection point; This represents the maximum effective observation distance between the industrial camera and the target checkpoint.

[0035] Optional constraints may also include: laser projection constraints and visibility constraints.

[0036] In practical applications, multi-objective constraint optimization algorithms can automatically solve for the optimal observation angle and distance that balances obstacle avoidance safety, imaging clarity, and laser projection accuracy, eliminating the subjectivity and collision risk of manual test flights, thereby improving the standardization level of inspection viewpoints and the efficiency of task execution.

[0037] S3. Control the ring-shaped adjustable laser module to first emit a point laser towards the target inspection point, and then emit and adjust a ring-shaped laser with the laser point illuminating the target inspection point as the center.

[0038] Furthermore, step S3 specifically includes: S31. Control the ring-shaped adjustable laser module to emit a point laser first, and use the industrial camera to locate the center of the laser spot in real time.

[0039] S32. By fine-tuning the pose of the laser calibration UAV using a servo, the laser point is adjusted to the target inspection point.

[0040] S33. After the laser spot is locked, control the ring-shaped adjustable laser module to turn on the ring laser.

[0041] S34. Align the center of the ring laser with the center of the laser spot.

[0042] S4. Control the ring-shaped adjustable laser module to adjust the diameter of the ring laser until the outline of the target inspection point is completely defined.

[0043] In practical applications, defining the outline of the target inspection point can eliminate interference from complex backgrounds and obstructions, ensuring that only the target to be inspected is retained within the shooting area.

[0044] S5. Under simultaneous illumination by point laser and ring laser, the industrial camera captures inspection images of the target inspection points.

[0045] Furthermore, step S5 specifically includes: S51. Real-time acquisition of images with ring lasers via an industrial camera.

[0046] S52. Extract the inner and outer edges of the ring laser and the actual contour edges of the target checkpoint in the image with the ring laser using an edge detection algorithm.

[0047] S53. Real-time calculation of the fit between the inner and outer edges of the ring laser and the actual contour edge of the target inspection point.

[0048] S54. Based on the fit, adjust the diameter of the ring laser by using PID control of the zoom lens group until the inner diameter of the ring laser completely defines the outline of the target inspection point.

[0049] S6. Based on the range of the ring laser in the inspection image, the inspection image is cropped using the edge detection method to obtain the cropped inspection image.

[0050] Furthermore, step S6 specifically includes: S61. Detect the inner and outer boundaries of the ring laser in the inspection image, and use an active contour model to fit the ring boundary to generate a mask.

[0051] S62. The inspection image is cropped using the minimum bounding rectangle of the ring laser, and the background outside the mask is removed to obtain the cropped inspection image.

[0052] Furthermore, before step S61, the method further includes: if the ring laser produces partial boundary blurring or partial line loss due to the surface curvature of the target detection point, then the center of the laser spot is used as a geometric constraint, and the missing part of the ring laser is fitted by the RANSAC algorithm.

[0053] In practical applications, the ring laser forms a clear bright ring at the edge of the target contour. The onboard processor of the laser calibration UAV first converts the image into a grayscale image, then uses the Canny edge detection operator to extract continuous edge pixels of the ring laser, and finally obtains the minimum bounding rectangle of the ring region through contour fitting. Finally, using this rectangle as the cropping boundary, the original image is automatically cropped, retaining the target area within the ring and removing the external background, thus obtaining the cropped inspection image. Images framed by the ring laser are suitable for automated image cropping using edge detection methods, removing cluttered backgrounds and obstructions without manual intervention, allowing subsequent fault analysis to focus only on the target contour. This reduces image data volume and computational costs, while avoiding misjudgments of defect detection due to background noise.

[0054] Therefore, in this embodiment, no background processing is required, and the subsequent fault analysis process can be realized solely through the UAV processor, reducing data transmission costs and background processing costs, while also ensuring the real-time nature of fault analysis.

[0055] S7. Based on the cropped inspection image, using the laser point as the positioning reference point, the improved YOLOv8 algorithm is used to perform fault detection analysis on the target inspection point to obtain the analysis results; the improved YOLOv8 algorithm is the addition of a spatial attention mechanism to the YOLOv8 algorithm.

[0056] Furthermore, step S7 specifically includes: S71. Use the center coordinates of the laser point as the positioning reference point of the cropped inspection image.

[0057] S72. By improving the YOLOv8 algorithm, the location of the positioning reference point is used as the center prior of the spatial attention map, and the fault detection analysis of the target inspection point is performed based on the cropped inspection image to obtain the analysis results.

[0058] Furthermore, the fault types corresponding to the fault detection and analysis process include: corrosion, cracks, deformation, arc burns, loose connections, and foreign matter adhesion.

[0059] Furthermore, after obtaining the analysis results, the target checkpoints that have malfunctioned are repaired based on the analysis results.

[0060] In practical applications, such as Figure 5 As shown, in addition to introducing a spatial attention mechanism, this embodiment also introduces a dedicated small target detection head for the P2 layer (used to detect small targets such as bolts and nuts), which improves the detection accuracy of small targets by utilizing high-resolution feature maps.

[0061] Specifically, the standard YOLOv8 architecture employs a sequential neck design, progressively downsampling and enhancing feature maps through convolutional layers (Conv), C2f layers (enhanced residual modules), and SPPF layers (spatial pyramid pooling fusion). Its head module concatenates multi-scale features (P3, P4, P5) and performs target prediction through detection layers. To enhance the detection capability of small targets in UAV imagery, this embodiment introduces an additional detection head at the P2 layer of the network. Unlike the standard detection heads (P3, P4, P5 layers) which correspond to deeper, more abstract features and have lower spatial resolution, the P2 layer processes high-resolution feature maps extracted from the early stages of the backbone network. These shallow structures retain more refined spatial information, such as object boundaries, textures, and edge features. The introduction of the P2 detection layer is based on the inherent characteristics of small objects: they often occupy only a few pixels in high-altitude UAV imagery and are easily submerged or diluted during downsampling in deep networks. Although the P3 layer already has moderate resolution, stride convolution can still lead to the loss of crucial local details. In contrast, P2 retains finer spatial granularity and captures lower-level visual features that are better suited for recognizing small targets, especially in complex backgrounds and occluded scenes.

[0062] The processed P2 feature map is processed by the ScalSeq module, which aggregates multi-scale information and enhances the representation of small targets. Finally, the enhanced P2 output is integrated with the existing P3-P5 detection heads to form a multi-scale detection head set. This ensures comprehensive coverage of target scales, particularly improving the model's sensitivity and accuracy in detecting small, densely distributed targets.

[0063] The technical effects of this application are as follows: This application utilizes a ring-shaped adjustable laser module and an industrial camera mounted on a drone to guide the laser calibration drone in a 360° circular flight around a lightning rod inspection point, acquiring surrounding environmental information. Based on this 360° circular image, a multi-objective constraint optimization algorithm is employed to analyze the optimal observation angle and distance under safe conditions, obtaining the optimal safe viewpoint and ensuring inspection accuracy and flight safety. The ring-shaped adjustable laser module first emits a point laser, then emits a ring laser centered on that point, dynamically adjusting the diameter of the ring laser until it completely outlines the target inspection point. During this process, active laser calibration overcomes interference from obstructions in complex environments, enhancing the location and inspection capabilities of the inspection point. Under simultaneous illumination by the point and ring lasers, the industrial camera captures inspection images. Subsequently, based on the ring laser range, edge detection is used to quickly crop the image, retaining only the defined area, significantly reducing the difficulty and computational cost of subsequent fault analysis. Finally, using the laser point as the positioning reference point, the improved YOLOv8 algorithm with the introduction of spatial attention mechanism is used to perform accurate fault detection analysis on the cropped image, which improves the ability to identify small or typical defects.

[0064] Example 2: This example provides a computer system, which can be a server or a terminal, and its internal structure diagram can be as follows. Figure 6 As shown, the computer system includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and databases. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media to run. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements the methods described above.

[0065] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer system to which the present application is applied. A specific computer system may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0066] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0067] The processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited thereto.

[0068] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0069] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting the flight of an unmanned aerial vehicle around a lightning rod based on laser calibration, characterized in that, The laser calibration drone is a drone equipped with a ring-shaped adjustable laser module and an industrial camera, and the method includes: The laser calibration UAV is guided to fly around the target inspection point of the lightning rod, and collect 360° circular images of the environment around the target inspection point; Based on 360° ring imagery, a multi-objective constraint optimization algorithm is used to analyze the optimal observation angle and optimal observation distance of the laser calibration UAV under safe conditions, and the optimal safe viewpoint is obtained. The ring-shaped adjustable laser module is controlled to first emit a point laser towards the target inspection point, and then emit and adjust a ring laser with the laser point illuminating the target inspection point as the center. The diameter of the ring-shaped laser is adjusted by controlling the ring-shaped adjustable laser module until the outline of the target inspection point is completely defined. Under simultaneous illumination by point laser and ring laser, the industrial camera captures inspection images of the target inspection points; Based on the range of the ring laser in the inspection image, the inspection image is cropped using the edge detection method to obtain the cropped inspection image. Based on the cropped inspection image, using the laser point as the positioning reference point, the improved YOLOv8 algorithm is used to perform fault detection analysis on the target inspection point, and the analysis results are obtained; the improved YOLOv8 algorithm is the addition of a spatial attention mechanism to the YOLOv8 algorithm.

2. The method for detecting the unmanned aerial vehicle flying around the lightning rod based on the laser calibration according to claim 1, characterized in that, The laser calibration drone flies around the target inspection point of the lightning rod while maintaining a dynamic safe distance. The dynamic safe distance is obtained by the laser calibration drone dynamically adjusting its posture after real-time ranging of the target inspection point by the ring adjustable laser module. 3.The laser calibration based unmanned aerial vehicle around-pylon flight patrol detection method according to claim 1, characterized in that, The constraints of the multi-objective constrained optimization algorithm include at least geometric constraints and observation constraints.

4. The method for detecting the unmanned aerial vehicle flying around the lightning rod according to claim 1, wherein, The control of the ring-shaped adjustable laser module first emits a point laser towards the target inspection point, and then emits and adjusts a ring-shaped laser with the laser point illuminating the target inspection point as the center. Specifically, this includes: The ring-shaped adjustable laser module is controlled to emit a point laser first, and the center of the laser spot is located in real time using the industrial camera. The laser point is adjusted to the target inspection point by servo fine-tuning the pose of the laser calibration UAV. After the laser spot is locked, the ring-shaped adjustable laser module is controlled to turn on the ring laser. Align the center of the ring laser with the center of the laser spot.

5. The method for detecting the unmanned aerial vehicle flying around the lightning rod according to the laser calibration based on the unmanned aerial vehicle, characterized in that, Controlling the annular adjustable laser module to adjust the diameter of the annular laser until the outline of the target inspection point is completely defined includes: Images with ring lasers are acquired in real time using an industrial camera; The inner and outer edges of the ring laser and the actual contour edges of the target checkpoint in the image with the ring laser are extracted using an edge detection algorithm. Real-time calculation of the fit between the inner and outer edges of the ring laser and the actual contour edge of the target inspection point; Based on the fit, the diameter of the ring laser is adjusted by using a PID control zoom lens group until the inner diameter of the ring laser completely defines the outline of the target inspection point.

6. The method for aerial inspection and detection around lightning rods based on laser-calibrated UAVs according to claim 1, characterized in that, Based on the range of the ring-shaped laser in the inspection image, an edge detection method is used to crop the inspection image, resulting in a cropped inspection image, specifically including: The inner and outer boundaries of the ring-shaped laser in the inspection image are detected, and the ring boundary is fitted using an active contour model to generate a mask; The inspection image is cropped using the minimum bounding rectangle of the ring laser, and the background outside the mask is removed to obtain the cropped inspection image.

7. The method according to claim 6, wherein, Before detecting the inner and outer boundaries of the ring-shaped laser in the inspection image, and fitting the ring boundary using an active contour model to generate the mask, the process also includes: If the ring laser produces partial boundary blurring or partial line loss due to the surface curvature of the target detection point, the center of the laser spot is used as a geometric constraint, and the missing part of the ring laser is fitted using the RANSAC algorithm. 8.The laser calibration based unmanned aerial vehicle around-pylon flight patrol detection method according to claim 1, wherein, Based on the cropped inspection image, using laser points as positioning reference points, an improved YOLOv8 algorithm is used to perform fault detection analysis on the target inspection points, yielding analysis results, including: The center coordinates of the laser point are used as the positioning reference point for the cropped inspection image; By improving the YOLOv8 algorithm, the location of the reference point is used as the center prior of the spatial attention map, and the fault detection analysis of the target inspection point is performed based on the cropped inspection image to obtain the analysis results.

9. The method for aerial inspection and detection around lightning rods based on laser-calibrated UAVs according to claim 1, characterized in that, The fault types corresponding to the fault detection and analysis process include: rust, cracks, deformation, arc burns, loose connections, and foreign matter adhesion.

10. A computer system comprising: The memory, the processor, and the computer program stored in the memory and executable on the processor are characterized in that the processor executes the computer program to implement the method for aerial inspection and detection of a laser-calibrated UAV around a lightning rod, as described in any one of claims 1-9.

Images