Rapid butt joint and separation method, system and equipment for unmanned aerial vehicle and insulator detection mechanism and medium

Abstract

The invention relates to the technical field of insulator detection, and discloses an unmanned aerial vehicle and insulator detection mechanism rapid butt joint and separation method, system, equipment and medium, and the method comprises the steps: building a comprehensive basic sensing system through collecting various sensor data, carrying out the optimization processing of multi-source data through employing a dynamic weight fusion algorithm, and carrying out the rapid butt joint and separation of the unmanned aerial vehicle and the insulator detection mechanism; high-precision recognition and pose calculation of an insulator support target are achieved by combining SIFT feature matching, ICP point cloud registration and infrared temperature contour information, then the distance between the unmanned aerial vehicle and the target support is calculated in real time based on a power transmission tower three-dimensional map and a three-level decision mechanism, and an action control instruction is output. The flight stability and the accurate docking capability of the unmanned aerial vehicle in the complex electromagnetic environment are ensured, and the reliability and the robustness of the system in the complex electromagnetic environment are improved by receiving a guide pipe contact signal and executing mechanical locking and releasing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of insulator detection, and particularly relates to a method, system and device for quick docking and separation of a UAV and an insulator detection mechanism and a medium. BACKGROUND

[0002] With the continuous expansion of the scale of the power system and the improvement of the intelligent level, the application of UAVs in the field of power inspection is increasingly widespread. Power inspection work usually needs to be carried out in complex terrain and harsh environments, including the inspection of high-voltage transmission lines, substation equipment and distribution facilities. UAVs, with their flexibility, efficiency and non-contact operation characteristics, have become an important tool for power inspection. However, in actual application, UAVs face the challenge of complex electromagnetic environments, especially near high-voltage transmission lines, where strong electromagnetic fields can significantly interfere with the navigation, positioning and communication systems of the UAV. This interference not only affects the flight stability of the UAV, but also can lead to the accumulation of positioning errors, thereby affecting the accuracy and safety of the inspection operation. In addition, UAVs need to complete high-precision docking tasks in complex electromagnetic environments, such as automatic loading and unloading of equipment or precise coupling with ground equipment, which puts higher requirements on the dynamic positioning and docking reliability of UAVs. Therefore, how to achieve high-precision dynamic positioning and reliable physical coupling of UAVs in a strong electromagnetic interference environment has become a key technical problem that needs to be solved in the current field of power inspection.

[0003] Existing UAV positioning and docking technologies mainly rely on single sensors or traditional mechanical lock mechanisms, and these technologies are difficult to meet actual needs in complex electromagnetic environments. For example, laser radar-based positioning systems may experience signal attenuation or misjudgment under strong electromagnetic interference, resulting in a decrease in positioning accuracy; visual sensors may be affected by changes in light or environmental reflections, and cannot provide stable positioning information. In addition, traditional mechanical lock mechanisms are prone to slow response or failure under strong electromagnetic interference, especially in dynamic docking scenarios that require fast response, and their reliability is difficult to guarantee. These technical limitations not only reduce the efficiency of power inspection operations, but also increase the risk of equipment damage and personnel injury. Therefore, there is an urgent need for a technical solution that can achieve high-precision dynamic positioning and reliable physical coupling in complex electromagnetic environments to improve the application effect and safety of UAVs in power inspection. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a method, system, device and medium for quick docking and separation of a UAV and an insulator detection mechanism, solving the technical problem that current UAVs are difficult to stably position near a strong magnetic field.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a method for rapid docking and separation of a UAV and an insulator detection mechanism, comprising:

[0008] Collecting electromagnetic field intensity, visible light / infrared image, millimeter wave radar, and inertial measurement unit raw data as basic perception data;

[0009] Optimizing the basic perception data, combining feature matching, point cloud registration, and infrared temperature contour information to achieve high-precision identification of the insulator support target and output the accurate pose of the target support relative to the UAV;

[0010] Based on the accurate pose of the target support relative to the UAV and the known three-dimensional map of the transmission tower, the distance between the UAV and the target support is calculated in real time, a three-level decision mechanism is started according to the distance threshold, and the action control instruction is output and the UAV is guided to the target;

[0011] Receiving the guide tube contact signal, performing mechanical locking and releasing operations, and obtaining the task execution result;

[0012] Based on the locking / separation timestamp, position error, electromagnetic intensity, and execution time timer, the preset threshold is compared to trigger an alarm and diagnose the fault type, and the Kalman filter parameters are adaptively adjusted to obtain the fault handling instruction and the optimized sensor weight.

[0013] As a preferred scheme of the method for rapid docking and separation of a UAV and an insulator detection mechanism, wherein: the three-level decision mechanism according to the distance threshold comprises:

[0014] When the distance between the UAV and the target support is greater than the first threshold, the far-field control stage is started;

[0015] When the distance between the UAV and the target support is between the first threshold and the second threshold, the near-field control stage is started;

[0016] When the distance between the UAV and the target support is less than the second threshold, the critical docking control stage is started.

[0017] The beneficial effects of the preferred technical scheme are: the distance between the UAV and the target support is calculated in real time and the action control instruction is output, ensuring the flight stability and precise docking capability of the UAV in a complex electromagnetic environment.

[0018] As a preferred scheme of the method for rapid docking and separation of a UAV and an insulator detection mechanism, wherein: the far-field control stage comprises:

[0019] The first planning algorithm is used for path planning, and a cost function of path planning is calculated;

[0020] The actual moving cost, the Euclidean distance and the influence of the electromagnetic field strength are comprehensively considered, the weight factor is adjusted according to the real-time change of the electromagnetic field strength, and the path selection is optimized;

[0021] The normal vector of the target support is calculated through the pitch angle and the yaw angle of the target support;

[0022] By comparing the current attitude of the unmanned aerial vehicle with the target attitude, the angle difference to be adjusted is calculated, and the yaw and pitch angle velocity instructions of the unmanned aerial vehicle are generated to adjust the attitude of the unmanned aerial vehicle, so that the unmanned aerial vehicle gradually aligns with the target.

[0023] As a preferred scheme of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method, the near-field control stage includes:

[0024] The feature point data of the binocular vision system is used to determine the target support precise docking point coordinates;

[0025] The target support precise docking point coordinates are converted to the coordinate system of the unmanned aerial vehicle, and the position deviation vector relative to the unmanned aerial vehicle is calculated;

[0026] The position deviation vector is used to generate a speed control instruction to adjust the position of the unmanned aerial vehicle;

[0027] The attitude deviation is used to generate an angular velocity control instruction to adjust the attitude of the unmanned aerial vehicle;

[0028] The three-axis speed instructions and three-axis angular velocity instructions of the unmanned aerial vehicle are output.

[0029] As a preferred scheme of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method, the near-field control stage includes:

[0030] It is detected whether the position and attitude deviation meets the triggering condition of the aerodynamic guiding mechanism, and if so, the aerodynamic guiding mechanism is started;

[0031] According to the dynamics model of the pneumatic system, the extension force of the guide tube is calculated to ensure that the guide tube is smoothly extended;

[0032] The contact force is monitored in real time through the pressure sensor, the impedance control mode is entered, the position correction amount is calculated according to the deviation between the contact force and the expected force, the position of the unmanned aerial vehicle is adjusted, and the position correction instruction is output.

[0033] As a preferred scheme of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method, the optimization processing of the basic perception data includes:

[0034] Analyze the meridional velocity of each target in the radar point cloud, filter out dynamic points with a speed exceeding a threshold, and retain a static target point cloud; extract coarse-grained three-dimensional coordinates of the insulator support from the static target point cloud;

[0035] Extract SURF feature points based on the visible light image, and remove false matching points through geometric consistency verification; identify high-temperature mutation regions for the infrared image, locate the support contour in combination with edge detection, calculate the depth of feature points based on binocular vision, and generate 3D space coordinates of the support;

[0036] Calculate the initial attitude angle by integrating the angular velocity of the inertial measurement unit, compare it with the pitch / yaw angle estimated by vision, and dynamically compensate the integral drift according to the attitude deviation between vision and the inertial measurement unit, and output the stable attitude angle.

[0037] As a preferred scheme of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method, the output accurate pose of the target support relative to the unmanned aerial vehicle comprises:

[0038] Real-time monitoring of the environmental electromagnetic field intensity, based on the electromagnetic field weighted fusion weight distribution for radar, vision and inertial measurement unit;

[0039] The support coordinates extracted by the radar and the 3D coordinates reconstructed by the vision are weighted and averaged according to the assigned fusion weight, and the three-dimensional coordinates of the support relative to the unmanned aerial vehicle are fused and output;

[0040] The support attitude angle estimated by vision and the corrected attitude angle of the inertial measurement unit are fused according to the assigned fusion weight, and the pitch angle and yaw angle of the support are output;

[0041] The fused support position is converted from the camera coordinate system to the unmanned aerial vehicle body coordinate system, and the camera installation position offset is compensated; based on this, the three-dimensional position and relative attitude angle of the support relative to the unmanned aerial vehicle are output.

[0042] The beneficial effects of the preferred technical scheme are: realizing high-precision recognition and pose calculation of the insulator support target, and effectively overcoming the performance limitations of single sensor in a strong electromagnetic environment.

[0043] In the second aspect, the present application provides a rapid docking and separation system of unmanned aerial vehicle and insulator detection mechanism, comprising:

[0044] The data acquisition module is used for acquiring electromagnetic field intensity, visible light / infrared image, millimeter wave radar and inertial measurement unit raw data as basic perception data;

[0045] A processing and identifying module is configured to perform optimization processing on the basic perception data, combine feature matching, point cloud registration and infrared temperature profile information, realize high-precision identification of the insulator support target, and output the accurate pose of the target support relative to the unmanned aerial vehicle.

[0046] A judgment and control module is configured to calculate the distance between the unmanned aerial vehicle and the target support in real time based on the accurate pose of the target support relative to the unmanned aerial vehicle and the known three-dimensional map of the power transmission tower, start a three-level decision mechanism according to a distance threshold, output an action control instruction and guide the unmanned aerial vehicle to fly to the target.

[0047] An operation execution module is configured to receive a guide tube contact signal, perform mechanical locking and release operations, and obtain a task execution result.

[0048] A diagnosis and response module is configured to trigger an alarm and diagnose a fault type based on a comparison between a locking / separation timestamp, a position error, an electromagnetic intensity and an execution time timer and a preset threshold, and to obtain a fault processing instruction and an optimized sensor weight through adaptive adjustment of Kalman filtering parameters.

[0049] In a third aspect, the present application provides an electronic device, comprising a memory and a processor; the memory is used to store computer executable instructions, and the processor realizes the steps of the method for rapid docking and separation of the unmanned aerial vehicle and the insulator detection mechanism when executing the computer executable instructions.

[0050] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions realize the steps of the method for rapid docking and separation of the unmanned aerial vehicle and the insulator detection mechanism when executed by a processor.

[0051] Compared with the prior art, the present application provides a UAV and insulator detection mechanism rapid docking and separation method, system, device and medium, a comprehensive basic perception system is constructed by collecting electromagnetic field intensity, visible light / infrared image, millimeter wave radar and IMU and other sensor data. On this basis, the multi-source data is optimized by using a dynamic weight fusion algorithm, and combined with SIFT feature matching, ICP point cloud registration and infrared temperature contour information, high-precision recognition and pose calculation of the insulator support target are realized, and the performance limitations of single sensor in strong electromagnetic environment are effectively overcome. At the same time, based on the three-dimensional map of the transmission tower and the three-level decision mechanism, the distance between the UAV and the target support is calculated in real time and the action control instruction is output, so as to ensure the flight stability and precise docking ability of the UAV in the complex electromagnetic environment. In addition, by receiving the contact signal of the guide tube and executing mechanical locking and releasing, combined with real-time monitoring and comparison of parameters such as locking / separation timestamp, position error and electromagnetic intensity, the decision tree is used to diagnose the fault type, and the Kalman filter parameter is used to adaptively adjust and optimize the sensor weight table, so as to further improve the reliability and robustness of the system in the complex electromagnetic environment. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0053] Figure 1 The overall flow logic diagram of the UAV and insulator detection mechanism rapid docking and separation method described in an embodiment of the present application.

[0054] Figure 2 The brief step block diagram of the UAV and insulator detection mechanism rapid docking and separation method described in an embodiment of the present application.

[0055] Figure 3 The three-level decision mechanism schematic diagram of the UAV and insulator detection mechanism rapid docking and separation method described in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0057] Embodiment 1, Reference Figure 1 For an embodiment of the present application, a method for rapid docking and separation of a UAV and an insulator detection mechanism is provided, as shown in Figure 1 Specifically includes the following steps:

[0058] S100: Collect electromagnetic field intensity, visible light / infrared image, millimeter wave radar and inertial measurement unit raw data as basic perception data;

[0059] S200: Optimize the basic perception data, combine feature matching, point cloud registration and infrared temperature contour information, realize high-precision identification of the insulator support target, and output the accurate pose of the target support relative to the UAV;

[0060] S300: Based on the accurate pose of the target support relative to the UAV and the known three-dimensional map of the power transmission tower, the distance between the UAV and the target support is calculated in real time, the three-level decision mechanism is started according to the distance threshold, the action control instruction is output and the UAV is guided to the target;

[0061] S400: Receive the guide tube contact signal, perform mechanical locking and release operation, and obtain the task execution result;

[0062] S500: Trigger alarm by comparing preset threshold based on locking / separation timestamp, position error, electromagnetic intensity and execution time timer, diagnose fault type, and adjust sensor weight through Kalman filter parameters to obtain fault handling instruction and optimized sensor weight.

[0063] It should be noted that, in order to solve the technical problem that the current UAV is difficult to stabilize positioning near strong magnetic field, the above steps S100-S500 collect electromagnetic field intensity, visible light / infrared image, millimeter wave radar and IMU and other sensor data to build a comprehensive basic perception system. On this basis, a dynamic weight fusion algorithm is used to optimize the multi-source data, and combined with SIFT feature matching, ICP point cloud registration and infrared temperature contour information, high-precision identification and pose calculation of the insulator support target are realized, effectively overcoming the performance limitations of single sensor in strong electromagnetic environment. At the same time, based on the three-dimensional map of the power transmission tower and the three-level decision mechanism, the distance between the UAV and the target support is calculated in real time and the action control instruction is output, ensuring the flight stability and precise docking ability of the UAV in complex electromagnetic environment. In addition, by receiving the guide tube contact signal and performing mechanical locking and release, combining real-time monitoring and comparison of locking / separation timestamp, position error, electromagnetic intensity and other parameters, using decision tree to diagnose fault type, and adjusting sensor weight table through Kalman filter parameters, the reliability and robustness of the system in complex electromagnetic environment are further improved.

[0064] Embodiment 2, refer to Figure 2 and Figure 3 Based on the previous embodiment, the embodiment provides a specific implementation of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method, which is used to illustrate the technical means adopted in the method.

[0065] In the embodiment of the present application, the above-mentioned step S100 collects electromagnetic field intensity, visible light / infrared image, millimeter wave radar and inertial measurement unit raw data as basic perception data, including:

[0066] Specifically, the radar can effectively measure the target distance by transmitting a 77GHz frequency-modulated continuous wave (FMCW) with a bandwidth of 1GHz and the ability to resist electromagnetic interference. The detection distance is from 0.1m to 5m, and the anti-interference degree to battery interference reaches 100V / m.

[0067] Specifically, a visual system based on a visible light camera (1920x1080 resolution, 30fps) and an infrared thermal imager (temperature range -20℃ to 150℃, accuracy ±2℃) is used to obtain environmental image data, help extract feature points and perform three-dimensional positioning.

[0068] Specifically, the inertial measurement unit IMU has a 3-axis accelerometer and a 3-axis gyroscope, with a zero offset stability less than 0.5° / h. The IMU provides angular velocity and acceleration information for positioning and is the basis for dynamic attitude estimation. Based on the above-mentioned collected data, the time stamp of the sensor is aligned using a hardware trigger signal to ensure that the sensor data is synchronized in time, and the synchronization error is controlled within 1ms.

[0069] It should be noted that the above-mentioned step S100 constructs a multi-source perception system, effectively overcoming the data loss or distortion problem of a single sensor in a strong electromagnetic environment, providing comprehensive and reliable basic data support for subsequent target recognition and pose calculation, and significantly improving the perception robustness of the system in a complex electromagnetic environment.

[0070] In the embodiment of the present application, as shown in Figure 2 The above-mentioned step S200 optimizes the basic perception data, combines feature matching, point cloud registration and infrared temperature contour information, realizes high-precision identification of the insulator support target, and outputs the accurate pose of the target support relative to the unmanned aerial vehicle, including the following sub-steps B1-B2:

[0071] In B1: the optimization processing of the basic perception data includes:

[0072] Analyze the meridian velocity of each target in the radar point cloud, filter out dynamic points with a speed exceeding a threshold, and retain static target point clouds; extract the coarse-grained three-dimensional coordinates of the insulator support from the static target point clouds;

[0073] SURF feature points are extracted based on the visible light image, and false matching points are removed through geometric consistency verification; a high-temperature mutation region is identified for the infrared image, a bracket contour is located through edge detection, feature point depth is calculated based on binocular vision, and 3D space coordinates of the bracket are generated;

[0074] An initial attitude angle is calculated by integrating the angular velocity of an inertial measurement unit, and is compared with a pitch / yaw angle estimated by vision, and integral drift is dynamically compensated according to the attitude deviation between vision and the inertial measurement unit, and a stable attitude angle is output.

[0075] In an optional implementation, the optimized processing of the basic perception data can further include: directly performing semantic segmentation on the radar point cloud by using a PointNet++ or VoxelNet network, automatically identifying and filtering out dynamic targets through end-to-end training, and outputting accurate three-dimensional coordinates of the insulator bracket, thereby avoiding precision loss of traditional threshold filtering.

[0076] In another optional implementation, the optimized processing of the basic perception data can further include: projecting a YOLOv5 detection box of a visible light image, a hot spot region of an infrared image, and a radar point cloud to a unified coordinate system, fusing target position information of the three types of sensors through Kalman filtering, and jointly optimizing 3D coordinates and attitude angles by using graph optimization, thereby improving the robustness of spatial positioning.

[0077] In B2, the accurate pose of the target bracket relative to the unmanned aerial vehicle includes:

[0078] The electromagnetic field intensity of the environment is monitored in real time, and fusion weights are assigned to the radar, vision, and inertial measurement unit based on the electromagnetic field;

[0079] The bracket coordinates extracted by the radar and the 3D coordinates reconstructed by the vision are weighted and averaged according to the assigned fusion weights, and the three-dimensional coordinates of the bracket relative to the unmanned aerial vehicle are fused and output;

[0080] The bracket attitude angles estimated by the vision and the attitude angles corrected by the inertial measurement unit are fused according to the assigned fusion weights, and the pitch angle and the yaw angle of the bracket are output;

[0081] The fused bracket position is converted from the camera coordinate system to the unmanned aerial vehicle body coordinate system, and the camera installation position offset is compensated; based on this, the three-dimensional position and the relative attitude angle of the bracket relative to the unmanned aerial vehicle are output.

[0082] Specifically, the point cloud of the insulator bracket is separated from the radar point cloud data, and noise introduced by electromagnetic interference of the power transmission line is effectively printed:

[0083]

[0084] P filteredP represents the radar point cloud data obtained after the filtering operation radar P represents the point cloud data returned by the original radar sensor P represents an indicator function for determining whether the point cloud data is a static target; if the distance change rate of the target is less than a set speed threshold v th , the point cloud data is retained, otherwise it is deleted; if is true, the indicator function is 1, indicating that the point cloud is retained, otherwise it is 0 and discarded; p represents the distance between the target object and the radar; represents the derivative of the target distance p with respect to time t;

[0085] Specifically, for binocular vision features, the SURF (Speeded Up Robust Features) algorithm is used for feature point detection, and the RANSAC algorithm is used to remove false matching points; for infrared images, edge detection is performed by temperature gradient calculation to identify the local overheating area of the insulator as a feature point; the matching point pairs in the visible light binocular image are stereo matched to estimate the three-dimensional coordinates:

[0086]

[0087] where f represents the camera focal length, b represents the binocular camera baseline length, d disp represents the disparity of the left and right images, and the spatial position p vis is obtained by combining the pixel position and depth in the image v v v T ;

[0088] In an optional embodiment, the feature point detection can also use the ORB algorithm, combined with improved FAST corner detection and rotation-invariant BRIEF descriptor, which significantly improves the calculation efficiency while maintaining similar SURF robustness, and is more suitable for unmanned aerial vehicle vision systems with high real-time requirements.

[0089] In another optional embodiment, the feature point detection can also use a deep learning method, which jointly outputs feature point positions and descriptors through self-supervised training of an end-to-end network, can directly extract high-repetition feature points from visible light / infrared images, has stronger adaptability to light changes and noise, and supports feature matching of heterogeneous sensors.

[0090] Specifically, for IMU, the angular velocity data of the IMU is integrated to obtain the attitude angle, and the closed-loop correction is performed using the angle estimated by vision:

[0091]

[0092] where θ​​​vis represents the pitch / yaw angle of visual estimation, K represents the Kalman gain matrix, and θ imu (t) represents the attitude angle changing over time, and ω imu represents the angular velocity data measured by the IMU sensor, θ0 represents the initial attitude angle, and θ vis represents the pitch / yaw angle of visual estimation, and θ imu represents the measured attitude angle.

[0093] Specifically, the electromagnetic environment adaptive fusion positioning includes:

[0094] According to the real-time electric field intensity, the weights of each sensor are dynamically adjusted:

[0095]

[0096] wherein, E RMS represents the effective value of the real-time electromagnetic field intensity, and a represents the attenuation coefficient, which is set to 0.001. If the electromagnetic field intensity E RMS is greater than 50 V / m, the weight of vision will increase.

[0097] Exemplarily, the allocation rules are as follows:

[0098] When E RMS < 30 V / m, the weight is: w radar = 0.6, w vis = 0.3, and w imu = 0.1.

[0099] When E RMS ≥ 30 V / m, the weight is: w radar = 0.3, w vis = 0.6, and w imu = 0.1.

[0100] wherein, w radar represents the fusion weight of radar data, w vis represents the fusion weight of visual data, and w imu represents the fusion weight of the IMU sensor.

[0101] Specifically, the fusion position estimation is: p fuse = w radar · p radar + w vis · p vis .

[0102] wherein, p fuse represents the fused 3D position coordinates.

[0103] It should be noted that the above step S200 realizes high-precision recognition and pose calculation of the insulator support target, solves the problem of inaccurate positioning of a single sensor such as a traditional vision or radar sensor under strong magnetic field interference, and provides key spatial attitude information for precise approach of the unmanned aerial vehicle to the target.

[0104] In the embodiment of the present application, the above step S300 calculates the distance between the unmanned aerial vehicle and the target support in real time based on the accurate pose of the target support relative to the unmanned aerial vehicle and the known three-dimensional map of the power transmission tower, and starts a three-level decision mechanism according to the distance threshold value, outputs the action control instruction and guides the unmanned aerial vehicle to fly to the target, including:

[0105] Specifically, as shown in Figure 3 starting a three-level decision mechanism according to the distance threshold value, including:

[0106] When the distance between the unmanned aerial vehicle and the target support is greater than the first threshold value, a far-field control stage is started;

[0107] When the distance between the unmanned aerial vehicle and the target support is between the first threshold value and the second threshold value, a near-field control stage is started;

[0108] When the distance between the unmanned aerial vehicle and the target support is less than the second threshold value, a critical docking control stage is started.

[0109] It should be noted that in the embodiment, the first threshold value is 2m and the second threshold value is 0.5m.

[0110] It should be noted that the first threshold value is determined based on the stable detection range of the conventional sensor of the unmanned aerial vehicle (such as the effective recognition distance of the RGB camera being 1.5-3m) and the response delay of the flight control system, to ensure that there is enough space for trajectory correction in the far-field stage; the second threshold value is set according to the physical operation radius of the mechanical arm / docking mechanism and the air disturbance influence range, to avoid collision risk caused by air disturbance in the near-field stage. Both are verified by experiments, and when the positioning error is <5cm at 2m and the control accuracy is ±1cm at 0.5m, the efficiency and safety can be balanced.

[0111] In the embodiment of the present application, the far-field control stage (>2m) includes:

[0112] a first planning algorithm is used for path planning, and a cost function of the path planning is calculated;

[0113] Considering the actual movement cost, the Euclidean distance and the influence of the electromagnetic field strength, the weight factor is adjusted according to the real-time change of the electromagnetic field strength, and the path selection is optimized;

[0114] The normal vector of the target support is calculated through the pitch angle and the yaw angle of the target support;

[0115] By comparing the current attitude of the unmanned aerial vehicle and the target attitude, the angle difference to be adjusted is calculated, and the yaw and pitch angle velocity instructions of the unmanned aerial vehicle are generated to adjust the attitude of the unmanned aerial vehicle, so that the unmanned aerial vehicle gradually aligns with the target.

[0116] Specifically, a first planning algorithm, i.e., an A* algorithm, is used to calculate the optimal path from the current position of the unmanned aerial vehicle to the target point, avoiding the guide lines and the tower, wherein the cost function is as follows:

[0117] f(n)=g(n)+h(n)+λ·E map (n)

[0118] Wherein, g(n) represents the actual moving cost from the starting point to node n, h(n) represents the Euclidean distance from node n to the target point, λ represents the electromagnetic field path weight factor, and E map (n) represents the electric field intensity at node n.

[0119] Specifically, the normal vector of the support is calculated The normal vector of the support is derived from the pitch angle and the yaw angle of binocular vision; the attitude of the unmanned aerial vehicle is represented by a quaternion q drone , which is converted into Euler angles (θ d , ψ d ):

[0120] Δθ=k p ·(θ v -θ d ), Δψ=k p ·(ψ v -ψ d )

[0121] Wherein, k p represents the attitude control proportion coefficient, θ v represents the pitch angle of the target support, (θ d , ψ d ) represents the pitch angle and the yaw angle of the unmanned aerial vehicle, ψ v represents the yaw angle of the target support, Δθ represents the error correction value of the pitch angle, and Δψ represents the error correction value of the yaw angle.

[0122] Specifically, the yaw angle velocity instruction of the unmanned aerial vehicle is obtained The pitch angle velocity instruction of the unmanned aerial vehicle is obtained

[0123] In an optional embodiment, the first planning algorithm can also be an improved RRT* (rapidly-exploring random tree star) algorithm, by introducing a dynamic sampling strategy and an electromagnetic field strength constraint condition, evaluating the node electromagnetic environment risk value in real time during the random tree expansion process, preferentially selecting a low field strength area for path expansion, and simultaneously using a gradually optimal rewiring mechanism to optimize path smoothness.

[0124] In another optional embodiment, the first planning algorithm can also be an artificial potential field method combined with an electromagnetic field model, constructing a composite potential field function including a wire repulsive potential field, a target point attractive potential field and an electromagnetic field strength constraint potential field, solving the optimal path by a gradient descent method, and introducing a dynamic adjustment coefficient to automatically adjust the weight of each potential field according to the real-time electromagnetic field strength change.

[0125] In the embodiments of the present application, the near-field control stage (0.5m-2m) includes:

[0126] The feature point data of the binocular vision system is used to determine the accurate docking point coordinates of the target support;

[0127] The accurate docking point coordinates of the target support are converted to the coordinate system of the UAV, and a position deviation vector relative to the UAV is calculated;

[0128] A speed control instruction is generated based on the position deviation vector to adjust the position of the UAV;

[0129] An angular velocity control instruction is generated based on the attitude deviation to adjust the attitude of the UAV;

[0130] The three-axis speed instruction and the three-axis angular velocity instruction of the UAV are output.

[0131] Specifically, the pose of the UAV is composed of a rotation matrix and a translation vector, representing the attitude of the UAV relative to the world coordinate system;

[0132] The pose deviation matrix is represented as:

[0133]

[0134] wherein, represents the position of the target point in the coordinate system of the UAV, R -1 represents the inverse of the rotation matrix, P vision represents the support docking point coordinates, and t represents the position of the UAV;

[0135] The deviation vector is represented as:

[0136]

[0137] wherein, P ref represents the desired docking point position;

[0138] The angle deviation is expressed as:

[0139]

[0140] wherein R target represents the rotation matrix of the target point, R drone represents the rotation matrix of the UAV, and Δθ represents the angle deviation between the target and the UAV.

[0141] Specifically, the fine-tuning instruction is generated based on a PID controller:

[0142]

[0143] wherein k p = 1.2; k d = 0.3; k i = 0.05; and v cmd represents the speed instruction of the UAV.

[0144] Angle control:

[0145]

[0146] wherein k p,θ = 0.9, k d,θ = 0.2, which are the proportional and differential gains of the angle control, represents the rate of change of the pitch angle error, and Δθ represents the error between the target angle and the current angle.

[0147] In the embodiments of the present application, the critical docking control stage (<0.5 m) includes:

[0148] Detecting whether the position and attitude deviation satisfies the triggering condition of the pneumatic guiding mechanism, and starting the pneumatic guiding mechanism if the triggering condition is satisfied;

[0149] According to the dynamics model of the pneumatic system, the extension force of the guiding tube is calculated to ensure that the guiding tube is smoothly extended;

[0150] The contact force is monitored in real time by a pressure sensor, the impedance control mode is entered, the position correction amount is calculated according to the deviation between the contact force and the expected force, the position of the UAV is adjusted, and the position correction instruction is output.

[0151] Specifically, the triggering condition of the pneumatic guiding mechanism is that when ||ΔP||<0.05 m and Δθ<5°, the pneumatic guiding mechanism is triggered.

[0152] Specifically, the extension dynamics model of the guiding tube is as follows:

[0153] F extend = P air ·A piston -k springx-b-v

[0154] wherein P air represents the pressure of compressed air, A piston represents the cross-sectional area of the piston, k spring represents the stiffness of the return spring, b represents the damping coefficient, and v represents the speed of the guide tube during extension, F extend represents the extension force acting on the guide tube during extension of the guide tube;

[0155] During docking, the contact force F contact is sensed by the force sensor and fed back to the system, and through impedance control, the system converts the contact force into fine-tuning instructions to adjust the extension speed of the guide tube or the unknown:

[0156] Delta X adjust = K virtual * F contact

[0157] wherein Delta X adjust represents the adjusted position deviation, K virtual represents the virtual stiffness coefficient, and F contact represents the contact force;

[0158] The fine force control system adjusts the contact force depending on the real-time feedback of the guide tube extension process, and when the guide tube is about to contact the target, the changes in damping and spring force will affect F extend , thereby affecting the change in contact force, and the fine force control system adjusts in real time according to the change to ensure that the guide tube generates appropriate docking pressure when reaching the target position; by monitoring the relationship between the extension force and the contact force in real time, the fine force control can accurately adjust the extension process of the guide tube, avoid excessive or insufficient contact force, and thus achieve flexible docking.

[0159] It should be noted that the above step S300 calculates the distance in real time based on the accurate pose and the three-dimensional map of the power transmission tower, and dynamically adjusts the flight strategy through a three-level decision mechanism, to ensure that the unmanned aerial vehicle can still fly stably and approach the target accurately under strong electromagnetic interference, significantly reduce the collision risk, and improve the reliability of autonomous navigation in complex environments.

[0160] In the embodiments of the present application, the above step S400 receives the guide tube contact signal, performs mechanical locking and release operations, and obtains the task execution result, including:

[0161] In the critical docking stage, when the unmanned aerial vehicle is less than 0.5 meters away from the insulator support, the aerodynamic guide mechanism is started, and the flexible conical guide tube is extended, and the extension length of the guide tube is calculated by the following formula:

[0162] L = L0 + k d

[0163] Wherein, L represents the guide tube extension length; L0 represents the guide tube initial length; k represents the extension proportion coefficient; d represents the distance between the unmanned aerial vehicle and the support;

[0164] The material of the guide tube is conductive rubber, which has good wear resistance and anti-electromagnetic interference performance. The contact signal is detected by the pressure sensor. When the guide tube contacts the support, the pressure sensor outputs a signal to trigger the pneumatic piston. The starting piston is driven by compressed air, and the driving force is calculated by the following formula:

[0165] F=P·A

[0166] Wherein, F represents the driving force, P represents the compressed air pressure, and A represents the piston area;

[0167] The pneumatic piston pushes the lock tongue into the support slot, and the displacement of the lock tongue is monitored by the displacement sensor in real time. When the displacement of the lock tongue reaches the preset value, the displacement sensor outputs a signal. The locking state is verified by the pressure and position sensors. When the pressure sensor detects that the pressure P lock ≥50N and the displacement sensor detects that the displacement S lock ≥S threshold , the locking is successful, wherein S threshold represents the displacement threshold of the locking position;

[0168] After the locking state verification is successful, the system sends a locking confirmation signal to the unmanned aerial vehicle, and the unmanned aerial vehicle stops moving and enters a stable state. When the unmanned aerial vehicle completes the detection task, it sends a separation command to trigger the pneumatic valve to release pressure, and the release speed is controlled by the following formula:

[0169]

[0170] Wherein, k p represents the release rate constant;

[0171] After the pneumatic valve is released, the spring mechanism quickly retracts the lock tongue, and the retraction time is determined by the spring constant and mass. During the retraction of the lock tongue, the displacement sensor monitors its position in real time. When the lock tongue is completely retracted to the position, the displacement sensor outputs a separation completion signal. After receiving the separation completion signal, the unmanned aerial vehicle immediately leaves, completing the separation process.

[0172] The abnormal handling of the docking process includes: if the locking confirmation signal is not received within the preset time, the system triggers the unmanned aerial vehicle to retreat strategy, and restarts the docking process.

[0173] The abnormal handling of the separation process: if the lock tongue is not completely retracted to the position, the system starts the standby battery iron module to temporarily absorb the lock tongue and force it to retract.

[0174] It should be noted that the above step S400 triggers the mechanical locking and releasing operation through the guide tube contact signal, realizes the physical docking of the unmanned aerial vehicle and the target support, completes the preset task action, and the modular design takes into account the operation accuracy and execution efficiency, and provides reliable execution result feedback for subsequent fault diagnosis.

[0175] In the embodiment of the application, the above step S500 triggers an alarm based on the locking / separation timestamp, position error, electromagnetic intensity and execution time timer compared with the preset threshold, and diagnoses the fault type, and the fault processing instruction and the optimized sensor weight are obtained through adaptive adjustment of the Kalman filter parameter, including:

[0176] Specifically, the decision tree fault diagnosis is used to detect and handle abnormal situations during docking and separation, to ensure that the system can respond in time when problems occur, including:

[0177] Locking timeout: trigger the unmanned aerial vehicle retreat strategy, adjust the attitude to move away from the insulator support, restart the docking process, re-perform sensing and positioning, and if multiple attempts fail, record fault information and return to standby state;

[0178] Separation failure: start the standby electromagnet strong retreat module, short-circuit power absorption lock tongue, ensure that it is retracted in place, resend the separation instruction, check the lock tongue state, and if it is not successful, record fault information and return to standby state;

[0179] Record the environmental parameters and system performance of each docking, including electromagnetic intensity, positioning error, execution time, locking state and separation state;

[0180] Optimize the sensor filter parameter using an adaptive filter algorithm, and the calculation formula of the filter weight w is as follows:

[0181] w new =w old -α(d-d target )

[0182] Wherein, α represents the learning rate, d represents the current positioning error, d target represents the expected positioning error; w old represents the filter parameter before updating, w new represents the updated filter parameter;

[0183] The new filter weight calculated is updated to the filter of the millimeter wave radar sensor, and in the next signal processing, the filter uses the updated filter weight to more effectively denoise and process the signal, improving detection accuracy and stability.

[0184] It should be noted that the above step S500 is combined with the multi-parameter threshold comparison and the decision tree fault diagnosis to monitor the system state in real time and identify the abnormality, the Kalman filter parameter is adaptively adjusted to optimize the sensor weight, the system anti-interference ability is dynamically improved, the stability and fault tolerance of the task execution are ensured, and the failure rate in the strong magnetic field environment is effectively reduced.

[0185] In embodiment 3, a UAV and insulator detection mechanism rapid docking and separation system is provided, comprising:

[0186] A data acquisition module is configured to acquire electromagnetic field intensity, visible light / infrared image, millimeter wave radar and inertial measurement unit raw data as basic perception data.

[0187] A processing and recognition module is configured to optimize the basic perception data, combine feature matching, point cloud registration and infrared temperature contour information, realize high-precision recognition of the insulator support target, and output the accurate pose of the target support relative to the UAV.

[0188] A judgment and control module is configured to calculate the distance between the UAV and the target support in real time based on the accurate pose of the target support relative to the UAV and the known three-dimensional map of the power transmission tower, start a three-level decision mechanism according to the distance threshold, output the action control instruction and guide the UAV to fly to the target.

[0189] An operation execution module is configured to receive the guide tube contact signal, execute the mechanical locking and release operation, and obtain the task execution result.

[0190] A diagnosis response module is configured to trigger an alarm and diagnose the fault type by comparing the locking / separation timestamp, position error, electromagnetic intensity and execution time timer with the preset threshold, and adaptively adjust the Kalman filter parameter to obtain the fault handling instruction and the optimized sensor weight.

[0191] It should be noted that the technical scheme of the UAV and insulator detection mechanism rapid docking and separation system belongs to the same concept as the technical scheme of the above-mentioned UAV and insulator detection mechanism rapid docking and separation method. The technical scheme of the UAV and insulator detection mechanism rapid docking and separation system in this embodiment is not described in detail, and the description of the technical scheme of the above-mentioned UAV and insulator detection mechanism rapid docking and separation method can be referred to.

[0192] The above-mentioned unit modules can be embedded in or independent of the processor in the electronic device in hardware form, or stored in the memory in the electronic device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned modules.

[0193] The embodiment also provides an electronic device, which comprises a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, an operator network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement the method for rapid docking and separation of the unmanned aerial vehicle and the insulator detection mechanism. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen. The input device of the electronic device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the electronic device, or an external keyboard, touchpad or mouse, etc.

[0194] The embodiment also provides a computer readable storage medium, which stores a computer program. The program is executed by a processor to implement the method proposed in the above embodiment.

[0195] The storage medium proposed in the embodiment and the method proposed in the above embodiment belong to the same inventive concept. The technical details not described in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0196] From the above description about the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary universal hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solutions of the present application or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH memory, a hard disk or an optical disk, etc., and includes a number of instructions to make an electronic device (which can be a personal computer, a server or a network device, etc.) execute the method of the embodiment of the present application.

[0197] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A method for rapid docking and separation of a drone and an insulator detection mechanism, characterized in that, The method comprises the following steps: Collecting electromagnetic field intensity, visible light / infrared image, millimeter wave radar and inertial measurement unit raw data as basic perception data; Optimizing the basic perception data, combining feature matching, point cloud registration and infrared temperature contour information to realize high-precision identification of insulator support target and output accurate pose of the target support relative to the unmanned aerial vehicle; Based on the accurate pose of the target support relative to the unmanned aerial vehicle and the known three-dimensional map of the power transmission tower, the distance between the unmanned aerial vehicle and the target support is calculated in real time, and a three-level decision mechanism is started according to the distance threshold value to output action control instructions and guide the unmanned aerial vehicle to fly to the target; Receiving the contact signal of the guiding tube, executing the mechanical locking and releasing operation, and obtaining the task execution result; Based on the locking / separation timestamp, position error, electromagnetic intensity and execution time timer, comparing with the preset threshold value to trigger alarm and diagnose fault type, and through Kalman filter parameter adaptive adjustment to obtain fault handling instructions and optimized sensor weight.

2. The method of claim 1, wherein the drone and the insulator detection mechanism are quickly coupled and decoupled. The three-level decision mechanism according to the distance threshold value comprises: When the distance between the unmanned aerial vehicle and the target support is greater than the first threshold value, the far-field control stage is started; When the distance between the unmanned aerial vehicle and the target support is between the first threshold value and the second threshold value, the near-field control stage is started; When the distance between the unmanned aerial vehicle and the target support is less than the second threshold value, the critical docking control stage is started.

3. The method of claim 2, wherein the drone and the insulator detection mechanism are quickly coupled and decoupled. The far-field control stage comprises: Using a first planning algorithm to plan a path and calculate a cost function of the path planning; Considering the actual moving cost, Euclidean distance and electromagnetic field intensity, adjusting the weight factor according to the real-time change of the electromagnetic field intensity to optimize the path selection; Calculating the normal vector of the target support through the pitch angle and yaw angle of the target support; By comparing the current attitude and target attitude of the unmanned aerial vehicle, calculating the angle difference to be adjusted, and generating yaw and pitch angle velocity instructions of the unmanned aerial vehicle to adjust the attitude of the unmanned aerial vehicle, so that the unmanned aerial vehicle gradually aligns with the target.

4. The method of claim 3, wherein the drone and the insulator detection mechanism are quickly coupled and decoupled. The near-field control stage comprises: Using feature point data of a binocular vision system to determine the accurate docking point coordinates of the target support; Converting the accurate docking point coordinates of the target support to the coordinate system of the unmanned aerial vehicle to calculate the position deviation vector relative to the unmanned aerial vehicle; Generating a speed control instruction based on the position deviation vector to adjust the position of the unmanned aerial vehicle; Generating an angular velocity control instruction based on the attitude deviation to adjust the attitude of the unmanned aerial vehicle; Outputting the three-axis speed instruction and three-axis angular velocity instruction of the unmanned aerial vehicle.

5. The method of claim 4, wherein the drone and the insulator detection mechanism are quickly coupled and decoupled. The critical docking control stage comprises: Detecting whether the position and attitude deviation meets the triggering condition of the aerodynamic guiding mechanism, and if it meets, starting the aerodynamic guiding mechanism; According to the dynamics model of the aerodynamic system, calculating the extension force of the guiding tube to ensure the smooth extension of the guiding tube; Through the real-time monitoring of the contact force by the pressure sensor, entering the impedance control mode, calculating the position correction amount according to the deviation between the contact force and the expected force, adjusting the position of the unmanned aerial vehicle, and outputting the position correction instruction.

6. The method of claim 1, wherein the drone and insulator detection mechanism are quickly coupled and decoupled. The optimization processing of the basic perception data comprises: Analyze the meridional velocity of each target in the radar point cloud, filter out dynamic points with a speed exceeding a threshold, and retain the static target point cloud; extract coarse-grained three-dimensional coordinates of the insulator support from the static target point cloud; Extract SURF feature points based on the visible light image, and remove false matching points through geometric consistency verification; identify high-temperature mutation regions for the infrared image, locate the support contour through edge detection, calculate the depth of the feature points based on binocular vision, and generate 3D space coordinates of the support; Integrate the initial attitude angle calculated by the angular velocity of the inertial measurement unit, compare it with the pitch / yaw angle estimated by vision, and dynamically compensate the integral drift according to the attitude deviation between vision and the inertial measurement unit to output a stable attitude angle.

7. The method of claim 6, wherein the drone and the insulator detection mechanism are quickly coupled and decoupled. The accurate pose of the target support relative to the unmanned aerial vehicle includes: Real-time monitoring of the environmental electromagnetic field intensity, and assigning fusion weights to the radar, vision, and inertial measurement unit based on the electromagnetic field weighting; Weighting and averaging the support coordinates extracted by the radar and the 3D coordinates reconstructed by vision according to the assigned fusion weights, and fusing to output the three-dimensional coordinates of the support relative to the unmanned aerial vehicle; Fusing the support attitude angle estimated by vision and the attitude angle corrected by the inertial measurement unit according to the assigned fusion weights, and outputting the pitch angle and yaw angle of the support; Converting the fused support position from the camera coordinate system to the unmanned aerial vehicle body coordinate system to compensate for the camera installation position offset; based on this, outputting the three-dimensional position and relative attitude angle of the support relative to the unmanned aerial vehicle.

8. The unmanned aerial vehicle and insulator detection mechanism rapid docking and separation system, applying the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method according to any one of claims 1-7, characterized in that, It includes: A data acquisition module for acquiring electromagnetic field intensity, visible light / infrared images, millimeter wave radar, and inertial measurement unit raw data as basic perception data; A processing and recognition module for optimizing the basic perception data, combining feature matching, point cloud registration, and infrared temperature contour information to achieve high-precision recognition of insulator support targets, and outputting the accurate pose of the target support relative to the unmanned aerial vehicle; A judgment and control module for calculating the distance between the unmanned aerial vehicle and the target support in real time based on the accurate pose of the target support relative to the unmanned aerial vehicle and the known three-dimensional map of the power transmission tower, starting a three-level decision mechanism according to the distance threshold, outputting action control instructions, and guiding the unmanned aerial vehicle to fly towards the target; An operation execution module for receiving a contact signal of the guide tube, performing mechanical locking and release operations, and obtaining task execution results; A diagnosis response module for triggering an alarm by comparing preset threshold values based on locking / separation timestamps, position errors, electromagnetic intensities, and execution time timers, diagnosing fault types, and adjusting fault handling instructions and optimized sensor weights through Kalman filtering parameters. 9.An electronic device comprising a memory and a processor, the electronic device characterized by: The memory is used to store computer executable instructions, and the processor executes the computer executable instructions to realize the steps of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method of any one of claims 1-7.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: The computer executable instructions are executed by the processor to realize the steps of the unmanned aerial vehicle and insulator detection mechanism rapid docking and separation method of any one of claims 1-7.

Images