Power transmission line simulation flight intelligent inspection method based on three-dimensional laser radar
Through intelligent inspection methods based on three-dimensional lidar, three-dimensional spatial point cloud data and visual image data are collected and processed in real time, wires and tree barriers are identified, attitude dynamic response and flight collision factors of the drone, and flight control is performed, which solves the problem that drones are difficult to avoid obstacles in complex environments in traditional inspection methods, and improves patrol safety and stability.
Patent Information
- Application Number
- CN202510213453.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-26
- Publication Date
- 2025-03-28
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional manual inspection and drone flight inspection are difficult to achieve independent obstacle avoidance flight in complex environments, which poses a collision risk and is difficult to meet the needs of inspection work.
The intelligent inspection method of imitation line flight of power transmission lines based on three-dimensional lidar is adopted. By collecting three-dimensional spatial point cloud data and visual image data in real time, real-time wire identification and tree barrier identification are performed, the attitude dynamic response and flight collision factors of patrol drones are calculated, and flight control is performed based on these data.
It has realized the collision risks of patrol drones in advance during the inspection process, improved the safety of flight patrols, and enhanced the autonomous decision-making ability and flight stability of the drone.
Smart Images

Figure CN119717881A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of flight inspection control technology, and more specifically, to a transmission line simulation flight intelligent inspection method based on three-dimensional laser radar. Background Art
[0002] As a key facility for power transmission, transmission lines are exposed to the natural environment for a long time and are affected by internal pressure and external aggression, which can easily lead to defects and hidden dangers in equipment such as poles, towers and conductors. Line inspection is the basic work to effectively ensure the safe and stable operation of transmission lines.
[0003] However, the traditional manual inspection method has problems such as low efficiency and great safety hazards, which cannot meet the needs of inspection work. Inspection drones cannot actively perceive changes in the on-site environment during flight inspections, and it is difficult to achieve autonomous obstacle avoidance flight. In complex environments such as mountainous areas, collision accidents are prone to occur when encountering obstacles such as intersections and trees, and lack adaptive adjustment capabilities. Therefore, how to pre-identify the collision risk of inspection drones during the inspection process to improve the safety of flight inspections is a difficult problem faced by the industry. Summary of the invention
[0004] The present application provides a transmission line flight intelligent inspection method based on three-dimensional laser radar, which can pre-identify the collision risk of inspection drones during the inspection process to improve the safety of flight inspection.
[0005] The present application provides a transmission line flight intelligent inspection method based on three-dimensional laser radar, the inspection method comprising the following steps: Real-time collection of 3D point cloud data and visual image data within the transmission line area; Conductor real-time identification is performed based on the three-dimensional point cloud data and the visual image data, thereby obtaining the spatial characteristics of the inspection conductor of the transmission line, and the attitude dynamic responsiveness of the inspection drone during the inspection process is determined based on the spatial characteristics of the inspection conductor; Obtaining an initial positioning result of the inspection drone, and determining a relative collision feature between the inspection wire and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial feature; Tree obstacle recognition is performed through the three-dimensional space point cloud data and the visual image data, thereby obtaining inspection tree obstacle information in the transmission line area, and flight collision prediction is performed based on the inspection tree obstacle information to obtain the flight collision factor of the inspection drone; The inspection UAV is controlled in flight based on the relative collision characteristics and the flight collision factor.
[0006] In this embodiment, the laser radar device and visual sensor carried by the inspection drone scan and collect three-dimensional space point cloud data and visual image data in the transmission line area in real time.
[0007] In this embodiment, the wire real-time identification is performed based on the three-dimensional point cloud data and the visual image data, and then the inspection wire spatial features of the transmission line are obtained, which specifically include: Performing wire point cloud extraction on the three-dimensional space point cloud data to obtain a point cloud candidate area; Performing wire structure fitting on the point cloud candidate area to obtain three-dimensional point cloud data of the inspection wire; Marking the boundaries of the wires on the visual image data to obtain a marked visual image; Fusing the three-dimensional point cloud data and the annotated visual image to obtain a wire fused image; The spatial features of the conductor fusion image are extracted to obtain the spatial features of the inspection conductor of the transmission line.
[0008] In this embodiment, the wire point cloud is extracted from the three-dimensional space point cloud data to obtain the point cloud candidate area, which specifically includes: Highly screening the three-dimensional space point cloud data, and then clustering the highly screened three-dimensional space point cloud data to obtain an inspection wire point cloud; Determining the geometric variation of each point in the inspection wire point cloud; Points whose geometric changes are greater than the set change threshold are eliminated to obtain the point cloud candidate area.
[0009] In this embodiment, determining the attitude dynamic responsiveness of the inspection drone during the inspection process according to the inspection wire spatial characteristics specifically includes: Extract all inspection sections of the inspection wire from the inspection wire spatial features; Determine the inspection steering vector of the inspection UAV between adjacent inspection sections; The dynamic responsiveness of the attitude of the inspection UAV during the inspection process is determined through all inspection steering vectors.
[0010] In this embodiment, obtaining the initial positioning result of the inspection drone specifically includes: based on the data of the inertial navigation system and the global navigation satellite system, using a deep coupling positioning algorithm to perform accurate initial positioning of the inspection drone, and then obtaining the initial positioning result of the inspection drone.
[0011] In this embodiment, determining the relative collision characteristics between the inspection wire and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial characteristics specifically includes: Determine the relative spatial characteristics between the inspection wire and the inspection drone according to the initial positioning result and the spatial characteristics of the inspection wire; The relative collision characteristics between the inspection wire and the inspection drone are determined by the attitude dynamic responsiveness and the relative spatial characteristics.
[0012] In this embodiment, tree obstacle recognition is performed using the three-dimensional space point cloud data and the visual image data, which is inputting the three-dimensional space point cloud data and the visual image data into a deep learning target detection algorithm for tree obstacle recognition.
[0013] In this embodiment, flight collision prediction is performed based on the inspection tree obstacle information to obtain the flight collision factor of the inspection drone, which specifically includes: Construct the trajectory equation of the inspection drone; Determine the collision limit distance between the inspection drone and the inspection tree barrier according to the trajectory equation and the inspection tree barrier information; The flight collision factor of the inspection UAV is determined by the collision limit distance and the set inspection safety radius.
[0014] In this embodiment, the flight control of the inspection drone based on the relative collision feature and the flight collision factor specifically includes: determining a collision risk response index according to the relative collision characteristic and the flight collision factor; The inspection UAV is controlled in flight according to the collision risk response index and a given response interval.
[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: The invention collects three-dimensional point cloud data and visual image data in the area of the power transmission line in real time; performs real-time conductor identification based on the three-dimensional point cloud data and the visual image data, thereby obtaining the spatial characteristics of the inspection conductor of the power transmission line, and determines the attitude dynamic responsiveness of the inspection drone during the inspection process based on the spatial characteristics of the inspection conductor; obtains the initial positioning result of the inspection drone, and determines the relative collision characteristics between the inspection conductor and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the spatial characteristics of the inspection conductor; performs tree obstacle identification based on the three-dimensional point cloud data and the visual image data, thereby obtaining the inspection tree obstacle information in the area of the power transmission line, and performs flight collision prediction based on the inspection tree obstacle information, thereby obtaining the flight collision factor of the inspection drone; and performs flight control on the inspection drone based on the relative collision characteristics and the flight collision factor.
[0016] It can be seen that in this application, first, the wire is identified in real time based on the three-dimensional point cloud data and visual image data, and the spatial characteristics and attitude dynamic responsiveness of the inspection wire are calculated based on these data. The inspection UAV can adjust the flight attitude to maintain flight stability according to the changes in the inspection path; then, by determining the relative collision characteristics based on the initial positioning results of the inspection UAV, the attitude dynamic responsiveness and the spatial characteristics of the inspection wire, the relative position changes of the UAV and the wire can be monitored in real time during the flight, and potential collision risks can be identified in time; further, through the calculation of tree obstacle identification and flight collision factors, the inspection UAV's ability to predict and respond to potential collision risks during the inspection process can be significantly improved; finally, by controlling the flight of the inspection UAV based on the collision risk response index and a given response interval, the collision risk can be dynamically evaluated and processed during the flight. The collision risk response index can quantify the collision risk, which can help the flight control system to respond appropriately in different risk environments, thereby improving the flight safety of the inspection UAV in complex environments.
[0017] In summary, the technical solution adopted in this application can pre-identify the collision risk of inspection drones during the inspection process to improve the safety of flight inspections. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0019] Figure 1 It is an exemplary flow chart of the intelligent inspection method for power transmission line simulation flight based on three-dimensional laser radar provided in this application; Figure 2 is an exemplary flow chart for determining the dynamic responsiveness of the attitude of an inspection drone during an inspection process provided by the present application; Figure 3 It is an exemplary flow chart for determining the relative collision characteristics between the inspection wire and the inspection drone provided in the present application. DETAILED DESCRIPTION
[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0021] The embodiment of the present application provides a transmission line simulation flight intelligent inspection method based on three-dimensional laser radar, the core of which is to collect three-dimensional space point cloud data and visual image data in the transmission line area in real time; to identify the wire in real time based on the three-dimensional space point cloud data and the visual image data, and then obtain the spatial characteristics of the inspection wire of the transmission line, and determine the attitude dynamic responsiveness of the inspection drone during the inspection process based on the inspection wire spatial characteristics; to obtain the initial positioning result of the inspection drone, and determine the relative collision characteristics between the inspection wire and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial characteristics; to identify tree obstacles through the three-dimensional space point cloud data and the visual image data, and then obtain the inspection tree obstacle information in the transmission line area, and perform flight collision prediction based on the inspection tree obstacle information to obtain the flight collision factor of the inspection drone; and to control the flight of the inspection drone based on the relative collision characteristics and the flight collision factor. The above scheme can be used to pre-identify the collision risk of the inspection drone during the inspection process to improve the safety of flight inspection.
[0022] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in the figure, this figure is an exemplary flow chart of a transmission line flight intelligent inspection method based on three-dimensional laser radar according to this embodiment of the present application, and the inspection method includes the following steps: In step S1, three-dimensional space point cloud data and visual image data in the transmission line area are collected in real time.
[0023] In specific implementation, the laser radar device and visual sensor carried by the inspection drone scan and collect three-dimensional spatial point cloud data and visual image data in the transmission line area in real time; the laser radar (LiDAR) device carried by the inspection drone usually includes a scanning laser radar and multiple laser emitters, which can scan the environment in all directions. During the laser radar scanning process, the inspection drone will scan the ground and air objects according to the flight trajectory and sensor direction, so as to obtain three-dimensional spatial point cloud data in the transmission line area; the inspection drone is usually also equipped with visual sensors (such as high-definition cameras, infrared cameras, etc.). The inspection drone uses the camera to take real-time photos of the transmission line and the surrounding environment to obtain high-resolution image data, that is, visual image data in the transmission line area.
[0024] The position coding is introduced to describe the point coordinate information and construct the point cloud vector. The point cloud is labeled and classified. The center value of the intensity interval is determined based on the reflection intensity of the denoised point cloud coordinates through the adaptive clustering method, which is expressed as: ; in represents the central value of the j-th interval intensity, The total number of intervals representing the range of reflection intensity, determined based on historical data, and Respectively represent the maximum and minimum values of the reflection intensity; Calculate the number of points in each intensity interval for histogram counting, expressed as: ; ; in The reflection intensity of the interval is The number of point clouds, Represents the total number of point cloud data, represents the i-th reflection intensity value, represents the center value of the jth reflection intensity interval, represents the indicator function, It represents the ratio of the difference between the maximum and minimum values of the reflection intensity to the total number of intervals B; based on Discretize the reflection intensity to form a histogram, and calculate the proportion of the intensity range in the entire point cloud according to the number of point clouds with different reflection intensities based on the total number of point clouds; Based on the sum of the proportion mean and standard deviation as the proportion threshold, the number of interval reflection intensity point clouds greater than the proportion threshold is As the window center for initializing mean shift clustering, the mean shift algorithm is used for adaptive clustering, in which the weighted mean in the window is calculated by the density and function, expressed as: ; in represents the updated kth window center, n represents the total number of reflection intensities, Indicates that the density kernel function uses a Gaussian kernel, , x represents the Gaussian kernel calculation term, represents the kth window center of mean shift clustering, W represents the window width, which is determined based on historical values; Iterate and update the window center. When the difference of the updated window center is no longer obvious during the continuous iteration, stop the iteration and classify the point cloud data into the closest cluster center through the attribution relationship. The Transformer model is used to introduce position coding to describe the point coordinate information, which is encoded using trigonometric functions and expressed as: ; in Represents the position encoding of the i-th point cloud coordinates, , and are the three-dimensional point cloud coordinate information of the i-th point cloud, Represents the frequency parameter, which is usually set to , where d is the feature dimension of the Transformer model, usually set to 64, 128, or 256, and k represents the index of the encoding channel at the current position; Construct a point cloud vector, including point cloud coordinates, reflection intensity, cluster labels, and position encoding, and use the self-attention mechanism based on the Transformer model to determine the query, key, and value. , , , where Q, K, and V represent query, key, and value vectors respectively. , as well as Represent the weights of query, key, and value matrices respectively, Represents the point cloud vector and calculates the attention relationship between the point clouds, expressed as: ; ; in represents the attention weight between the i-th point cloud and the j-th point cloud, and Represents the query and key of the i-th point cloud and the j-th point cloud respectively. The T in represents the transpose calculation, represents the dimension of the Key matrix, represents the Transformer output of the i-th point cloud, Represents the value of the j-th point cloud; Use multi-layer perceptron MLP to After the final classification, different classification labels are obtained, including label classification of poles, power lines, vegetation, ground and buildings.
[0025] Use multi-layer perceptron MLP to After the final classification, different classification labels are obtained, including label classification of poles, power lines, vegetation, ground and buildings.
[0026] By using the reflection intensity of the denoised point cloud coordinates based on the adaptive clustering method to determine the center value of the intensity interval, the discreteness of the reflection intensity value can be reduced, the classification accuracy of the point cloud data can be improved, the number of points in each intensity interval is calculated and the histogram is counted, making the distribution of the point cloud data more interpretable, and providing the initial window center for the subsequent mean shift clustering, improving the stability of the clustering algorithm, and using the mean shift Mean-Shift algorithm for adaptive clustering. The density kernel function is used to calculate the weighted mean in the window, so that the clustering process can adapt to the changes in the point cloud density in different areas, thereby improving the clustering accuracy and avoiding the problem that the traditional clustering method needs to preset the number of categories. The Transformer model processes point cloud information and uses trigonometric functions for position encoding, which can better describe point cloud coordinate information in high-dimensional space and avoid numerical instability caused by directly using coordinate values. The self-attention mechanism calculates the attention relationship between point clouds, which can mine the correlation between point clouds, improve the accuracy of point cloud classification, and enhance the adaptability to point cloud data in complex environments. Finally, the multi-layer perceptron MLP is used to classify the output of the Transformer, making the classification of poles, power lines, vegetation, ground and buildings more accurate, improving the generalization ability of the classification model, and reducing the classification error rate caused by noisy point clouds. The adaptive clustering method is combined with the mean shift algorithm to make the interval division of the point cloud reflection intensity more adaptive, ensuring that the initialization of the cluster center is based on the data distribution characteristics rather than fixed rules, thereby improving the accuracy and stability of clustering. The mean shift clustering algorithm calculates the weighted mean through the density kernel function, which can ensure that the point cloud cluster center is continuously optimized during the iteration process, so that the final classification result is more in line with the actual distribution of the data. On this basis, the Transformer model introduces position encoding, and gives spatial information to the point cloud data through trigonometric functions, so that the network can learn the relative position information between point cloud points, thereby improving the robustness of point cloud classification. In addition, the Transformer model is combined with point cloud data modeling, so that high-dimensional features can be further classified through the multi-layer perceptron MLP, ensuring that the final classification result is more stable, and avoiding the failure of traditional Euclidean distance-based classification methods when point cloud data is sparse or complex.
[0027] In step S2, the conductor is identified in real time based on the three-dimensional point cloud data and the visual image data, thereby obtaining the spatial characteristics of the inspection conductor of the transmission line, and the dynamic responsiveness of the attitude of the inspection drone during the inspection process is determined based on the spatial characteristics of the inspection conductor.
[0028] The following method can be used to perform real-time conductor recognition based on the three-dimensional point cloud data and the visual image data, and then obtain the spatial characteristics of the inspection conductor of the transmission line, namely: Performing wire point cloud extraction on the three-dimensional space point cloud data to obtain a point cloud candidate area; Performing wire structure fitting on the point cloud candidate area to obtain three-dimensional point cloud data of the inspection wire; Marking the boundaries of the wires on the visual image data to obtain a marked visual image; Fusing the three-dimensional point cloud data and the annotated visual image to obtain a wire fused image; The spatial features of the conductor fusion image are extracted to obtain the spatial features of the inspection conductor of the transmission line.
[0029] In the specific implementation, first, the wire point cloud can be extracted from the three-dimensional point cloud data to obtain a point cloud candidate area, wherein the point clouds included in the point cloud candidate area are all point clouds of the inspection wire; then, the wire structure can be fitted in the point cloud candidate area, that is, in the point cloud candidate area, the RANSAC algorithm is used to fit the linear structure and the outliers are removed. If the wire point cloud is broken due to occlusion, the point cloud interpolation algorithm (such as B-spline curve) is used to repair the wire continuity, thereby obtaining the three-dimensional point cloud data of the inspection wire.
[0030] In addition, in the specific implementation, first, the wire boundaries of the visual image data can be annotated, that is, the Canny edge detection algorithm is used to perform edge detection on the visual image data, possible wire boundaries are extracted, and the wire boundaries are annotated in the visual image, so as to obtain the annotated visual image; then, the 3D point cloud data and the annotated visual image can be fused, that is, the 3D point cloud data and the annotated visual image are spatially aligned, and the mapping relationship can be obtained through the positioning system (such as GNSS, IMU, etc.) of the inspection drone and the visual inertial sensor (VIO), and then the 3D point cloud data can be mapped to the annotated visual image according to the mapping relationship, and the annotated visual image can be obtained. to the conductor fusion image; finally, the spatial feature extraction of the conductor fusion image can be performed, that is, the spatial features of the inspection conductor can be extracted using geometric analysis methods (such as calculating curvature, curve length, conductor spacing, etc.), including the curvature, straightness, straight segment length, etc. of the conductor. The inspection conductor can be segmented according to its geometric features and spatial position, and each inspection segment is represented as a continuous path of the conductor, thereby obtaining the spatial features of the inspection conductor of the transmission line, wherein the spatial features of the inspection conductor include the three-dimensional position, shape, length, multiple inspection segments and inspection segment length of the inspection conductor, which can provide a basis for subsequent inspection tasks, flight control, and obstacle avoidance.
[0031] In this embodiment, the wire point cloud is extracted from the three-dimensional point cloud data to obtain the point cloud candidate area in the following manner, namely: Highly screening the three-dimensional space point cloud data, and then clustering the highly screened three-dimensional space point cloud data to obtain an inspection wire point cloud; Determining the geometric variation of each point in the inspection wire point cloud; Points whose geometric changes are greater than the set change threshold are eliminated to obtain the point cloud candidate area.
[0032] In specific implementation, first, the three-dimensional point cloud data can be highly screened according to a preset flight altitude range, and the point cloud data below the flight altitude range can be filtered to eliminate interference from ground objects; then, the three-dimensional point cloud data after the height screening can be clustered, and the DBSCAN algorithm can be used for density clustering to identify slender structures with high point cloud density, and preliminarily screen out possible wire point clouds, thereby obtaining the inspection wire point cloud; secondly, the geometric variation degree of each point in the inspection wire point cloud can be determined, wherein the geometric variation degree represents the degree of geometric variation of the point cloud, and the local curvature of the point cloud can be calculated, and the calculation result can be used as the geometric variation degree of the point cloud; finally, the points whose geometric variation degree is greater than the set variation threshold can be eliminated, obstacles with large curvature changes can be removed, only the point cloud with linear features is retained, and the area composed of the remaining points is used as the point cloud candidate area.
[0033] Preferably, in this embodiment, reference Figure 2 As shown, this figure is an exemplary flow chart of determining the dynamic responsiveness of the attitude of the inspection drone during the inspection process in an embodiment of the present application. In this embodiment, the dynamic responsiveness of the attitude of the inspection drone during the inspection process is determined according to the spatial characteristics of the inspection wire, which can be implemented by the following steps: First, in step S21, all inspection sections of the inspection wire are extracted from the inspection wire spatial features; Then, in step S22, the inspection steering vector of the inspection drone between adjacent inspection sections is determined; Finally, in step S23, the dynamic responsiveness of the attitude of the inspection UAV during the inspection process is determined through all the inspection steering vectors.
[0034] In specific implementation, the inspection steering vector of the inspection drone between adjacent inspection segments can be determined, where the inspection steering vector is a vector representing the steering angle and direction of the inspection drone. Assuming that there are adjacent inspection segments u and v, and the adjacent inspection segments u and v are composed of inspection nodes a, b and c, there is a vector and , thereby calculating the steering angle of the inspection drone between adjacent inspection segments u and v for , the inspection steering vector between adjacent inspection segments u and v can be obtained. The inspection steering vector of the inspection UAV between adjacent inspection segments can be obtained in the above manner; finally, the attitude dynamic responsiveness of the inspection UAV during the inspection process is determined through all the inspection steering vectors, wherein the attitude dynamic responsiveness is an indicator used to measure the responsiveness of the inspection UAV to the steering of each inspection segment during flight. The number of inspection steering vectors whose values exceed the set threshold in all inspection steering vectors can be counted, and the ratio of the statistical number to the number of all inspection steering vectors can be used as the attitude dynamic responsiveness of the inspection UAV during the inspection process.
[0035] It should be noted that the real-time identification of wires is based on three-dimensional point cloud data and visual image data, and the spatial characteristics and attitude dynamic responsiveness of the inspection wires are calculated based on these data. The inspection UAV can adjust its flight attitude to maintain flight stability according to changes in the inspection path (such as sharp turns or irregular paths), which can significantly improve the flight stability, path accuracy and safety of the inspection UAV.
[0036] In step S3, the initial positioning result of the inspection drone is obtained, and the relative collision characteristics between the inspection wire and the inspection drone are determined based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial characteristics.
[0037] In this embodiment, the initial positioning result of the inspection UAV can be obtained specifically in the following manner, namely: based on the data of the inertial navigation system and the global navigation satellite system, a deep coupling positioning algorithm is used to perform accurate initial positioning of the inspection UAV, thereby obtaining the initial positioning result of the inspection UAV.
[0038] In specific implementation, first, the data of the inertial navigation system and the global navigation satellite system can be obtained. The inertial navigation system (INS) measures the acceleration and angular velocity of the inspection UAV through inertial sensors (such as accelerometers and gyroscopes) and calculates the speed, position, attitude and other information of the UAV. The global navigation satellite system (GNSS) provides global positioning data by receiving satellite signals; then, the deep coupling positioning algorithm can be used for accurate initial positioning of the inspection UAV. The deep coupling positioning algorithm refers to the real-time fusion of INS and GNSS data, using their respective advantages to complement each other to improve positioning accuracy and reliability, and using INS to calculate the position, speed and attitude of the UAV based on the acceleration and angular velocity provided by the inertial sensor, so as to use the GNSS positioning information as a correction signal, and combine the estimation result of INS to correct the position, and finally obtain the initial positioning result of the inspection UAV.
[0039] Preferably, in this embodiment, reference Figure 3As shown, this figure is an exemplary flow chart of determining the relative collision characteristics between the inspection wire and the inspection drone in an embodiment of the present application. In this embodiment, the relative collision characteristics between the inspection wire and the inspection drone are determined based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial characteristics, which can be implemented by the following steps: First, in step S31, the relative spatial characteristics between the inspection wire and the inspection drone are determined according to the initial positioning result and the spatial characteristics of the inspection wire; Then, in step S32, the relative collision characteristics between the inspection wire and the inspection drone are determined by the attitude dynamic responsiveness and the relative spatial characteristics.
[0040] In specific implementation, first, the relative position between the inspection UAV and the inspection wire can be calculated by inputting the initial positioning result and the spatial feature of the inspection wire into the edge computing unit of the inspection UAV, and the spatial direction of the inspection wire (such as the heading and pitch angle of the wire) is calculated, and the flight attitude of the inspection UAV (obtained through the attitude dynamic responsiveness) is compared with it to obtain the relative flight direction, thereby obtaining the relative spatial feature between the inspection wire and the inspection UAV, that is, the relative spatial feature includes the relative position and relative direction between the inspection wire and the inspection UAV; then, the product of the attitude dynamic responsiveness and the relative position in the relative spatial feature and the ratio of the relative position can be calculated, and the absolute value of the difference between the result and 1 is used as the relative collision feature between the inspection wire and the inspection UAV, wherein the relative collision feature is a feature representing the collision probability between the inspection wire and the inspection UAV.
[0041] It should be noted that by determining the relative collision characteristics based on the initial positioning results of the inspection UAV, the dynamic responsiveness of the attitude, and the spatial characteristics of the inspection wires, the relative position changes of the UAV and the wires can be monitored in real time during the flight, and potential collision risks can be identified in a timely manner.
[0042] In step S4, tree obstacle recognition is performed using the three-dimensional point cloud data and the visual image data to obtain inspection tree obstacle information within the power transmission line area, and flight collision prediction is performed based on the inspection tree obstacle information to obtain the flight collision factor of the inspection UAV.
[0043] In this embodiment, tree obstacle recognition is performed using the three-dimensional point cloud data and the visual image data, namely, inputting the three-dimensional point cloud data and the visual image data into a deep learning target detection algorithm for tree obstacle recognition, thereby obtaining inspection tree obstacle information within the transmission line area.
[0044] In the specific implementation, you can choose a suitable deep learning target detection model, such as YOLO (You Only LookOnce), Faster R-CNN, RetinaNet, etc. These models are good at quickly and accurately detecting target objects from images, and can locate objects in images (that is, provide bounding boxes); it should be noted that in this application, the inspection tree barrier information includes the tree barrier location and the tree barrier point cloud area.
[0045] In this embodiment, the flight collision prediction is performed based on the inspection tree obstacle information to obtain the flight collision factor of the inspection UAV in the following manner, namely: Construct the trajectory equation of the inspection drone; Determine the collision limit distance between the inspection drone and the inspection tree barrier according to the trajectory equation and the inspection tree barrier information; The flight collision factor of the inspection UAV is determined by the collision limit distance and the set inspection safety radius.
[0046] In the specific implementation, first, the trajectory equation of the inspection drone can be constructed. The inspection drone will fly along a certain path during the inspection process, usually along the transmission line. The trajectory equation of the drone is constructed to describe its motion path and provide a basis for subsequent collision prediction. The trajectory equation can be in the following form: r(t)=(x(t), y(t), z(t)), where r(t) represents the position of the drone at time t, x(t), y(t) and z(t) are the displacement functions of the drone in the x, y and z directions respectively; then, the collision limit distance between the inspection drone and the inspection tree barrier can be determined based on the trajectory equation and the inspection tree barrier information, where the collision limit distance It represents the shortest distance between the inspection UAV and the inspection tree barrier during the inspection process. The existing distance calculation method (such as Euclidean distance) can be used to calculate the distance between each point on the UAV flight path and the tree barrier, and determine the minimum relative distance, that is, the collision limit distance; finally, the flight collision factor of the inspection UAV can be determined by the collision limit distance and the set inspection safety radius, where the flight collision factor is an indicator to measure the flight collision risk between the inspection UAV and the inspection tree barrier. The inspection safety radius can be subtracted from the collision limit distance, and the absolute value of the difference between the ratio of the calculation result and the inspection safety radius and 1 is used as the flight collision factor of the inspection UAV. The larger the value of the flight collision factor, the higher the collision risk.
[0047] It should be noted that through the calculation of tree obstacle identification and flight collision factors, the inspection UAV's ability to predict and respond to potential collision risks during the inspection process can be significantly improved. It not only optimizes flight safety and avoids the risk of colliding with obstacles, but also enhances the UAV's autonomous decision-making ability, making flight inspections more efficient and intelligent, and with strong emergency handling capabilities, thereby improving the safety and reliability of flight inspections.
[0048] In step S5, the inspection UAV is flight controlled based on the relative collision feature and the flight collision factor.
[0049] In this embodiment, the flight control of the inspection drone based on the relative collision characteristics and the flight collision factor can be specifically performed in the following manner, namely: determining a collision risk response index according to the relative collision characteristic and the flight collision factor; The inspection UAV is controlled in flight according to the collision risk response index and a given response interval.
[0050] In specific implementation, the collision risk response index can be determined based on the relative collision characteristics and the flight collision factor, wherein the collision risk response index represents the possibility of collision risk of the inspection UAV during the inspection process. The relative collision characteristics and the flight collision factor can be weighted and summed to obtain the result as the collision risk response index, wherein the corresponding weight can be preset based on historical experiments and data analysis; then, a given response interval is obtained, which can be divided into multiple levels: Low risk range: If the collision risk response index is within this range, it indicates that the inspection drone is in a safe flight state and the current flight path can be maintained unchanged.
[0051] Medium risk range: When the collision risk response index is in the medium risk range, the flight control system needs to issue an early warning and appropriately adjust the flight speed or altitude to avoid possible collisions.
[0052] High-risk interval: If the collision risk response index exceeds the set high-risk threshold, it means that the inspection drone faces a greater collision risk, and the flight control system should actively adjust the flight path, reduce the speed, or temporarily change the flight direction to avoid obstacles.
[0053] It should be noted that by controlling the flight of the inspection UAV based on the collision risk response index and the given response range, the collision risk can be dynamically evaluated and handled during the flight. The collision risk response index can quantify the collision risk and help the flight control system respond appropriately in different risk environments, thereby improving the flight safety of the inspection UAV in complex environments.
[0054] It can be seen that in this application, first, the wire is identified in real time based on the three-dimensional point cloud data and visual image data, and the spatial characteristics and attitude dynamic responsiveness of the inspection wire are calculated based on these data. The inspection UAV can adjust the flight attitude to maintain flight stability according to the changes in the inspection path; then, by determining the relative collision characteristics based on the initial positioning results of the inspection UAV, the attitude dynamic responsiveness and the spatial characteristics of the inspection wire, the relative position changes of the UAV and the wire can be monitored in real time during the flight, and potential collision risks can be identified in time; further, through the calculation of tree obstacle identification and flight collision factors, the inspection UAV's ability to predict and respond to potential collision risks during the inspection process can be significantly improved; finally, by controlling the flight of the inspection UAV based on the collision risk response index and a given response interval, the collision risk can be dynamically evaluated and processed during the flight. The collision risk response index can quantify the collision risk, which can help the flight control system to respond appropriately in different risk environments, thereby improving the flight safety of the inspection UAV in complex environments.
[0055] In summary, the technical solution adopted in this application can pre-identify the collision risk of inspection drones during the inspection process to improve the safety of flight inspections.
[0056] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0057] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.
[0058] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
Claims
1. A transmission line flight intelligent inspection method based on three-dimensional laser radar, characterized in that: The inspection method comprises the following steps: Real-time collection of 3D point cloud data and visual image data within the transmission line area; Conductor real-time identification is performed based on the three-dimensional point cloud data and the visual image data, thereby obtaining the spatial characteristics of the inspection conductor of the transmission line, and the attitude dynamic responsiveness of the inspection drone during the inspection process is determined based on the spatial characteristics of the inspection conductor; Obtaining an initial positioning result of the inspection drone, and determining a relative collision feature between the inspection wire and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial feature; Tree obstacle recognition is performed through the three-dimensional space point cloud data and the visual image data, thereby obtaining inspection tree obstacle information in the transmission line area, and flight collision prediction is performed based on the inspection tree obstacle information to obtain the flight collision factor of the inspection UAV; The inspection UAV is controlled in flight based on the relative collision characteristics and the flight collision factor.
2. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: The laser radar device and visual sensor carried by the inspection drone scans and collects three-dimensional spatial point cloud data and visual image data in the transmission line area in real time.
3. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: The wire is identified in real time based on the three-dimensional point cloud data and the visual image data, and then the inspection wire spatial features of the transmission line are obtained, which specifically include: Performing wire point cloud extraction on the three-dimensional space point cloud data to obtain a point cloud candidate area; Performing wire structure fitting on the point cloud candidate area to obtain three-dimensional point cloud data of the inspection wire; Marking the boundaries of the wires on the visual image data to obtain a marked visual image; Fusing the three-dimensional point cloud data and the annotated visual image to obtain a wire fused image; The spatial features of the conductor fusion image are extracted to obtain the spatial features of the inspection conductor of the transmission line.
4. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 3, characterized in that: Conductor point cloud extraction is performed on the three-dimensional space point cloud data to obtain point cloud candidate areas, which specifically include: Highly screening the three-dimensional space point cloud data, and then clustering the highly screened three-dimensional space point cloud data to obtain an inspection wire point cloud; Determining the geometric variation of each point in the inspection wire point cloud; Points whose geometric changes are greater than the set change threshold are eliminated to obtain the point cloud candidate area.
5. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: Determining the attitude dynamic responsiveness of the inspection drone during the inspection process according to the inspection wire spatial characteristics specifically includes: Extract all inspection sections of the inspection wire from the inspection wire spatial features; Determine the inspection steering vector of the inspection UAV between adjacent inspection sections; The dynamic responsiveness of the attitude of the inspection UAV during the inspection process is determined through all inspection steering vectors.
6. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: Obtaining the initial positioning result of the inspection UAV specifically includes: based on the data of the inertial navigation system and the global navigation satellite system, using a deep coupling positioning algorithm to perform accurate initial positioning of the inspection UAV, and then obtaining the initial positioning result of the inspection UAV.
7. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: Determining the relative collision characteristics between the inspection wire and the inspection drone based on the attitude dynamic responsiveness, the initial positioning result and the inspection wire spatial characteristics specifically includes: Determine the relative spatial characteristics between the inspection wire and the inspection drone according to the initial positioning result and the spatial characteristics of the inspection wire; The relative collision characteristics between the inspection wire and the inspection drone are determined by the attitude dynamic responsiveness and the relative spatial characteristics.
8. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: Identifying tree obstacles through the three-dimensional space point cloud data and the visual image data is to input the three-dimensional space point cloud data and the visual image data into a deep learning target detection algorithm for tree obstacle identification.
9. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: According to the inspection tree obstacle information, flight collision prediction is performed to obtain the flight collision factors of the inspection UAV, which specifically include: Construct the trajectory equation of the inspection drone; Determine the collision limit distance between the inspection drone and the inspection tree barrier according to the trajectory equation and the inspection tree barrier information; The flight collision factor of the inspection UAV is determined by the collision limit distance and the set inspection safety radius.
10. The method for intelligent inspection of power transmission lines based on three-dimensional laser radar as claimed in claim 1, characterized in that: The flight control of the inspection UAV based on the relative collision feature and the flight collision factor specifically includes: determining a collision risk response index according to the relative collision characteristic and the flight collision factor; The inspection UAV is controlled in flight according to the collision risk response index and a given response interval.
Citation Information
Cited By
Unmanned aerial vehicle model real-time identification method and system based on laser radar point cloud features
CN120428249A
Unmanned aerial vehicle autonomous line-imitating flight control system for power transmission line
CN120949810A
