A method for detecting a power distribution network tree barrier by a UAV and related device

Abstract

The application discloses a UAV detection method for a tree barrier in a distribution network and a related device. Based on the idea of statistical function parameter estimation, two-dimensional point cloud data collected by a single laser radar is merged into three-dimensional point cloud data. On the basis of the three-dimensional point cloud data, an improved DBSCAN clustering algorithm is used to achieve high-precision clustering effect. The number of nearest neighbor point clouds of the DBSCAN clustering algorithm does not need to be manually set, the manual trial and error cost is reduced, tree barrier misjudgment is reduced, and tree barrier detection accuracy is improved. Thus, the technical problems of the prior art, such as the need for manual selection of the number of nearest neighbor point clouds, the inability to adaptively adjust parameters to data, low tree barrier detection accuracy, and low efficiency, are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of distribution network line inspection, and in particular to a UAV detection method for distribution network tree barriers and related devices. BACKGROUND

[0002] With the acceleration of distribution network construction, the contradiction between rapid development of distribution network and lack of personnel is becoming more and more prominent. The length of line operation and maintenance per capita increases year by year, and the traditional manual operation and maintenance mode may cause the discovery of line defects to lag behind the demand for safe operation of the line. Therefore, it is an inevitable choice to use modern scientific and technological means to improve the operation and maintenance technology level of the power transmission line. Since the power transmission line needs to pass through various complex terrains and vegetation, it brings many difficulties to the detection of the power line. According to the relevant provisions of the high-voltage power grid line: within a certain distance of the diameter of the ultra-high voltage power grid line, if there is an object in direct contact with the ground, it is possible to cause a safety accident such as electric leakage and discharge. At present, the distance detection between the high-altitude power grid and the surrounding objects is an important indicator of the safety detection of the ultra-high voltage power grid line. The power line is exposed in the wild for a long time, and the growth of trees may cause discharge of the power transmission line, which must be handled in time. Therefore, it is a very important task to detect the distance between the high-voltage power line and the trees.

[0003] With the continuous development of science and technology, the use of UAVs for inspection of power transmission lines has gradually become a research hotspot. Compared with the traditional inspection mode, the UAV is light, small in size, easy to carry, can be remotely controlled, stable in attitude, easy to take off and land, and flexible to operate. It can perform efficient and non-contact inspection of the power transmission line to ensure safety. In the UAV tree barrier detection technology, how to identify the tree barrier is the key technology. In the existing technology, a laser radar is used to scan the power transmission line and establish a three-dimensional point cloud map, and on this basis, the tree barrier is detected. Improving the clustering accuracy of the tree barrier point cloud data can improve the recognition rate of the tree barrier. The existing clustering algorithm mostly depends on the initialization setting of its algorithm parameters, such as the need for manual selection of the number of nearest neighbor point clouds in the DBSCAN clustering algorithm. It cannot adaptively adjust the parameters to the data, resulting in low detection accuracy and low efficiency of the tree barrier. SUMMARY

[0004] The present application provides a UAV detection method for distribution network tree barriers and related devices, which solves the technical problem that the existing technology needs manual selection of the number of nearest neighbor point clouds, cannot adaptively adjust the parameters to the data, and has low detection accuracy and low efficiency of the tree barrier.

[0005] In view of this, the first aspect of this application provides a method for detecting tree obstacles in a power distribution network using a drone, which is applied to a detection device consisting of a drone, a radar, a computer, a positioning module, a wireless transmission module, and a ground control module; wherein the radar is disposed on the lower side of the drone, and the wireless transmission module, the computer, and the positioning module are disposed on the upper side of the drone.

[0006] The method includes:

[0007] With the drone as the origin, the right and bottom of the drone are the positive x-axis and positive y-axis, respectively. When the drone is flying between towers, a two-dimensional point cloud dataset is collected by the radar, and the position information of the drone is collected by the positioning module.

[0008] The two-dimensional point cloud dataset is merged into a three-dimensional point cloud map, and the three-dimensional point cloud map is segmented into ground point cloud to obtain a ground point cloud map and a non-ground point cloud map.

[0009] The optimal number of nearest neighbors (K) is determined by estimating the K parameter in the KNN process through function fitting and differentiation.

[0010] The minimum and maximum number of point clouds for clustering are set based on the optimal number of nearest neighbors. DBSCAN clustering is performed on the non-ground point cloud map to obtain a cluster set.

[0011] Based on the relationship between the average number of point clouds of clustered objects on the y-axis and a preset wire distance threshold, the clustered objects are classified as wires or objects below wires.

[0012] The highest point of the object below the guide is determined. Based on the difference between the highest point and the average number of the guide point cloud on the y-axis, and the relationship with the preset tree barrier distance threshold, the object below the guide is classified as a tree barrier object or a non-tree barrier object, thereby obtaining tree barrier point cloud data.

[0013] Optionally, the estimation of the nearest neighbor number K parameter in the KNN process through function fitting and differentiation to determine the optimal nearest neighbor number specifically includes:

[0014] Set an initial number of nearest neighbors and construct a discrete function for the number of nearest neighbors;

[0015] By fitting the discrete function of the number of nearest neighbors in the form of a quadratic function, a continuous function of the number of nearest neighbors is obtained based on the characteristic of uniform distribution of the point cloud.

[0016] Based on the aforementioned characteristics, the relationship between the number of nearest neighbors and the number of point clouds covered by the tree barrier diameter is analyzed to determine the optimal number of nearest neighbors.

[0017] Optionally, determining the highest point of the object below the guide wire, and classifying the object below the guide wire into a tree-obstacle object or a non-tree-obstacle object based on the difference between the highest point and the average of the guide wire point cloud along the y-axis, and a preset tree obstacle distance threshold, thereby obtaining tree obstacle point cloud data, specifically includes:

[0018] The point cloud with the smallest y-axis position of the object below the conductor is taken as the highest point of the object below the conductor.

[0019] Determine whether the difference between the highest point and the average number of the traverse point cloud on the y-axis is greater than a preset tree barrier distance threshold. If so, determine that the object below the traverse is a tree barrier object; otherwise, determine that there is no tree barrier object between the towers.

[0020] Optionally, the step of determining the highest point of the object below the guide, and classifying the object below the guide as a tree-obstacle object or a non-tree-obstacle object based on the difference between the highest point and the average of the guide point cloud along the y-axis, and a preset tree obstacle distance threshold, thereby obtaining tree obstacle point cloud data, further includes:

[0021] The wireless transmission module transmits the tree obstacle point cloud data, the corresponding acquisition time, and the corresponding location information to the ground control module.

[0022] A second aspect of this application provides a drone detection system for tree obstacles in power distribution networks, the system comprising:

[0023] The data acquisition unit is used to collect a two-dimensional point cloud dataset by radar and the location information of the drone by using the drone as the origin, with the right and bottom of the drone as the positive x-axis and positive y-axis, respectively.

[0024] The segmentation unit is used to merge the two-dimensional point cloud dataset into a three-dimensional point cloud map, and to perform ground point cloud segmentation on the three-dimensional point cloud map to obtain a ground point cloud map and a non-ground point cloud map.

[0025] The estimation unit is used to estimate the number of nearest neighbors K in the KNN process by means of function fitting and differentiation, and to determine the optimal number of nearest neighbors;

[0026] A clustering unit is used to set the minimum and maximum number of point clouds for clustering based on the number of best nearest neighbors, and to perform DBSCAN clustering on the non-ground point cloud map to obtain a cluster set.

[0027] The classification unit is used to classify the clustered objects into wires or objects below wires based on the relationship between the average number of point clouds of clustered objects on the y-axis and a preset wire distance threshold.

[0028] The detection unit is used to determine the highest point of the object below the guide line, and classify the object below the guide line into a tree obstacle object or a non-tree obstacle object based on the difference between the highest point and the average number of the guide line point cloud on the y-axis and the relationship with a preset tree obstacle distance threshold, thereby obtaining tree obstacle point cloud data.

[0029] Optionally, the estimation unit is specifically used for:

[0030] Set an initial number of nearest neighbors and construct a discrete function for the number of nearest neighbors;

[0031] By fitting the discrete function of the number of nearest neighbors in the form of a quadratic function, a continuous function of the number of nearest neighbors is obtained based on the characteristic of uniform distribution of the point cloud.

[0032] Based on the aforementioned characteristics, the relationship between the number of nearest neighbors and the number of point clouds covered by the tree barrier diameter is analyzed to determine the optimal number of nearest neighbors.

[0033] Optionally, the detection unit is specifically used for:

[0034] The point cloud with the smallest y-axis position of the object below the conductor is taken as the highest point of the object below the conductor.

[0035] Determine whether the difference between the highest point and the average number of the traverse point cloud on the y-axis is greater than a preset tree barrier distance threshold. If so, determine that the object below the traverse is a tree barrier object; otherwise, determine that there is no tree barrier object between the towers.

[0036] Optionally, it also includes: a transmitting unit;

[0037] The transmitting unit is used to transmit the tree obstacle point cloud data, the corresponding acquisition time, and the corresponding location information to the ground control module via a wireless transmission module.

[0038] A third aspect of this application provides a drone detection device for tree obstacles in power distribution networks, the device comprising a processor and a memory:

[0039] The memory is used to store program code and transmit the program code to the processor;

[0040] The processor is configured to execute the steps of the UAV detection method for tree obstacles in the power distribution network as described in the first aspect above, according to the instructions in the program code.

[0041] A fourth aspect of this application provides a computer-readable storage medium for storing program code for performing the method described in the first aspect above.

[0042] As can be seen from the above technical solutions, this application has the following advantages:

[0043] This application provides a method for detecting tree obstacles in power distribution networks using unmanned aerial vehicles (UAVs). The method includes: using the UAV as the origin, with the positive x-axis to the right and the positive y-axis to the bottom of the UAV, respectively; when the UAV is flying between towers, acquiring a two-dimensional point cloud dataset via radar and acquiring the UAV's position information via a positioning module; merging the two-dimensional point cloud dataset into a three-dimensional point cloud map, and performing ground point cloud segmentation on the three-dimensional point cloud map to obtain a ground point cloud map and a non-ground point cloud map; estimating the nearest neighbor number K parameter in the KNN process using function fitting and differentiation to determine the optimal nearest neighbor number; setting the minimum and maximum number of point clouds for clustering based on the optimal nearest neighbor number, and performing DBSCAN clustering on the non-ground point cloud map to obtain a cluster set; classifying the clustered objects as either objects below the conductor or objects below the conductor based on the relationship between the average number of point clouds on the y-axis of the clustered objects and a preset conductor distance threshold; determining the highest point of the objects below the conductor, and classifying the objects below the conductor as tree obstacle objects or non-tree obstacle objects based on the difference between the highest point and the average number of point clouds on the y-axis of the conductor and the preset tree obstacle distance threshold, thereby obtaining tree obstacle point cloud data.

[0044] Compared with existing technologies, this invention, based on the idea of ​​estimating parameters using statistical functions, first merges the two-dimensional point cloud data collected by single-line lidar into three-dimensional point cloud data. Based on this three-dimensional point cloud data, an improved DBSCAN clustering algorithm is used to achieve high-precision clustering results. This eliminates the need for manually setting the number of nearest neighbor points in the DBSCAN clustering algorithm, thereby reducing manual trial-and-error costs, minimizing false tree obstacle detection, and improving tree obstacle detection accuracy. This solves the technical problem of existing technologies requiring manual selection of an appropriate number of nearest neighbor points, which cannot adaptively adjust parameters based on data, resulting in low accuracy and efficiency in tree obstacle detection. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating an embodiment of a method for detecting tree obstacles in a power distribution network provided in this application.

[0046] Figure 2 This is a schematic diagram of the structure of a drone detection device provided in the embodiments of this application;

[0047] Figure 3 This is a diagram of a drone inspection flight model provided in the embodiments of this application;

[0048] Figure 4 This is a schematic diagram of the structure of an embodiment of a UAV detection system for tree obstacles in a power distribution network provided in this application. Detailed Implementation

[0049] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0050] The present application provides a method for detecting tree obstacles in a power distribution network using a drone, which is applied to a detection device consisting of a drone, radar, computer, positioning module, wireless transmission module and ground control module; wherein the radar is located on the lower side of the drone, and the wireless transmission module, computer and positioning module are located on the upper side of the drone.

[0051] It should be noted that the drone in this embodiment is a multi-rotor drone, the computer is an embedded computer, and the radar is a single-line lidar; wherein, the single-line lidar is configured on the underside of the multi-rotor drone, and the embedded computer, wireless transmission module, and positioning module are configured on the upper side of the multi-rotor drone. The multi-rotor drone conducts inspection flights over the power distribution network, and the detection device is as follows: Figure 2 As shown. Please refer to [the original text]. Figure 3 1 represents the drone, 2 represents the guide wire, and 3 represents the scanning range of the single-line lidar. The single-line lidar scans the environment below the multi-rotor drone and transmits the point cloud data to an embedded computer for processing. The embedded computer performs tree obstacle detection based on an improved DBSCAN on the inter-tower point cloud data. The detection method is as follows:

[0052] Please see Figure 1 The methods include:

[0053] Step 101: With the UAV as the origin, the right and bottom of the UAV are the positive x-axis and positive y-axis, respectively. When the UAV is flying between towers, a two-dimensional point cloud dataset is collected by radar, and the UAV's position information is collected by the positioning module.

[0054] It should be noted that the two-dimensional point cloud data is defined using a coordinate system with the UAV's position as the origin, the right side of the UAV as the positive x-axis, and the bottom of the UAV as the positive y-axis. The two-dimensional point cloud data acquired by the embedded computer at time t from the single-line LiDAR is defined as follows: The location information of the UAV collected by the positioning module is L(t), where I(t) is the number of point clouds collected in the two-dimensional point cloud data at time t. When the UAV completes the inspection flight between the power distribution towers, it enters the tree obstacle detection between the towers. The process is as follows:

[0055] The two-dimensional point cloud data collected during the flight between towers were merged into a three-dimensional point cloud map. Where T s Te M and M represent the start time, end time, and number of points in the 3D point cloud map, respectively, for the inter-tower flight.

[0056] Step 102: Merge the two-dimensional point cloud dataset into a three-dimensional point cloud map, and perform ground point cloud segmentation on the three-dimensional point cloud map to obtain a ground point cloud map and a non-ground point cloud map.

[0057] It should be noted that when performing ground point cloud segmentation on the 3D point map, based on the power distribution network inspection environment, the point cloud with the largest y-axis value collected at time t is definitely the ground point cloud. Based on this ground point cloud... The point cloud collected at time t is filtered, when When the value is greater than the threshold θ, the point cloud If the point cloud is not ground-based, it is considered a non-ground point cloud; otherwise, it is considered a ground-based point cloud. The final result is a non-ground point cloud map. Where m is the number of point clouds in the non-ground point cloud map.

[0058] Step 103: Estimate the number of nearest neighbors K in the KNN process by using function fitting and differentiation to determine the optimal number of nearest neighbors;

[0059] It should be noted that the DBSCAN clustering algorithm requires pre-setting the parameter K, the number of nearest neighbors in the KNN process. To improve clustering accuracy, K is estimated using a function fitting and differentiation method. The estimation process is as follows:

[0060] Step 1031: Set the initial number of nearest neighbors K = 60, and construct the nearest neighbor discrete function g(k) = {(2,D2),(3,D3),...,(K,D...} K )}, where D K Let k be the average maximum distance between the nearest neighbors. Its calculation formula is:

[0061]

[0062]

[0063] in Let k be the maximum nearest neighbor distance of point cloud i. Let m be the distance between point cloud i and its j-th nearest point cloud.

[0064] Step 1032: Fit the nearest neighbor discrete function g(k) to a quadratic function to obtain the nearest neighbor number continuous function. Based on the uniform distribution of the point cloud, when the number of nearest neighbors is greater than the number of point clouds covered by the tree barrier diameter, the derivative of the nearest neighbor number continuous function G(k) is greater than 1. When the number of nearest neighbors is less than the number of point clouds covered by the tree barrier diameter, the derivative of the nearest neighbor number continuous function G(k) is less than 1 for the value of k. Therefore, when the derivative of the nearest neighbor number continuous function G(k0) is 1, if k0 ≤ K, k0 is the optimal number of nearest neighbors, proceed to step 104; otherwise, let K = K + 1, proceed to step 1031.

[0065] Step 104: Set the minimum and maximum number of point clouds for clustering based on the number of best nearest neighbors, perform DBSCAN clustering on the non-ground point cloud map, and obtain the cluster set;

[0066] It should be noted that the minimum and maximum number of point clouds in the cluster are set with the optimal nearest neighbor number k0. min N max DBSCAN clustering is performed on the non-ground point cloud map to obtain the cluster set V = {V j |1≤j≤B}, where V j Let B be the point cloud of the clustered objects, and let B be the number of clusters.

[0067] Step 105: Based on the relationship between the average number of point clouds of clustered objects on the y-axis and the preset wire distance threshold, classify the clustered objects into wires or objects below wires.

[0068] It should be noted that, because the clearance distance between the drone and the guide wire is less than the clearance distance between the drone and the object below the guide wire, based on this feature, when the average number of points in the point cloud of a clustered object on the y-axis is less than or equal to the guide wire distance threshold, the clustered object is determined to be a guide wire; otherwise, the clustered object is an object below the guide wire.

[0069] Step 106: Determine the highest point of the object below the guide. Based on the difference between the highest point and the average number of the guide point cloud on the y-axis, and the relationship with the preset tree barrier distance threshold, classify the object below the guide as a tree barrier object or a non-tree barrier object, thereby obtaining the tree barrier point cloud data.

[0070] It should be noted that the point cloud with the smallest y-axis position for all objects below the conductor is taken as the highest point of that object. Determine the highest point of the object The difference between the point cloud data and the average point cloud data along the y-axis is checked. If the difference is greater than the tree barrier distance threshold, the object is considered a tree barrier object. Further, the tree barrier point cloud data and the acquisition time T = {t|t∈O(V)} are considered. treeThe location information G(t), t∈T, corresponding to O(V) is sent to the ground control module, where O(V) tree The data set represents the time frame for collecting tree obstacle point cloud data. Otherwise, it is assumed that there are no tree obstacles between the towers. If the tree obstacle detection has not yet ended, the next round of flying tree obstacle detection between towers will begin; otherwise, the tree obstacle detection will end.

[0071] The above is one embodiment provided in this application, and the following is another embodiment provided in this application.

[0072] Please see Figure 2 The UAV detection system for tree obstacles in power distribution networks provided in this application embodiment includes:

[0073] The acquisition unit 201 is used to collect a two-dimensional point cloud dataset by radar when the UAV is flying between towers, with the UAV as the origin, the right side of the UAV and the bottom of the UAV as the positive x-axis and the positive y-axis, respectively. The UAV's position information is also collected by the positioning module.

[0074] The segmentation unit 202 is used to merge the two-dimensional point cloud dataset into a three-dimensional point cloud map, and to segment the three-dimensional point cloud map into ground point cloud to obtain a ground point cloud map and a non-ground point cloud map.

[0075] The estimation unit 203 is used to estimate the number of nearest neighbors K parameter in the KNN process by means of function fitting and differentiation, and to determine the optimal number of nearest neighbors;

[0076] Clustering unit 204 is used to set the minimum and maximum number of point clouds in the cluster with the best nearest neighbor number, and to perform DBSCAN clustering on the non-ground point cloud map to obtain a cluster set.

[0077] The classification unit 205 is used to classify clustered objects into wires or objects below wires based on the relationship between the average number of point clouds of clustered objects in the cluster set on the y-axis and a preset wire distance threshold.

[0078] The detection unit 206 is used to determine the highest point of the object below the guide. Based on the difference between the highest point and the average number of the guide point cloud on the y-axis, and the relationship with the preset tree obstacle distance threshold, the object below the guide is classified as a tree obstacle object or a non-tree obstacle object, thereby obtaining tree obstacle point cloud data.

[0079] Furthermore, this application embodiment also provides a drone detection device for tree obstacles in power distribution networks, the device including a processor and a memory:

[0080] The memory is used to store program code and transmit the program code to the processor;

[0081] The processor is used to execute the steps of the UAV detection method for tree obstacles in the power distribution network as described in the above method embodiments, according to the instructions in the program code.

[0082] Furthermore, this application embodiment also provides a computer-readable storage medium for storing program code, which is used to execute the UAV detection method for power distribution network tree obstacles described in the above method embodiment.

[0083] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0084] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0085] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0086] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0087] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0088] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0089] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0090] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for detecting a power distribution tree barrier by a UAV, characterized in that, Be applied to the detection device constituted by unmanned aerial vehicle, radar, computer, positioning module, wireless transmission module and ground control module, wherein the radar is arranged on the lower side of the unmanned aerial vehicle, the wireless transmission module, the computer and the positioning module are arranged on the upper side of the unmanned aerial vehicle; The method comprises: With the unmanned aerial vehicle as the origin, the right and lower sides of the unmanned aerial vehicle are x positive axis and y positive axis respectively, when the unmanned aerial vehicle flies between towers, two-dimensional point cloud data set is collected through the radar, and the position information of the unmanned aerial vehicle is collected through the positioning module; The two-dimensional point cloud data set is combined into a three-dimensional point cloud graph, and ground point cloud segmentation is performed on the three-dimensional point cloud graph to obtain a ground point cloud graph and a non-ground point cloud graph; The number of nearest neighbors in KNN process is determined by function fitting derivation method Parameter estimation is performed to determine the optimal number of nearest neighbors The best nearest neighbor number is set to determine the minimum and maximum point cloud quantities of clustering, DBSCAN clustering is performed on the non-ground point cloud graph to obtain a clustering set; According to the relationship between the average number of point clouds of the clustered objects in the clustering set on the y axis and the preset conductor distance threshold, the clustered objects are classified as conductors or objects below the conductors; The highest point of the object below the conductor is determined, and according to the difference between the average number of point clouds of the highest point on the y axis and the conductor point cloud and the preset tree barrier distance threshold, the object below the conductor is classified as a tree barrier object or a non-tree barrier object, thereby obtaining tree barrier point cloud data; The number of nearest neighbors in the KNN process is determined by the function fitting derivation method The number of nearest neighbors in the KNN process is determined by the function fitting derivation method An initial nearest neighbor number is set, and a nearest neighbor number discrete function is constructed; The nearest neighbor number discrete function is fitted in the form of a quadratic function to obtain a nearest neighbor number continuous function according to the characteristics of uniform distribution of point clouds; Based on the characteristics, the relationship between the nearest neighbor number and the number of point clouds covered by the tree barrier diameter is analyzed to determine the best nearest neighbor number. 2.The method of claim 1, wherein, The determination of the highest point of the object below the conductor, according to the difference between the average number of point clouds of the highest point on the y axis and the conductor point cloud and the preset tree barrier distance threshold, classifies the object below the conductor as a tree barrier object or a non-tree barrier object, thereby obtaining tree barrier point cloud data, specifically comprising: The point cloud of the object below the conductor on the y axis is the highest point of the object below the conductor; Determine whether the difference between the average number of point clouds of the highest point on the y axis and the conductor point cloud is greater than the preset tree barrier distance threshold, if yes, determine that the object below the conductor is a tree barrier object, otherwise, determine that there is no tree barrier object between the towers. 3.The method of claim 1, wherein, The determination of the highest point of the object below the conductor, according to the difference between the average number of point clouds of the highest point on the y axis and the conductor point cloud and the preset tree barrier distance threshold, classifies the object below the conductor as a tree barrier object or a non-tree barrier object, thereby obtaining tree barrier point cloud data, and then further comprising: The tree barrier point cloud data, the corresponding collection time of the tree barrier point cloud data and the corresponding position information are sent to the ground control module through the wireless transmission module.

4. A drone detection system for tree obstructions in power distribution networks, characterized in that, It comprises: The acquisition unit is used for taking the unmanned aerial vehicle as the origin, the right and lower sides of the unmanned aerial vehicle are x positive axis and y positive axis respectively, when the unmanned aerial vehicle flies between towers, two-dimensional point cloud data set is collected through the radar, and the position information of the unmanned aerial vehicle is collected through the positioning module; The segmentation unit is configured to combine the two-dimensional point cloud data set into a three-dimensional point cloud map, and perform ground point cloud segmentation on the three-dimensional point cloud map to obtain a ground point cloud map and a non-ground point cloud map. An estimation unit is configured to estimate the number of nearest neighbors in a KNN process by function fitting derivation Parameters are estimated to determine the optimal number of nearest neighbors. The clustering unit is configured to set clustering minimum and maximum point cloud numbers according to the optimal nearest neighbor number, perform DBSCAN clustering on the non-ground point cloud map to obtain a clustering set. The classification unit is configured to classify the clustered objects in the clustering set into conductors or objects under conductors according to a relationship between an average value of point clouds of the clustered objects in the y-axis and a preset conductor distance threshold. The detection unit is configured to determine a highest point of the objects under conductors, and classify the objects under conductors into tree barrier objects or non-tree barrier objects according to a relationship between a difference between the highest point and an average value of conductor point clouds in the y-axis and a preset tree barrier distance threshold, so as to obtain tree barrier point cloud data. The estimation unit is specifically configured to: set an initial nearest neighbor number, and construct a nearest neighbor number discrete function; perform function fitting on the nearest neighbor number discrete function in a quadratic function form to obtain a nearest neighbor number continuous function according to a characteristic of uniform distribution of point clouds; and analyze a relationship between the nearest neighbor number and a number of point clouds covered by a tree barrier diameter based on the characteristic, and determine the optimal nearest neighbor number.

5. The drone detection system of claim 4, wherein, The detection unit is specifically configured to: take a point cloud with the smallest y-axis value of the objects under conductors as the highest point of the objects under conductors; determine whether a difference between the highest point and an average value of conductor point clouds in the y-axis is greater than a preset tree barrier distance threshold, and if yes, determine that the objects under conductors are tree barrier objects, and if not, determine that there is no tree barrier object between towers.

6. The drone detection system of claim 4, wherein, The device further includes a sending unit. The sending unit is configured to send the tree barrier point cloud data, a collection time corresponding to the tree barrier point cloud data, and position information corresponding to the tree barrier point cloud data to a ground control module through a wireless transmission module. The device includes a processor and a memory.

7. An unmanned aerial vehicle detection device for networked tree barriers, comprising: The memory is configured to store program code and transmit the program code to the processor. The processor is configured to execute the method for detecting a tree barrier of a distribution network by a UAV according to instructions in the program code. The computer readable storage medium is configured to store program code for executing the method for detecting a tree barrier of a distribution network by a UAV.

8. A computer-readable storage medium, characterized in that, ​

Images