Method and device for processing inspection image of power transmission line

Abstract

The embodiment of the invention relates to a power transmission line inspection image processing method and device. The method comprises the following steps: installing an inspection device on an inspection aircraft; the inspection device receives a line electric tower data set and a line high-precision map sent by the server; confirming electric tower acquisition parameters according to the data set to obtain an acquisition parameter sequence; according to the high-precision map, the starting coordinate and the ending coordinate of the inspection are identified; an inspection route configuration instruction carrying the starting and ending coordinates and the high-precision map is sent to the flight control system; after an inspection take-off instruction sent by the flight control system is received, inspection images of all the electric towers are acquired according to the navigation positioning system, the high-definition camera and the acquisition parameter sequence, and an image set sequence is obtained and stored; and the image set sequence is sent to the server when an inspection landing instruction sent by the flight control system is received. According to the invention, the inspection period can be shortened, the inspection efficiency is improved, the inspection real-time performance is enhanced, and the inspection success rate is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and device for processing transmission line inspection images. Background Art

[0002] Inspection of power towers (referred to as power towers) in transmission lines is a major inspection task in transmission line inspection. At present, there are two conventional processing methods for this type of transmission line inspection tasks: conventional inspection methods based on manual work and conventional inspection methods based on aircraft. The conventional inspection method based on manual work is to assign staff to drive along the power transmission direction of the transmission line to manually collect images of each power tower on the line. The conventional inspection method based on aircraft (such as unmanned drones, manned helicopters, etc.) is to input the inspection route into the flight control system of the inspection aircraft to dynamically plan the inspection path, and then the aircraft performs a human-machine interactive semi-manual inspection of each power tower on the transmission line based on the dynamically planned inspection path; the human-machine interactive semi-manual inspection here means that the aircraft automatically hovers near a power tower and sends the location information to the background operator at the same time, and then the background operator controls the aircraft through remote control to collect images of each position of the current power tower.

[0003] Both of the above-mentioned conventional inspection methods have some defects: 1) The conventional inspection method based on manual work has a long inspection cycle, low efficiency, and poor real-time performance. Especially for some power towers with special installation locations (such as mountains, deserts, etc.), its inspection intensity is even more difficult to meet the high-frequency inspection requirements; 2) The conventional inspection method based on aircraft requires manual operation and requires the aircraft to be able to hover for a long time. This not only has high requirements on the communication network quality of the inspection environment, but also has high requirements on the endurance of the aircraft. In some inspection environments with poor communication network quality or too long transmission lines, the inspection failure rate of this inspection method will also increase, resulting in a decrease in inspection efficiency and weakened inspection real-time performance. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device, electronic device and computer-readable storage medium for processing inspection images of power transmission lines in view of the defects of the prior art. The present invention installs an inspection device on an inspection aircraft and connects it to a remote server and a flight control system, a navigation and positioning system and a high-definition camera on the aircraft; and before the flight inspection, the inspection device performs data preprocessing based on the tower data set + high-precision map of the power transmission line sent by the remote server to obtain the corresponding tower acquisition parameter sequence and the inspection start / end coordinates; and sends the inspection start / end coordinates + high-precision map to the flight control system for flight trajectory planning and flight control; and during the flight inspection, all towers of the transmission line are inspected and processed according to the navigation and positioning system, the high-definition camera and the tower acquisition parameter sequence to generate a corresponding image set sequence and save it; and at the end of the flight inspection, the saved image set sequence is sent to the remote server. The present invention provides a new aircraft inspection mechanism, which can be used without human-computer interaction or hovering during the inspection. Compared with the conventional manual inspection method, the inspection cycle of the present invention is short, the efficiency is high, and the real-time performance is strong. For some power towers with special installation locations (such as mountains, deserts and other areas), the inspection intensity can also meet the high-frequency inspection requirements. Compared with the conventional aircraft-based inspection method, the inspection cycle of the present invention is shorter, the efficiency is higher, and the real-time performance is stronger. In the inspection environment with poor network quality, the tower image acquisition can be completed based on the offline operation of the inspection device, and the inspection success rate is higher. Under the condition of the same endurance, the inspection route of the inspection mechanism of the present invention is longer.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present invention provides a method for processing a transmission line inspection image, the method comprising:

[0006] The first inspection device is installed on the corresponding first inspection aircraft; the first inspection aircraft includes at least a drone and a helicopter; the first inspection device is connected to a remote first server via wireless communication, and is connected to a flight control system, a navigation and positioning system, and a high-definition camera of the first inspection aircraft via data bus communication; the high-definition camera is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function;

[0007] The first inspection device receives the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map; and confirms the tower acquisition parameters according to the first data set to obtain the corresponding first acquisition parameter sequence; and identifies the starting and ending coordinates of this inspection according to the first high-precision map to obtain the corresponding first starting coordinates and the first ending coordinates; and sends the inspection route configuration instruction carrying the first starting coordinates, the first ending coordinates and the first high-precision map to the flight control system; and after receiving the inspection take-off instruction sent by the flight control system, performs inspection image acquisition processing on all towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain the corresponding first image set sequence and save it; and when receiving the inspection landing instruction sent by the flight control system, sends the saved first image set sequence to the first server.

[0008] Preferably, the first transmission line includes a plurality of first towers T i , 1≤tower index i≤N, N is the total number of towers of the first transmission line;

[0009] The first data set includes a plurality of first electric tower data D i ; The first tower data D i With the first electric tower T i One to one correspondence;

[0010] The first electric tower data D i It includes a first tower point cloud and a first tower coordinate; the first tower point cloud is the corresponding first tower X i The laser radar scanning point cloud of the first tower, the coordinates of each point in the point cloud of the first tower are the coordinates of the world coordinate system; the coordinates of the first tower are the corresponding X i The center point coordinates of

[0011] The coordinates of the starting and ending points of the center line of the first transmission line correspond to the coordinates of the center points of a power output mechanism and a power receiving mechanism respectively. i The coordinates of the first electric tower correspond to the coordinates of an intermediate point on the center line of the line;

[0012] The first high-precision map is a strip-shaped high-precision map; the map centerline of the first high-precision map coincides with the centerline of the line; the strip map width W of the first high-precision map is twice the standard transmission line safety distance, and the standard transmission line safety distance is the vertical safety distance from the centerline of the transmission line required by the power specification;

[0013] The first acquisition parameter sequence is composed of a plurality of first acquisition parameters A i The first acquisition parameter A i With the first electric tower T i One-to-one correspondence; the first acquisition parameter A i It is composed of the corresponding coordinates of the first power tower and a first area point cloud sequence; the first area point cloud sequence is formed by sequentially sorting a plurality of first area point clouds;

[0014] The first image set sequence consists of a plurality of first image sets P i The first image set P is sorted in sequence; i With the first electric tower T i One-to-one correspondence; the first image set P i It is composed of multiple first image information sequences; the first image information sequences correspond one-to-one to the first area point cloud; the first image information sequence is formed by sequentially sorting one or more first image information; the first image information includes a first acquired image and a first camera posture.

[0015] Preferably, the step of confirming the tower acquisition parameters according to the first data set to obtain a corresponding first acquisition parameter sequence specifically includes:

[0016] The first inspection device detects each of the first tower data D of the first data set. i The first tower point cloud is subjected to denoising and downsampling processing to obtain a corresponding first point cloud; and each of the first point clouds is subjected to ground point cloud filtering processing to obtain a corresponding second point cloud; and each of the second point clouds is subjected to tower edge point cloud extraction processing to obtain a corresponding third point cloud;

[0017] and using a preset point cloud target recognition model to perform corresponding tower type classification and recognition processing according to each of the second point clouds to obtain a corresponding first prediction vector; and taking the tower type corresponding to the first prediction probability with the largest probability in the first prediction vector as the corresponding first type; the first prediction vector includes a plurality of the first prediction probabilities, each of which corresponds to one tower type; the tower types include at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well-shaped type, and a door type;

[0018] And performing point cloud confirmation processing of the tower area according to each of the first types and the corresponding third point cloud to obtain the corresponding first area point cloud sequence;

[0019] And each of the first tower data D i The corresponding first tower coordinates and the first regional point cloud sequence constitute the corresponding first acquisition parameter A i; and all the first acquisition parameters A obtained i Form the corresponding first acquisition parameter sequence.

[0020] Furthermore, the point cloud target recognition model includes at least a PointNet model, a PointNet++ model, a VoxelNet model, a PointPillars model, and a PV-RCNN model;

[0021] The point cloud target recognition model is used to perform corresponding tower type classification and recognition processing according to the input tower point cloud and output a corresponding tower type prediction vector; the tower type prediction vector includes a plurality of tower type prediction probabilities, each of which corresponds to one tower type;

[0022] Before using the point cloud target recognition model, it is necessary to perform model training on the point cloud target recognition model in a supervised training manner based on a preset model data set; the model data set includes multiple first data records; the first data record includes a first training point cloud and a first labeled tower type; the first training point cloud is a pre-collected tower point cloud; the first labeled tower type is the tower type corresponding to the first training point cloud.

[0023] Further preferably, the step of performing model training on the point cloud target recognition model in a supervised training manner based on a preset model data set specifically includes:

[0024] Step 51, the first inspection device randomly divides the model data set into two sub-data sets as a corresponding first training set and a first evaluation set based on a preset first training evaluation ratio;

[0025] Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first training evaluation ratio;

[0026] Step 52, extracting the first first data record of the first training set as the corresponding current training record;

[0027] Step 53, and based on the vector length of the tower type prediction vector of the point cloud target recognition model, set an all-zero vector of equal length as the corresponding first training label vector; and reset the vector data corresponding to the first label tower type of the current training record in the first training label vector to 1;

[0028] Step 54, inputting the first training point cloud of the current training record into the point cloud target recognition model to perform corresponding tower type classification and recognition processing to obtain a corresponding first training prediction vector;

[0029] Step 55, and bring the first training prediction vector obtained this time and the corresponding first training label vector into a preset first model loss function to calculate and obtain a corresponding first loss value;

[0030] Wherein, the first model loss function includes L1 loss function, L2 loss function and cross entropy loss function;

[0031] Step 56, and identify whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identify whether the current training record is the last first data record of the first training set, if so, go to step 57, if not, extract the next first data record of the first training set as the new current training record and return to step 53; if the first loss value does not satisfy the first loss value range, then based on a preset first model parameter optimizer, perform a round of model parameter modulation on the point cloud target recognition model in the direction of minimizing the first model loss function, and return to step 54 at the end of this round of parameter modulation;

[0032] Wherein, the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer;

[0033] Step 57, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and set a full zero vector of equal length as the corresponding first evaluation label vector based on the vector length of the tower type prediction vector of the point cloud target recognition model; and reset the vector data corresponding to the first label tower type of the current evaluation record in the first evaluation label vector to 1; and input the first training point cloud of the current evaluation record into the point cloud target recognition model for corresponding tower type classification and recognition processing to obtain the corresponding first evaluation prediction vector; and form a corresponding first prediction-label pair with the first evaluation prediction vector obtained this time and the corresponding first evaluation label vector; and at the end of this round of traversal, bring all the first prediction-label pairs obtained into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0034] Wherein, the first model evaluation function at least includes an RMSE function;

[0035] Step 58, and identify whether the first evaluation value meets the preset first evaluation value range; if the first evaluation value does not meet the first evaluation value range, return to step 51 to continue training; if the first evaluation value meets the first evaluation value range, stop training and confirm that the model training of the point cloud target recognition model is completed.

[0036] Further, performing the tower area point cloud confirmation processing according to each of the first types and the corresponding third point cloud to obtain the corresponding first area point cloud sequence specifically includes:

[0037] The first inspection device uses each of the first types as the corresponding current type; and uses the third point cloud corresponding to the current type as the corresponding current point cloud;

[0038] And based on the current type, a preset first correspondence table is queried, and the first area segmentation strategy field of the first correspondence record in which the first tower type field in the first correspondence table matches the current type is extracted as the corresponding current strategy; wherein the first correspondence table is a correspondence table reflecting the correspondence between the tower type and the tower image acquisition area segmentation strategy; the first correspondence table includes multiple first correspondence records; the first correspondence record includes the first tower type field and the first area segmentation strategy field; the first tower type field includes at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well shape and a door type; the first area segmentation strategy field includes a vertical and horizontal cutting strategy, a tower top cutting strategy, a tower body cutting strategy and a tower base cutting strategy; the vertical and horizontal cutting strategy includes vertical cutting, horizontal cutting and vertical and horizontal cutting; the tower top, tower body and tower base cutting strategies are all composed of a height percentage range and an area segmentation height; the height percentage range is composed of a pair of minimum and maximum height percentages;

[0039] And take the XOY plane passing through the origin O in the world coordinate system as the first plane; and project each point in the current point cloud onto the first plane to obtain the corresponding first projection point; and identify the projection edge points of all the first projection points to obtain the corresponding multiple first projection edge points; and perform projection edge fitting on all the first projection edge points to obtain a closed first projection edge curve; and perform curve point sampling on the first projection edge curve to obtain the corresponding multiple first sampling points; and record the projection point of the first tower coordinate corresponding to the current point cloud on the first plane as the corresponding first center point; and use the connecting straight line between the first center point and each of the first sampling points as the corresponding first straight line; and use the line segment length where each of the first straight lines intersects with the first projection edge curve as the corresponding first line segment length, and use the first straight line corresponding to the longest first line segment length as the corresponding first transverse tangent; and on the first plane, the perpendicular line through the first center point to the first transverse tangent is recorded as the corresponding first longitudinal tangent;

[0040] And the vertical and horizontal cutting strategy, the tower top cutting strategy, the tower body cutting strategy and the tower base cutting strategy of the current strategy are used as the corresponding current vertical and horizontal cutting strategy, the current tower top cutting strategy, the current tower body cutting strategy and the current tower base cutting strategy;

[0041] and identifying the current longitudinal and transverse cutting strategy; if the current longitudinal and transverse cutting strategy is longitudinal cutting, then making an orthogonal plane of the first plane along the first longitudinal cutting line in the world coordinate system is recorded as a corresponding first section; if the current longitudinal and transverse cutting strategy is transverse cutting, then making an orthogonal plane of the first plane along the first transverse cutting line in the world coordinate system is recorded as a corresponding first section; if the current longitudinal and transverse cutting strategy is longitudinal and transverse cutting, then making an orthogonal plane of the first plane along the first longitudinal cutting line and the first transverse cutting line in the world coordinate system respectively to obtain two corresponding first sections;

[0042] The minimum and maximum coordinates in the Z-axis direction of the current point cloud are extracted to obtain the corresponding minimum height coordinate z min and the maximum height coordinate z max ; and based on the minimum height coordinate z min and the maximum height coordinate z max Calculate the corresponding current tower height = |z max -z min|; and on the current section, based on the height percentage range of the current tower height and the current tower top segmentation strategy, the current tower body segmentation strategy and the current tower base segmentation strategy, define the corresponding tower top height range, tower body height range and tower base height range; and extract the point clouds whose height coordinates in the current point cloud satisfy the tower top, tower body and tower base height ranges respectively as the corresponding tower point cloud, tower body point cloud and tower base point cloud; wherein the tower top, tower body and tower base height ranges are each composed of a pair of minimum and maximum heights;

[0043] And use each of the first sections obtained as the corresponding current section; use the tower top, tower body and tower base point clouds as the corresponding current local point clouds in sequence; use the projection area of ​​the current local point cloud on the current section as the corresponding current projection area; use the area segmentation height corresponding to the current projection area as the corresponding current segmentation height; use the absolute difference between the maximum and minimum heights of the tower top, tower body or tower base height range corresponding to the current projection area as the corresponding current projection area height; use the absolute difference between the maximum and minimum coordinates of the current projection area on the horizontal coordinate axis of the current section as the corresponding current projection area width; and based on the current projection area height and the current projection area width degree, a rectangle parallel to the horizontal coordinate axis of the current section and completely covering the current projection area is made on the current section plane and recorded as the corresponding current projection area rectangle; and the current projection area rectangle is longitudinally segmented based on the current segmentation height and the preset adjacent segmentation area overlap to obtain a plurality of corresponding first segmentation rectangles; and the sub-point clouds corresponding to each of the first segmentation rectangles in the current local point cloud are extracted as the corresponding first area point clouds; wherein the height of each first segmentation rectangle matches the current segmentation height, the width matches the current projection area width, and the overlap of every two adjacent first segmentation rectangles is greater than or equal to the adjacent segmentation area overlap, which is less than 1;

[0044] All the obtained first region point clouds are sequentially sorted to form a corresponding first region point cloud sequence.

[0045] Preferably, identifying the starting and ending coordinates of the inspection according to the first high-precision map to obtain the corresponding first starting coordinates and first ending coordinates specifically includes:

[0046] The first inspection device extracts the starting and ending point coordinates of the map center line of the first high-precision map as the corresponding first starting coordinates and the first ending coordinates.

[0047] Preferably, performing inspection image acquisition processing on all the towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain and save a corresponding first image set sequence specifically includes:

[0048] The first inspection device sets a current positioning coordinate with a three-dimensional coordinate initialized to all zeros; sets a current tower spacing initialized to 0; sets a current tower index initialized to 0; initializes an empty sequence as the corresponding current center coordinate sequence; sets a current shooting state initialized to stop shooting; sets a current processing state initialized to processing; the current processing state includes processing and stopping processing; the current shooting state includes starting shooting and stopping shooting;

[0049] And when the current processing state is processing, regularly send real-time positioning query instructions to the navigation and positioning system according to a preset first time frequency; and receive the real-time positioning coordinates sent back by the navigation and positioning system; and calculate the real-time coordinates of the axis of the high-definition camera based on the real-time positioning coordinates received this time and the relative coordinate offset between the axis of the high-definition camera and the center of mass of the first inspection aircraft, and update the current positioning coordinates based on the real-time coordinates of the axis obtained this time; wherein, the real-time positioning coordinates are the real-time coordinates of the center of mass of the first inspection aircraft;

[0050] Each time the current positioning coordinate is updated, the current positioning coordinate and each of the first acquisition parameters A of the first acquisition parameter sequence are compared. i The straight-line distance of the first electric tower coordinates is calculated to obtain the corresponding first distance; and the first distance with the smallest distance is used as the corresponding current shortest distance; and the first acquisition parameter A corresponding to the current shortest distance is used as the first acquisition parameter A corresponding to the first electric tower coordinates. i The index i is used as the corresponding current index; and the current tower spacing is updated based on the current shortest distance obtained this time; and whether the current index matches the current tower index is identified, and if not, the current tower index is updated based on the current index;

[0051] When the current tower index is updated once, the first acquisition parameter A corresponding to the current tower index is updated. iThe first regional point cloud sequence is used as the corresponding current regional point cloud sequence; and the point cloud center coordinates of each of the first regional point clouds in the current regional point cloud sequence are identified to obtain the corresponding first center coordinates; and the straight-line distance between the current positioning coordinates and each of the first center coordinates is calculated to obtain the corresponding second distance; and all the first center coordinates are sequentially sorted in the order of the second distance from small to large to obtain the corresponding first center coordinate sequence; and the current center coordinate sequence is updated once based on the first center coordinate sequence obtained this time;

[0052] And when the current distance between the electric towers is updated each time, it is identified whether the current distance between the electric towers is less than or equal to the preset first distance threshold; if so, the current shooting state is updated to start shooting; if not, the current shooting state is updated to stop shooting;

[0053] And each time the current shooting state is updated from stop shooting to start shooting, a round of traversal is performed on all the first center coordinates of the current center coordinate sequence; and in this round of traversal, the first center coordinate currently traversed is used as the corresponding current center coordinate; and the axial direction of the high-definition camera is maintained in the direction toward the current center coordinate; and the first area point cloud corresponding to the current center coordinate is used as the corresponding current area point cloud; and the point coordinates in the current area point cloud whose Z-axis coordinate is consistent with the Z-axis component coordinate of the current center coordinate and whose Y-axis coordinate is smaller than the Y-axis component coordinate of the current center coordinate are marked as corresponding first reference coordinates to form a corresponding first reference coordinate sequence; and each of the first reference coordinates in the first reference coordinate sequence is used as the corresponding current reference coordinate in turn; and the latest current positioning coordinate is compared with the current reference coordinate. The straight-line distance is calculated to obtain the corresponding current distance; and the high-definition camera is called to use the current center coordinate as the visual axis center coordinate and automatically focus according to the current distance, and at the end of this automatic focusing, the image is taken based on the principle of rapid reflection mirror shift compensation with the current reference coordinate as the rapid reflection mirror tracking coordinate to obtain the corresponding first acquisition image; and the real-time posture of the high-definition camera at the current moment is used as the corresponding first camera posture; and the first acquisition image and the corresponding first camera posture form a corresponding first image information; and all the first image information corresponding to the current first reference coordinate sequence are sequentially sorted to form a corresponding first image information sequence; and at the end of this round of traversal, all the first image information sequences obtained in this round of traversal form a first image set P whose index i matches the current tower index. i and save;

[0054] And in the Nth first image set Pi When the saving is successful, the current processing state is updated to stop processing;

[0055] Each time the current processing state is updated from being processed to stopped processing, all the first image sets P currently saved are i The first image set sequence is sorted in sequence to form a corresponding first image set sequence and saved; and when the first image set sequence is saved successfully, all the first image sets P currently saved are i delete.

[0056] The method further comprises:

[0057] The flight control system is connected to the navigation and positioning system via a data bus communication method;

[0058] The flight control system is used to extract the corresponding first starting coordinates, the first ending coordinates and the first high-precision map from the inspection route configuration instruction when receiving the inspection route configuration instruction sent by the first inspection device; and use the first starting coordinates as the corresponding current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the corresponding current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection take-off instruction to the first inspection device at the time of successful take-off; and during the flight, use the real-time positioning coordinates regularly pushed by the navigation and positioning system as the latest current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the new current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection landing instruction to the first inspection device at the time of successful landing;

[0059] The navigation and positioning system is a combined navigation system, which is composed of a gyroscope, an accelerometer and a GNSS system; the navigation and positioning system is used to measure the real-time coordinates of the center of mass of the first inspection aircraft in real time and refresh the real-time positioning coordinates based on the latest measurement results; and regularly push the latest real-time positioning coordinates to the flight control system according to a preset second time frequency; and upon receiving the real-time positioning query instruction sent by the first inspection device, send the latest real-time positioning coordinates back to the first inspection device;

[0060] The first server is used to send the line tower data set and the line high-precision map of the first transmission line specified by the operator to the first inspection device; and is also used to identify the working conditions of each tower of the first transmission line according to the first image set sequence received this time after receiving the first image set sequence sent by the first inspection device.

[0061] A second aspect of an embodiment of the present invention provides a device for implementing the method for processing a power transmission line inspection image according to the first aspect, the device comprising: a remote data transceiver module, a first data preprocessing module, a second data preprocessing module and an inspection image acquisition module;

[0062] The device is installed on the corresponding first inspection aircraft; the first inspection aircraft includes at least a drone and a helicopter; the device is connected to a remote first server via wireless communication, and is connected to the flight control system, navigation and positioning system and high-definition camera of the first inspection aircraft via data bus communication; the high-definition camera is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function;

[0063] The remote data transceiver module is used to receive the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map;

[0064] The first data preprocessing module is used to confirm the tower acquisition parameters according to the first data set to obtain a corresponding first acquisition parameter sequence;

[0065] The second data preprocessing module is used to identify the starting and ending coordinates of the inspection according to the first high-precision map to obtain the corresponding first starting coordinates and first ending coordinates; and send the inspection route configuration instruction carrying the first starting coordinates, the first ending coordinates and the first high-precision map to the flight control system;

[0066] The inspection image acquisition module is used to, after receiving the inspection takeoff instruction sent by the flight control system, perform inspection image acquisition processing on all the towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain and save the corresponding first image set sequence;

[0067] The remote data transceiver module is further configured to send the stored first image set sequence to the first server upon receiving a patrol landing instruction sent by the flight control system.

[0068] A third aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0069] The processor is used to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

[0070] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0071] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.

[0072] The embodiment of the present invention provides a method, device, electronic device and computer-readable storage medium for processing inspection images of power transmission lines. As can be seen from the above content, the embodiment of the present invention installs an inspection device on the inspection aircraft, and connects it to a remote server and the flight control system, navigation and positioning system and high-definition camera on the aircraft; and before the flight inspection, the inspection device performs data preprocessing based on the tower data set + high-precision map of the power transmission line sent by the remote server to obtain the corresponding tower acquisition parameter sequence and inspection start / end coordinates; and sends the inspection start / end coordinates + high-precision map to the flight control system for flight trajectory planning and flight control; and during the flight inspection, all towers of the transmission line are inspected and processed according to the navigation and positioning system, high-definition camera and tower acquisition parameter sequence to generate a corresponding image set sequence and save it; and at the end of the flight inspection, the saved image set sequence is sent to the remote server. The embodiment of the present invention provides a new aircraft inspection mechanism, which can perform inspections without human-computer interaction or hovering operations. Compared with conventional manual inspection methods, the inspection cycle of the embodiment of the present invention is significantly shortened, the efficiency is significantly improved, and the real-time performance is significantly enhanced. For some power towers with special installation locations (such as mountains, deserts, etc.), the inspection intensity can also meet the high-frequency inspection requirements. Compared with conventional aircraft-based inspection methods, the inspection cycle of the embodiment of the present invention will be shorter, the efficiency will be higher, and the real-time performance will be stronger. In an inspection environment with poor network quality, the tower image acquisition can be completed based on the offline operation of the inspection device, so that the inspection success rate is higher. Under the same endurance, the inspection route of the inspection mechanism of the embodiment of the present invention will also be longer. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic diagram of a method for processing a power transmission line inspection image provided in Embodiment 1 of the present invention;

[0074] Figure 2 A schematic diagram of a first power transmission line and a first high-precision map provided in Embodiment 1 of the present invention;

[0075] Figure 3 A module structure diagram of a transmission line inspection image processing device provided in Embodiment 2 of the present invention;

[0076] Figure 4 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0078] Embodiment 1 of the present invention provides a method for processing a transmission line inspection image, such as Figure 1 As shown in the schematic diagram of a method for processing a power transmission line inspection image provided in the first embodiment of the present invention, the method mainly comprises the following steps:

[0079] Step 1: Install the first inspection device on the corresponding first inspection aircraft.

[0080] Here, the first inspection aircraft in the embodiment of the present invention includes at least a UAV and a helicopter, and of course may also include any other type of flying equipment.

[0081] The first inspection device of the embodiment of the present invention is a module, device, terminal, equipment, computer, server or system that can implement the method steps of the embodiment of the present invention; the first inspection device is connected to the remote first server through wireless communication, and is connected to the flight control system, navigation and positioning system and high-definition camera of the first inspection aircraft through data bus communication; the flight control system is connected to the navigation and positioning system through data bus communication; the wireless communication method here includes at least Bluetooth, NFC, WiFi, 4G, 5G, LTE, IOT and other communication methods; the data bus communication method here includes at least serial bus, CAN bus, etc.; the navigation and positioning system here is a combined navigation system, composed of a gyroscope, an accelerometer and a GNSS system; the high-definition camera here is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function.

[0082] Step 2, the first inspection device receives the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map; and confirms the tower acquisition parameters according to the first data set to obtain the corresponding first acquisition parameter sequence; and identifies the start and end coordinates of this inspection according to the first high-precision map to obtain the corresponding first start coordinates and first end coordinates; and sends the inspection route configuration instruction carrying the first start coordinates, the first end coordinates and the first high-precision map to the flight control system; and after receiving the inspection takeoff instruction sent by the flight control system, according to the navigation positioning system, the high-definition camera and the first acquisition parameter sequence, the inspection image acquisition processing is performed on all the towers of the first transmission line to obtain the corresponding first image set sequence and save it; and when receiving the inspection landing instruction sent by the flight control system, the saved first image set sequence is sent to the first server;

[0083] The first image set sequence consists of a plurality of first image sets P i The first image set P is sorted in sequence; i With the first electric tower T i One-to-one correspondence; the first image set P i It is composed of a plurality of first image information sequences; the first image information sequences correspond to the first area point cloud one by one; the first image information sequence is formed by sequentially sorting one or more first image information; the first image information includes a first acquired image and a first camera pose;

[0084] The current step 2 specifically includes:

[0085] Step 21, the first inspection device receives the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map;

[0086] The first transmission line includes a plurality of first towers T i , 1≤tower index i≤N, N is the total number of towers of the first transmission line;

[0087] The first data set includes a plurality of first electric tower data D i ; First Tower Data D i With the first electric tower T i One-to-one correspondence; each first tower data D i Including the first tower point cloud and the first tower coordinates; the first tower point cloud is the corresponding first tower X i The laser radar scanning point cloud, the coordinates of each point in the first tower point cloud are the coordinates of the world coordinate system; the coordinates of the first tower are the corresponding X i The center point coordinates of the first tower are also world coordinate system coordinates;

[0088] It should be noted that if Figure 2 As shown in the schematic diagram of the first transmission line and the first high-precision map provided in the first embodiment of the present invention: the coordinates of the starting and ending points of the center line of the first transmission line correspond to the coordinates of the center points of a power output mechanism and a power receiving mechanism respectively, and each first tower T on the first transmission line i The coordinates of the first electric tower correspond to the coordinates of an intermediate point on the center line of the line; the first high-precision map is a strip-shaped high-precision map, the map center line of the first high-precision map coincides with the center line of the line, and the strip map width W of the first high-precision map is twice the standard transmission line safety distance, and the standard transmission line safety distance is the vertical safety distance from the center line of the transmission line required by the power specification;

[0089] Step 22, confirming the tower acquisition parameters according to the first data set to obtain a corresponding first acquisition parameter sequence;

[0090] The first acquisition parameter sequence is composed of a plurality of first acquisition parameters A i The first acquisition parameter A i With the first electric tower T i One-to-one correspondence; first acquisition parameter A i It is composed of the corresponding first power tower coordinates and the first area point cloud sequence; the first area point cloud sequence is formed by sequentially sorting a plurality of first area point clouds;

[0091] Specifically, step 221 includes: the first inspection device checks each first tower data D of the first data set. i The first power tower point cloud is subjected to denoising and downsampling processing to obtain the corresponding first point cloud; and each first point cloud is subjected to ground point cloud filtering processing to obtain the corresponding second point cloud; and each second point cloud is subjected to power tower edge point cloud extraction processing to obtain the corresponding third point cloud;

[0092] Here, the embodiment of the present invention adopts conventional point cloud denoising, point cloud downsampling, ground point cloud recognition and filtering, edge point cloud recognition and extraction processing methods to perform corresponding denoising, downsampling, ground point cloud filtering and tower edge point cloud extraction operations on the first tower point cloud;

[0093] Step 222, using a preset point cloud target recognition model to perform corresponding tower type classification and recognition processing according to each second point cloud to obtain a corresponding first prediction vector; and taking the tower type corresponding to the first prediction probability with the largest probability in the first prediction vector as the corresponding first type;

[0094] The first prediction vector includes a plurality of first prediction probabilities, each of which corresponds to a tower type; the tower types include at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well-shaped type, and a door type;

[0095] Here, the point cloud target recognition model of the embodiment of the present invention includes at least a PointNet model, a PointNet++ model, a VoxelNet model, a PointPillars model, and a PV-RCNN model; the point cloud target recognition model is used to perform corresponding tower type classification and recognition processing according to the input tower point cloud and output a corresponding tower type prediction vector; here, the tower type prediction vector output by the model includes multiple tower type prediction probabilities, and each tower type prediction probability corresponds to a tower type;

[0096] It should be noted that, before using the point cloud target recognition model, the first inspection device of the embodiment of the present invention needs to perform model training on the point cloud target recognition model in a supervised training manner based on a preset model data set; the model data set here is a data set pre-constructed through data collection, and is composed of multiple first data records; each first data record includes a first training point cloud and a first label electric tower type; the first training point cloud is a pre-collected electric tower point cloud; the first label electric tower type is the electric tower type corresponding to the first training point cloud;

[0097] It should also be noted that the processing steps of the first inspection device of the embodiment of the present invention for training the point cloud target recognition model in a supervised training manner based on the preset model data set are shown in the following steps S1-S8:

[0098] Step S1, the first inspection device randomly divides the model data set into two sub-data sets as a corresponding first training set and a first evaluation set based on a preset first training evaluation ratio;

[0099] The first training evaluation ratio is a preset ratio value, such as 8:2; the first training set and the first evaluation set both include a plurality of first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first training evaluation ratio;

[0100] Step S2, extracting the first first data record of the first training set as the corresponding current training record;

[0101] Step S3, and based on the vector length of the tower type prediction vector of the point cloud target recognition model, set an all-zero vector of equal length as the corresponding first training label vector; and reset the vector data corresponding to the first label tower type of the current training record in the first training label vector to 1;

[0102] Step S4, inputting the first training point cloud of the current training record into the point cloud target recognition model to perform corresponding tower type classification and recognition processing to obtain the corresponding first training prediction vector;

[0103] Step S5, bringing the first training prediction vector obtained this time and the corresponding first training label vector into the preset first model loss function to calculate and obtain the corresponding first loss value;

[0104] Among them, the first model loss function includes L1 loss function, L2 loss function and cross entropy loss function;

[0105] Step S6, and identify whether the first loss value meets the preset first loss value range; if the first loss value meets the first loss value range, identify whether the current training record is the last first data record of the first training set, if so, go to step S7, if not, extract the next first data record of the first training set as the new current training record and return to step S3; if the first loss value does not meet the first loss value range, then based on the preset first model parameter optimizer, perform a round of model parameter modulation on the point cloud target recognition model in the direction of minimizing the first model loss function, and return to step S4 at the end of this round of parameter modulation;

[0106] The first loss value range is a preset loss value range; the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer;

[0107] Step S7, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and set a full zero vector of equal length as the corresponding first evaluation label vector based on the vector length of the tower type prediction vector of the point cloud target recognition model; and reset the vector data corresponding to the first label tower type of the current evaluation record in the first evaluation label vector to 1; and input the first training point cloud of the current evaluation record into the point cloud target recognition model for corresponding tower type classification and recognition processing to obtain the corresponding first evaluation prediction vector; and form a corresponding first prediction-label pair from the first evaluation prediction vector obtained this time and the corresponding first evaluation label vector; and at the end of this round of traversal, bring all the obtained first prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0108] Wherein, the first model evaluation function at least includes the RMSE function;

[0109] Step S8, and identify whether the first evaluation value meets the preset first evaluation value range; if the first evaluation value does not meet the first evaluation value range, return to step S1 to continue training; if the first evaluation value meets the first evaluation value range, stop training and confirm that the model training of the point cloud target recognition model is completed;

[0110] Wherein, the first evaluation value range is a preset evaluation value range;

[0111] Step 223, performing point cloud confirmation processing on the tower area according to each first type and the corresponding third point cloud to obtain a corresponding first area point cloud sequence;

[0112] Specifically, it includes: step 2231, the first inspection device takes each first type as the corresponding current type; and takes the third point cloud corresponding to the current type as the corresponding current point cloud;

[0113] Step 2232, based on the current type, query the preset first correspondence table, extract the first area segmentation strategy field of the first correspondence record matching the first tower type field with the current type in the first correspondence table as the corresponding current strategy;

[0114] Among them, the first correspondence table is a correspondence table reflecting the correspondence between the tower type and the tower image acquisition area segmentation strategy; the first correspondence table includes multiple first correspondence records; the first correspondence record includes a first tower type field and a first area segmentation strategy field; the first tower type field includes at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well shape, and a door type; the first area segmentation strategy field includes a vertical and horizontal segmentation strategy, a tower top segmentation strategy, a tower body segmentation strategy, and a tower base segmentation strategy; the vertical and horizontal segmentation strategy includes vertical segmentation, horizontal segmentation, and vertical and horizontal segmentation; the tower top, tower body, and tower base segmentation strategies are all composed of a height percentage range and an area segmentation height; the height percentage range is composed of a pair of minimum and maximum height percentages;

[0115] Step 2233, and take the XOY plane through the origin O in the world coordinate system as the first plane; and project each point in the current point cloud onto the first plane to obtain the corresponding first projection point; and perform projection edge point identification on all the obtained first projection points to obtain corresponding multiple first projection edge points; and perform projection edge fitting on all the obtained first projection edge points to obtain a closed first projection edge curve; and perform curve point sampling on the first projection edge curve to obtain corresponding multiple first sampling points; and record the projection point of the first tower coordinate corresponding to the current point cloud on the first plane as the corresponding first center point; and use the connecting straight line between the first center point and each first sampling point as the corresponding first straight line; and use the line segment length where each first straight line intersects the first projection edge curve as the corresponding first line segment length, and use the first straight line corresponding to the longest first line segment length as the corresponding first transverse tangent; and on the first plane, the perpendicular line of the first transverse tangent through the first center point is recorded as the corresponding first longitudinal tangent;

[0116] Step 2234, and use the vertical and horizontal cutting strategy, tower top cutting strategy, tower body cutting strategy and tower base cutting strategy of the current strategy as the corresponding current vertical and horizontal cutting strategy, current tower top cutting strategy, current tower body cutting strategy and current tower base cutting strategy;

[0117] Step 2235, and identify the current vertical and horizontal cutting strategy; if the current vertical and horizontal cutting strategy is vertical cutting, then make an orthogonal plane of the first plane along the first vertical cutting line in the world coordinate system and record it as a corresponding first section; if the current vertical and horizontal cutting strategy is horizontal cutting, then make an orthogonal plane of the first plane along the first horizontal cutting line in the world coordinate system and record it as a corresponding first section; if the current vertical and horizontal cutting strategy is vertical and horizontal cutting, then make an orthogonal plane of the first plane along the first vertical cutting line and the first horizontal cutting line in the world coordinate system respectively to obtain two corresponding first sections;

[0118] Step 2236, extract the minimum and maximum coordinates in the Z-axis direction of the current point cloud to obtain the corresponding minimum height coordinate z min and the maximum height coordinate z max ; and based on the minimum height coordinate z min and the maximum height coordinate z max Calculate the corresponding current tower height = |z max -z min |; and on the current section, based on the current tower height and the height percentage range of the current tower top segmentation strategy, the current tower body segmentation strategy and the current tower base segmentation strategy, define the corresponding tower top height range, tower body height range and tower base height range; and extract the point clouds whose height coordinates in the current point cloud meet the tower top, tower body and tower base height ranges respectively as the corresponding tower top point cloud, tower body point cloud and tower base point cloud;

[0119] Among them, the height ranges of the tower top, tower body and tower base are all composed of a pair of minimum and maximum heights;

[0120] Step 2237, and use each obtained first section as the corresponding current section; and use the tower top, tower body and tower base point clouds as the corresponding current local point clouds in sequence; and use the projection area of ​​the current local point cloud on the current section as the corresponding current projection area; and use the regional segmentation height corresponding to the current projection area as the corresponding current segmentation height; and use the absolute difference between the maximum and minimum heights of the tower top, tower body or tower base height range corresponding to the current projection area as the corresponding current projection area height; and use the absolute difference between the maximum and minimum coordinates of the current projection area on the horizontal coordinate axis of the current section as the corresponding current projection area width; and based on the current projection area height and the current projection area width, make a rectangle on the current section that is parallel to the horizontal coordinate axis of the current section and can completely cover the current projection area, and record it as the corresponding current projection area rectangle; and based on the current segmentation height and the preset adjacent segmentation area overlap, perform longitudinal segmentation processing on the current projection area rectangle to obtain corresponding multiple first segmentation rectangles; and extract the sub-point clouds corresponding to each first segmentation rectangle in the current local point cloud as the corresponding first regional point cloud;

[0121] The overlap of adjacent segmentation regions is a preset percentage, and the preset value of the overlap of adjacent segmentation regions is less than 1; the height of each first segmentation rectangle matches the current segmentation height, and the width matches the current projection area width, and the overlap of every two adjacent first segmentation rectangles is greater than or equal to the overlap of adjacent segmentation regions, which is less than 1;

[0122] Step 2238, all the first region point clouds obtained are sequentially sorted to form a corresponding first region point cloud sequence;

[0123] Step 224, and by each first tower data D i The corresponding first tower coordinates and the first area point cloud sequence constitute the corresponding first acquisition parameter A i ; and all the first acquisition parameters A are obtained i forming a corresponding first acquisition parameter sequence;

[0124] Step 23, identifying the starting and ending coordinates of this inspection according to the first high-precision map to obtain the corresponding first starting coordinates and first ending coordinates;

[0125] Specifically, the first inspection device extracts the coordinates of the start and end points of the center line of the first high-precision map as the corresponding first start coordinates and first end coordinates;

[0126] Step 24, and sending the inspection route configuration instruction carrying the first starting coordinates, the first ending coordinates and the first high-precision map to the flight control system;

[0127] Step 25, after receiving the inspection takeoff instruction sent by the flight control system, performing inspection image acquisition processing on all the towers of the first transmission line according to the navigation positioning system, the high-definition camera and the first acquisition parameter sequence to obtain the corresponding first image set sequence and save it;

[0128] Specifically, the method comprises: step 251, the first inspection device sets a current positioning coordinate with a three-dimensional coordinate initialized to all zeros; sets a current tower spacing initialized to 0; sets a current tower index initialized to 0; initializes an empty sequence as the corresponding current center coordinate sequence; sets a current shooting state initialized to stop shooting; and sets a current processing state initialized to processing;

[0129] The current processing state includes processing and stopping processing; the current shooting state includes starting shooting and stopping shooting;

[0130] Step 252, when the current processing state is processing, regularly send a real-time positioning query instruction to the navigation and positioning system according to the preset first time frequency; and receive the real-time positioning coordinates sent back by the navigation and positioning system; and calculate the real-time coordinates of the axis of the high-definition camera based on the real-time positioning coordinates received this time and the relative coordinate offset between the axis of the high-definition camera and the center of mass of the first inspection aircraft, and update the current positioning coordinates based on the real-time coordinates of the axis obtained this time;

[0131] The first time frequency is a preset frequency value; the real-time positioning coordinates are the real-time coordinates of the center of mass of the first inspection aircraft;

[0132] Step 253, when the current positioning coordinates are updated once, the current positioning coordinates and each first acquisition parameter A of the first acquisition parameter sequence are compared. i The straight-line distance of the first electric tower coordinates is calculated to obtain the corresponding first distance; and the first distance with the smallest distance is taken as the corresponding current shortest distance; and the first acquisition parameter A corresponding to the current shortest distance is taken as i The index i of the tower is used as the corresponding current index; and the current tower spacing is updated based on the current shortest distance obtained this time; and whether the current index matches the current tower index is identified, and if not, the current tower index is updated based on the current index;

[0133] Step 254, each time the current tower index is updated, the first acquisition parameter A corresponding to the current tower index is updated. iThe first area point cloud sequence is used as the corresponding current area point cloud sequence; the point cloud center coordinates of each first area point cloud in the current area point cloud sequence are identified to obtain the corresponding first center coordinates; the straight-line distance between the current positioning coordinates and each first center coordinate is calculated to obtain the corresponding second distance; all the first center coordinates are sequentially sorted in the order of the second distance from small to large to obtain the corresponding first center coordinate sequence; and the current center coordinate sequence is updated once based on the first center coordinate sequence obtained this time;

[0134] Step 255, when the current tower spacing is updated once, it is identified whether the current tower spacing is less than or equal to the preset first distance threshold; if so, the current shooting state is updated to start shooting; if not, the current shooting state is updated to stop shooting;

[0135] Here, the first distance threshold is a preset length parameter;

[0136] Step 256, each time the current shooting state is updated from stop shooting to start shooting, all the first center coordinates of the current center coordinate sequence are traversed once; and in this traversal process, the first center coordinate currently traversed is used as the corresponding current center coordinate; and the axial direction of the high-definition camera is maintained in the direction toward the current center coordinate; and the first area point cloud corresponding to the current center coordinate is used as the corresponding current area point cloud; and the point coordinates in the current area point cloud whose Z-axis coordinate is consistent with the Z-axis component coordinate of the current center coordinate and whose Y-axis coordinate is smaller than the Y-axis component coordinate of the current center coordinate are marked as corresponding first reference coordinates to form a corresponding first reference coordinate sequence; and each first reference coordinate of the first reference coordinate sequence is used as the corresponding current reference coordinate in turn; and the latest current positioning coordinate is compared with the current The straight-line distance of the reference coordinate is calculated to obtain the corresponding current distance; and the high-definition camera is called to use the current center coordinate as the visual axis center coordinate and automatically focus according to the current distance, and at the end of this automatic focusing, the current reference coordinate is used as the fast reflector tracking coordinate based on the fast reflector mirror shift compensation principle to capture the image to obtain the corresponding first acquisition image; and the real-time posture of the high-definition camera at the current moment is used as the corresponding first camera posture; and the first acquisition image and the corresponding first camera posture form a corresponding first image information; and all the first image information corresponding to the current first reference coordinate sequence are sequentially sorted to form a corresponding first image information sequence; and at the end of this round of traversal, all the first image information sequences obtained in this round of traversal form a first image set P with an index i matching the current tower index. i and save;

[0137] Step 257, and in the Nth first image set P i When saving is successful, the current processing status is updated to stop processing;

[0138] Step 258, each time the current processing state is updated from being processed to stopped processing, all the first image sets P currently saved are i The first image set sequence is sorted in sequence to form a corresponding first image set sequence and saved; and when the first image set sequence is saved successfully, all the first image sets P currently saved are i delete;

[0139] Step 26: When receiving the inspection and landing instruction sent by the flight control system, the saved first image set sequence is sent to the first server.

[0140] The above steps 1-2 provide a description of the processing steps of the first inspection device of an embodiment of the present invention. In the above steps, there are corresponding data interaction processes between the first inspection device and the flight control system, the navigation and positioning system, and the first server. In order to have a comprehensive connection with the technical solution of the embodiment of the present invention, some processing steps of the flight control system, the navigation and positioning system, and the first server in the embodiment of the present invention are also briefly described below.

[0141] The flight control system of an embodiment of the present invention is used to extract the corresponding first starting coordinates, the first ending coordinates and the first high-precision map from the inspection route configuration instruction when receiving the inspection route configuration instruction sent by the first inspection device; and use the first starting coordinates as the corresponding current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the corresponding current flight trajectory; and perform flight control based on the current flight trajectory; and at the moment of successful take-off, send an inspection take-off instruction to the first inspection device; and during the flight, use the real-time positioning coordinates regularly pushed by the navigation and positioning system as the latest current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain a new current flight trajectory; and perform flight control based on the current flight trajectory; and at the moment of successful landing, send an inspection landing instruction to the first inspection device.

[0142] The navigation and positioning system of an embodiment of the present invention is used to measure the real-time coordinates of the center of mass of the first inspection aircraft in real time and refresh the real-time positioning coordinates based on the latest measurement results; and regularly push the latest real-time positioning coordinates to the flight control system at a preset second time frequency; and upon receiving the real-time positioning query instruction sent by the first inspection device, send the latest real-time positioning coordinates back to the first inspection device; the second time frequency here is a pre-set frequency value.

[0143] The first server of the embodiment of the present invention is used to send the line tower data set and the line high-precision map of the first transmission line specified by the operator to the first inspection device; and is also used to identify the working conditions of each tower of the first transmission line according to the first image set sequence received this time after receiving the first image set sequence sent by the first inspection device. Here, when the first server identifies the working conditions of each tower of the first transmission line according to the first image set sequence received this time, it can be identified by manual identification, machine identification, or intelligent model identification.

[0144] Figure 3 The module structure diagram of a transmission line inspection image processing device provided in the second embodiment of the present invention is a terminal device or server that implements the function of the first inspection device in the aforementioned method embodiment, and can also be a device that enables the aforementioned terminal device or server to implement the function of the first inspection device in the aforementioned method embodiment. For example, the device can be a device or chip system of the aforementioned terminal device or server. Figure 3 As shown, the transmission line inspection image processing device 1014 provided in the second embodiment of the present invention includes: a remote data transceiver module 10141, a first data preprocessing module 10142, a second data preprocessing module 10143 and an inspection image acquisition module 10144.

[0145] The transmission line inspection image processing device 1014 is installed on the corresponding first inspection aircraft 101; the first inspection aircraft 101 includes at least a drone and a helicopter; the transmission line inspection image processing device 1014 is connected to the remote first server 102 through wireless communication, and is connected to the flight control system 1011, navigation and positioning system 1012 and high-definition camera 1013 of the first inspection aircraft 101 through data bus communication; the high-definition camera 1013 is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function.

[0146] The remote data transceiver module 10141 is used to receive the line tower data set and the line high-precision map of the first transmission line sent by the first server 102 as the corresponding first data set and the first high-precision map. The first data preprocessing module 10142 is used to confirm the tower acquisition parameters according to the first data set to obtain the corresponding first acquisition parameter sequence. The second data preprocessing module 10143 is used to identify the start and end coordinates of this inspection according to the first high-precision map to obtain the corresponding first start coordinates and first end coordinates; and send the inspection route configuration instruction carrying the first start coordinates, the first end coordinates and the first high-precision map to the flight control system 1011. The inspection image acquisition module 10144 is used to, after receiving the inspection takeoff instruction sent by the flight control system 1011, perform inspection image acquisition processing on all the towers of the first transmission line according to the navigation positioning system 1012, the high-definition camera 1013 and the first acquisition parameter sequence to obtain the corresponding first image set sequence and save it. The remote data transceiver module 10141 is further configured to send the stored first image set sequence to the first server 102 upon receiving a patrol landing instruction sent by the flight control system 1011 .

[0147] In addition, the flight control system 1011 is connected to the navigation and positioning system 1012 through a data bus communication method. The flight control system 1011 is used to extract the corresponding first starting coordinates, the first ending coordinates and the first high-precision map from the inspection route configuration instruction when receiving the inspection route configuration instruction sent by the processing device 1014 of the power line inspection image; and use the first starting coordinates as the corresponding current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the corresponding current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection takeoff instruction to the processing device 1014 of the power line inspection image at the time of successful takeoff; and during the flight, use the real-time positioning coordinates regularly pushed by the navigation and positioning system 1012 as the latest current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain a new current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection landing instruction to the processing device 1014 of the power line inspection image at the time of successful landing.

[0148] The navigation and positioning system 1012 is a combined navigation system, which is composed of a gyroscope, an accelerometer and a GNSS system. The navigation and positioning system 1012 is used to measure the real-time coordinates of the center of mass of the first inspection aircraft 101 in real time and refresh the real-time positioning coordinates based on the latest measurement results; and regularly push the latest real-time positioning coordinates to the flight control system 1011 according to the preset second time frequency; and when receiving the real-time positioning query instruction sent by the processing device 1014 of the transmission line inspection image, the latest real-time positioning coordinates are sent back to the device.

[0149] The first server 102 is used to send the line tower data set and line high-precision map of the first transmission line specified by the operator to the transmission line inspection image processing device 1014; it is also used to identify the working conditions of each tower of the first transmission line according to the first image set sequence received this time after receiving the first image set sequence sent by the transmission line inspection image processing device 1014.

[0150] An embodiment of the present invention provides a transmission line inspection image processing device, which can execute the method steps of the first inspection device function in the above method embodiment. Its implementation principle and technical effect are similar and will not be repeated here.

[0151] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the remote data transceiver module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The function of the above-mentioned module is determined. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0152] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital signal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module above is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0153] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the above method embodiments are generated. The above computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above-mentioned computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the above-mentioned computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) methods. The above-mentioned computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The above-mentioned available medium can be a magnetic medium (such as a floppy disk, a hard disk, a tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0154] Figure 4 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. The electronic device may be a terminal device or a server for implementing the method of the aforementioned embodiment, or may be a terminal device or a server for implementing the method of the aforementioned embodiment connected to the aforementioned terminal device or server. Figure 4As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver 303. Various instructions may be stored in the memory 302 to complete various processing functions and implement the processing steps described in the aforementioned embodiment method. Preferably, the electronic device involved in the embodiment of the present invention also includes: a power supply 304, a system bus 305 and a communication port 306. The system bus 305 is used to realize the communication connection between components. The above-mentioned communication port 306 is used for connecting and communicating between the electronic device and other peripherals.

[0155] exist Figure 4 The system bus 305 mentioned in the figure can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include random access memory (RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk storage.

[0156] The above-mentioned processor can be a general-purpose processor, including a central processing unit CPU, a network processor (Network Processor, NP), a graphics processing unit (Graphics Processing Unit, GPU), etc.; it can also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0157] It should be noted that an embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the method and processing process provided in the above embodiments.

[0158] The embodiment of the present invention provides a method, device, electronic device and computer-readable storage medium for processing inspection images of power transmission lines. As can be seen from the above content, the embodiment of the present invention installs an inspection device on the inspection aircraft, and connects it to a remote server and the flight control system, navigation and positioning system and high-definition camera on the aircraft; and before the flight inspection, the inspection device performs data preprocessing based on the tower data set + high-precision map of the power transmission line sent by the remote server to obtain the corresponding tower acquisition parameter sequence and inspection start / end coordinates; and sends the inspection start / end coordinates + high-precision map to the flight control system for flight trajectory planning and flight control; and during the flight inspection, all towers of the transmission line are inspected and processed according to the navigation and positioning system, high-definition camera and tower acquisition parameter sequence to generate a corresponding image set sequence and save it; and at the end of the flight inspection, the saved image set sequence is sent to the remote server. The embodiment of the present invention provides a new aircraft inspection mechanism, which can perform inspections without human-computer interaction or hovering operations. Compared with conventional manual inspection methods, the inspection cycle of the embodiment of the present invention is significantly shortened, the efficiency is significantly improved, and the real-time performance is significantly enhanced. For some power towers with special installation locations (such as mountains, deserts, etc.), the inspection intensity can also meet the high-frequency inspection requirements. Compared with conventional aircraft-based inspection methods, the inspection cycle of the embodiment of the present invention will be shorter, the efficiency will be higher, and the real-time performance will be stronger. In an inspection environment with poor network quality, the tower image acquisition can be completed based on the offline operation of the inspection device, so that the inspection success rate is higher. Under the same endurance, the inspection route of the inspection mechanism of the embodiment of the present invention will also be longer.

[0159] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0160] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0161] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for processing transmission line inspection images, characterized in that: The method comprises: The first inspection device is installed on the corresponding first inspection aircraft; the first inspection aircraft includes at least a drone and a helicopter; the first inspection device is connected to a remote first server via wireless communication, and is connected to a flight control system, a navigation and positioning system, and a high-definition camera of the first inspection aircraft via data bus communication; the high-definition camera is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function; The first inspection device receives the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map; and confirms the tower acquisition parameters according to the first data set to obtain the corresponding first acquisition parameter sequence; and identifies the starting and ending coordinates of this inspection according to the first high-precision map to obtain the corresponding first starting coordinates and the first ending coordinates; and sends the inspection route configuration instruction carrying the first starting coordinates, the first ending coordinates and the first high-precision map to the flight control system; and after receiving the inspection take-off instruction sent by the flight control system, performs inspection image acquisition processing on all towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain the corresponding first image set sequence and save it; and when receiving the inspection landing instruction sent by the flight control system, sends the saved first image set sequence to the first server.

2. The method for processing transmission line inspection images according to claim 1, characterized in that: The first transmission line includes a plurality of first power towers T i , 1≤tower index i≤N, N is the total number of towers of the first transmission line; The first data set includes a plurality of first electric tower data D i ; The first tower data D i With the first electric tower T i One to one correspondence; The first electric tower data D i It includes a first tower point cloud and a first tower coordinate; the first tower point cloud is the corresponding first tower X i The laser radar scanning point cloud of the first tower, the coordinates of each point in the point cloud of the first tower are the coordinates of the world coordinate system; the coordinates of the first tower are the corresponding X i The center point coordinates of The coordinates of the starting and ending points of the center line of the first transmission line correspond to the coordinates of the center points of a power output mechanism and a power receiving mechanism respectively. i The coordinates of the first electric tower correspond to the coordinates of an intermediate point on the center line of the line; The first high-precision map is a strip-shaped high-precision map; the map centerline of the first high-precision map coincides with the centerline of the line; the strip map width W of the first high-precision map is twice the standard transmission line safety distance, and the standard transmission line safety distance is the vertical safety distance from the centerline of the transmission line required by the power specification; The first acquisition parameter sequence is composed of a plurality of first acquisition parameters A i The first acquisition parameter A i With the first electric tower T i One-to-one correspondence; the first acquisition parameter A i It is composed of the corresponding coordinates of the first power tower and a first area point cloud sequence; the first area point cloud sequence is formed by sequentially sorting a plurality of first area point clouds; The first image set sequence consists of a plurality of first image sets P i The first image set P is sorted in sequence; i With the first electric tower T i One-to-one correspondence; the first image set P i It is composed of multiple first image information sequences; the first image information sequences correspond one-to-one to the first area point cloud; the first image information sequence is formed by sequentially sorting one or more first image information; the first image information includes a first acquired image and a first camera posture.

3. The method for processing transmission line inspection images according to claim 2, characterized in that: The step of confirming the tower acquisition parameters according to the first data set to obtain a corresponding first acquisition parameter sequence specifically includes: The first inspection device detects each of the first tower data D of the first data set. i The first tower point cloud is subjected to denoising and downsampling processing to obtain a corresponding first point cloud; and each of the first point clouds is subjected to ground point cloud filtering processing to obtain a corresponding second point cloud; and each of the second point clouds is subjected to tower edge point cloud extraction processing to obtain a corresponding third point cloud; and using a preset point cloud target recognition model to perform corresponding tower type classification and recognition processing according to each of the second point clouds to obtain a corresponding first prediction vector; and taking the tower type corresponding to the first prediction probability with the largest probability in the first prediction vector as the corresponding first type; the first prediction vector includes a plurality of the first prediction probabilities, each of which corresponds to one tower type; the tower types include at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well-shaped type, and a door type; And performing point cloud confirmation processing of the tower area according to each of the first types and the corresponding third point cloud to obtain the corresponding first area point cloud sequence; And each of the first tower data D i The corresponding first tower coordinates and the first regional point cloud sequence constitute the corresponding first acquisition parameter A i ; and all the first acquisition parameters A obtained i Form the corresponding first acquisition parameter sequence.

4. The method for processing transmission line inspection images according to claim 3, characterized in that: The point cloud target recognition model at least includes a PointNet model, a PointNet++ model, a VoxelNet model, a PointPill ars model, and a PV-RCNN model; The point cloud target recognition model is used to perform corresponding tower type classification and recognition processing according to the input tower point cloud and output a corresponding tower type prediction vector; the tower type prediction vector includes a plurality of tower type prediction probabilities, each of which corresponds to one tower type; Before using the point cloud target recognition model, the point cloud target recognition model needs to be trained in a supervised training manner based on a preset model data set; the model data set includes a plurality of first data records; the first data record includes a first training point cloud and a first label electric tower type; the first training point cloud is a pre-collected electric tower point cloud; The first labeled tower type is the tower type corresponding to the first training point cloud.

5. The method for processing transmission line inspection images according to claim 4, characterized in that: The step of training the point cloud target recognition model in a supervised training manner based on a preset model data set specifically includes: Step 51, the first inspection device randomly divides the model data set into two sub-data sets as a corresponding first training set and a first evaluation set based on a preset first training evaluation ratio; Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first training evaluation ratio; Step 52, extracting the first first data record of the first training set as the corresponding current training record; Step 53, and based on the vector length of the tower type prediction vector of the point cloud target recognition model, set an all-zero vector of equal length as the corresponding first training label vector; and reset the vector data corresponding to the first label tower type of the current training record in the first training label vector to 1; Step 54, inputting the first training point cloud of the current training record into the point cloud target recognition model to perform corresponding tower type classification and recognition processing to obtain a corresponding first training prediction vector; Step 55, and bring the first training prediction vector obtained this time and the corresponding first training label vector into a preset first model loss function to calculate and obtain a corresponding first loss value; Wherein, the first model loss function includes L1 loss function, L2 loss function and cross entropy loss function; Step 56, and identify whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identify whether the current training record is the last first data record of the first training set, if so, go to step 57, if not, extract the next first data record of the first training set as the new current training record and return to step 53; if the first loss value does not satisfy the first loss value range, then based on a preset first model parameter optimizer, perform a round of model parameter modulation on the point cloud target recognition model in the direction of minimizing the first model loss function, and return to step 54 at the end of this round of parameter modulation; Wherein, the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer; Step 57, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and set a full zero vector of equal length as the corresponding first evaluation label vector based on the vector length of the tower type prediction vector of the point cloud target recognition model; and reset the vector data corresponding to the first label tower type of the current evaluation record in the first evaluation label vector to 1; and input the first training point cloud of the current evaluation record into the point cloud target recognition model for corresponding tower type classification and recognition processing to obtain the corresponding first evaluation prediction vector; and form a corresponding first prediction-label pair with the first evaluation prediction vector obtained this time and the corresponding first evaluation label vector; and at the end of this round of traversal, bring all the first prediction-label pairs obtained into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value; Wherein, the first model evaluation function at least includes an RMSE function; Step 58, and identify whether the first evaluation value meets the preset first evaluation value range; if the first evaluation value does not meet the first evaluation value range, return to step 51 to continue training; if the first evaluation value meets the first evaluation value range, stop training and confirm that the model training of the point cloud target recognition model is completed.

6. The method for processing power transmission line inspection images according to claim 3, characterized in that: The step of performing tower area point cloud confirmation processing according to each of the first types and the corresponding third point clouds to obtain the corresponding first area point cloud sequence specifically includes: The first inspection device uses each of the first types as the corresponding current type; and uses the third point cloud corresponding to the current type as the corresponding current point cloud; And based on the current type, a preset first correspondence table is queried, and the first area segmentation strategy field of the first correspondence record in which the first tower type field in the first correspondence table matches the current type is extracted as the corresponding current strategy; wherein the first correspondence table is a correspondence table reflecting the correspondence between the tower type and the tower image acquisition area segmentation strategy; the first correspondence table includes multiple first correspondence records; the first correspondence record includes the first tower type field and the first area segmentation strategy field; the first tower type field includes at least a wine glass type, a cat head type, an upper character type, a dry character type, a T type, an F type, a Japanese character type, a well shape and a door type; the first area segmentation strategy field includes a vertical and horizontal cutting strategy, a tower top cutting strategy, a tower body cutting strategy and a tower base cutting strategy; the vertical and horizontal cutting strategy includes vertical cutting, horizontal cutting and vertical and horizontal cutting; the tower top, tower body and tower base cutting strategies are all composed of a height percentage range and an area segmentation height; the height percentage range is composed of a pair of minimum and maximum height percentages; And take the XOY plane passing through the origin O in the world coordinate system as the first plane; and project each point in the current point cloud onto the first plane to obtain the corresponding first projection point; and identify the projection edge points of all the first projection points to obtain the corresponding multiple first projection edge points; and perform projection edge fitting on all the first projection edge points to obtain a closed first projection edge curve; and perform curve point sampling on the first projection edge curve to obtain the corresponding multiple first sampling points; and record the projection point of the first tower coordinate corresponding to the current point cloud on the first plane as the corresponding first center point; and use the connecting straight line between the first center point and each of the first sampling points as the corresponding first straight line; and use the line segment length where each of the first straight lines intersects with the first projection edge curve as the corresponding first line segment length, and use the first straight line corresponding to the longest first line segment length as the corresponding first transverse tangent; and on the first plane, the perpendicular line through the first center point to the first transverse tangent is recorded as the corresponding first longitudinal tangent; And the vertical and horizontal cutting strategy, the tower top cutting strategy, the tower body cutting strategy and the tower base cutting strategy of the current strategy are used as the corresponding current vertical and horizontal cutting strategy, the current tower top cutting strategy, the current tower body cutting strategy and the current tower base cutting strategy; and identifying the current longitudinal and transverse cutting strategy; if the current longitudinal and transverse cutting strategy is longitudinal cutting, then making an orthogonal plane of the first plane along the first longitudinal cutting line in the world coordinate system is recorded as a corresponding first section; if the current longitudinal and transverse cutting strategy is transverse cutting, then making an orthogonal plane of the first plane along the first transverse cutting line in the world coordinate system is recorded as a corresponding first section; if the current longitudinal and transverse cutting strategy is longitudinal and transverse cutting, then making an orthogonal plane of the first plane along the first longitudinal cutting line and the first transverse cutting line in the world coordinate system respectively to obtain two corresponding first sections; The minimum and maximum coordinates in the Z-axis direction of the current point cloud are extracted to obtain the corresponding minimum height coordinate z min and the maximum height coordinate z max ; and based on the minimum height coordinate z min and the maximum height coordinate z max Calculate the corresponding current tower height = |z max -z min |; and on the current section, based on the height percentage range of the current tower height and the current tower top segmentation strategy, the current tower body segmentation strategy and the current tower base segmentation strategy, define the corresponding tower top height range, tower body height range and tower base height range; and extract the point clouds whose height coordinates in the current point cloud satisfy the tower top, tower body and tower base height ranges respectively as the corresponding tower point cloud, tower body point cloud and tower base point cloud; wherein the tower top, tower body and tower base height ranges are each composed of a pair of minimum and maximum heights; And use each of the first sections obtained as the corresponding current section; use the tower top, tower body and tower base point clouds as the corresponding current local point clouds in sequence; use the projection area of ​​the current local point cloud on the current section as the corresponding current projection area; use the area segmentation height corresponding to the current projection area as the corresponding current segmentation height; use the absolute difference between the maximum and minimum heights of the tower top, tower body or tower base height range corresponding to the current projection area as the corresponding current projection area height; use the absolute difference between the maximum and minimum coordinates of the current projection area on the horizontal coordinate axis of the current section as the corresponding current projection area width; and based on the current projection area height and the current projection area width degree, a rectangle parallel to the horizontal coordinate axis of the current section and completely covering the current projection area is made on the current section plane and recorded as the corresponding current projection area rectangle; and the current projection area rectangle is longitudinally segmented based on the current segmentation height and the preset adjacent segmentation area overlap to obtain a plurality of corresponding first segmentation rectangles; and the sub-point clouds corresponding to each of the first segmentation rectangles in the current local point cloud are extracted as the corresponding first area point clouds; wherein the height of each first segmentation rectangle matches the current segmentation height, the width matches the current projection area width, and the overlap of every two adjacent first segmentation rectangles is greater than or equal to the adjacent segmentation area overlap, which is less than 1; All the obtained first region point clouds are sequentially sorted to form a corresponding first region point cloud sequence.

7. The method for processing power transmission line inspection images according to claim 2, characterized in that: The identifying the starting and ending coordinates of the inspection according to the first high-precision map to obtain the corresponding first starting coordinates and first ending coordinates specifically includes: The first inspection device extracts the starting and ending point coordinates of the map center line of the first high-precision map as the corresponding first starting coordinates and the first ending coordinates.

8. The method for processing power transmission line inspection images according to claim 2, characterized in that: The performing inspection image acquisition processing on all the towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain a corresponding first image set sequence and save it specifically includes: The first inspection device sets a current positioning coordinate with a three-dimensional coordinate initialized to all zeros; sets a current tower spacing initialized to 0; sets a current tower index initialized to 0; initializes an empty sequence as the corresponding current center coordinate sequence; sets a current shooting state initialized to stop shooting; sets a current processing state initialized to processing; the current processing state includes processing and stopping processing; the current shooting state includes starting shooting and stopping shooting; And when the current processing state is processing, regularly send real-time positioning query instructions to the navigation and positioning system according to a preset first time frequency; and receive the real-time positioning coordinates sent back by the navigation and positioning system; and calculate the real-time coordinates of the axis of the high-definition camera based on the real-time positioning coordinates received this time and the relative coordinate offset between the axis of the high-definition camera and the center of mass of the first inspection aircraft, and update the current positioning coordinates based on the real-time coordinates of the axis obtained this time; wherein, the real-time positioning coordinates are the real-time coordinates of the center of mass of the first inspection aircraft; Each time the current positioning coordinate is updated, the current positioning coordinate and each of the first acquisition parameters A of the first acquisition parameter sequence are compared. i The straight-line distance of the first electric tower coordinates is calculated to obtain the corresponding first distance; and the first distance with the smallest distance is used as the corresponding current shortest distance; and the first acquisition parameter A corresponding to the current shortest distance is used as the first acquisition parameter A corresponding to the first electric tower coordinates. i The index i is used as the corresponding current index; and the current tower spacing is updated based on the current shortest distance obtained this time; and whether the current index matches the current tower index is identified, and if not, the current tower index is updated based on the current index; When the current tower index is updated once, the first acquisition parameter A corresponding to the current tower index is updated. i The first regional point cloud sequence is used as the corresponding current regional point cloud sequence; and the point cloud center coordinates of each of the first regional point clouds in the current regional point cloud sequence are identified to obtain the corresponding first center coordinates; and the straight-line distance between the current positioning coordinates and each of the first center coordinates is calculated to obtain the corresponding second distance; and all the first center coordinates are sequentially sorted in the order of the second distance from small to large to obtain the corresponding first center coordinate sequence; and the current center coordinate sequence is updated once based on the first center coordinate sequence obtained this time; And when the current distance between the electric towers is updated each time, it is identified whether the current distance between the electric towers is less than or equal to the preset first distance threshold; if so, the current shooting state is updated to start shooting; if not, the current shooting state is updated to stop shooting; And each time the current shooting state is updated from stop shooting to start shooting, a round of traversal is performed on all the first center coordinates of the current center coordinate sequence; and in this round of traversal, the first center coordinate currently traversed is used as the corresponding current center coordinate; and the axial direction of the high-definition camera is maintained in the direction toward the current center coordinate; and the first area point cloud corresponding to the current center coordinate is used as the corresponding current area point cloud; and the point coordinates in the current area point cloud whose Z-axis coordinate is consistent with the Z-axis component coordinate of the current center coordinate and whose Y-axis coordinate is smaller than the Y-axis component coordinate of the current center coordinate are marked as corresponding first reference coordinates to form a corresponding first reference coordinate sequence; and each of the first reference coordinates in the first reference coordinate sequence is used as the corresponding current reference coordinate in turn; and the latest current positioning coordinate is compared with the current reference coordinate. The straight-line distance is calculated to obtain the corresponding current distance; and the high-definition camera is called to use the current center coordinate as the visual axis center coordinate and automatically focus according to the current distance, and at the end of this automatic focusing, the image is taken based on the principle of rapid reflection mirror shift compensation with the current reference coordinate as the rapid reflection mirror tracking coordinate to obtain the corresponding first acquisition image; and the real-time posture of the high-definition camera at the current moment is used as the corresponding first camera posture; and the first acquisition image and the corresponding first camera posture form a corresponding first image information; and all the first image information corresponding to the current first reference coordinate sequence are sequentially sorted to form a corresponding first image information sequence; and at the end of this round of traversal, all the first image information sequences obtained in this round of traversal form a first image set P whose index i matches the current tower index. i and save; And in the Nth first image set P i When the saving is successful, the current processing state is updated to stop processing; Each time the current processing state is updated from being processed to stopped processing, all the first image sets P currently saved are i The first image set sequence is sorted in sequence to form a corresponding first image set sequence and saved; and when the first image set sequence is saved successfully, all the first image sets P currently saved are i delete.

9. The method for processing power transmission line inspection images according to claim 8, characterized in that: The method further comprises: The flight control system is connected to the navigation and positioning system via a data bus communication method; The flight control system is used to extract the corresponding first starting coordinates, the first ending coordinates and the first high-precision map from the inspection route configuration instruction when receiving the inspection route configuration instruction sent by the first inspection device; and use the first starting coordinates as the corresponding current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the corresponding current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection take-off instruction to the first inspection device at the time of successful take-off; and during the flight, use the real-time positioning coordinates regularly pushed by the navigation and positioning system as the latest current starting coordinates; and plan the aircraft trajectory according to the current starting coordinates, the first ending coordinates and the first high-precision map to obtain the new current flight trajectory; and perform flight control based on the current flight trajectory; and send the inspection landing instruction to the first inspection device at the time of successful landing; The navigation and positioning system is a combined navigation system, which is composed of a gyroscope, an accelerometer and a GNSS system; the navigation and positioning system is used to measure the real-time coordinates of the center of mass of the first inspection aircraft in real time and refresh the real-time positioning coordinates based on the latest measurement results; and regularly push the latest real-time positioning coordinates to the flight control system according to a preset second time frequency; and upon receiving the real-time positioning query instruction sent by the first inspection device, send the latest real-time positioning coordinates back to the first inspection device; The first server is used to send the line tower data set and the line high-precision map of the first transmission line specified by the operator to the first inspection device; and is also used to identify the working conditions of each tower of the first transmission line according to the first image set sequence received this time after receiving the first image set sequence sent by the first inspection device.

10. A device for executing the method for processing a power transmission line inspection image according to any one of claims 1 to 8, characterized in that: The device comprises: a remote data transceiver module, a first data preprocessing module, a second data preprocessing module and an inspection image acquisition module; The device is installed on the corresponding first inspection aircraft; the first inspection aircraft includes at least a drone and a helicopter; the device is connected to a remote first server via wireless communication, and is connected to the flight control system, navigation and positioning system and high-definition camera of the first inspection aircraft via data bus communication; the high-definition camera is a visible light camera with automatic focusing function and fast reflection mirror shift compensation function; The remote data transceiver module is used to receive the line tower data set and the line high-precision map of the first transmission line sent by the first server as the corresponding first data set and the first high-precision map; The first data preprocessing module is used to confirm the tower acquisition parameters according to the first data set to obtain a corresponding first acquisition parameter sequence; The second data preprocessing module is used to identify the starting and ending coordinates of the inspection according to the first high-precision map to obtain the corresponding first starting coordinates and first ending coordinates; and send the inspection route configuration instruction carrying the first starting coordinates, the first ending coordinates and the first high-precision map to the flight control system; The inspection image acquisition module is used to, after receiving the inspection takeoff instruction sent by the flight control system, perform inspection image acquisition processing on all the towers of the first transmission line according to the navigation and positioning system, the high-definition camera and the first acquisition parameter sequence to obtain and save the corresponding first image set sequence; The remote data transceiver module is further configured to send the stored first image set sequence to the first server upon receiving a patrol landing instruction sent by the flight control system.

11. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 8; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 8.

Images