Unmanned aerial vehicle autonomous inspection method for electric power tower inspection

Abstract

The invention relates to an unmanned aerial vehicle autonomous inspection method for electric power tower inspection. A first unmanned aerial vehicle, a second unmanned aerial vehicle and an inspection control platform are included. According to the method, the line analysis model is established, the line is inspected through the first unmanned aerial vehicle to judge whether the tower is deformed or inclined or not and obtain the temperature value of the power transmission cable, and the basic parameters and the temperature value of the power transmission cable of each inspection section are input into the line analysis model; the line analysis model automatically outputs simulation state data of the power transmission cable, and adaptively corrects the second flight path according to the simulation state data, so that the second unmanned aerial vehicle can adapt to the sag change of the power transmission cable, each inspection section of the power transmission cable is ensured to be always in a relatively central area of an image picture which can be shot by the second unmanned aerial vehicle, and target loss is avoided; and details of each point location can be better shot.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power inspection, and in particular to an autonomous inspection method of an unmanned aerial vehicle for electric power tower inspection. Background Art

[0002] Power poles and transmission cables are vital components of the power transmission system. They carry the transmission lines and ensure that electricity can be transmitted stably and efficiently. However, since power poles and transmission cables are usually built in remote, high-altitude, and harsh environments, they are exposed to the natural environment for a long time and are easily affected by various external factors, resulting in damage and failure. In order to ensure the safe and stable operation of the power grid, it is crucial to conduct regular inspections of power poles and transmission cables.

[0003] Traditional inspections of power poles and transmission cables usually combine manual inspections with aerial inspections. Manual inspections require a lot of manpower and material resources, and there are certain safety hazards. Although aerial inspections can improve efficiency, they do not achieve automatic inspections and still require drone pilots to operate.

[0004] The Chinese patent application number 202110525293.5 discloses a method and system for planning the route of a power tower drone inspection, including obtaining point cloud data of key inspection components of the tower according to the tower type and converting it into an image as a target sample class; converting the entire power tower into image data according to the geographical location of the tower; retrieving the image data with the target sample class as the target to obtain the location information of the key inspection components; and planning the inspection trajectory of the drone using the location information of the key inspection parts. This patent realizes the automatic inspection of the line by drones and can accurately identify the key components on the line.

[0005] The Chinese patent application number 202010490245.2 discloses a method for detecting broken strands of ground wires on distribution line towers using drone images, including the following steps: A) Establishing a dataset of broken strands of ground wires, marking and locating large targets in the image; B) Transmitting the image data taken by the drone to the PC via a wired connection, entering the typical defect intelligent analysis module, identifying the area to be analyzed, establishing an analysis frame, manually marking defects, and obtaining sample data; C) Constructing a Cascade R-CNN network, using sample data to train and test the Cascade R-CNN network; D) Obtaining images taken by drone patrols, identifying the area to be analyzed, importing the Cascade R-CNN network, and obtaining detection results. The substantial effects of this invention are: improving the performance of small target detection; and improving the accuracy of fault detection of broken strands of ground wires on distribution line towers.

[0006] Although the above disclosed technology can realize the automatic inspection of pole towers and transmission cables by drones, it does not take into account the impact of temperature on transmission cables. Since the distance between two adjacent sets of pole towers is relatively far, and the cable material is ductile, the thermal expansion and contraction phenomenon has a significant impact on the transmission cables.

[0007] like Figure 2 As shown, in the existing UAV autonomous inspection solution, the flight path of the UAV is pre-designed, which can ensure that the key parts of the transmission cables and towers are in a relatively centered position in the picture during the UAV flight shooting process. The dotted box in the figure is a schematic diagram of the UAV's shooting screen range.

[0008] Since the effect of temperature on the transmission cable was not considered when designing the flight route, when the temperature of the transmission cable changes a lot (mainly due to factors such as air temperature, sunlight and power transmission heating that cause the cable to heat up and cool down), the length and sag of the transmission cable will change significantly (such as Figure 2 When the temperature of the power transmission cable changes greatly, and the drone still inspects according to the default flight path, the power transmission cable may not be found in the captured image, resulting in target loss and affecting the normal inspection. Summary of the invention

[0009] Based on the above description, the present invention provides an autonomous inspection method for power pole towers using drones to solve the problem of power transmission cables disappearing from the shooting screen during the inspection process, thereby ensuring the normal progress of the inspection.

[0010] The technical solution of the present invention to solve the above technical problems is as follows: An autonomous inspection method for power pole towers using a drone, marking a positioning point A on the top of each group of pole towers, and marking a positioning point B on the cross arm of each group of pole towers, measuring the longitude and latitude and altitude of point A and point B, selecting a point at a certain distance directly above the positioning point A of each group of pole towers as an upper path point, and connecting each upper path point to obtain a first flight path; Define the power line between two adjacent groups of towers as an inspection section, and measure the benchmark parameters of the transmission cables of each inspection section under the selected standard temperature, including the benchmark suspension point height, benchmark span, and benchmark sag; evenly select a number of benchmark target points on the transmission cables of each inspection section, calculate and calibrate the position coordinates of each benchmark and the benchmark tension it bears, save the position coordinates and benchmark tension data of each benchmark target point, select a point at a certain shooting distance L directly outside the suspension point and each target point as a side path point, and connect each side path point to obtain the default second flight path; During the line inspection operation, the first UAV flies along the first flight path in front, and the second UAV flies along the default second flight path in the back; the first UAV photographs the transmission cables and towers from above to obtain natural light images and infrared thermal images, and uploads the image data and real-time flight data obtained from the shooting to the inspection control platform; The inspection control platform includes a first image processing unit, a second image processing unit, a line analysis model, a path planning unit, an instruction sending unit and an abnormal alarm unit; The first image processing unit identifies the positioning point A and the positioning point B of each pole tower in the natural light image uploaded by the first drone, and determines the longitude and latitude of point A and point B according to the longitude and latitude of the first drone during shooting, the resolution of the image, and the relative positions of point A and point B in the image; compares the obtained longitude and latitude of point A and point B with the respective measured longitude and latitude, and calculates the longitude and latitude offset. When the longitude and latitude offset is less than the preset offset, it is considered that the current pole tower and cross arm are not tilted, otherwise it is considered that the current pole tower and cross arm are tilted; At the same time, the first image processing unit identifies and locates the highlighted area in the infrared thermal image uploaded by the first UAV to obtain the abnormally high temperature area on the transmission cable and the longitude and latitude of its center point; the first image processing unit also obtains the current temperature value of the transmission cable by processing the infrared thermal image; The line analysis model judges whether two adjacent groups of towers and crossarms are tilted in turn. If it is judged that the two adjacent groups of towers and crossarms are not tilted, the suspension point height, reference span, reference sag, reference target point position, reference tension and current temperature value of the transmission cable between the two groups of towers are further obtained; the line analysis model is pre-set with line simulation state data of the above-mentioned dimensional parameters at different values, and the line simulation state data includes the simulated sag of the transmission cable, the tension currently borne by each target point and the current position coordinates of each target point; the line analysis model automatically outputs the current line simulation state data according to the acquired parameters, and then the line analysis model fits the simulated position coordinates of each target to obtain a fitting curve, and the line analysis model then brings the longitude and latitude of the center point of the abnormally high temperature area into the fitting curve to obtain the height data of the center point, thereby obtaining the position coordinates of the center point of the abnormally high temperature area; At the same time, the line analysis model calculates the coordinate change of each target point in the vertical plane in the current state, that is, the vertical offset vector, and calculates the tension change value of each target point, and then compares the calculated tension change value with the preset reference value. If the former is greater than the latter, it is considered that the tension change of the target point is too large, and the target point is classified as a high-risk target point, otherwise, the target point is classified as a normal target point; The route analysis model sends the calculated vertical offset vectors of the coordinates of each target point, the classification information of each target point, and the coordinates of the center point of the abnormally high temperature area to the path planning unit; The route planning unit vertically corrects the coordinates of the lateral path points according to the vertical offset vectors of each target point, and then selects a point at a certain shooting distance L just outside the center point of the abnormally high temperature area as a newly added lateral path point; the route planning unit also corrects the shooting distances of normal target points, high-risk target points, abnormally high temperature areas and hanging points, assigning a correction coefficient of 1 to normal target points, a correction coefficient of k1 to high-risk target points, a correction coefficient of k2 to abnormally high temperature areas, and a correction coefficient of k3 to hanging points; after correcting the shooting distance, the shooting distance of normal target points is still L; the shooting distance of high-risk shooting points is L×k1; the shooting distance of abnormally high temperature areas is L×k2; the shooting distance of hanging points is L×k3; The coordinate change of each lateral path point in the horizontal plane, i.e., the horizontal offset vector, is calculated according to the shooting distance change of each lateral path point; the path planning unit then corrects the coordinates of each lateral path point in the horizontal direction according to the horizontal offset vector; finally, a corrected second flight path is generated according to the corrected coordinates of each lateral path point; the instruction sending unit generates a control instruction according to the second flight path, and sends the control instruction to the second UAV; The second UAV flies according to the corrected second flight path. During the flight, the second UAV takes natural light images and infrared thermal images of the pole tower and the transmission cable from the side, and uploads the captured image data to the inspection control platform; The second image processing unit retrieves the image data uploaded by the second UAV, and determines whether there is a broken strand phenomenon by performing image processing on the natural light image, and determines whether an abnormally high temperature situation occurs by identifying and locating the abnormally bright area in the infrared thermal image; when a broken strand or abnormally high temperature occurs, the abnormal alarm unit generates an alarm signal and sends the alarm signal to the operation and maintenance personnel, reminding the operation and maintenance personnel to take timely measures.

[0011] As a preferred solution: the process of identifying the highlight area is that the first image processing unit performs grayscale processing on the infrared thermal image to obtain a grayscale image and retains the original infrared thermal image, determines the area with the largest bright spot by reading the grayscale values ​​of each area and comparing them, then extracts the edge contour of the area with the maximum brightness, overlaps the grayscale image with the original infrared thermal image, and crops the original infrared thermal image according to the extracted edge contour to intercept the image area within the contour, that is, to obtain the target image area; then the first image processing unit randomly scans a number of pixels in the target image area and reads the brightness value of each pixel, then calculates the average brightness value, and finally compares the average brightness with a preset brightness threshold. If the former is greater than the latter, the target image area is considered to be a highlight area, otherwise it is not a highlight area.

[0012] As a preferred solution: the process of locating the center point of the highlight area is to first scan each pixel point on the edge of the target image area, and infer the longitude and latitude of each pixel point in the image according to the longitude and latitude of the drone when the image was taken and the resolution of the image; then filter out the point with the largest longitude, the point with the smallest longitude, the point with the largest latitude and the point with the smallest latitude from these pixel points, a total of four pixel points; then connect two groups of relative pixel points to obtain two connecting lines, define the intersection of the two connecting lines as the center point of the highlight area, and finally calculate the longitude and latitude of the center point according to the longitude and latitude of the four pixel points, thereby realizing the positioning of the highlight area.

[0013] As a preferred solution: when selecting reference target points, the number of reference target points is determined according to the reference span and reference sag of the transmission cable in the inspection section. The larger the reference span and the larger the reference sag, the greater the number of reference target points.

[0014] As a preferred solution: the first UAV and the second UAV include a microcontroller module, and also include a data acquisition module, an obstacle avoidance sensor, a posture sensor, a positioning and navigation module, a wireless communication module, a power motor, a storage module and a power supply module connected to the microcontroller module; and also include a camera module and an infrared thermal imaging module, and the output ends of the camera module and the infrared thermal imaging module are connected to the input end of the data acquisition module.

[0015] Compared with the prior art, the technical solution of the present application has the following beneficial technical effects: the present application establishes a line analysis model, firstly uses the first UAV to inspect the line to determine whether the pole tower is deformed or tilted and obtain the temperature value of the transmission cable, and inputs the basic parameters and temperature values ​​of the transmission cables in each inspection section into the line analysis model. The line analysis model automatically outputs the simulated state data of the transmission cables, and adaptively corrects the second flight path accordingly, so that the second UAV can adapt to the changes in the sag of the transmission cables, ensuring that the transmission cables in each inspection section are always in the relatively central area of ​​the image that can be captured by the second UAV, avoiding loss of the target, and better capturing the details of each point. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a system diagram of the inspection control platform and drone; Figure 2 Schematic diagram of the tower and transmission cables; Figure 3 It is a schematic diagram of the sag of the transmission cable and the change of the position of the target point; Figure 4 Side view and top view of the second flight path.

[0017] In the accompanying drawings, the components represented by the reference numerals are listed as follows: 1. Pole tower; 2. Transmission cable; 3. Cross arm; 4. Insulator; 5. Suspension point; 6. First drone; 7. Second drone. DETAILED DESCRIPTION

[0018] Reference Figure 1 , an autonomous inspection method for power tower inspection using unmanned aerial vehicles, comprising a first unmanned aerial vehicle 6, a second unmanned aerial vehicle 7 and an inspection control platform.

[0019] The drone in this embodiment includes a microcontroller module, and also includes a data acquisition module, an obstacle avoidance sensor, a posture sensor, a positioning and navigation module, a wireless communication module, a power motor, a storage module and a power module connected to the microcontroller module.

[0020] The drone in this embodiment is equipped with a camera module and an infrared thermal imaging module, and the output ends of the camera module and the infrared thermal imaging module are connected to the input end of the data acquisition module.

[0021] The obstacle avoidance sensor is used to sense obstacles in the drone's surrounding environment and feed back the sensed obstacle information to the microcontroller module. During the flight, the microcontroller module controls the drone to avoid obstacles; the attitude sensor is used to detect the drone's flight attitude and feed back the attitude information to the microcontroller module. During the flight, the microcontroller module controls the drone's flight attitude; the positioning and navigation module is used to receive the flight path information sent by the inspection control platform and generate navigation data, and transmit the navigation data to the microcontroller module. The microcontroller module controls the drone's power motor according to the navigation data, so that the drone flies according to the flight path; the wireless communication module is used for communication and data transmission between the drone and the cloud platform.

[0022] The inspection control platform includes a data receiving unit, a data storage unit, a first image processing unit, a second image processing unit, a line analysis model, a path planning unit, an instruction sending unit and an abnormal alarm unit.

[0023] Reference Figure 2 The working principle of the inspection system is: Mark the positioning point A at the top of each group of towers 1, and mark the positioning point B on the cross arm 3 of each group of towers 1, measure the longitude, latitude and altitude of points A and B, select a point at a certain distance directly above the positioning point A of each group of towers 1 as the upper path point, and connect the upper path points to obtain the first flight path.

[0024] The power line between two adjacent groups of towers 1 is defined as an inspection section. Under the selected standard temperature (e.g., 25 degrees), the reference parameters of the transmission cable 2 of each inspection section are measured and calculated, including the reference suspension point height, reference span, and reference sag. The reference parameters of the transmission cable are stored in the data storage unit. According to the reference span and reference sag of the transmission cable, several reference target points (e.g., Figure 2 The hollow dots on the cable in the middle are shown), and the tension (reference tension) borne by the position coordinates of each reference target point is calculated and calibrated, and the coordinate data and reference tension data of each reference target point are stored in the data storage unit. A point is selected as a side path point at a certain shooting distance L directly outside the suspension point 5 and each target point, and the default second flight path can be obtained by connecting each side path point.

[0025] The path data of the first flight path and the second flight path are stored in the path planning unit.

[0026] In this embodiment, when selecting reference target points, the number of reference target points is determined according to the reference span and reference sag of the transmission cable in the inspection section. The larger the reference span and the larger the reference sag, the more reference target points there are.

[0027] During the line inspection process, the first UAV flies along the first flight path in the front, and the second UAV flies along the default second flight path in the back; the first UAV photographs the transmission cables and towers from above to obtain natural light images and infrared thermal images, and uploads the captured image data and real-time flight data to the inspection control platform, and the uploaded data is stored in the data storage unit.

[0028] The first image processing unit retrieves the natural light image data captured by the first image processing unit from the data storage unit, identifies the positioning point A and the positioning point B of each pole tower in the image, and determines the longitude and latitude of point A and point B according to the longitude and latitude of the first drone when the image is captured, the resolution of the image, and the relative positions of point A and point B in the image; compares the obtained longitude and latitude of point A and point B with the respectively measured longitude and latitude, and calculates the longitude and latitude offset (i.e., the modulus of the offset vector); when the longitude and latitude offset is less than the preset offset, it indicates that the current pole tower and cross arm are not significantly tilted; otherwise, it indicates that the current pole tower and cross arm are significantly tilted.

[0029] At the same time, the first image processing unit retrieves the infrared thermal image data taken by the first drone from the data storage unit, and the first image processing unit identifies and locates the highlight area in the infrared thermal image, so as to obtain the abnormally high temperature area on the transmission cable and the longitude and latitude of its center point. Specifically, in this embodiment, the identification and positioning process of the highlight area is as follows: the first image processing unit performs grayscale processing on the infrared thermal image to obtain a grayscale image and retains the original infrared thermal image, and the areas with different brightness can be clearly distinguished from the grayscale image. By reading the grayscale values ​​of each area and comparing them, the area with the largest bright spot can be determined, and then the edge contour of the area with the largest brightness is extracted, the grayscale image is overlapped with the original infrared thermal image, and the original infrared thermal image is cropped according to the extracted edge contour to intercept the image area within the contour, that is, to obtain the target image area. Subsequently, the first image processing unit randomly scans a number of pixels in the target image area and reads the brightness value of each pixel, and then calculates the average brightness value, and finally compares the average brightness with the preset brightness threshold. If the former is greater than the latter, the target image area is considered to be a highlight area, otherwise it is not a highlight area.

[0030] Then the center point of the highlighted area is located. The process is as follows: scan each pixel point on the edge of the target image area, and calculate the longitude and latitude of each pixel point in the image according to the longitude and latitude of the first drone when the image was taken and the resolution of the image; then filter out the point with the largest longitude, the point with the smallest longitude, the point with the largest latitude, and the point with the smallest latitude from these pixels, a total of four pixels; then connect two groups of relative pixel points to obtain two connecting lines, and define the intersection of the two connecting lines as the center point of the highlighted area; finally, calculate the longitude and latitude of the center point based on the longitude and latitude of the four pixel points, thereby realizing the positioning of the highlighted area (i.e., the abnormally high temperature area).

[0031] Figure 4 (a) represents the side view of the transmission cable in the inspection section. The hollow dots in the figure represent the target points, and the solid dots (i.e., point R) in the figure represent the center point of the abnormally high temperature area. Since the first UAV takes the infrared thermal image from above the line, the coordinates of the center point of the abnormally high temperature area obtained by processing the infrared thermal image uploaded by the first UAV only have longitude and latitude data, but no altitude data. The probability of failure in the abnormally high temperature area is relatively high, and it needs to be photographed and analyzed in detail. Therefore, it is necessary to add side path points to the abnormally high temperature area to ensure that the second UAV can take a clear image of the abnormally high temperature area.

[0032] Figure 4 (b) is a top view of each target point in the inspection section. The solid line with an arrow in the figure represents the flight path of the second UAV, and L in the figure is the default shooting distance (i.e., the horizontal distance between the flight path and the transmission cable).

[0033] On the other hand, the first image processing unit also reads the image brightness of the transmission cables in each group of infrared thermal images taken in each inspection section and retrieves the corresponding temperature value based on the image brightness (there is a corresponding relationship between the image brightness and temperature in the infrared thermal image), and then calculates the average temperature value, which is defined as the current temperature value of the transmission cables in the inspection section.

[0034] The first image processing unit feeds back the image processing result to the line analysis model.

[0035] The line analysis model judges whether two adjacent groups of towers and crossarms are tilted in turn. If it is judged that neither of the two adjacent groups of towers nor the crossarms are tilted, the suspension point height, reference span, reference sag, reference target point position, reference tension and current temperature value of the transmission cable between the two groups of towers are further obtained.

[0036] The line analysis model is pre-set with line simulation state data of the above-mentioned dimensional parameters (i.e., suspension point height, reference span, reference sag, reference target point position, reference tension and current temperature value) at different values. The line simulation state data includes the simulated sag of the transmission cable, the tension currently borne by each target point and the current position coordinates of each target point. The line analysis model automatically outputs the current line simulation state data according to the acquired parameters, and then the line analysis model fits the simulated position coordinates of each current target point to obtain a fitting curve. The line analysis model then brings the longitude and latitude of the center point of the abnormally high temperature area into the fitting curve to obtain the height data of the center point, thereby obtaining the position coordinates of the center point of the abnormally high temperature area.

[0037] The vertical offset vector of the target point here is the vertical offset vector that should be corrected for each corresponding path point on the second flight path.

[0038] Reference Figure 3 , the transmission cable above (shown by the solid line) in the figure is the reference state, and the transmission cable below (shown by the dotted line) is the state after the temperature rises. It can be seen that the cable will stretch and its sag will increase after the temperature rises. When the cable length and sag increase, the coordinates of each target point and the tension received will also change accordingly. In the figure, Ⅰ represents the position of a reference target point on the transmission cable of the inspection section, and Ⅱ is the position of the target point after offset. Vector a in the figure is the vertical offset vector of the target point, that is, the vertical offset vector of each path point of the second drone. This ensures that when the second drone shoots the transmission cable from the side, the transmission cable is always in the center of the picture to avoid losing the target.

[0039] The line analysis model calculates the coordinate change of each target point in the vertical plane under the current state, that is, the vertical offset vector, and calculates the tension change value of each target point at the same time, and then compares the calculated tension change value with the preset reference value. If the former is greater than the latter, it is considered that the tension change on which the target point is subjected is too large, and the target point is classified as a high-risk target point, otherwise the target point is classified as a normal target point.

[0040] The risk of scattered and broken strands at high-risk target points increases, and the target points need to be photographed at close range to obtain clearer image details to facilitate subsequent image recognition.

[0041] For transmission lines, abnormally high temperature parts may be short circuits or arc burns, affecting the safe operation of the cable. Therefore, for abnormally high temperature parts, its center point needs to be added as a shooting point, and it needs to be shot at a close distance to obtain clearer and richer image details for subsequent image recognition.

[0042] In addition, the hanging points are also places where broken and scattered strands are more likely to occur, so the hanging points also need to be photographed at a close distance.

[0043] In this embodiment, the shooting distances of normal target points, high-risk target points, abnormally high temperature areas and hanging points are also corrected.

[0044] The route analysis model sends the calculated vertical offset vectors of the coordinates of each target point, the classification information of each target point, and the coordinates of the center point of the abnormally high temperature area to the path planning unit; The route planning unit vertically corrects the coordinates of the lateral path points according to the vertical offset vectors of each target point, and then selects a point at a certain shooting distance L just outside the center point of the abnormally high temperature area as a newly added lateral path point; the route planning unit also corrects the shooting distances of normal target points, high-risk target points, abnormally high temperature areas and hanging points, assigning a correction coefficient of 1 to normal target points, a correction coefficient of k1 to high-risk target points, a correction coefficient of k2 to abnormally high temperature areas, and a correction coefficient of k3 to hanging points; after correcting the shooting distance, the shooting distance of normal target points is still L; the shooting distance of high-risk shooting points is L×k1; the shooting distance of abnormally high temperature areas is L×k2; the shooting distance of hanging points is L×k3; like Figure 4 As shown in (c), the coordinate change of each side path point in the horizontal plane, that is, the horizontal offset vector b, is calculated according to the change of the shooting distance of each side path point; the path planning unit then corrects the coordinates of each side path point in the horizontal direction according to the horizontal offset vector; finally, a corrected second flight path is generated according to the corrected coordinates of each side path point; the instruction sending unit generates a control instruction according to the second flight path, and sends the control instruction to the second UAV.

[0045] In this way, the second UAV can inspect according to the corrected second flight path, thereby ensuring that the transmission cables in each inspection section are always in the relatively central area of ​​the image that can be captured by the second UAV, avoiding loss of the target and better capturing the details of each point.

[0046] The second UAV flies according to the corrected second flight path. During the flight, the second UAV takes natural light images and infrared thermal images of the pole tower and the transmission cable from the side, and uploads the captured image data to the inspection control platform; The second image processing unit retrieves the image data uploaded by the second UAV, and determines whether there is a broken strand phenomenon by performing image processing on the natural light image, and determines whether an abnormally high temperature situation occurs by identifying and locating the abnormally bright area in the infrared thermal image; when a broken strand or abnormally high temperature occurs, the abnormal alarm unit generates an alarm signal and sends the alarm signal to the operation and maintenance personnel, reminding the operation and maintenance personnel to take timely measures.

[0047] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for autonomous inspection of power towers by drones, characterized in that: Mark the positioning point A on the top of each group of towers, and mark the positioning point B on the cross arm of each group of towers. Measure the longitude, latitude and altitude of point A and point B. Select a point at a certain distance directly above the positioning point A of each group of towers as the upper path point. Connect the upper path points to obtain the first flight path. Define the power line between two adjacent groups of towers as an inspection section, and measure the benchmark parameters of the transmission cables in each inspection section under the selected standard temperature, including the benchmark suspension point height, benchmark span, and benchmark sag; Select several reference target points evenly on the transmission cables of each inspection section, calculate and calibrate the position coordinates of each reference and the reference tension it bears, save the position coordinates and reference tension data of each reference target point, select a point at a certain shooting distance L directly outside the hanging point and each target point as a side path point, and connect each side path point to obtain the default second flight path; During the line inspection operation, the first UAV flies along the first flight path in front, and the second UAV flies along the default second flight path in the back; the first UAV photographs the transmission cables and towers from above to obtain natural light images and infrared thermal images, and uploads the image data and real-time flight data obtained from the shooting to the inspection control platform; The inspection control platform includes a first image processing unit, a second image processing unit, a line analysis model, a path planning unit, an instruction sending unit and an abnormal alarm unit; The first image processing unit identifies the positioning point A and the positioning point B of each tower in the natural light image uploaded by the first drone, and determines the longitude and latitude of point A and point B according to the longitude and latitude of the first drone when shooting, the resolution of the image, and the relative positions of point A and point B in the image; The obtained longitude and latitude of point A and point B are compared with the measured longitude and latitude, and the longitude and latitude offset is calculated. When the longitude and latitude offset is less than the preset offset, it is considered that the current tower and cross arm are not tilted, otherwise it is considered that the current tower and cross arm are tilted; At the same time, the first image processing unit identifies and locates the highlighted area in the infrared thermal image uploaded by the first UAV to obtain the abnormally high temperature area on the transmission cable and the longitude and latitude of its center point; the first image processing unit also obtains the current temperature value of the transmission cable by processing the infrared thermal image; The line analysis model determines whether two adjacent groups of towers and crossarms are tilted in turn. If it is determined that the two adjacent groups of towers and crossarms are not tilted, the suspension point height, reference span, reference sag, reference target point position, reference tension and current temperature value of the transmission cable between the two groups of towers are further obtained; The line analysis model is pre-set with line simulation state data of the above-mentioned dimensional parameters at different values, and the line simulation state data includes the simulated sag of the transmission cable, the tension currently borne by each target point, and the current position coordinates of each target point; The line analysis model automatically outputs the current line simulation state data according to the acquired parameters, and then the line analysis model fits the simulated position coordinates of each target to obtain a fitting curve. The line analysis model then brings the longitude and latitude of the center point of the abnormally high temperature area into the fitting curve to obtain the height data of the center point, thereby obtaining the position coordinates of the center point of the abnormally high temperature area; At the same time, the line analysis model calculates the coordinate change of each target point in the vertical plane in the current state, that is, the vertical offset vector, and calculates the tension change value of each target point, and then compares the calculated tension change value with the preset reference value. If the former is greater than the latter, it is considered that the tension change of the target point is too large, and the target point is classified as a high-risk target point, otherwise, the target point is classified as a normal target point; The route analysis model sends the calculated vertical offset vectors of the coordinates of each target point, the classification information of each target point, and the coordinates of the center point of the abnormally high temperature area to the path planning unit; The route planning unit vertically corrects the coordinates of the lateral path points according to the vertical offset vectors of each target point, and then selects a point at a certain shooting distance L just outside the center point of the abnormally high temperature area as a newly added lateral path point; the route planning unit also corrects the shooting distances of normal target points, high-risk target points, abnormally high temperature areas and hanging points, assigning a correction coefficient of 1 to normal target points, a correction coefficient of k1 to high-risk target points, a correction coefficient of k2 to abnormally high temperature areas, and a correction coefficient of k3 to hanging points; after correcting the shooting distance, the shooting distance of normal target points is still L; the shooting distance of high-risk shooting points is L×k1; the shooting distance of abnormally high temperature areas is L×k2; the shooting distance of hanging points is L×k3; The coordinate change of each lateral path point in the horizontal plane, i.e., the horizontal offset vector, is calculated according to the shooting distance change of each lateral path point; the path planning unit then corrects the coordinates of each lateral path point in the horizontal direction according to the horizontal offset vector; finally, a corrected second flight path is generated according to the corrected coordinates of each lateral path point; the instruction sending unit generates a control instruction according to the second flight path, and sends the control instruction to the second UAV; The second UAV flies according to the corrected second flight path. During the flight, the second UAV takes natural light images and infrared thermal images of the pole tower and the transmission cable from the side, and uploads the captured image data to the inspection control platform; The second image processing unit retrieves the image data uploaded by the second UAV, and determines whether there is a broken strand phenomenon by performing image processing on the natural light image, and determines whether an abnormally high temperature situation occurs by identifying and locating the abnormally bright area in the infrared thermal image; when a broken strand or abnormally high temperature occurs, the abnormal alarm unit generates an alarm signal and sends the alarm signal to the operation and maintenance personnel, reminding the operation and maintenance personnel to take timely measures.

2. The autonomous inspection method for power tower inspection using a drone according to claim 1 is characterized by: The process of identifying the highlight area is that the first image processing unit performs grayscale processing on the infrared thermal image to obtain a grayscale image and retains the original infrared thermal image, determines the area with the largest bright spot by reading the grayscale values ​​of each area and comparing them, then extracts the edge contour of the area with the maximum brightness, overlaps the grayscale image with the original infrared thermal image, and crops the original infrared thermal image according to the extracted edge contour to intercept the image area within the contour, that is, to obtain the target image area; then the first image processing unit randomly scans a number of pixels in the target image area and reads the brightness value of each pixel, and then calculates the average brightness value, and finally compares the average brightness with a preset brightness threshold. If the former is greater than the latter, the target image area is considered to be a highlight area, otherwise it is not a highlight area.

3. The autonomous inspection method for power tower inspection using a drone according to claim 2 is characterized by: The process of locating the center point of the highlight area is to first scan each pixel point on the edge of the target image area, and then calculate the longitude and latitude of each pixel point in the image according to the longitude and latitude of the drone when the image was taken and the resolution of the image; Then, the point with the largest longitude, the point with the smallest longitude, the point with the largest latitude and the point with the smallest latitude are screened out from these pixels, which is a total of four pixels. Two groups of relative pixels are then connected to obtain two connecting lines, and the intersection of the two connecting lines is defined as the center point of the highlight area. Finally, the longitude and latitude of the center point are calculated based on the longitude and latitude of the four pixels, thereby realizing the positioning of the highlight area.

4. The autonomous inspection method for power tower inspection using a drone according to claim 1 is characterized by: When selecting reference target points, the number of reference target points is determined according to the reference span and reference sag of the transmission cable in the inspection section. The larger the reference span and the larger the reference sag, the greater the number of reference target points.

5. The autonomous inspection method for power tower inspection using a drone according to claim 1 is characterized by: The first UAV and the second UAV include a microcontroller module, and also include a data acquisition module, an obstacle avoidance sensor, a posture sensor, a positioning and navigation module, a wireless communication module, a power motor, a storage module and a power supply module connected to the microcontroller module; and also include a camera module and an infrared thermal imaging module, and the output ends of the camera module and the infrared thermal imaging module are connected to the input end of the data acquisition module.

Images