Method for measuring ground wire sag based on unmanned aerial vehicle image recognition

Abstract

A kind of ground wire sag measurement method based on unmanned aerial vehicle image recognition, including setting the GPS mark base point position of adjacent tower A, B;Using unmanned aerial vehicle to carry out image sampling operation flow to ground wire;Image recognition positioning is carried out to the sampling point of ground wire sag;Mathematical model analysis is carried out to sampling point;The sag equation finally obtained is used in business application.The measurement method can carry out real-time, rapid sag measurement, using the high mobility and portability of unmanned aerial vehicle can be closer to the target that is complex and difficult to measure, improve the standardization, efficiency and measurement accuracy of sag measurement operation.The present application solves the technical problem that ground wire sag acceptance is difficult due to unfavorable measurement factors such as non-ideal measurement angle, tower location is remote, large span of sag, etc., so that sag measurement is more simple, reliable and accurate.

Description

Technical Field

[0001] This invention belongs to the fields of power transmission lines and unmanned aerial vehicle (UAV) application technology. Background Technology

[0002] In the field of power transmission lines, conductor sag is a core indicator for the safe operation of transmission lines, directly determining the safe distance between the line and the ground, conductor tension, and weather resistance stability. Excessive sag can easily lead to safety accidents such as conductor-to-ground discharge and wind-induced flashover, while insufficient sag may cause the conductor tension to exceed the standard, resulting in breakage under extreme weather conditions. Therefore, it is necessary to regularly and accurately measure the conductor sag to ensure that the transmission lines meet the relevant industry standards.

[0003] Traditional methods for measuring conductor sag are primarily manual, including theodolite measurement, total station trigonometric leveling, and laser distance measurement. These methods require maintenance personnel to be on-site, deploy measuring instruments on the tower or ground, and manually aim, read, and calculate the sag value. However, they have significant limitations: First, due to the limitations of terrain conditions, it is difficult to set up instruments in complex terrain areas such as canyons, dense forests, and mountainous areas, and the measurement line of sight is easily blocked, making it impossible to complete effective observations. Secondly, the work efficiency is low. Single-span measurement requires multiple people to work together, which is time-consuming and labor-intensive, making it difficult to meet the needs of rapid inspection of large-scale transmission lines. Third, the safety risks are high. When working in dangerous areas such as high altitudes and steep slopes, it is easy to cause safety accidents such as people falling and equipment damage. Fourth, the measurement accuracy is easily affected by human error and environmental interference, making it difficult to guarantee data consistency.

[0004] To address the shortcomings of traditional methods, the power system industry has gradually explored automated measurement technologies. Unmanned aerial vehicle (UAV) technology, due to its flexibility, mobility, and wide coverage, has been introduced into the transmission line field. Early UAV sag measurement relied heavily on specialized payloads such as lidar and infrared thermal imagers, but these were costly, complex to operate, and involved cumbersome data processing, hindering large-scale application. Furthermore, some UAV-based aerial image measurement methods only used simple edge detection algorithms to extract conductor contours, failing to incorporate precise modeling based on the catenary's mechanical properties. In addition, non-standard image acquisition and the lack of reference objects led to significant measurement errors, especially in scenarios with image distortion or conductor / ground wire obstruction, making it difficult to meet engineering accuracy requirements. There is an urgent need to avoid these measurement problems, fully leverage the flexible acquisition advantages of UAVs and the precise analysis capabilities of computer vision algorithms, break through the bottlenecks of traditional measurement technologies, and improve measurement reliability and portability. Summary of the Invention

[0005] The purpose of this invention is to provide a method for measuring conductor sag based on UAV image recognition, which allows for initial calibration by setting GPS marker base points on adjacent towers, standardizing the sampling process, and enabling a single person to complete the measurement task.

[0006] The steps of this invention are: S1. Set the GPS marker base points of adjacent towers A and B as the initial calibration positions in the airborne camera images; S2. Operational procedures for using drones to sample images of ground wires; S3. Use an image recognition algorithm to locate the sampling points of the conductor sag. S4. Perform mathematical model analysis on the sampling points, including two steps: three-dimensional mapping and curve fitting. Accurately represent the positioning information of all sampling points in the form of sag equations in the plane coordinate system formed by the two suspension points and the lowest point. S5. The final obtained sag equation is used in business applications, including calculating characteristic indicators and verifying engineering errors.

[0007] The GPS marker base point location described in this invention is the positioning point of the marker object at the beginning and end of the conductor. The marker object includes, but is not limited to, signs, marker poles, etc., that can be clearly photographed by the drone's onboard camera. The setting of the marker base point location must be combined with real-time differential positioning equipment.

[0008] The image sampling process of the S2 UAV described in this invention includes the following steps: S21. Start sampling. Manually control the drone to take off and fly to the GPS marker base point of Tower A. Manually mark the GPS marker base point of Tower A as waypoint A1 and take a picture of the base point location. S22. Manually control the drone to fly to the GPS marker base point of Tower B, manually mark Tower B as waypoint B1, and take an image of the base point location. S23. Make the UAV fly at a constant speed from waypoint B1 to A1, and at the same time take side-view ground line sag images according to the performance of the airborne camera, and fly to the GPS marker base point position of tower A. S24. Manually control the drone to fly directly above the sag of the conductor wire of tower A, manually mark waypoint A2, and take a picture of the base point position; S25. Manually control the drone to fly directly above the sag of the conductor wire of Tower B, manually mark waypoint B2, and take a picture of the base point position. S26. Make the UAV fly at a constant speed from waypoint B2 to A2, and at the same time take pictures of the overhead ground wire at regular intervals according to the performance of the airborne camera, and fly to the direct top of the sag of the ground wire of tower A. S27. End sampling, manually remotely control the drone to fly to the landing position for recovery, and export the collected image data.

[0009] The image sampling process of the S2 UAV described in this invention involves the UAV using onboard RTK for positioning and navigation throughout the entire process, without relying on the onboard camera for positioning and navigation through images, and all obtained images contain RTK positioning information.

[0010] The S24 ground wire arc directly above the S24 ground wire in this invention refers to a marker containing the ground wire being measured at a vertical position in the downward view of the airborne camera. If necessary, the flight altitude or view of the UAV can be adjusted.

[0011] The sampling points for the sag of the conductor in S3 of this invention include, but are not limited to, points on the line or several points in the interval of the conductor image. Different image recognition techniques use different methods for determining the points, such as center point, side point, or point family.

[0012] The image recognition algorithm in S3 of this invention includes the following steps: S31. Through feature extraction, process the markers in the GPS marker base point image of the iron tower, then identify the shape of the markers and mark their relative positions in the image; S32. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the side-view guide line sag image and the UAV, and then obtain the absolute height position information and latitude-longitude relative position information of the sampling point in the side-view guide line sag image. S33. By extracting features, process the markers in the image directly above the sag of the tower conductor and ground wire, then identify the shape of the markers and mark their relative positions in the image; S34. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the top-view guide wire sag image and the UAV, and then obtain the absolute latitude and longitude position information and relative height position information of the sampling points in the top-view guide wire sag image. S35. Combine the absolute and relative position information described in S33 and S34 with the drone's constant-speed cruise flight speed and image timestamp to calculate the absolute position information of the sampling point.

[0013] S36. Repeat steps S31-S35 for all conductor images until the absolute position information of all sampling points is obtained.

[0014] The feature extraction and recognition in S31 and S33 of this invention refer to advanced algorithms in the field of computer vision. Different algorithm strategies can be selected for different test objects, which will not be elaborated in this invention.

[0015] The absolute altitude position information in S32 of this invention refers to the fact that in the side view image, the UAV and the grounding wire only have overlapping information in the two dimensions of altitude and horizontality, so only absolute altitude position information can be obtained; the absolute latitude and longitude position information in S34 refers to the fact that in the top view image, the UAV and the grounding wire only have overlapping information in the two dimensions of latitude and longitude, so only absolute latitude and longitude position information can be obtained.

[0016] In the present invention, the three-dimensional mapping in S4 refers to solving the latitude, longitude, and altitude coordinates of all sampling points, combined with the suspension points and the lowest point at both ends of the guide wire, onto the same plane, retaining only two dimensions of information, setting a coordinate system, and finally visually displaying the transformed X and Y coordinates; the curve fitting in S4 is to obtain the fitting equation for the target of interest, including but not limited to methods such as visualization analysis, model selection, linear interpolation, and error minimization target fitting, etc. The target of interest includes but is not limited to minimum error, minimum absolute error, weighted least squares, etc.

[0017] This invention leverages the flexible and mobile image sampling capabilities of drones, eliminating the need for on-site instrument setup and completely resolving the problems of obstructed lines of sight and difficult instrument deployment in traditional manual measurements. This significantly expands the applicable scenarios for sag measurement. By setting GPS markers at adjacent towers for initial calibration and standardizing the sampling process, this invention greatly reduces the difficulty of identifying conductors and ground wires in subsequent image recognition algorithms, significantly improving accuracy. Furthermore, this invention eliminates the need for multiple personnel on-site; a single drone operator can control the drone and process the data on a computer to complete the measurement task. This significantly reduces human intervention, lowers safety risks in high-altitude and steep-slope areas, and drastically shortens the time required for a single measurement, meeting the needs of rapid inspection of large-scale transmission lines and effectively reducing maintenance costs. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a schematic diagram of the UAV image sampling operation process method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the sampling point localization and image recognition algorithm according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the mathematical model analysis and business application of sampling points in an embodiment of the present invention; Figure 5 This is a side-view photograph of the sag of the ground guide wire according to an embodiment of the present invention; Figure 6 This is a top-view photograph of the sag of the grounding conductor, according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be described in more detail, clearly, and completely below with reference to the accompanying drawings: like Figure 1 As shown, the conductor sag measurement method of this embodiment includes the following steps: S1. Set the GPS marker base points of adjacent towers A and B as the initial calibration positions in the airborne camera images.

[0020] In the power grid tower construction project acceptance scenario of this embodiment, the core objective is to verify the length and sag of the conductor installed between two adjacent towers. The project construction drawings have already marked the GPS location of the towers and the catenary equation for the sag, but it is also necessary to set GPS marker base points related to the start and end positions of the conductor. During the initial measurement task, it is necessary to manually carry the RTK positioning equipment to the start and end positions of the conductor on the tower, fix and position the bright red reflective conductor protective sleeve marker, and use it as the initial calibration position in the airborne camera image in subsequent steps. After the first task is completed, the marker is not retrieved and can be used as a permanent measurement point in the next measurement, improving measurement efficiency.

[0021] S2. Operational procedures for using drones to sample images of ground wires; like Figure 2 The process for performing image sampling of the sag of the conductor being tested outdoors in this embodiment includes the following steps: S21. Start sampling. Manually control the drone to take off and fly to the GPS marker base point of Tower A. Manually mark the GPS marker base point of Tower A as waypoint A1 and take a picture of the base point location.

[0022] In this embodiment, the surveyor manually remotely controls the drone to take off and fly into the field of view of the onboard camera. The bright red reflective conductor protection sleeve marker of Tower A is positioned in the center of the image, and a photo is taken. Figure 5 Use the image as the base point location, and then manually mark the GPS marker base point of Tower A as waypoint A1.

[0023] S22. Manually control the drone to fly to the GPS marker base point of Tower B, manually mark Tower B as waypoint B1, and take an image of the base point location.

[0024] In this embodiment, after completing S21, the surveyor manually remotely controls the drone to fly to the center of the image where the bright red reflective wire protective sleeve marker of Tower B is located, and takes a picture as the base point position image. Then, the surveyor manually marks the GPS marker base point of Tower B as waypoint B1.

[0025] S23. Make the UAV fly at a constant speed from waypoint B1 to A1, and at the same time take side-view guide line sag images according to the performance of the airborne camera, and fly to the GPS marker base point position of tower A.

[0026] In this embodiment, the UAV automatically plans the route from B1 to A1. The route planning algorithm can be any advanced algorithm. The planned route must completely include the sag of the ground wire being measured. If the automatically planned route cannot completely include the object being measured, manual track recording can be used when determining the route. In this embodiment, when the UAV is manually controlled to fly in S21 and S22, the flight trajectory has been recorded and it has a certain track-following flight function. Therefore, when flying at a constant speed at waypoints, it can automatically and completely photograph the sag of the ground wire being measured.

[0027] S24. Manually control the drone to fly directly above the sag of the conductor wire of tower A, manually mark waypoint A2, and take a picture of the base point position.

[0028] In this embodiment, after completing the side view image sampling, the top view image sampling immediately begins. The surveyor manually remotely controls the drone to fly directly above the sag of the conductor and ground wire of tower A, positions the bright red reflective conductor protective sleeve marker in the center of the onboard camera's image, and takes a picture. Figure 6 Use the image as the base point location, and then manually mark waypoint A2 at that location.

[0029] S25. Manually control the drone to fly directly above the sag of the conductor wire of Tower B, manually mark waypoint B2, and take a picture of the base point position.

[0030] In this embodiment, after completing S24, the surveyor manually controls the drone to fly directly above the sag of the conductor and ground wire of tower B, places the bright red reflective conductor protective sleeve marker in the center of the image of the airborne camera, takes a picture as the base point position image, and then manually marks the departure point B2.

[0031] S26. Make the UAV fly at a constant speed from waypoint B2 to A2, and at the same time take pictures of the overhead guide wire at regular intervals according to the performance of the airborne camera, and fly to the top of the sag of the guide wire of tower A.

[0032] In this embodiment, since the sag of the conductor is vertical, the orthogonal projection of the sag falls completely on the straight line from B2 to A2. The UAV automatically plans the route from B2 to A2, and the surveyor issues an instruction for the UAV to automatically fly at a constant speed from B2 to A2, automatically and completely capturing the image of the sag of the conductor being measured.

[0033] S27. End sampling, manually remotely control the drone to fly to the landing position for recovery, and export the collected image data.

[0034] In this embodiment, the surveyor manually remotely controls the drone to fly to the landing position and retrieve it. All the collected image data is exported via a data cable, the data is preprocessed, and then imported into a prepared computer. The preprocessing operation can be any advanced algorithm or method for image inspection, data filtering, classification, etc.

[0035] The image sampling process of the S2 UAV involves the UAV using onboard RTK for positioning and navigation throughout the entire process, without relying on the onboard camera for positioning and navigation through images, and all the obtained images contain RTK positioning information.

[0036] The S24 ground wire sags directly above the marker in the downward view of the airborne camera, which contains the ground wire being measured at a vertical position. If necessary, the UAV's flight altitude or view can be adjusted.

[0037] S3. The sampling points of the conductor sag are located using an image recognition algorithm.

[0038] In this embodiment, the sampling points of the conductor sag include, but are not limited to, points on the line or several points in the interval of the conductor image. Different image recognition techniques use different methods to determine the points, such as center point, side point, point family, etc.

[0039] like Figure 3 In this embodiment, the steps for fixed-point positioning in the specific implementation of the image recognition algorithm include: S31. By extracting features, process the markers in the GPS marker base point image of the iron tower, then identify the shape of the markers and mark their relative positions in the image.

[0040] S32. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the side-view guide line sag image and the UAV, and then obtain the absolute height position information and latitude-longitude relative position information of the sampling point in the side-view guide line sag image.

[0041] S33. By extracting features, process the markers in the image directly above the sag of the tower conductor and ground wire, then identify the shape of the markers and mark their relative positions in the image.

[0042] S34. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the top-view guide wire sag image and the UAV, and then obtain the absolute latitude and longitude position information and relative altitude position information of the sampling points in the top-view guide wire sag image.

[0043] S35. Combine the absolute and relative position information described in S33 and S34 with the drone's constant-speed cruise flight speed and image timestamp to calculate the absolute position information of the sampling point.

[0044] S36. Repeat steps S31-S35 for all conductor images until the absolute position information of all sampling points is obtained.

[0045] The feature extraction and recognition in S31 and S33 refer to advanced algorithms in the field of computer vision. Different algorithm strategies can be selected for different test objects, which will not be elaborated in this invention.

[0046] The absolute altitude position information in S32 refers to the fact that in the side view image, the UAV and the grounding wire only have overlapping information in the two dimensions of altitude and horizontality, so only absolute altitude position information can be obtained; the absolute latitude and longitude position information in S34 refers to the fact that in the top view image, the UAV and the grounding wire only have overlapping information in the two dimensions of latitude and longitude, so only absolute latitude and longitude position information can be obtained.

[0047] S4. Perform mathematical model analysis on the sampling points, including two steps: three-dimensional mapping and curve fitting. Accurately represent the positioning information of all sampling points in the form of sag equations in the plane coordinate system formed by the two suspension points and the lowest point.

[0048] like Figure 4 In this embodiment, performing three-dimensional mapping on the positioning information of all sampling points means combining the latitude, longitude, and elevation coordinates of all sampling points with the suspension points and lowest point at both ends of the ground wire, calculating them onto the same plane, retaining only two dimensions of information, setting a coordinate system, and finally visually displaying the transformed X and Y coordinates.

[0049] In this embodiment, curve fitting is performed on the sampling points in three-dimensional space to obtain the fitting equation for the target of interest. The processing methods include, but are not limited to, visualization analysis, model selection, linear interpolation, and error minimization target fitting. This embodiment uses linear interpolation and error minimization target fitting. The target of interest includes, but is not limited to, minimum error, minimum absolute error, and weighted least squares. In this embodiment, the target of interest is the lowest point with minimum error, focusing on the sag height.

[0050] S5. The final obtained sag equation is used in business applications, including calculating characteristic indicators and verifying engineering errors.

[0051] In this embodiment, the obtained sag equation is further compared with the catenary equation marked on the project construction drawings for engineering error verification, sag minimum point characteristic index comparison, and other business applications to evaluate the construction quality of the tested conductor.

Claims

1. A method for measuring the sag of a ground wire based on UAV image recognition, characterized in that: The steps are as follows: S1. Set the GPS marker base points of adjacent towers A and B as the initial calibration positions in the airborne camera images; S2. Operational procedures for using drones to sample images of ground wires; S3. Use an image recognition algorithm to locate the sampling points of the conductor sag. S4. Perform mathematical model analysis on the sampling points, including two steps: three-dimensional mapping and curve fitting. Accurately represent the positioning information of all sampling points in the form of sag equations in the plane coordinate system formed by the two suspension points and the lowest point. S5. The final obtained sag equation is used in business applications, including calculating characteristic indicators and verifying engineering errors.

2. The method for measuring conductor sag based on UAV image recognition according to claim 1, characterized in that: The GPS marker base point location is the positioning point of the marker object at the beginning and end of the conductor. The marker object includes, but is not limited to, signs, marker poles, etc. that can be clearly photographed by the drone's onboard camera. The setting of the marker base point location must be combined with real-time differential positioning equipment.

3. The method for measuring the sag of a ground wire based on UAV image recognition according to claim 1, characterized in that: The image sampling process of the S2 UAV includes the following steps: S21. Start sampling. Manually control the drone to take off and fly to the GPS marker base point of Tower A. Manually mark the GPS marker base point of Tower A as waypoint A1 and take a picture of the base point location. S22. Manually control the drone to fly to the GPS marker base point of Tower B, manually mark Tower B as waypoint B1, and take an image of the base point location. S23. Make the UAV fly at a constant speed from waypoint B1 to A1, and at the same time take side-view ground line sag images according to the performance of the airborne camera, and fly to the GPS marker base point position of tower A. S24. Manually control the drone to fly directly above the sag of the conductor wire of tower A, manually mark waypoint A2, and take a picture of the base point position; S25. Manually control the drone to fly directly above the sag of the conductor wire of Tower B, manually mark waypoint B2, and take a picture of the base point position. S26. Make the UAV fly at a constant speed from waypoint B2 to A2, and at the same time take pictures of the overhead ground wire at regular intervals according to the performance of the airborne camera, and fly to the direct top of the sag of the ground wire of tower A. S27. End sampling, manually remotely control the drone to fly to the landing position for recovery, and export the collected image data.

4. The method for measuring conductor sag based on UAV image recognition according to claim 1, characterized in that: The image sampling process of the S2 UAV involves the UAV using onboard RTK for positioning and navigation throughout the entire process, without relying on the onboard camera for positioning and navigation through images, and all the obtained images contain RTK positioning information.

5. The method for measuring conductor sag based on UAV image recognition according to claim 3, characterized in that: The S24 ground wire sags directly above the marker in the downward view of the airborne camera, which contains the ground wire being measured at a vertical position. If necessary, the UAV's flight altitude or view can be adjusted.

6. The method for measuring conductor sag based on UAV image recognition according to claim 1, characterized in that: The sampling points for the sag of the conductor in S3 include, but are not limited to, points on the line or several points in the interval of the conductor image. Different image recognition techniques use different methods for determining the points, such as center point, side point, or point family.

7. The method for measuring conductor sag based on UAV image recognition according to claim 1, characterized in that: The image recognition algorithm in S3 includes the following steps: S31. Through feature extraction, process the markers in the GPS marker base point image of the iron tower, then identify the shape of the markers and mark their relative positions in the image; S32. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the side-view guide line sag image and the UAV, and then obtain the absolute height position information and latitude-longitude relative position information of the sampling point in the side-view guide line sag image. S33. By extracting features, process the markers in the image directly above the sag of the tower conductor and ground wire, then identify the shape of the markers and mark their relative positions in the image; S34. Calculate the relative position of the marker with the current image UAV positioning information and the marker's GPS information to obtain the true distance between the top-view guide wire sag image and the UAV, and then obtain the absolute latitude and longitude position information and relative height position information of the sampling points in the top-view guide wire sag image. S35. Combine the absolute and relative position information described in S33 and S34 with the drone's constant-speed cruise flight speed and image timestamp to calculate the absolute position information of the sampling point; S36. Repeat steps S31-S35 for all conductor images until the absolute position information of all sampling points is obtained.

8. The method for measuring conductor sag based on UAV image recognition according to claim 7, characterized in that: The feature extraction and recognition in S31 and S33 refer to advanced algorithms in the field of computer vision. Different algorithm strategies can be selected for different test objects, which will not be elaborated in this invention.

9. The method for measuring conductor sag based on UAV image recognition according to claim 7, characterized in that: The absolute altitude position information in S32 refers to the fact that in the side view image, the UAV and the grounding wire only have overlapping information in the two dimensions of altitude and horizontality, so only absolute altitude position information can be obtained; the absolute latitude and longitude position information in S34 refers to the fact that in the top view image, the UAV and the grounding wire only have overlapping information in the two dimensions of latitude and longitude, so only absolute latitude and longitude position information can be obtained.

10. The method for measuring conductor sag based on UAV image recognition according to claim 1, characterized in that: In S4, the three-dimensional mapping refers to solving the latitude, longitude, and altitude coordinates of all sampling points, combined with the suspension points and lowest points at both ends of the ground wire, onto the same plane, retaining only two dimensions of information, setting up a coordinate system, and finally visually displaying the transformed X and Y coordinates. In S4, the curve fitting is to obtain the fitting equation for the target of interest, including but not limited to methods such as visualization analysis, model selection, linear interpolation, and error minimization target fitting. The target of interest includes but is not limited to minimum error, minimum absolute error, and weighted least squares.

Images