An Automated Measurement Device and Method for Overhead Conductor Sag Based on Multi-Sensor Fusion

Abstract

This application discloses an automated measurement device and method for overhead conductor sag based on multi-sensor fusion. The device includes a vision computing platform, a pan-tilt control platform, a data acquisition platform, and a human-machine interaction platform. Specifically: the vision computing platform identifies the suspension points of the conductor on the local side, the opposite side, and the tangent position using image recognition technology, and sends the target position information to the pan-tilt control platform; the pan-tilt control platform automatically controls the pan-tilt to align with the target based on the target position information; the data acquisition platform collects the required data for sag calculation after pan-tilt alignment, including the height of the conductor suspension point on the local side, the angle between the measuring device and the conductor tangent, and the straight-line distance and angle between the measuring device and the opposite side suspension point; the measuring device calculates the conductor sag based on the collected data; and the human-machine interaction platform displays the measurement results of the conductor sag. This application helps to achieve highly automated and high-precision sag measurement, improving measurement reliability and engineering applicability.

Description

Technical Field

[0001] This application relates to the field of power engineering measurement technology, and relates to, but is not limited to, an automated measurement device and method for overhead conductor sag based on multi-sensor fusion. Background Technology

[0002] The sag value of transmission lines is a crucial factor in their design and maintenance. This factor is easily affected by environmental factors. When temperatures change, conductors often expand or contract. Rising temperatures stretch the conductors, increasing sag, while falling temperatures tighten them, decreasing sag. In winter, ice or snow covering the conductors increases their load, further increasing sag. Frequent strong winds also cause changes in sag. The appropriateness of the sag value is closely related to the stability of the transmission line. If the sag is less than the specified range, the internal tension of the conductor increases, making conductor breakage or insulator string failure more likely. If the sag is greater than the specified range, the conductor's clearance from the ground is too small, increasing the risk of conductor-to-ground discharge, which can severely paralyze the power supply system.

[0003] In existing technologies, sag detection methods include manual inspection and observation. Observation methods have lower measurement accuracy and are significantly affected by factors such as the observer's eyesight, experience, and observation angle. While manual inspection methods offer high accuracy, the inspection cycle is too long, and manual inspections in the field are difficult, risky, and prone to safety hazards. The current standard DL / T1367-2014, "Technical Guidelines for Transmission Line Inspection," uses traditional equipment such as theodolites and distance measuring instruments for measurement. This method is cumbersome, difficult to operate, and unsuitable for rapid and flexible inspection in vast and complex terrains, especially in mountainous areas, deserts, and other areas with poor transportation access.

[0004] Therefore, there is an urgent need for a more accurate automated measurement device and method for overhead conductor sag to solve the problems of existing measurement methods being easily affected by the subjective factors of the surveyors, cumbersome measurement steps, and low efficiency, so as to achieve automated processing of engineering measurement data and a significant improvement in the accuracy of sag calculation. Summary of the Invention

[0005] This application provides an automated measurement device and method for overhead conductor sag based on multi-sensor fusion.

[0006] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide an automated measurement device for overhead conductor sag based on multi-sensor fusion. The device includes a vision computing platform, a gimbal control platform, a data acquisition platform, and a human-computer interaction platform. The vision computing platform includes a camera and a Jetson circuit board. The gimbal control platform includes a two-degree-of-freedom control gimbal and a gimbal servo motor. The data acquisition platform includes an attitude sensor, a laser rangefinder, and a control circuit board, wherein: The visual computing platform is used to identify the suspension points of the conductor on this side, the conductor suspension points on the opposite side, and the positions of the conductor tangent using image recognition technology, and sends the target position information to the pan-tilt control platform. The pan-tilt control platform is used to automatically control the pan-tilt to align with the target based on the target position information. The data acquisition platform is used to acquire the data required for sag calculation after the pan-tilt is aligned. The acquired data includes the height of the conductor suspension point on this side, the inclination angle between the measuring device and the conductor tangent, the straight-line distance between the measuring device and the conductor suspension point on the opposite side, and the inclination angle. The measuring device calculates the conductor sag based on the acquired data. The formula for calculating the conductor sag is expressed as follows: ; In the formula, The acute angle perpendicular to the point of tangency between the observation point and the overhead conductor; The acute angle between the observation point and the suspension point of the overhead conductor; The distance between the suspension point of the overhead wire at the end of the device and the horizontal and vertical distance between the instrument and the device. The distance is the straight-line distance from the observation point to the far-end suspension point; the human-computer interaction platform is used to display the measurement results of the conductor sag.

[0007] The technical solution provided in this application identifies the suspension points of the conductor on the local side, the suspension points of the conductor on the opposite side, and the positions of the conductor tangents in a visual computing platform using image recognition technology, and sends the target position information to the gimbal control platform. An improved YOLOv8n-OBB target detection model is used, inserting convolutional attention modules with different weights into the Backbone and Neck layers, thereby significantly improving the detection mPA. By introducing a conductor contour enhancement subnetwork, edge detection and line fitting are used to correct and strengthen the conductor contour features, improving the tangent point detection recall rate by 5.1%. A dynamic anchor box adjustment strategy is employed, automatically clustering and generating 3-6 sets of anchor boxes adapted to the scene based on the input image scale, improving the small target detection accuracy by 4.3%, thus achieving high-precision identification of conductor suspension points and tangents in complex backgrounds. In the gimbal control platform, the target position information is automatically... The gimbal is aligned with the target using a three-loop PID control strategy combined with fuzzy PID parameter adaptive tuning. Through the coordinated control of the position loop, velocity loop, and current loop, and with the inclusion of a deviation signal dead zone and piecewise integration processing, rapid and accurate alignment with the target is achieved. The gimbal's response speed and control accuracy are significantly improved, ensuring stable target tracking even in complex environments. After gimbal alignment, a data acquisition platform collects the data required for sag calculation. This data includes the height of the conductor suspension point on this side, the inclination angle between the measuring device and the conductor tangent, and the straight-line distance and inclination angle between the measuring device and the conductor suspension point on the opposite side. The data acquisition platform integrates an attitude sensor and a laser rangefinder, simultaneously collecting multi-source data after precise gimbal alignment to provide high-precision basic data for sag calculation. The measuring device calculates the conductor sag based on the collected high-precision data, and the measurement results are displayed through a human-machine interface platform. The measuring device provided in this application realizes fully automated and high-precision measurement of the sag of overhead transmission lines through multi-sensor fusion and intelligent control technology. Under standard experimental conditions, the sag measurement error can be controlled within ±2cm, which is better than the traditional manual observation method. A single sag measurement is completed automatically in about 2 minutes, which is more than 5 times more efficient than the traditional manual operation.

[0008] Optionally, the visual computing platform employs an improved YOLOv8n-OBB object detection model to identify wire suspension points and wire tangent positions, and inserts convolutional attention modules with different weights into its Backbone and Neck layers respectively; the object detection model adopts a dynamic anchor box adjustment strategy, automatically clustering and generating 3-6 sets of anchor boxes adapted to the scene based on the input image scale; the visual computing platform also includes a wire contour enhancement subnetwork, which, after processing by the convolutional attention module, corrects and enhances the wire contour features through edge detection and line fitting.

[0009] Optionally, the gimbal control platform adopts a three-loop PID control strategy to control the gimbal servo motor, including a position loop, a speed loop, and a current loop; the three-loop PID control strategy adopts digital incremental PID and is equipped with a dead zone for the deviation signal and piecewise integral processing; the gimbal control platform performs parameter adaptive tuning through a fuzzy PID controller, taking the system deviation and the deviation change rate as input, using the Mamdani model and triangular membership functions for fuzzy inference, and outputting the adjustment amount of the PID parameters.

[0010] Optionally, the measuring device is fixed by a tripod, and the laser rangefinder and the camera's shooting center point are located on the same horizontal plane; the camera communicates with the Jetson circuit board via USB; the Jetson circuit board communicates with the control circuit board, the human-machine interface platform, and the control circuit board and the laser rangefinder via serial communication and exchanges data according to the Modbus communication protocol.

[0011] Secondly, embodiments of this application provide an automated measurement method for overhead conductor sag based on multi-sensor fusion. This method is applied to an automated measurement device for overhead conductor sag based on multi-sensor fusion. The method includes: selecting a sag observation station; fixing the measurement device below a nearby tower; and initializing the measurement device to a horizontal state; automatically aligning the device with the suspension point of the conductor on the local side using a visual computing platform and a gimbal control platform; acquiring the height of the suspension point of the conductor on the local side using a data acquisition platform; automatically aligning the device with the tangent position of the conductor using a visual computing platform and a gimbal control platform; acquiring the inclination angle between the measurement device and the tangent of the conductor using a data acquisition platform; automatically aligning the device with the suspension point of the conductor on the opposite side using a visual computing platform and a gimbal control platform; acquiring the straight-line distance and inclination angle between the measurement device and the suspension point of the conductor on the opposite side using a data acquisition platform; calculating the sag based on the height of the suspension point of the conductor on the local side, the inclination angle between the measurement device and the tangent of the conductor, and the straight-line distance and inclination angle between the measurement device and the suspension point of the conductor on the opposite side; and displaying the measurement results of the conductor sag through a human-computer interaction platform.

[0012] Optionally, the initialization process of the measuring device and the automatic alignment process of the gimbal specifically include: adjusting the measuring device to a horizontal state through the attitude sensor and the two-degree-of-freedom gimbal, and controlling the shooting direction of the camera to be perpendicular to the tower on this side; the camera captures the target image and transmits it to the visual computing platform; the visual computing platform identifies the target position through the target detection model and transmits the target position information to the gimbal control platform; the gimbal control platform plans the motion path and controls the gimbal to align with the target.

[0013] Optionally, the target detection model is trained using the following methods: constructing a two-level training system containing a basic dataset and scene subdivision datasets for different terrains, and introducing domain-adaptive training; employing a semi-supervised training method that combines a small amount of manual annotation with a large amount of pseudo-annotation, wherein the pseudo-annotation is automatically generated by the trained model and manually corrected; and performing strong noise enhancement on the training data, wherein the strong noise includes at least one of salt-and-pepper noise, Gaussian blur and random occlusion, as well as synthetic image enhancement simulating fog, rain and snow weather.

[0014] Optionally, the automatic alignment of the conductor tangent position includes: determining the tangent point by segmenting and labeling the conductor segments, fitting a quadratic curve, and finding the extreme points.

[0015] Optionally, if the gimbal fails to align with the target within a preset time, the camera exposure parameters will be adjusted and the gimbal will be re-aligned. If the number of re-alignments exceeds a preset number, the system will switch to manual assistance mode.

[0016] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the steps in the above-described automated measurement method for overhead conductor sag based on multi-sensor fusion.

[0017] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in the above-described automated measurement method for overhead conductor sag based on multi-sensor fusion.

[0018] The beneficial effects of the technical solutions provided in this application include at least the following: This application provides an automated measurement device and method for overhead conductor sag based on multi-sensor fusion. In a visual computing platform, image recognition technology is used to identify the suspension points of the conductor on the local side, the suspension points of the conductor on the opposite side, and the positions of the conductor tangents. The target position information is then sent to a gimbal control platform. An improved YOLOv8n-OBB target detection model is employed, inserting convolutional attention modules with different weights into the backbone and neck layers, significantly improving the detection mPA. By introducing a conductor contour enhancement subnetwork, edge detection and line fitting are used to correct and strengthen the conductor contour features, improving the tangent point detection recall rate by 5.1%. Furthermore, a dynamic anchor box adjustment strategy is used to automatically cluster and generate 3-6 sets of anchor boxes adapted to the scene based on the input image scale, improving the small target detection accuracy by 4.3%. This achieves high-precision identification of conductor suspension points and tangent points in complex backgrounds. In the gimbal control platform… The system automatically controls the gimbal to align with the target based on the target's location information. It employs a three-loop PID control strategy combined with fuzzy PID parameter adaptive tuning. Through coordinated control of the position loop, velocity loop, and current loop, and with a deviation signal dead zone and piecewise integration processing, it achieves rapid and accurate alignment with the target. This significantly improves the gimbal's response speed and control precision, ensuring stable target tracking even in complex environments. After gimbal alignment, a data acquisition platform collects the necessary data for sag calculation. This data includes the height of the conductor suspension point on this side, the inclination angle between the measuring device and the conductor tangent, and the straight-line distance and inclination angle between the measuring device and the opposite conductor suspension point. The data acquisition platform integrates an attitude sensor and a laser rangefinder, simultaneously collecting multi-source data after precise gimbal alignment. This provides high-precision basic data for sag calculation. The measuring device calculates the conductor sag based on the collected high-precision data, and the measurement results are displayed through a human-machine interface platform. The measuring device provided in this application realizes fully automated and high-precision measurement of the sag of overhead transmission lines through multi-sensor fusion and intelligent control technology. Under standard experimental conditions, the sag measurement error can be controlled within ±2cm, which is better than the traditional manual observation method. A single sag measurement is completed automatically in about 2 minutes, which is more than 5 times more efficient than the traditional manual operation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 A schematic diagram of an automated measurement device for overhead conductor sag based on multi-sensor fusion provided in an embodiment of this application; Figure 2A schematic diagram of the structure of an improved YOLOv8n-OBB target detection model provided in an embodiment of this application; Figure 3 A control principle and PID adjustment flowchart of a gimbal control platform provided in this application embodiment; Figure 4 A schematic diagram of a sag measurement calculation model provided in an embodiment of this application; Figure 5 A schematic diagram illustrating the working process and fault tolerance mechanism of a measuring device provided in this application embodiment; Figure 6 A flowchart of an automated measurement method for overhead conductor sag based on multi-sensor fusion provided in this application embodiment; Figure 7 This is a schematic diagram of the hardware entity of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0022] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific order of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0023] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have a meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0024] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0025] In view of the current problems in the measurement of overhead conductor sag in the field of power engineering measurement technology, this application provides an automated measurement device and method for overhead conductor sag based on multi-sensor fusion.

[0026] The technical solution of this application is described below, starting with the device embodiment.

[0027] Please refer to Figure 1 It shows a schematic diagram of an automated overhead conductor sag measurement device based on multi-sensor fusion provided in an embodiment of this application, such as... Figure 1 As shown, the device includes a visual computing platform 01, a gimbal control platform 02, a data acquisition platform 03, and a human-computer interaction platform 04. The visual computing platform 01 includes a camera and a Jetson circuit board. The gimbal control platform 02 includes a two-degree-of-freedom control gimbal and a gimbal servo motor. The data acquisition platform 03 includes an attitude sensor, a laser rangefinder sensor, and a control circuit board. The visual computing platform 01 is used to identify the suspension point of the conductor on this side, the suspension point of the conductor on the opposite side, and the position of the conductor tangent using image recognition technology, and sends the target position information to the gimbal control platform 02. The gimbal control platform 02 is used to automatically control the gimbal to align with the target based on the target position information. The data acquisition platform 03 is used to collect the data required for sag calculation after the gimbal is aligned. The collected data includes the height of the suspension point of the conductor on this side, the inclination angle between the measuring device and the conductor tangent, the straight-line distance between the measuring device and the suspension point of the conductor on the opposite side, and the inclination angle. The measuring device calculates the conductor sag based on the collected data. The human-computer interaction platform 04 is used to display the measurement results of the conductor sag.

[0028] In this embodiment, the visual computing platform 01 is used to identify the suspension point of the conductor on this side, the suspension point of the conductor on the opposite side, and the position of the conductor tangent using image recognition technology, and sends the target position information to the gimbal control platform 02. Specifically, a camera captures a photo of the target measurement position and transmits the image to the visual computing platform 01 via USB. The visual computing platform 01 automatically identifies the relevant position information of the target measurement position, including the conductor suspension point and the conductor tangent position, based on the improved YOLOv8n-OBB target positioning algorithm. Through serial communication between the visual computing platform 01 and the gimbal control platform 02, the target position information is transmitted from the visual computing platform 01 to the gimbal control platform 02 to provide target position information for the subsequent alignment process of the gimbal control platform 02.

[0029] In a specific embodiment, the visual computing platform 01 uses an improved YOLOv8n-OBB object detection model to identify the conductor suspension point and conductor tangent position. The conductor tangent position refers to the point where the camera's optical axis is tangent to the lowest point of the overhead conductor's sag, which is represented as the extreme point of the conductor's outline in the image. First, a dataset is constructed using the acquired basic image data, comprising approximately 5,000 high-definition transmission line images collected on-site, covering different weather, lighting, background, and conductor posture scenarios. The image resolution is 640×640 to accommodate the model's input size. Furthermore, the dataset is categorized using rotated bounding boxes with an angle range of [-90°, 90°] to accurately represent the spatial orientation of the conductor suspension points and tangent positions. The rotation angle is determined through the following process: fitting the catenary equation to the pixel set between conductor suspension points to obtain a fitted curve; calculating the tangent direction of each sampling point on the fitted curve; and using the tangent direction as the rotation angle of the corresponding rotated bounding box. The categorized annotations specifically include conductor suspension points, conductor tangent positions, tower structures, insulator strings, and other background interference objects. Among these, conductor suspension points and conductor tangent positions are key detection targets. Each image in the dataset is annotated with a refined bounding box, providing center point coordinates and orientation angle information. Notably, the annotation of the conductor tangent position is achieved through a geometric fitting method. Specifically, piecewise OBBs are used to annotate the continuous line segments of the conductor in the image. A quadratic curve is fitted using the conductor point set according to the least squares method to establish the geometric model of the conductor from that perspective. The actual lowest point of the conductor, i.e., the tangent point, is determined by calculating the extreme points of the fitted curve.

[0030] Furthermore, various data augmentation processes were applied to the dataset. Multiple noise types were randomly added, including at least one of salt-and-pepper noise, Gaussian blur, and random occlusion, with each noise type having an addition probability of 0.15. In image augmentation, brightness and contrast were randomly adjusted to simulate lighting changes, enhancing the model's adaptability under different conditions. Simultaneously, to address common meteorological interference in power line inspections, synthetic image augmentation techniques simulating fog, rain, and snow were introduced. Power scene adaptive data augmentation was also performed, simulating at least one of the following interference types: corona discharge halo effect, conductor vibration blurring in light winds, and partial occlusion of insulator strings, to enhance the model's robustness in harsh environments. After image augmentation, the dataset size was effectively expanded, with approximately half of the dataset having noise added, resulting in a dataset containing 7500 images. The dataset was then divided into training, validation, and test sets in a 7:2:1 ratio to improve the training effect and the objectivity of the evaluation results.

[0031] In specific embodiments, since tangent points lack explicit texture in the image, traditional annotation tools cannot directly capture them. The flexible deformation of the conductor leads to significant differences in anchor frame sizes within the same distance range, causing static anchor frames to fail. Furthermore, the accuracy requirements for sag measurement differ by orders of magnitude from the mAP metric of general detection. These reasons prevent the annotation strategies of general target detection models from being directly applied to sag measurement. Therefore, this application provides an improved YOLOv8n-OBB target detection model for identifying conductor suspension points and tangent positions. Please refer to... Figure 2 It shows a schematic diagram of the structure of an improved YOLOv8n-OBB target detection model provided in an embodiment of this application, as follows. Figure 2 As shown, based on the YOLOv8n-OBB network structure, convolutional attention modules with different weights are inserted after the C2f module in the Backbone layer and before and after the SPPF module in the Neck layer, respectively. Channel attention and spatial attention are calculated sequentially. The spatial attention submodule uses a 7×7 convolutional kernel to aggregate contextual information, enhancing the model's perception of the local structure and spatial relationships of the conductor, thereby more accurately locating the conductor's suspension points and tangent positions in complex backgrounds. The technical solution provided in this application, by inserting convolutional attention modules with different weights after the C2f module in the Backbone layer and before and after the SPPF module in the Neck layer, allows the Backbone layer to focus on channel feature filtering to suppress interference from background vegetation and clouds, while the Neck layer focuses on spatial feature enhancement to accurately locate the conductor's fine tangent points.

[0032] In a specific embodiment, the target detection model is trained using the following method to improve detection accuracy and robustness under different terrains and complex environments: First, a two-level training dataset is constructed. For power transmission line scenarios in different terrains such as mountains, coastlines, and deserts, a two-level training system is established, consisting of a basic dataset and scene-specific datasets. The basic dataset contains 7500 images, covering targets such as conductor suspension points, tangent points, and towers in general scenarios. The scene-specific dataset contains 2000 images for each specific terrain (e.g., mountains, coastlines, deserts) to specifically optimize the model's performance in specific environments. During training, domain-adaptive training is introduced to reduce transfer errors between different scenarios. Cross-scenario measurement error comparison experiments show that after adopting this training system, the measurement error in the desert scenario decreased from ±2.0cm to ±1.5cm. Secondly, a semi-supervised training strategy is adopted to reduce the cost of dataset construction. Specifically, a training method of a small amount of manual annotation plus a large amount of pseudo-annotation is used. A small number of manually annotated images are used to train the initial model, and this model is used to automatically annotate a large number of unannotated images to generate pseudo-annotations. The pseudo-annotations are then manually corrected, with the correction rate controlled to not exceed 10%, ensuring annotation quality while significantly reducing the workload of manual annotation. Furthermore, to improve the model's robustness under real-world conditions, strong noise enhancement processing is added during training. Strong noise includes at least one of salt-and-pepper noise, Gaussian blur, and random occlusion. Through this strong noise training, the model's resistance to real-world interference such as ice accumulation on the conductor surface and bird droppings occlusion is significantly improved, thus ensuring accurate identification of conductor suspension points and tangent positions even in harsh environments. The model training method provided in this application effectively solves the problems of cross-scene measurement errors and real-world interference, laying a solid foundation for subsequent high-precision sag measurement. After training, the center coordinates of feature points can be output through the trained YOLOv8n-OBB target detection model. To verify the effectiveness of the introduced attention mechanism, an ablation experiment was conducted. The results showed that, under the same training conditions, after removing the convolutional attention module, the model's mPA for detecting wire tangent points decreased by about 4.7%, and the localization error for dangling points increased. This indicates that the attention mechanism has a significant improvement effect in detecting complex backgrounds and multi-scale targets.

[0033] In a specific embodiment, for the rotating bounding box detection task, the target detection model adopts a dynamic anchor box adjustment strategy. Based on the scale of the input image, it automatically clusters and generates 3-6 sets of anchor boxes adapted to the scene. It analyzes in real time the aspect ratio, scale, and angle distribution of the bounding boxes of wires and suspending points in the input image, automatically clustering and generating 3-6 sets of anchor boxes adapted to the current scene to adapt to the morphological changes of the wires under different viewpoints, improving detection recall and localization accuracy, and reducing the false negative rate of small-scale tangent points. Comparison data between dynamic and fixed anchor boxes shows a 4.3% improvement in small target detection accuracy. The YOLOv8n-OBB target detection model output... Each rotating detection box contains information such as center point coordinates, width, height, and rotation angle. These detection boxes are first processed by non-maximum suppression to remove redundant boxes. Then, a secondary screening is performed based on the catenary geometric constraints. For any two detection boxes, if the difference in their rotation angles is less than 15 degrees and the intersection-union ratio is greater than 0.7, the suspending point detection box with higher confidence is retained first. After the screening is completed, all wire detection boxes are sorted according to the ordinate of their center points, and parabolic fitting is performed using the least squares method. If the fitting residual exceeds 3 pixels, an anomaly alarm is triggered, indicating that there may be a deviation in the detection results.

[0034] In a specific embodiment, the visual computing platform 01 also includes a wire contour enhancement subnetwork. After the convolutional attention module has finished processing, the wire contour enhancement subnetwork strengthens the wire contour features through edge detection and line fitting. Specifically, after the convolutional attention module has finished processing, a 3×3 convolution and a histogram of oriented gradients are added for feature extraction to enhance the grayscale difference between the wire and the background, thereby solving the problem of blurred wire contours in foggy and snowy environments. After the addition of the wire contour enhancement subnetwork, the tangent point detection recall rate is improved by 5.1%.

[0035] In this embodiment, the gimbal control platform 02 is used to automatically control the gimbal to align with the target based on the target position information identified by the vision computing platform 01. Specifically, the gimbal control platform 02 calculates the target position information and plans the motion path of the two-degree-of-freedom gimbal as the desired position. It controls the gimbal servo motor so that the measuring device can automatically and accurately align with the target position. If the gimbal fails to align with the target within a preset time (e.g., 30 seconds), it adjusts the camera exposure parameters and realigns. After the number of realignments exceeds a preset number (e.g., 3 times), it switches to manual assistance mode.

[0036] In a specific embodiment, the gimbal control platform 02 employs a three-loop PID control strategy to control the gimbal servo motor, including a position loop, a speed loop, and a current loop. Please refer to [reference needed]. Figure 3This document illustrates the control principle and PID adjustment flowchart of a gimbal control platform 02 provided in this application embodiment. The position loop PID converts the error between the desired and actual positions into speed commands, while the speed loop PID converts the difference between the desired and actual speeds into current commands. The current loop detects the coil current and, through a current regulator and PWM control, makes the actual current approach the desired current, thereby achieving precise motor control. The three-loop PID control strategy uses digital incremental PID as the control basis and introduces piecewise integral and deviation signal dead-zone control to improve and optimize the basic PID control model, enhancing the stability and anti-interference capability of the control system. Furthermore, the gimbal control platform 02 performs real-time adaptive parameter tuning through a fuzzy PID controller. Using system deviation and deviation change rate as input, it employs the Mamdani model and triangular membership functions for fuzzy inference, outputting the adjustment amount of the PID parameters.

[0037] In one embodiment, the fuzzy control rule table is shown in Table 1. A fuzzy PID controller is introduced based on this rule table for real-time online adaptive parameter tuning. Specifically, the inputs to the fuzzy PID controller are designed to be the system deviation e and the rate of change of deviation. Output PID parameters through fuzzy inference , , Adjustment amount , , Furthermore, the fuzzy sets and universes of discourse of the input and output variables are determined, where the universe of discourse of the input e is [-5, 5]. The domain of discourse is [-140, 140], and the output is... , , The universe of discourse is set according to actual needs, and the fuzzy set linguistic values ​​are {NB (negative large), NM (negative medium), NS (negative small), ZO (zero), PS (positive small), PM (positive medium), PB (positive large)}. Triangular membership functions (trimf) are used to fuzzify the input and output variables. Based on the fast response and small overshoot characteristics of the gimbal control, a fuzzy control rule table is designed, and a Mamdani-type inference model is used for fuzzy inference. Finally, the centroid method is used for defuzzification to obtain the precise adjustment amount of the PID parameters. , , The real-time adjusted PID parameters are obtained by superimposing the integer output of the fuzzy PID controller with the initial PID parameters. The formula for calculating the real-time adjusted PID parameters is expressed as follows: ; In the formula, , , These are the adaptively corrected PID parameters; , , These are the initial PID parameters; , , This refers to the adjustment amount of the PID parameters output after processing by the fuzzy PID controller. The technical solution provided in this application can dynamically adjust the PID parameters according to the real-time status of the system, improving the control accuracy and robustness of the gimbal in complex environments.

[0038] Table 1 (Fuzzy Control Rule Table) In this embodiment, the data acquisition platform 03 is used to collect the data required for sag calculation after the gimbal is aligned. This includes the height of the suspension point of the guide wire on this side, the inclination angle between the measuring device and the tangent of the guide wire, and the straight-line distance and inclination angle between the measuring device and the suspension point of the guide wire on the opposite side. Specifically, the above data is collected by an attitude sensor and a laser rangefinder. If the straight-line distance measurement between the measuring device and the suspension point of the guide wire on the opposite side fails, the measurement is switched to a long-distance measurement mode and remeasured. After all data is collected, the measuring device calculates the guide wire sag based on the collected data. Specifically, the collected data is substituted into the sag calculation formula, thereby achieving automated and high-precision measurement of the sag. Please refer to [reference needed]. Figure 4 It shows a schematic diagram of a sag measurement calculation model provided in an embodiment of this application. Figure 4 We can obtain: , , Therefore, the formula for calculating sag is expressed as follows: ; In the formula, Indicates sag; The acute angle representing the perpendicular angle between the observation point and the point of tangency of the overhead conductor; This represents the acute angle between the observation point and the suspension point of the overhead conductor. The acute angle is the acute angle of the line connecting the observation point and the target point relative to the horizontal plane. This indicates the horizontal and vertical distance from the suspension point of the overhead wire at the end of the device to the instrument. This represents the straight-line distance from the observation point to the far-end suspension point.

[0039] In a specific embodiment, the measuring device is fixed to a tripod with screws, and the laser rangefinder and the camera's shooting center point are located on the same horizontal plane to ensure that the laser ranging direction is consistent with the camera's optical axis, thereby improving measurement accuracy. Regarding communication connectivity, the camera communicates with the Jetson circuit board via USB, while the Jetson circuit board communicates with the control circuit board, the human-machine interface platform 04, and the laser rangefinder via serial communication according to the Modbus communication protocol. This multi-layered communication architecture ensures efficient and reliable data transmission between platforms.

[0040] In one embodiment, please refer to Figure 5 It illustrates a schematic diagram of the working process and fault tolerance mechanism of a measuring device provided in an embodiment of this application, such as... Figure 5As shown, the initialization process begins, including adjusting the gimbal to a horizontal position using an attitude sensor and a two-degree-of-freedom gimbal, and ensuring the camera's shooting direction is perpendicular to the tower on this side, establishing a benchmark for subsequent measurements. Then, the first positioning is performed, locating the conductor suspension point of tower 1 (this side). The visual computing platform identifies the suspension point on this side using an improved target detection model and sends the target position information to the gimbal control platform. The gimbal control platform automatically aligns the gimbal with the target. During this process, the system continuously checks the positioning success rate. If successful, it proceeds to the next step; if unsuccessful, it checks if the number of positioning attempts is less than three. If less than three, it automatically adjusts the camera exposure parameters and repositions; if three attempts fail, it switches to manual alignment mode, requiring manual alignment by the operator. After successful positioning, the gimbal is automatically controlled to maintain alignment, and a laser rangefinder measures the height distance 'a' of the suspension point on this side. A second positioning process is then performed to locate the tangent point of the conductor. The vision computing platform identifies the tangent point (determined through geometric fitting), and the gimbal control platform aligns the gimbal with the tangent point. This positioning process also includes automatic exposure adjustment and manual alignment mechanisms. Upon successful positioning, the gimbal automatically follows the system, and the attitude sensor measures the tilt angle between the gimbal and the conductor tangent. A third positioning process is then performed to locate the suspension point of the conductor on the opposite side of Tower 2. The vision computing platform identifies the opposite suspension point, and the gimbal control platform aligns the gimbal. If positioning fails, there are also automatic exposure adjustment and manual alignment mechanisms. If the number of positioning attempts is less than three, the exposure is automatically adjusted and the attempt is retried; otherwise, manual alignment is performed. Upon successful positioning, the gimbal automatically follows the system, and the laser rangefinder measures the straight-line distance D between the gimbal and the opposite suspension point. Simultaneously, the attitude sensor measures the tilt angle in that direction. Finally, based on the collected data of the suspension point height a, the tangent tilt angle, the opposite distance D, and the opposite tilt angle, the sag value is calculated using the sag calculation formula, and the measurement results are displayed through a human-machine interface platform. Throughout the process, visual recognition and gimbal closed-loop control are used for each positioning to ensure alignment accuracy. In addition, the fault-tolerant mechanism of automatic exposure adjustment and manual alignment ensures measurement reliability in complex environments, thereby realizing fully automated high-precision measurement of overhead conductor sag.

[0041] In summary, the automated measurement device for overhead conductor sag based on multi-sensor fusion provided in this application includes a vision computing platform, a gimbal control platform, a data acquisition platform, and a human-computer interaction platform. The vision computing platform includes a camera and a Jetson circuit board. The gimbal control platform includes a two-degree-of-freedom control gimbal and a gimbal servo motor. The data acquisition platform includes an attitude sensor, a laser rangefinder sensor, and a control circuit board. In the visual computing platform, image recognition technology is used to identify the suspension points of the conductors on the local side, the suspension points of the conductors on the opposite side, and the positions of the conductor tangents. The target position information is then sent to the gimbal control platform. An improved YOLOv8n-OBB target detection model is employed, inserting convolutional attention modules with different weights into the Backbone and Neck layers, significantly improving the detection mPA. By introducing a conductor contour enhancement subnetwork, edge detection and line fitting are used to correct and strengthen the conductor contour features, improving the tangent point detection recall rate by 5.1%. Furthermore, a dynamic anchor box adjustment strategy is used to automatically cluster and generate 3-6 sets of anchor boxes adapted to the scene based on the input image scale, improving the small target detection accuracy by 4.3%. This achieves high-precision identification of conductor suspension points and tangent points in complex backgrounds. In the gimbal control platform, the gimbal is automatically controlled based on the target position information. The target alignment system employs a three-loop PID control strategy combined with fuzzy PID parameter adaptive tuning. Through the coordinated control of the position loop, velocity loop, and current loop, and with the inclusion of a deviation signal dead zone and piecewise integral processing, it achieves rapid and accurate target alignment, significantly improving the gimbal response speed and control precision, thereby ensuring stable target tracking in complex environments. After gimbal alignment, the system collects the necessary data for sag calculation via a data acquisition platform. This data includes the height of the conductor suspension point on this side, the inclination angle between the measuring device and the conductor tangent, and the straight-line distance and inclination angle between the measuring device and the opposite conductor suspension point. The data acquisition platform integrates an attitude sensor and a laser rangefinder, simultaneously collecting multi-source data after precise gimbal alignment, providing high-precision basic data for sag calculation. After the measuring device calculates the conductor sag based on the collected high-precision data, the measurement results are displayed through a human-machine interface platform. The measuring device provided in this application realizes fully automated and high-precision measurement of the sag of overhead transmission lines through multi-sensor fusion and intelligent control technology. Under standard experimental conditions, the sag measurement error can be controlled within ±2cm, which is better than the traditional manual observation method. A single sag measurement is completed automatically in about 2 minutes, which is more than 5 times more efficient than the traditional manual operation.

[0042] The above is a description of the apparatus embodiments of this application. Based on the foregoing embodiments, the method embodiments of this application are described below.

[0043] Please refer to Figure 6 It shows a flowchart of an automated measurement method for overhead conductor sag based on multi-sensor fusion provided in an embodiment of this application. This method is applied to, for example... Figure 1 The automated measurement device for overhead conductor sag based on multi-sensor fusion shown here is illustrated in the device embodiment. For details not disclosed in the method embodiment, please refer to the device embodiment. Figure 6 As shown, the method includes the following steps S610 to S660.

[0044] Step S610: Select the sag observation station, fix the measuring device under the near-side tower, and initialize the measuring device to a horizontal state; Step S620: Automatically align the suspension point of the conductor on this side through the vision computing platform and the gimbal control platform, and collect the height of the suspension point of the conductor on this side through the data acquisition platform; Step S630: The device is automatically aligned with the tangent position of the conductor through the vision computing platform and the gimbal control platform, and the inclination angle between the measuring device and the tangent of the conductor is collected through the data acquisition platform. Step S640: The device is automatically aligned with the suspension point of the opposite conductor through the vision computing platform and the gimbal control platform, and the straight distance and tilt angle between the measuring device and the suspension point of the opposite conductor are collected through the data acquisition platform. Step S650: Calculate the sag based on the height of the suspension point of the conductor on this side, the angle of inclination between the measuring device and the tangent of the conductor, the straight-line distance between the measuring device and the suspension point of the conductor on the opposite side, and the angle of inclination. Step S660: Display the measurement results of the conductor sag through the human-computer interaction platform.

[0045] In this embodiment of the application, the initialization process of the measuring device specifically includes: adjusting the measuring device to a horizontal state using an attitude sensor and a two-degree-of-freedom gimbal, and controlling the camera's shooting direction to be perpendicular to the tower on this side.

[0046] It should be noted that, in the embodiments of this application, if the above-mentioned automated measurement method for overhead conductor sag based on multi-sensor fusion is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a magnetic disk, or an optical disk. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0047] Correspondingly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements the steps in any of the above-described automated measurement methods for overhead conductor sag based on multi-sensor fusion. Correspondingly, embodiments of this application also provide a computer program product, which, when executed by a processor of an electronic device, is used to implement the steps in any of the above-described automated measurement methods for overhead conductor sag based on multi-sensor fusion.

[0048] Based on the same technical concept, this application provides an electronic device for implementing the automated measurement method for overhead conductor sag based on multi-sensor fusion described in the above method embodiments. Figure 7 This is a hardware entity diagram of an electronic device provided in an embodiment of this application, such as... Figure 7 As shown, the electronic device 700 includes a memory 710 and a processor 720. The memory 710 stores a computer program that can run on the processor 720. When the processor 720 executes the program, it implements the steps in any of the overhead conductor sag automated measurement methods based on multi-sensor fusion described in the embodiments of this application.

[0049] The memory 710 is configured to store instructions and applications executable by the processor 720, and can also cache data to be processed or already processed by the processor 720 and various modules in the electronic device (e.g., image data, audio data, voice communication data and video communication data), which can be implemented by flash memory or random access memory (RAM).

[0050] When the processor 720 executes the program, it implements the steps of the above-mentioned automated measurement method for overhead conductor sag based on multi-sensor fusion. The processor 720 typically controls the overall operation of the electronic equipment 700.

[0051] The aforementioned processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that other electronic devices can also implement the functions of the aforementioned processor, and this application does not specifically limit the specific implementation.

[0052] The aforementioned computer storage media / memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compass disc read-only memory (CD-ROM), etc.; or it can be various electronic devices that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0053] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0054] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0055] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0056] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0057] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0058] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0059] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause the device automatic test line to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0060] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0061] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0062] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An automated measurement device for overhead conductor sag based on multi-sensor fusion, characterized in that, The system includes a visual computing platform, a gimbal control platform, a data acquisition platform, and a human-computer interaction platform. The visual computing platform includes a camera and a Jetson circuit board. The gimbal control platform includes a two-degree-of-freedom control gimbal and a gimbal servo motor. The data acquisition platform includes an attitude sensor, a laser rangefinder, and a control circuit board. The visual computing platform is used to identify the suspension point of the conductor on this side, the suspension point of the conductor on the opposite side, and the position of the conductor tangent through image recognition technology, and send the target position information to the gimbal control platform. The gimbal control platform is used to automatically control the gimbal to align with the target based on the target location information; The data acquisition platform is used to collect the data required for sag calculation after the gimbal is aligned. The collected data includes the height of the suspension point of the conductor on this side, the inclination angle between the measuring device and the conductor tangent, the straight distance and inclination angle between the measuring device and the suspension point of the conductor on the opposite side. The measuring device calculates the conductor sag based on the collected data, and the formula for calculating the conductor sag is expressed as follows: ； In the formula, The acute angle perpendicular to the point of tangency between the observation point and the overhead conductor; The acute angle between the observation point and the suspension point of the overhead conductor; The distance between the suspension point of the overhead wire at the end of the device and the horizontal and vertical distance between the instrument and the device. The straight-line distance from the observation point to the far-end suspension point; The human-computer interaction platform is used to display the measurement results of conductor sag.

2. The apparatus according to claim 1, characterized in that, The visual computing platform uses an improved YOLOv8n-OBB object detection model to identify wire suspension points and wire tangent positions, and inserts convolutional attention modules with different weights into its Backbone and Neck layers respectively; the object detection model adopts a dynamic anchor box adjustment strategy, which automatically clusters according to the scale of the input image to generate 3-6 sets of anchor boxes adapted to the scene. The visual computing platform also includes a conductor contour enhancement subnetwork, which enhances conductor contour features through edge detection and line fitting after the convolutional attention module has finished processing.

3. The apparatus according to claim 1, characterized in that, The gimbal control platform uses a three-loop PID control strategy to control the gimbal servo motor, including a position loop, a speed loop, and a current loop; the three-loop PID control strategy uses digital incremental PID and is equipped with a deviation signal dead zone and piecewise integral processing. The gimbal control platform uses a fuzzy PID controller for adaptive parameter tuning. It takes the system deviation and the rate of change of deviation as input, uses the Mamdani model and triangular membership functions for fuzzy inference, and outputs the adjustment amount of the PID parameters.

4. The apparatus according to claim 1, characterized in that, The measuring device is fixed by a tripod, and the laser rangefinder and the shooting center point of the camera are located on the same horizontal plane; The camera communicates with the Jetson circuit board via USB; the Jetson circuit board communicates with the control circuit board, the human-machine interaction platform, and the laser rangefinder via serial communication and exchanges data according to the Modbus communication protocol.

5. An automated measurement method for overhead conductor sag based on multi-sensor fusion, characterized in that, include: Select a sag observation station, fix the measuring device near the bottom of the tower, and initialize the measuring device to a horizontal state; The visual computing platform and the gimbal control platform are used to automatically align with the suspension point of the conductor on this side, and the height of the suspension point of the conductor on this side is collected by the data acquisition platform. The device automatically aligns with the tangent of the conductor using a visual computing platform and a gimbal control platform, and collects the inclination angle between the measuring device and the tangent of the conductor using a data acquisition platform. The device automatically aligns with the suspension point of the opposite conductor using a visual computing platform and a gimbal control platform, and collects the straight distance and tilt angle between the measuring device and the suspension point of the opposite conductor using a data acquisition platform. The sag is calculated based on the height of the suspension point of the conductor on this side, the angle of inclination between the measuring device and the tangent of the conductor, the straight-line distance between the measuring device and the suspension point of the conductor on the opposite side, and the angle of inclination. The measurement results of conductor sag are displayed through a human-computer interaction platform.

6. The method according to claim 5, characterized in that, The initialization process of the measuring device and the automatic alignment process of the gimbal specifically include: The measurement device is adjusted to a horizontal state using an attitude sensor and a two-degree-of-freedom gimbal, and the camera's shooting direction is controlled to be perpendicular to the tower on this side. The camera captures an image of the target and transmits it to the visual computing platform. The visual computing platform identifies the target's location using a target detection model and transmits the target's location information to the gimbal control platform. The gimbal control platform plans the motion path and controls the gimbal to aim at the target.

7. The method according to claim 6, characterized in that, The target detection model was trained using the following method: A two-level training system is constructed, consisting of a basic dataset and scene-specific datasets for different terrains, and domain-adaptive training is introduced; A semi-supervised training method is adopted, which combines a small amount of manual annotation with a large amount of pseudo-annotation. The pseudo-annotation is automatically generated by the trained model and then manually corrected. The training data is enhanced with strong noise, which includes at least one of salt-and-pepper noise, Gaussian blur and random occlusion, as well as synthetic image enhancement simulating fog, rain and snow weather.

8. The method according to claim 5, characterized in that, The automatic alignment of the conductor tangent position includes: determining the tangent point by segmenting and labeling the conductor segments, fitting a quadratic curve, and finding the extreme points.

9. The method according to claim 5, characterized in that, If the gimbal fails to align with the target within the preset time, the camera exposure parameters will be adjusted and the gimbal will be re-aligned. After the preset number of re-alignments is exceeded, the system will switch to manual assistance mode.

Images