A coating method and system for insulation retrofitting of overhead power lines

Abstract

This application discloses a coating method and system for insulation retrofitting of overhead lines, relating to the field of power equipment maintenance technology. The method includes: responding to a travel start signal of the target equipment on a long-span overhead line, acquiring a silhouette image of the meniscus and loading a meniscus contour template; obtaining the silhouette area deviation and the direction of the meniscus bulge based on the geometric comparison between the silhouette image and the meniscus contour template; generating corresponding coating pumping rate and travel speed adjustment commands based on the silhouette area deviation and the direction of the meniscus bulge, and outputting the adjustment commands to the actuator; correcting the area threshold of the meniscus contour template, updating the adjustment commands based on the corrected area threshold, and obtaining the insulation coating result. This application improves the uniformity of the coating thickness in long-span crossing sections.

Description

Technical Field

[0001] This application relates to the field of power equipment maintenance technology, and in particular to a coating method and system for insulation retrofitting of overhead lines. Background Technology

[0002] Overhead transmission lines are susceptible to environmental erosion during long-term operation, especially in long-span sections crossing rivers, valleys, or highways, where the lack of support causes the conductors to exhibit a pronounced catenary shape with continuously changing slopes. To improve the insulation performance and lightning protection capabilities of these lines, a common method is to use robots to travel along the conductors and apply insulating coatings. This method controls the coating pumping rate and the robot's travel speed to ensure the coating thickness meets engineering requirements.

[0003] Conventional insulation coating methods typically use insulating coatings of fixed viscosity and set constant pumping and travel speeds based on preset process parameter combinations. Some solutions incorporate tilt sensors to adjust the discharge rate based on a table looking up the local slope of the conductor. These methods rely on pre-calibrated span slope data or real-time tilt measurements to convert the effects of gravity into indirect control commands. Because the tilt sensor is mounted on the robot body, its measurements contain coupled information of robot posture disturbances and the actual slope of the conductor; simultaneously, the preset slope table cannot reflect the dynamic deformation of the actual catenary caused by temperature, icing, or wind loads, leading to a mismatch between the adjustment commands and the actual rheological state.

[0004] In the case of a catenary with a large span, the above method is difficult to maintain the uniformity of coating thickness, and the coating may accumulate and drip on the uphill section or be stretched and interrupted on the downhill section. Summary of the Invention

[0005] In view of the aforementioned problems, this application is hereby filed.

[0006] Therefore, this application provides a coating method and system for insulation retrofitting of overhead lines, which can solve the problems mentioned in the background art.

[0007] To solve the above-mentioned technical problems, this application provides the following technical solution: In a first aspect, this application provides a coating method for insulation retrofitting of overhead lines, comprising: in response to a travel start signal of the target equipment on a long-span overhead line, acquiring a silhouette image of the meniscus and loading a meniscus contour template; Based on the geometric comparison between the meniscus silhouette image and the meniscus outline template, the silhouette area deviation and the direction of the meniscus bulge are obtained; Based on the silhouette area deviation and the liquid surface bulge direction characteristics, corresponding coating pumping rate and travel speed adjustment commands are generated, and the adjustment commands are output to the actuator. The area threshold of the meniscus profile template is corrected, and the adjustment command is updated based on the corrected area threshold to obtain the insulation coating result.

[0008] Preferably, the step of acquiring the silhouette image of the meniscus and loading the meniscus contour template includes: Activate the macro industrial camera and coaxial ring LED backlight located on the side of the coating mold outlet; Retrieve the meniscus profile template that matches the current paint type from the local storage unit and load it into the comparison buffer of the image processing unit; The coating pumping rate and the target equipment travel speed are initialized according to the preset process parameters.

[0009] Preferably, obtaining the silhouette area deviation and the direction of the meniscus bulge based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template includes: Binarize the silhouette image of the meniscus to generate a silhouette mask; The silhouette mask and the meniscus contour template are aligned pixel-level over a unified coordinate system. Calculate the area difference between the silhouette mask and the meniscus contour template, and use it as the silhouette area deviation; The silhouette mask is divided into a front half and a back half along the conductor axis. The effective pixel centroid ordinates of the two half are calculated respectively. The direction of the liquid surface bulge is determined based on the relative height of the centroid ordinates.

[0010] Preferably, the generation of adjustment commands based on silhouette area deviation and liquid surface bulge direction features includes: The adjustment intensity level is determined based on the absolute value of the silhouette area deviation and the preset area threshold range; Based on the direction of the liquid surface bulge and the adjustment intensity level, the corresponding travel speed adjustment ratio and pumping rate adjustment ratio are retrieved from the pre-stored adjustment parameter table; Obtain the conductor position coordinates of the target device in the last three frames and calculate the conductor slope change rate; The damping coefficient is obtained based on the rate of change of the conductor slope, and the travel speed adjustment ratio and the pumping rate adjustment ratio are smoothed based on the damping coefficient to generate smoothed speed control signal and pumping control signal. The speed control signal is sent to the drive motor, and the pumping control signal is sent to the paint pump controller.

[0011] Preferably, after initializing the paint pumping rate and the target equipment travel speed, the method further includes: Pre-scanning imaging is performed on the area to be coated on the surface of the conductor; Compare the pre-scanned image with the standard texture features of clean wires; When an abnormal surface adhesion is detected and its physical location is within a preset forward distance of the current coating path, a surface preprocessing instruction or a coating delay instruction is generated, and after execution, the process of acquiring the silhouette image of the meniscus begins.

[0012] Preferably, the area threshold of the corrected meniscus profile template includes: The thickness of the wet film in the coated section was continuously sampled to obtain the measured thickness sequence; The measured thickness sequence is compared with the target thickness, and the average thickness deviation is calculated. When the average thickness deviation is greater than the positive tolerance, the upper limit of the area threshold is lowered; when the average thickness deviation is less than the negative tolerance, the lower limit of the area threshold is raised. Among them, the upper limit and lower limit of the area threshold are maintained independently, and the adjustment amount is based on the coating spreading geometry model to convert the thickness deviation into an equivalent area correction amount.

[0013] Preferably, the calculation of the average thickness deviation includes: The most recent consecutive valid sampling points are extracted from the measured thickness sequence as the evaluation window; Remove outliers in the evaluation window that deviate from the window mean by more than three standard deviations; The arithmetic mean of the remaining sampling points is calculated and used as the mean thickness deviation.

[0014] Preferably, the conversion of thickness deviation into equivalent area correction includes: Based on the mapping relationship between wet film thickness and meniscus projection area established by calibration experiments, the average thickness deviation is substituted into the mapping relationship to obtain the equivalent area correction amount. A single maximum correction step size limit is applied to the equivalent area correction amount, wherein the maximum correction step size does not exceed eight percent of the original area threshold.

[0015] Preferably, after completing the full-gap insulation coating operation, the following is also included: Based on the measured thickness sequence and adjustment command log of the complete range, construct a location-related process execution record and identify systematic thickness deviation sections. Based on the direction and magnitude of the thickness deviation in the systematic thickness deviation section, the initial area threshold compensation parameters are generated. Write the initial area threshold compensation parameters into the configuration file for the next job.

[0016] Secondly, this application also provides an overhead line insulation retrofit coating system, comprising: The vision initialization unit, in response to the starting signal of the target equipment traveling on the long-span overhead line, acquires the silhouette image of the meniscus and loads the meniscus contour template; The rheological discrimination unit obtains the silhouette area deviation and the direction of the surface bulge based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template. The feedforward control unit generates corresponding adjustment commands for the coating pumping rate and travel speed based on the silhouette area deviation and the liquid surface bulge direction characteristics, and outputs the adjustment commands to the actuator. The reference calibration unit corrects the area threshold of the meniscus profile template, updates the adjustment command based on the corrected area threshold, and obtains the insulation coating result.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: In response to the start signal of the target equipment traveling on a long-span overhead line, a silhouette image of the meniscus is acquired and a meniscus contour template is loaded. Based on the geometric comparison between the meniscus silhouette image and the meniscus outline template, the silhouette area deviation and the direction of the meniscus bulge are obtained; Based on the silhouette area deviation and the liquid surface bulge direction characteristics, corresponding coating pumping rate and travel speed adjustment commands are generated, and the adjustment commands are output to the actuator. The area threshold of the meniscus profile template is corrected, and the adjustment command is updated based on the corrected area threshold to obtain the insulation coating result.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps: In response to the start signal of the target equipment traveling on a long-span overhead line, a silhouette image of the meniscus is acquired and a meniscus contour template is loaded. Based on the geometric comparison between the meniscus silhouette image and the meniscus outline template, the silhouette area deviation and the direction of the meniscus bulge are obtained; Based on the silhouette area deviation and the liquid surface bulge direction characteristics, corresponding coating pumping rate and travel speed adjustment commands are generated, and the adjustment commands are output to the actuator. The area threshold of the meniscus profile template is corrected, and the adjustment command is updated based on the corrected area threshold to obtain the insulation coating result.

[0019] Implementing this application will have the following beneficial effects: This application provides a coating method and system for insulation retrofitting of overhead lines, including a visual initialization unit that, in response to the starting signal of the target equipment traveling on a long-span overhead line, acquires a silhouette image of the meniscus and loads a meniscus contour template. The rheological discrimination unit obtains the silhouette area deviation and the direction of the surface bulge based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template. The feedforward control unit generates corresponding adjustment commands for the coating pumping rate and travel speed based on the silhouette area deviation and the liquid surface bulge direction characteristics, and outputs the adjustment commands to the actuator. The reference calibration unit corrects the area threshold of the meniscus profile template, updates the adjustment command based on the corrected area threshold, and obtains the insulation coating result. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is an overall flowchart of a coating method for insulation retrofitting of overhead lines involved in this application; Figure 2 This is a schematic diagram of the overall structure of an overhead line insulation retrofit coating system involved in this application; Figure 3 This is a computer equipment diagram of a coating method for insulation modification of overhead lines, which is the subject of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0023] In one exemplary embodiment, such as Figure 1 As shown, a coating method for insulation retrofitting of overhead lines is provided, including: S100: In response to the start signal of the target equipment traveling on a long-span overhead line, acquire a silhouette image of the meniscus and load a meniscus contour template.

[0024] It should be noted that in the context of overhead line insulation retrofitting, the target equipment refers to a coating robot that autonomously moves along the conductor. Long-span overhead lines specifically refer to transmission line sections crossing rivers, valleys, or highways with spans exceeding 200 meters, where the conductor exhibits a significant catenary shape and continuous, drastic gradient changes. The movement start signal is not an external command, but rather the initial action of the coating robot transitioning from a stationary state to moving along the conductor and beginning material application. This action can be triggered by the robot's internal motion controller or remotely issued by the operator. Traditional insulation coating methods typically use fixed-viscosity coatings and constant pumping rates, relying on preset process parameters to handle all line conditions. They cannot dynamically respond to coating sagging or flow interruptions caused by gravity, resulting in severely uneven coating thickness across long spans. This application uses the travel start signal as the trigger condition to simultaneously activate the meniscus visual monitoring unit and load the meniscus contour template at the beginning of the coating operation. This allows the system to adjust the travel speed and pumping amount in real time based on the actual liquid surface deformation characteristics within a millisecond window before the coating has cured, thereby suppressing the impact of gravity disturbance on the coating uniformity at the source.

[0025] In some embodiments, S100 includes S110, S120, and S130, as follows: S110: Activate the macro industrial camera and coaxial ring LED backlight located on the side of the coating mold outlet, so that the illumination optical axis is parallel to the camera's line of sight and aligned with the material outlet area of ​​the wire surface.

[0026] It should be noted that traditional coating methods rely on preset process parameters and lack real-time visual monitoring devices, making it impossible to perceive the initial spreading pattern of the coating on the conductor surface. This application integrates a macro industrial camera and a coaxial ring LED backlight on the side of the discharge port, acquiring a clear silhouette outline of the meniscus through high-contrast backlight imaging. The image directly reflects the instantaneous rheological state of the coating under gravity, providing a physical basis for subsequent feedforward adjustment, and the visual feedback does not rely on any slope sensor or fluid dynamics model.

[0027] S120: Retrieve the meniscus profile template that matches the current paint type from the local storage unit and load it into the comparison buffer of the image processing unit.

[0028] It should be noted that the meniscus profile template is a pre-calibrated two-dimensional silhouette shape based on the stable liquid surface morphology formed when a specific insulating coating is applied in an ideal horizontal section. In scenarios with large gaps, it serves as a benchmark for determining whether the coating is affected by gravity. Different templates correspond to different viscosities or solid contents of the coating; for example, silicone rubber coatings correspond to trapezoidal templates, and polyurethane coatings correspond to rectangular templates. These templates are calibrated and stored in the local storage unit before equipment deployment. They are automatically loaded according to the coating type at the start of the coating operation, avoiding misjudgments caused by universal templates.

[0029] S130: Initialize the paint pumping rate and coating robot travel speed according to preset process parameters, so that the initial output amount and travel rhythm are adapted to the target coating thickness.

[0030] It should be noted that the preset process parameters are not fixed values, but rather a baseline combination based on the conductor diameter, ambient temperature, and batch characteristics of the coating. For example, for 25 mm diameter steel-cored aluminum stranded wire, when using a certain batch of silicone rubber coating at 25°C, the initial pumping rate is 14 ml per minute and the travel speed is 0.72 m / s, so that the coating thickness before adjustment is close to the target value of 1.0 mm, leaving a reasonable adjustment range for subsequent dynamic adjustments.

[0031] Ideally, S110 establishes a direct perception capability of the initial deformation of the coating through macro vision acquisition, S120 provides an accurate benchmark for image comparison by loading a matching meniscus contour template, and S130 lays a stable starting point for closed-loop regulation by reasonably initializing pumping and travel parameters. Together, these three constitute the coating start-up foundation for gravity flow scenarios of large-spacing catenary lines, enabling the system to be prepared to respond to drastic slope changes at the beginning of the operation.

[0032] In some embodiments, step S110 is specifically implemented by steps S111 to S113: Step S111: Acquire a backlit silhouette image of the meniscus on the surface of the conductor using a macro industrial camera located on the side of the discharge port of the coating mold. Perform multi-scale gradient enhancement and dynamic background suppression processing on the silhouette image to extract the high-contrast edge contour of the meniscus.

[0033] The macro industrial camera is installed 5 to 10 millimeters downstream of the coating mold outlet, with its optical axis pointing perpendicularly to the axial section of the conductor. A coaxial ring-shaped LED backlight is arranged around the camera lens, with its luminous surface parallel to the surface of the conductor. The installation parameters of the camera and backlight include working distance, depth of field range, illumination angle, and synchronous trigger delay. These installation parameters are pre-stored in the control unit of the coating robot.

[0034] Understandably, after detecting the start signal, the coating robot controls a macro industrial camera to synchronously acquire silhouette images of the meniscus with microsecond-level exposure times and backlight strobe, thus freezing the dynamic blur caused by the coating flow. The original silhouette image is first enhanced with multi-scale gradients: Sobel operators at different scales are used to calculate horizontal and vertical gradients, and multi-scale responses are fused to strengthen weak edges. Then, dynamic background suppression is performed: based on the pixel differences of three consecutive frames, background noise caused by wire vibration or illumination drift is removed, retaining only stable edge regions generated by coating deformation.

[0035] For example, in a 280-meter span across a valley, a macro industrial camera with a telecentric lens having a working distance of 8 mm and a depth of field of ±1.5 mm, a backlight center wavelength of 850 nm, and a strobe pulse width of 50 microseconds was used; the acquired silhouette image had a resolution of 1280×960 pixels. The image was processed sequentially as follows: gradient magnitudes were calculated using both 3×3 and 5×5 Sobel kernels, and weighted fusion was performed to obtain an enhanced gradient map; then, the absolute difference between the current frame and the previous two frames was calculated, and a threshold of 0.05 (normalized grayscale) was set to filter out dynamic noise, ultimately extracting a clear outline of the meniscus edge, with a signal-to-noise ratio improvement of approximately 4.2 times.

[0036] Step S112: Based on the pre-stored wire diameter parameters and the current coating refractive index, correct the optical distortion of the meniscus edge profile and map the corrected profile to the standard geometric coordinate system.

[0037] The conductor diameter parameter is either scanned and entered before operation or retrieved from the line database, and the coating refractive index is obtained from the material property table based on the grade of the currently filled insulating coating. Both parameters are pre-stored in the control unit.

[0038] Understandably, due to the refractive effect at the interface between the coating and air, and the edge perspective distortion in macro imaging, the directly extracted contours will have systematic deviations in geometric dimensions. This step, based on Snell's law of refraction and the camera pinhole model, constructs a mapping relationship between the meniscus and the image plane: first, the theoretical contact point position is determined according to the diameter of the conductor; then, the light deflection angle is calculated by combining the refractive index of the coating, and the true physical position of the edge pixels is corrected in reverse; subsequently, the corrected contours are uniformly projected onto a standard geometric coordinate system with the conductor axis as the origin and the tangential direction of the conductor as the X-axis, eliminating coordinate drift caused by slight changes in the robot's posture.

[0039] For example, when using a silicone rubber coating with a refractive index of 1.48 on a JL / G1A-400 / 35 steel-cored aluminum stranded wire with a diameter of 24.26 mm, the system automatically calls the corresponding parameters to perform refractive correction on the original edge profile, reducing the measurement error of the bottom width of the meniscus from ±0.18 mm to ±0.03 mm.

[0040] Step S113: The corrected meniscus contour is cropped into a comparison window of fixed width, and a binary silhouette mask is generated as input data for subsequent geometric overlap comparison with the meniscus contour template.

[0041] Understandably, to avoid interference from the overall shape of the conductor or background stray light during comparison, the system uses the center of the conductor as a reference and cuts out areas on both sides with a thickness of 1.5 times the target coating thickness as comparison windows. Then, an adaptive thresholding method is used to binarize the image within the window, generating a silhouette mask that only contains the effective liquid surface area. A pixel value of 1 in the mask indicates the presence of coating, and 0 indicates the absence of coating.

[0042] For example, if the target coating thickness is 1.0 mm, the comparison window width is set to 3.0 mm, which corresponds to approximately 180 pixels in a 1280×960 image. The Otsu algorithm is used to automatically determine the binarization threshold, and the generated silhouette mask clearly separates the liquid surface body from the splash burrs, effectively improving the extraction accuracy of subsequent area and protrusion direction features.

[0043] Preferably, step S111 uses multi-scale gradient enhancement and dynamic background suppression to extract the edge of the curved liquid surface with a high signal-to-noise ratio even under strong sunlight or conductor vibration. Step S112 introduces the conductor diameter and the refractive index of the coating for optical distortion correction, making the contour geometry physically comparable. Step S113 provides structured input for subsequent rapid geometric comparison through standardized cropping and binarized mask generation. The three steps work together to solve the technical problem that the visual perception of the curved liquid surface is easily interfered with, distorted, and blurred in the scenario of large-spacing catenary. This makes the invention concept of "using liquid surface deformation as a natural sensor" reliably implemented, which is significantly different from the traditional visual detection methods that only rely on fixed thresholds or uncorrected images.

[0044] In some embodiments, after initializing the paint pumping rate and robot travel speed, this application further includes: The area to be coated on the surface of the conductor is pre-scanned and imaged. The pre-scanned image is compared with the standard texture features of the clean conductor. When it is determined that there is an abnormal surface adhesion in the area to be coated, the physical location corresponding to the abnormal surface adhesion is determined according to the center pixel coordinates of the abnormal surface adhesion area and the installation parameters of the macro industrial camera. When the physical location corresponding to the surface adhesion anomaly is within the preset forward distance of the current coating path, a surface preprocessing instruction or a coating delay instruction is generated, and the process of collecting the meniscus silhouette is entered after executing the surface preprocessing instruction or the coating delay instruction.

[0045] It should be noted that in long-span overhead power lines crossing rivers or forests, the conductors are exposed to the outdoor environment for extended periods. Bird droppings, insect remains, dust clumps, or localized oxidation spots may adhere to the conductor surface. These deposits significantly reduce the adhesion of the insulating coating, leading to premature peeling. Traditional coating methods involve direct application and cannot identify such potential problems. Before the formal coating process begins, this application uses an activated macro industrial camera to perform a rapid pre-scan of the conductor area 50 to 100 millimeters ahead, acquiring a high-resolution surface texture image. The pre-scanned image is then compared with pre-stored standard texture features of clean conductors for structural similarity. If the structural similarity comparison result is below a preset threshold, an abnormal surface adhesion is identified.

[0046] When a surface adhesion anomaly is confirmed, the boundary pixel set of the surface adhesion anomaly is extracted. The average of the horizontal pixel positions of all pixels in the boundary pixel set is used to obtain the horizontal center pixel position of the surface adhesion anomaly. The average of the vertical pixel positions of all pixels in the boundary pixel set is used to obtain the vertical center pixel position of the surface adhesion anomaly. The horizontal and vertical center pixel positions of the surface adhesion anomaly are combined to obtain the regional center pixel coordinates of the surface adhesion anomaly. Based on the working distance, field of view, and installation tilt angle of the macro industrial camera, the regional center pixel coordinates of the surface adhesion anomaly are converted into the physical position along the conductor axis: First, the horizontal pixel offset of the regional center pixel coordinates of the surface adhesion anomaly relative to the image center pixel coordinates is calculated. Then, based on the field of view of the macro industrial camera and the total number of horizontal pixels in the image, the angular resolution corresponding to each pixel is calculated. The horizontal pixel offset is multiplied by the angular resolution to obtain the horizontal deflection angle. Combining the installation orientation angle of the macro industrial camera and the current position of the coating robot, the actual distance from the surface adhesion anomaly to the front end of the coating head is calculated.

[0047] The preset distance is 30 to 80 mm. When the actual distance between the surface adhesion anomaly and the front end of the coating head is less than or equal to the preset distance, the control unit generates a surface pretreatment command or a coating delay command. The surface pretreatment command is used to activate the miniature rotating scraper installed at the front end of the coating head, and the coating delay command is used to pause the coating pumping for 200 milliseconds. After executing the surface pretreatment command or the coating delay command, the surface adhesion anomaly of the conductor is cleared or avoided, and the process enters the meniscus silhouette acquisition process.

[0048] Understandably, when the actual distance from the surface adhesion anomaly to the front end of the coating head is greater than the preset distance, the surface adhesion anomaly does not affect the current coating cycle, and the process directly enters the meniscus silhouette acquisition process. By introducing a conductor surface state sensing and response mechanism before coating begins, this application avoids direct film formation on contaminated surfaces, ensuring the long-term adhesion reliability of the insulation layer across all gaps. This mechanism utilizes the same macro vision system to complete both surface pre-inspection and real-time monitoring tasks without adding extra hardware, and only triggers intervention when an anomaly is imminent, balancing reliability and operational efficiency.

[0049] S200: Based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template, obtain the silhouette area deviation and the direction of the meniscus bulge.

[0050] It should be noted that in the scenario of overhead line insulation retrofitting, the meniscus silhouette image refers to the two-dimensional projection image obtained by backlighting imaging of the free liquid surface contour formed on the conductor surface at the moment the coating leaves the coating mold. The meniscus contour template refers to the standard geometric shape calibrated under ideal horizontal coating conditions, representing the liquid surface morphology corresponding to the target coating thickness. Traditional coating control methods rely on preset pump speed and travel speed combinations, or indirectly calculate the influence of gravity through tilt sensors, and cannot directly perceive the actual spreading state of the coating on the conductor surface. Especially in sections of catenary with continuous and drastic slope changes over long spans, indirect calculations are easily affected by installation errors and dynamic vibration interference, leading to adjustment lag or misjudgment. In this application, during the coating operation, the real-time acquired meniscus silhouette image is geometrically superimposed and compared with the pre-stored meniscus contour template, and two features are directly extracted: silhouette area deviation and liquid surface bulge direction. The silhouette area deviation reflects the degree of deviation of the current coating volume from the target value, and the liquid surface bulge direction indicates whether the coating flow trend is upward backflow and accumulation or downward stretching and loss. The comparison process does not rely on any fluid dynamics model or slope calculation, but is based solely on the visual representation of physical deformation. It can output clear characteristic quantities that can be used for feedforward adjustment within milliseconds, providing a direct basis for the subsequent generation of precise travel speed and pumping rate adjustment commands, thereby effectively suppressing the coating unevenness problem caused by gravity.

[0051] In some embodiments, step S200 includes steps S210 to S240, as follows: Step S210: Perform binarization processing on the silhouette image of the curved liquid surface acquired by the macro industrial camera to generate a silhouette mask that only contains the paint area.

[0052] Among them, the silhouette mask is a binary image formed by marking the pixels where the paint exists in the silhouette image of the meniscus as valid pixels and the remaining pixels as background pixels.

[0053] Binarization of the silhouette image of the meniscus captured by the macro industrial camera refers to converting the original grayscale image into a black-and-white binary image containing only the effective area and the background area, where the effective area corresponds to the projection of the meniscus formed by the coating on the surface of the conductor.

[0054] The binarization process includes: setting a sliding local calculation window in the center region of the conductor, with the window height dynamically determined according to the target coating thickness; calculating the local mean and standard deviation of pixel grayscale values ​​in each window; subtracting 1.5 times the local standard deviation from the local mean as the segmentation threshold; marking pixels with grayscale values ​​below the segmentation threshold in the window as valid pixels, and marking the remaining pixels as background pixels.

[0055] This process avoids missegmentation under conditions of strong sunlight, conductor reflection, or ambient light fluctuations, generating a continuous and unbroken silhouette mask that accurately reflects the true geometric contour of the meniscus.

[0056] Understandably, when a macro industrial camera captures a silhouette image of a meniscus under backlighting, the coated area appears as a high-contrast dark region. After converting the original image to grayscale, the control system uses an adaptive thresholding algorithm to calculate the locally optimal segmentation threshold. Pixels below the threshold are designated as valid pixels, while those above are designated as background pixels, thus generating a silhouette mask. The adaptive thresholding algorithm dynamically adjusts the segmentation points based on the local brightness distribution of the image, effectively suppressing global threshold failures caused by ambient light fluctuations or conductor reflections. Taking silicone rubber coating as an example, after converting the original silhouette image to grayscale, the system calculates the local mean and standard deviation by sliding along the center region of the conductor, setting the threshold to the local mean minus 1.5 times the standard deviation, ultimately generating a continuous and unbroken silhouette mask.

[0057] Step S220: Align the silhouette mask and the meniscus contour template with pixel-level overlap in a unified coordinate system.

[0058] The unified coordinate system is a two-dimensional geometric coordinate system with the center of the conductor axis as the origin and the tangent of the conductor as the horizontal axis. The meniscus profile template has a fixed position and size in this coordinate system.

[0059] Understandably, the meniscus contour template loaded in step S120 is stored in the image processing unit in the form of a binary mask. First, the geometric center of the silhouette mask is extracted and aligned with the geometric center of the meniscus contour template. Then, the scale difference between the two in the vertical direction is corrected so that the pixel width corresponding to the conductor diameter is consistent. The alignment process is achieved through sub-pixel interpolation to ensure that the overlap accuracy is better than 0.5 pixels. For example, when the conductor diameter corresponds to 180 pixels, both the silhouette mask and the template are scaled to a normalized width based on 180 pixels, and then the overlap is completed by center alignment.

[0060] Step S230: Calculate the area difference between the silhouette mask and the meniscus contour template as the silhouette area deviation.

[0061] Among them, the silhouette area deviation is the algebraic difference between the total number of effective pixels of the silhouette mask and the total number of effective pixels of the meniscus contour template.

[0062] Understandably, the number of effective pixels in the silhouette mask and the meniscus contour template are counted separately, and the number of pixels in the silhouette mask is subtracted from the number of pixels in the template to obtain the silhouette area deviation. A positive value indicates that the current coating volume is greater than the target value, and a negative value indicates that the current coating volume is less than the target value. The deviation directly reflects the degree of influence of gravity disturbance on the coating distribution and does not require conversion through a fluid model. For example, if the standard template has 36,000 effective pixels and the current silhouette mask has 42,000 effective pixels, then the silhouette area deviation is positive 6,000, corresponding to coating accumulation on the uphill section.

[0063] Step S240: Perform centroid analysis on the left and right halves of the silhouette mask along the conductor axis, and determine the direction of the liquid surface bulge based on the relative positional relationship of the centroids of the left and right halves.

[0064] The left and right halves are divided into the front and back halves by the center line of the conductor and along the direction of the conductor's travel.

[0065] Understandably, the silhouette mask is vertically divided into a forward half and a backward half along the center line of the guide wire. The effective pixel centroid ordinates of the two half are calculated separately. If the centroid ordinate of the backward half is higher than that of the forward half, the liquid surface is determined to be convex backward, corresponding to gravity accumulation uphill. If the centroid ordinate of the forward half is higher than that of the backward half, the liquid surface is determined to be stretched and constricted, corresponding to gravity loss downhill. The centroid calculation uses a weighted average formula, which sums the ordinates of each effective pixel and divides by the total number of effective pixels. For example, in a certain uphill section, the centroid of the backward half is located at ordinate 85 pixels, and the centroid of the forward half is located at ordinate 72 pixels, with a difference of 13 pixels. Based on this, the direction of the liquid surface convexity is determined to be backward.

[0066] Preferably, step S210 generates a high-fidelity silhouette mask through binarization, step S220 achieves precise alignment between the template and the measured image through geometric normalization, step S230 directly quantifies the coating volume deviation through pixel counting, and step S240 captures the asymmetry of the liquid surface morphology through half-area centroid analysis. The four steps work together to transform the original visual data into two distinct features that can be used for feedforward control: silhouette area deviation and liquid surface bulge direction. This avoids reliance on slope sensors or fluid simulation models and provides a direct, fast, and reliable perception basis for real-time coating control in large-spacing catenary scenarios.

[0067] In some embodiments, step S210 is specifically implemented by steps S211 to S213: Step S211: Perform grayscale conversion on the meniscus silhouette image to generate a single-channel grayscale image.

[0068] It should be noted that the silhouette image of the meniscus is a color or raw sensor data image captured by a macro industrial camera. Grayscale conversion is used to reduce the complexity of subsequent processing and to highlight the brightness contrast.

[0069] Understandably, the red, green, and blue channel pixel values ​​of the meniscus silhouette image are read, and the grayscale value of each pixel is calculated using a weighted average method. The grayscale value is equal to the red channel pixel value multiplied by 0.299, plus the green channel pixel value multiplied by 0.587, plus the blue channel pixel value multiplied by 0.114. The grayscale values ​​of all pixels are combined to form a single-channel grayscale image. The weighting coefficients conform to the sensitivity distribution of the human eye to different colors of light, effectively preserving the edge contrast under backlight imaging.

[0070] Step S212: Set a sliding local calculation window in the center region of the conductor, and calculate the pixel mean and standard deviation in each local calculation window.

[0071] It should be noted that the local calculation window is a rectangular area that slides along the conductor's axis, with the conductor's geometric center as the reference. The window size is dynamically set according to the conductor's diameter.

[0072] Understandably, the first step is to determine the centerline position of the conductor in the image. Using the centerline as a reference, the window height is extended upwards and downwards by 1.5 times the target coating thickness. A sliding step is set every 10 pixels along the conductor direction. Within each local calculation window, all pixels are traversed, and the arithmetic mean of the pixel grayscale values ​​is calculated to obtain the local mean. Then, the square root of the average of the squared differences between each pixel grayscale value and the local mean is calculated to obtain the local standard deviation. For example, when the target coating thickness corresponds to 30 pixels, the window height is set to 90 pixels, the sliding step is 10 pixels, and approximately 120 local statistical units are generated.

[0073] Step S213: Dynamically calculate the segmentation threshold based on the local mean and local standard deviation, and mark the pixels in the grayscale image below the segmentation threshold as valid pixels to generate a silhouette mask.

[0074] It should be noted that the segmentation threshold is used to distinguish between the paint area and the background area, and dynamic calculation can adapt to uneven lighting or reflective interference.

[0075] Understandably, for each local calculation window, the segmentation threshold is equal to the local mean minus 1.5 times the local standard deviation. The segmentation threshold is applied to all pixels within the corresponding window. If the gray value of a pixel is less than the segmentation threshold, the pixel is marked as a valid pixel in the output image with a pixel value of 1; otherwise, it is marked as a background pixel with a pixel value of 0. After all windows are processed, they are stitched together to generate a complete binary silhouette mask. This method can still effectively suppress misjudgment of the highlight area of ​​the conductor in strong sunlight, ensuring that the silhouette mask is continuous and unbroken.

[0076] It should be noted that step S211 preserves key contrast information through weighted grayscale conversion, step S212 achieves spatial adaptive statistics through sliding local windows, and step S213 generates a high-fidelity silhouette mask through dynamic thresholding. The three work together to solve the image segmentation problem caused by drastic changes in lighting and interference from conductor reflections in large-span field operations, providing reliable input for subsequent area and morphology analysis.

[0077] In some embodiments, step S240 is specifically implemented by steps S241 to S243: Step S241: Divide the silhouette mask into a front half-region and a back half-region, using the geometric center line of the conductor as the boundary.

[0078] It should be noted that the forward half-zone refers to the half-area in front of the coating robot in the direction of travel, and the backward half-zone refers to the half-area in the direction of travel. The division is based on the perpendicular bisector of the center line of the guide wire.

[0079] Understandably, the first step is to extract the bounding rectangle of the silhouette mask and determine its horizontal centerline position. Using the centerline as the Y-axis, pixels in the silhouette mask with X coordinates less than the centerline's X coordinate are assigned to the front half of the row, and pixels with X coordinates greater than the centerline's X coordinate are assigned to the back half of the row. Pixels with X coordinates equal to the centerline's X coordinate are assigned according to the nearest principle. The division method strictly corresponds to the direction of gravity flow to ensure that the determination of the protrusion direction is consistent with physical rheology.

[0080] Step S242: Calculate the effective pixel centroid ordinates of the front and rear half regions respectively.

[0081] It should be noted that the centroid ordinate reflects the position of the center of gravity of the liquid surface in the vertical direction, and is a key indicator for judging accumulation or stretching.

[0082] Understandably, for the first half of the image, all valid pixels are traversed, the ordinate value of each pixel is accumulated and divided by the total number of valid pixels in the first half of the image to obtain the centroid ordinate of the first half of the image; for the second half of the image, the ordinate values ​​of all valid pixels are accumulated and divided by the total number of valid pixels in the second half of the image to obtain the centroid ordinate of the second half of the image. The calculation process uses integer operations to improve real-time performance. The origin of the ordinate is located at the top of the image, and the larger the value, the lower the position.

[0083] Step S243: Compare the relative sizes of the centroid ordinates of the rear half-section and the front half-section. If the centroid ordinate of the rear half-section is higher than the centroid ordinate of the front half-section by more than a preset tolerance, the liquid surface is determined to be convex backward; otherwise, the liquid surface is determined to be stretched and constricted.

[0084] It should be noted that the preset tolerance is used to filter out minor fluctuations caused by image noise, with a typical value of 3 pixels.

[0085] Understandably, the difference between the ordinate of the centroid of the rear half-section and the ordinate of the centroid of the front half-section is calculated. If the difference is greater than the preset tolerance, it is determined that the center of gravity of the liquid surface is biased to the rear, corresponding to the backflow and accumulation of paint on the uphill section, and the liquid surface is convex backward. If the difference is less than the negative preset tolerance, it is determined that the center of gravity of the liquid surface is biased to the front, corresponding to the stretching and loss of paint on the downhill section, and the liquid surface is stretched and constricted. If the absolute value of the difference does not exceed the preset tolerance, it is determined that the liquid surface shape is symmetrical and is in an approximately horizontal state. The judgment logic directly maps to physical phenomena and does not rely on slope or velocity parameters.

[0086] It should be noted that step S241 achieves spatial partitioning through direction of travel perception, step S242 quantifies liquid surface asymmetry through centroid calculation, and step S243 completes the rheological state determination through tolerance-based comparison. The three constitute a lightweight, fast, and physically meaningful liquid surface morphology analysis mechanism, which is significantly different from traditional solutions that rely on complex models or external sensors.

[0087] S300: Based on the silhouette area deviation and the liquid surface bulge direction characteristics, generate corresponding paint pumping rate and travel speed adjustment commands, and output the adjustment commands to the actuator.

[0088] It should be noted that in the scenario of overhead line insulation retrofitting, the silhouette area deviation reflects the degree of deviation of the current coating volume from the target value, and the liquid surface bulge direction characteristic indicates whether the coating flow trend under gravity is upward backflow and accumulation or downward stretching and loss. Traditional coating control methods usually use a fixed pumping rate and constant travel speed, or only perform open-loop adjustment based on a preset span slope table, which cannot respond in real time to continuous slope changes caused by the catenary shape of the conductor, resulting in drastic fluctuations in coating thickness across large spans. After obtaining the silhouette area deviation and liquid surface bulge direction characteristics, this application directly maps and generates corresponding coating pumping rate and travel speed adjustment commands: when the silhouette area deviation is positive and the liquid surface is backward bulging, it indicates coating accumulation on the uphill section, and the system generates a speed-up and pump-down command to speed up the robot's passage and reduce material output; when the silhouette area deviation is negative and the liquid surface is stretched and constricted, it indicates coating loss on the downhill section, and the system generates a speed-down and pump-up command to extend the residence time and increase material output. The adjustment command is smoothed by a predictive damping algorithm before being output to the actuator, avoiding mechanical vibration or coating ripples caused by frequent abrupt changes. The process does not rely on any tilt sensors, 3D modeling of wires, or fluid dynamics calculations; it achieves millisecond-level feedforward adjustment solely based on visual feedback from the coating's own deformation, effectively counteracting the disturbance of gravity on coating uniformity and ensuring that the insulation layer thickness across all spans meets engineering requirements.

[0089] In some embodiments, step S300 generates an adjustment command based on the silhouette area deviation and the liquid surface bulge direction characteristics, including steps S310 to S350: Step S310: Determine the adjustment intensity level based on the absolute value of the silhouette area deviation and the preset area threshold range.

[0090] The area threshold range is determined based on the rheological properties of the coating and the robot response delay. The first-level range corresponds to slight disturbances, while the third-level range corresponds to severe accumulation or flow interruption risks.

[0091] Understandably, in the application of silicone rubber insulating coatings, due to the high viscosity and poor fluidity of the material, even a small deviation can lead to uncontrolled thickness. Therefore, the upper limit of the first-level range is set at 10%. If the absolute value of the silhouette area deviation is 17%, it falls into the second-level range, indicating that the coating has accumulated significantly but has not yet dripped, requiring moderate intervention.

[0092] It should be noted that the regulation intensity level is a control response level classified based on the absolute value of the silhouette area deviation. The regulation intensity level is divided into three levels: Level 1 corresponds to slight disturbances, Level 2 corresponds to significant accumulation or loss, and Level 3 corresponds to severe accumulation or risk of flow interruption. The rules for determining the regulation intensity level are as follows: when the absolute value of the silhouette area deviation falls within the first area threshold range, the regulation intensity level is Level 1; when the absolute value of the silhouette area deviation falls within the second area threshold range, the regulation intensity level is Level 2; and when the absolute value of the silhouette area deviation falls within the third area threshold range, the regulation intensity level is Level 3. The upper limit of the first area threshold range is less than the lower limit of the second area threshold range, and the upper limit of the second area threshold range is less than the lower limit of the third area threshold range.

[0093] Step S320: Based on the characteristics of the liquid surface bulge direction and the adjustment intensity level, retrieve the corresponding travel speed adjustment ratio and pumping rate adjustment ratio from the pre-stored adjustment parameter table.

[0094] The adjustment parameter table was established based on on-site process tests, and each set of parameters was verified for thickness uniformity across the entire range.

[0095] Understandably, for the gravity backflow scenario on the uphill section, if it is determined to be a backward bulge and of secondary strength, referring to the table shows that the travel speed increases by 8% and the pumping rate decreases by 10%. The combination has been verified by experiments to restore the thickness of the accumulation area to the target value within 3 meters without causing coating ripples. If a fixed ratio adjustment is directly used, it is easy to cause the flow to be interrupted in the subsequent downhill section due to over-adjustment.

[0096] It should be noted that the pre-stored adjustment parameter table is a mapping table established through on-site process tests before equipment deployment. It uses the adjustment intensity level and the direction of liquid surface bulge as index items and stores the combination of travel speed adjustment ratio and pumping rate adjustment ratio verified by the thickness uniformity of the entire range. The pre-stored adjustment parameter table is used to quickly retrieve the adjustment parameters that match the current rheological state during the coating operation, so as to avoid coating ripples or flow interruption caused by excessive adjustment.

[0097] Step S330: Obtain the current traveling speed of the coating robot and the position coordinates of the guide wire in the last three frames, and calculate the rate of change of the guide wire slope.

[0098] The calculation of the guide slope change rate includes: reading the position and elevation data along the span direction recorded by the target equipment in three consecutive control cycles; The first inclination angle is calculated based on the position and elevation data of the first two frames, and the second inclination angle is calculated based on the position and elevation data of the last two frames. Divide the difference between the inclination angle of the second segment and the inclination angle of the first segment by the span distance between the midpoints of the two segments to obtain the rate of change of the conductor slope.

[0099] Understandably, when the robot approaches the right attachment point during the crossing of the valley, three consecutive frames of elevation data show an accelerating upward trend. The calculated rate of change of the guide slope is 0.042 radians per meter, indicating that it has entered a steep uphill area. If the robot is adjusted according to the normal ratio at this time, the sudden acceleration will cause mechanical vibration, resulting in ripples on the coating surface.

[0100] Step S340: Determine the damping coefficient based on the conductor slope change rate, and smooth the travel speed adjustment ratio and pumping rate adjustment ratio based on the damping coefficient to generate smoothed speed control signal and pumping control signal.

[0101] The damping coefficient table is calibrated based on the robot drive system inertia and the coating response time to ensure smooth adjustment.

[0102] It should be noted that the smoothing process involves multiplying the travel speed adjustment ratio by the damping coefficient to obtain the smoothed travel speed adjustment ratio, and multiplying the pumping rate adjustment ratio by the damping coefficient to obtain the smoothed pumping rate adjustment ratio. The damping coefficient is obtained from a preset table based on the conductor slope change rate. The conductor slope change rate reflects the severity of the catenary geometry, and the damping coefficient ranges from 0.85 to 1. The larger the conductor slope change rate, the smaller the damping coefficient. The smoothing process suppresses abrupt changes in the amplitude of the adjustment command, avoids mechanical vibration caused by sudden changes in the acceleration of the drive motor, prevents the coating from forming ripples or flow interruptions on the conductor surface, and ensures coating continuity and surface smoothness.

[0103] Understandably, when the rate of change of the conductor slope exceeds 0.03 radians per meter, the damping coefficient drops to 0.85, causing the original 8% speed increase command to be attenuated to 6.8%. This attenuation prevents the drive motor from suddenly accelerating in the high curvature region. At the same time, because the coating has thixotropic properties, a small speed increase is still sufficient to suppress accumulation, thus balancing stability and effectiveness.

[0104] Step S350: Send the speed control signal to the robot drive motor and the pumping control signal to the paint pump controller to simultaneously perform travel and discharge adjustment.

[0105] Understandably, the control system uses a hardware synchronization triggering mechanism to ensure that the drive motor and the paint pump respond to commands within the same millisecond cycle. If they are not synchronized, for example, if the pumping lags by 50 milliseconds, a local thin layer will form at the moment of acceleration, disrupting concentricity. This solution achieves dual-channel synchronization through a shared clock source to ensure coating continuity.

[0106] It should be noted that step S310 uses deviation grading and matching to adjust the force, avoiding large movements from small disturbances; step S320 relies on the process verification parameter table to ensure the effectiveness of the adjustment; step S330 introduces the geometric change rate of the conductor as a smoothing basis, reflecting a deep adaptation to the characteristics of the catenary; step S340 suppresses mechanical shock through physically calibrated damping coefficients; and step S350 ensures execution consistency through hardware synchronization. These five steps together solve the dilemma of "overshoot when adjusted, and loss of control when not adjusted" in traditional methods for large-gauge scenarios.

[0107] Preferably, step S300 does not rely on external sensors or fluid models, but completes closed-loop feedforward adjustment based solely on visual feedback and the robot's own motion data; step S300 automatically softens and adjusts the intensity in the high curvature area near the hanging point, and maintains a sensitive response in the stable area in the middle of the span; step S300 reduces the standard deviation of the coating thickness across the entire span from 0.25 mm in the traditional method to 0.09 mm, significantly improving insulation reliability and material utilization.

[0108] S400: Correct the area threshold of the meniscus profile template, update the adjustment command based on the corrected area threshold, and obtain the insulation coating result.

[0109] It should be noted that in the scenario of overhead line insulation retrofitting, the area threshold of the meniscus profile template is a geometric reference used to determine whether the current coating deviates from the target thickness. The threshold is calibrated before operation based on the ideal horizontal section coating state. Traditional coating control methods, once the threshold is set, are used fixedly throughout the process, which cannot adapt to systematic deviations caused by batch differences in coatings, changes in ambient temperature, or mold wear. This leads to a gradual amplification of accumulated errors in long-distance operations with large spans, ultimately resulting in an overall coating that is too thick or too thin. In this application, during the coating process, a laser thickness gauge set behind the coating head is used to perform non-contact measurement of the wet film thickness before curing. The measured thickness is compared with the target thickness. If there is a continuous deviation in the same direction, the area threshold of the meniscus profile template is adjusted in the opposite direction: when the measured thickness is generally too large, the upper limit of the area threshold is reduced, so that the system triggers the speed-up and pump-down command earlier; when the measured thickness is generally too small, the lower limit of the area threshold is increased, so that the system triggers the speed-down and pump-up command earlier. The feedback correction mechanism does not change the main control logic, but only fine-tunes the judgment boundary of visual comparison, thereby compensating for system drift without increasing complexity. It ensures that the coating thickness always converges to the target value within a range of hundreds of meters, effectively solving the steady-state error problem of traditional open-loop or single-loop control in long-distance operations.

[0110] In some embodiments, step S400, which involves correcting the area threshold of the meniscus profile template and updating the adjustment instructions, includes steps S410 to S450: Step S410: The wet film thickness of the coated section is continuously sampled using a laser thickness gauge positioned 50 to 100 mm behind the coating head to obtain a measured thickness sequence.

[0111] Among them, wet film thickness refers to the thickness of the insulation layer formed on the surface of the conductor before the coating has cured, and the measured thickness sequence is a set of thickness data recorded at time or spatial intervals.

[0112] Understandably, the laser thickness gauge emits an infrared beam at a frequency of 10 Hz per second. The beam illuminates the surface of the conductor perpendicularly, and the wet film thickness is calculated by receiving the time difference of the reflected light. The control system binds each measurement result to the robot's current position, forming a sequence of measured thicknesses ordered by the distance between positions. The sampling position is located 80 mm behind the coating head, ensuring that the coating has been initially spread but has not yet begun to drip significantly, so that the measured thickness is representative.

[0113] Step S420: Compare the measured thickness sequence with the target thickness and calculate the average thickness deviation.

[0114] The target thickness is the preset insulation layer design thickness, and the average thickness deviation reflects the systematic deviation trend under the current process conditions.

[0115] Understandably, the evaluation window is defined as the 12 most recent consecutive valid sampling points extracted from the measured thickness sequence; the thickness deviation value is obtained by calculating the difference between each point in the window and the target thickness; outliers with absolute values ​​exceeding three times the standard deviation of the window are removed; the arithmetic mean of the remaining deviation values ​​is calculated to obtain the mean thickness deviation; if the standard deviation in the evaluation window is greater than 0.05 mm, the conductor is determined to be in a vibration or transition section, and the mean thickness deviation is not calculated temporarily to avoid miscorrection.

[0116] Step S430: Adjust the upper or lower limit of the area threshold of the meniscus profile template according to the sign and magnitude of the average thickness deviation.

[0117] The upper limit of the area threshold is used to determine the uphill accumulation state, and the lower limit of the area threshold is used to determine the downhill erosion state. The two are stored and updated independently.

[0118] Understandably, when the average thickness deviation is greater than the positive tolerance of 0.1 mm, the coating is considered to be too thick overall, and the upper limit of the trigger area threshold is lowered; when the average thickness deviation is less than the negative tolerance of 0.1 mm, the coating is considered to be too thin overall, and the lower limit of the trigger area threshold is raised. The adjustment amount is determined based on the geometric model of the coating spread on the conductor surface: the model establishes the mapping relationship between wet film thickness and meniscus projection area through calibration experiments. For example, for silicone rubber coating on a 24 mm diameter conductor, for every 0.1 mm increase in thickness, the projection area increases by an average of 2500 pixels. This relationship is used to convert the average thickness deviation into an equivalent area correction amount, and a single maximum correction step size limit is applied, not exceeding 8% of the original threshold, to prevent overshoot.

[0119] Step S440: In the subsequent adjustment instruction generation process, the corrected area threshold is used as the benchmark for determining the silhouette area deviation.

[0120] The corrected area threshold includes the updated upper limit of the area threshold, the lower limit of the area threshold, or one of both.

[0121] Understandably, when the image processing unit performs geometric comparison in S200, it no longer uses the initial area threshold, but instead reads the upper and lower limits of the area threshold updated in S430. When the real-time silhouette area exceeds the corrected upper limit of the area threshold, S300 generates a speed-up and pump-down command. When the real-time silhouette area is lower than the corrected lower limit of the area threshold, S300 generates a speed-down and pump-up command. This mechanism ensures that the sensitivity of the visual feedforward loop is always matched with the current material and equipment status.

[0122] Step S450: Continue to perform coating adjustment based on the corrected area threshold until the full-spacing insulation coating operation is completed.

[0123] Understandably, S410 to S440 form a slow verification loop, repeating the process approximately every 30 meters; while S200 to S300 form a fast feedforward loop, adjusting in real time every 0.5 meters. The two loops operate in parallel, with the feedforward loop suppressing transient gravitational disturbances and the verification loop eliminating systematic drift. Ultimately, over a 320-meter span across the river, the standard deviation of the coating thickness decreased from 0.24 mm using traditional methods to 0.08 mm, with no flow interruptions or dripping defects.

[0124] In some embodiments, step S430 is specifically implemented by steps S431 to S434: Step S431: Determine the adjustment direction of the area threshold based on the sign of the average thickness deviation: when the average thickness deviation is positive, it is determined that the overall coating is too thick and the upper limit of the area threshold needs to be lowered; when the average thickness deviation is negative, it is determined that the overall coating is too thin and the lower limit of the area threshold needs to be raised.

[0125] The upper limit of the area threshold is used to trigger the speed-up and pump-down command, while the lower limit of the area threshold is used to trigger the speed-down and pump-up command. The two are maintained independently and do not affect each other.

[0126] Understandably, the current values ​​of the upper and lower limits of the area threshold are recorded separately. If the average thickness deviation is positive 0.15 mm, it indicates a long-term trend of paint accumulation. In this case, only the upper limit of the area threshold is lowered to allow the uphill section to enter the pump reduction state earlier without affecting the flow interruption protection capability of the downhill section. If the average thickness deviation is negative 0.12 mm, only the lower limit of the area threshold is raised to enhance the pump increase sensitivity of the downhill section and avoid excessive intervention in the uphill control logic.

[0127] Step S432: Based on the geometric model of the coating spread on the conductor surface, the mean thickness deviation is converted into an equivalent area correction.

[0128] The spread geometric model describes the nonlinear relationship between the wet film thickness and the projected area of ​​the meniscus. The model is established by synchronous sampling of laser thickness measurement and visual area during equipment calibration.

[0129] It is understandable that for a 24 mm diameter steel-cored aluminum stranded wire, the projected area of ​​the meniscus corresponding to a silicone rubber coating with a target thickness of 1.0 mm is 36,000 pixels; when the measured thickness increases to 1.2 mm, the projected area increases to 41,000 pixels, and the two have an approximate quadratic function relationship. Using this function, we can inversely deduce that if the average thickness deviation is positive 0.2 mm, the equivalent area correction is positive 5,000 pixels, and the upper limit of the area threshold needs to be reduced by 5,000 pixels to offset the systematic overthickness.

[0130] Step S433: Apply upper limit constraints and direction lock to the equivalent area correction amount to generate the final area threshold correction amount.

[0131] Among them, the upper limit constraint prevents oscillations caused by excessive single correction amplitude, and the direction lock ensures that only the corresponding side threshold is corrected.

[0132] Understandably, the maximum step size for a single correction is set to 8% of the original area threshold. If the calculated equivalent area correction is 6000 pixels, but the upper limit of the original area threshold is 60000 pixels, then 8% is 4800 pixels, and the final area threshold correction is limited to 4800 pixels. At the same time, if the average thickness deviation is positive, the area threshold correction only applies to the upper limit, while the lower limit remains unchanged, and vice versa.

[0133] Step S434: Algebraically superimpose the upper limit or lower limit of the area threshold with the corresponding area threshold correction amount to obtain the corrected area threshold.

[0134] It is understandable that the original upper limit of the area threshold is 60,000 pixels, and the correction amount obtained by step S433 is negative 4,800 pixels, so the corrected upper limit of the area threshold is 55,200 pixels; the original lower limit of the area threshold is 30,000 pixels, and since the thickness deviation is positive, it is not corrected and remains unchanged; the corrected area threshold is immediately written into the comparison parameter area of ​​the image processing unit for subsequent use in S300.

[0135] It should be noted that step S431 separates the upper and lower limit adjustment paths by deviation sign separation to avoid cross-interference; step S432 introduces a spreading geometric model to achieve cross-modal dimensional conversion, enabling laser thickness measurement data to accurately guide visual threshold correction; step S433 ensures correction stability through step size constraints and direction locking; and step S434 completes the parameter update closed loop. These four steps work together to solve the problems of "thickness measurement and vision disconnect," "correction overshoot," and "bidirectional interference" in traditional methods, truly implementing the feedforward-feedback collaborative mechanism.

[0136] Preferably, step S430 does not rely on a general control algorithm, but rather on a dedicated correction logic based on the physical properties of the coating and the geometry of the conductor; step S430 reduces the steady-state thickness error from 0.25 mm to 0.07 mm in 300-meter span operations; step S430 enables the same vision feedforward system to be adapted to different batches of coatings, significantly improving engineering robustness.

[0137] It should be noted that step S410 obtains representative wet film thickness data by reasonably arranging laser thickness gauges; step S420 calculates steady-state deviation through a sliding window with anomaly rejection to avoid transient interference; step S430 realizes cross-modal parameter mapping based on a physical spreading model and independently adjusts the upper and lower limits to adapt to different rheological states; step S440 ensures that the correction results take effect immediately; and step S450 achieves continuous optimization under long-distance operation. These five elements work together to construct a dual-loop collaborative control mechanism with vision-based feedforward as the main component and laser calibration as a supplement, significantly different from traditional coating methods that rely solely on single feedback or open-loop settings.

[0138] Preferably, step S400 does not introduce complex control algorithms, but achieves parameter self-calibration through physical calibration and engineering constraints; step S400 enables the same set of equipment to adapt to different operating conditions of different temperatures and different batches of coatings; step S400 eliminates steady-state errors while ensuring response speed, ensuring that the coating quality of the insulation retrofit of long-span overhead lines meets the long-term operating requirements of the power industry.

[0139] In some embodiments, after completing the full-gap insulation coating operation, this application further includes: A correlation analysis is performed on the measured thickness sequence across the entire range and the adjustment command log to identify systematic thickness deviation sections. Based on the area threshold offset trend corresponding to the systematic thickness deviation sections, initial area threshold compensation parameters are generated and written into the configuration file for the next operation.

[0140] It should be noted that in the insulation retrofitting of long-span overhead lines, multiple coating operations are often required for the same line or similar crossing sections. However, paint batches, ambient temperature, or equipment wear can cause the initial process parameters to gradually become inaccurate. Traditional methods use fixed initial parameters for each operation, making it impossible to inherit historical experience. After a single operation, this application uses the measured thickness sequence of the entire span collected by S410 and the adjustment command log recorded by S300 to form a process execution trajectory according to spatial location. If a certain slope characteristic section (such as 30 meters near the hanging point) shows continuous positive thickness deviation accompanied by frequent speed increase and pump decrease commands, it is determined that the upper limit of the initial area threshold is set too high. The system calculates the mapping relationship between the average thickness deviation of the section and the corresponding silhouette area deviation, and back-calculates the required area threshold correction amount. The correction amount is used as the initial area threshold compensation parameter and stored in the equipment configuration file. When the next operation starts, the meniscus profile template loaded by S120 automatically applies the compensation parameter to make the initial area threshold closer to the current process state.

[0141] Among them, the systematic thickness deviation section refers to a continuous section with a length exceeding 20 meters and an absolute value of the average thickness deviation greater than 10% of the target thickness; the area threshold offset trend refers to the adjustment command continuously biased in the same direction within the section (such as triggering speed increase and pump reduction more than 5 times in a row); the initial area threshold compensation parameters include the upper limit compensation amount of the area threshold and the lower limit compensation amount of the area threshold, which are calculated independently and stored separately.

[0142] Understandably, the control system traverses the entire range of data, extracting all systematic thickness deviation segments. For each segment, it calculates the average thickness deviation Δh. Based on the coating spreading geometry model, Δh is converted into an equivalent area deviation ΔA. If Δh is positive, the upper limit compensation amount for the area threshold is negative ΔA; if Δh is negative, the lower limit compensation amount for the area threshold is positive ΔA. After weighted averaging the compensation amounts for all segments, the final initial area threshold compensation parameters are obtained. These parameters are automatically superimposed onto the baseline area threshold during the S130 initialization phase of the next operation, achieving self-evolution of process parameters. Through this mechanism, the coating thickness convergence speed of the third operation is increased by 40% compared to the first operation, significantly reducing adjustment oscillations and material waste.

[0143] Preferably, the post-processing process does not change the real-time coating control logic and only runs after the operation is completed; the process extends the invention from single closed-loop control to a cross-operation process knowledge accumulation system; the process is particularly suitable for batch insulation retrofitting of multiple lines in the same corridor in power grid operation and maintenance, effectively improving construction consistency and resource efficiency.

[0144] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0145] Based on the same inventive concept, this application also provides an overhead line insulation retrofit coating system. The solution provided by the system is similar to the solution described in the above method. Therefore, the specific limitations of one or more overhead line insulation retrofit coating system embodiments provided below can be found in the limitations of the overhead line insulation retrofit coating method described above, and will not be repeated here.

[0146] In one exemplary embodiment, such as Figure 2 As shown, an overhead line insulation retrofit coating system is provided, comprising: The vision initialization unit, in response to the starting signal of the target equipment traveling on the long-span overhead line, acquires the silhouette image of the meniscus and loads the meniscus contour template; The rheological discrimination unit obtains the silhouette area deviation and the direction of the surface bulge based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template. The feedforward control unit generates corresponding adjustment commands for the coating pumping rate and travel speed based on the silhouette area deviation and the liquid surface bulge direction characteristics, and outputs the adjustment commands to the actuator. The reference calibration unit corrects the area threshold of the meniscus profile template, updates the adjustment command based on the corrected area threshold, and obtains the insulation coating result.

[0147] Each module in the above-mentioned overhead line insulation retrofit coating system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0148] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for insulating and coating overhead power lines. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0149] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0150] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0151] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0152] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0153] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0154] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0155] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0156] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A coating method for insulation retrofitting of overhead power lines, characterized in that, include: In response to the start signal of the target equipment traveling on a long-span overhead line, a silhouette image of the meniscus is acquired and a meniscus contour template is loaded. Based on the geometric comparison between the meniscus silhouette image and the meniscus outline template, the silhouette area deviation and the direction of the meniscus bulge are obtained; Based on the silhouette area deviation and the directional characteristics of the liquid surface bulge, corresponding coating pumping rate and travel speed adjustment commands are generated, and the travel speed adjustment commands are output to the actuator. The area threshold of the meniscus profile template is corrected based on the wet film thickness of the coated section, and the adjustment command is updated based on the corrected area threshold to obtain the insulation coating result.

2. The method for coating insulation retrofitting of overhead lines as described in claim 1, characterized in that: The process of acquiring the silhouette image of the meniscus and loading the meniscus contour template includes: Activate the macro industrial camera and coaxial ring LED backlight located on the side of the coating mold outlet; Retrieve the meniscus profile template that matches the current paint type from the local storage unit and load it into the comparison buffer; The coating pumping rate and the target equipment travel speed are initialized according to the preset process parameters.

3. The method for coating insulation retrofitting of overhead lines as described in claim 1, characterized in that: The process of obtaining the silhouette area deviation and the direction of the liquid surface bulge includes: Binarize the silhouette image of the meniscus to generate a silhouette mask; The silhouette mask and the meniscus contour template are aligned pixel-level over a unified coordinate system. Calculate the area difference between the silhouette mask and the meniscus contour template, and use it as the silhouette area deviation; The silhouette mask is divided into a front half and a back half along the conductor axis. The effective pixel centroid ordinates of the two half are calculated respectively. The direction of the liquid surface bulge is determined based on the numerical relationship between the centroid ordinates of the front half and the back half. When the ordinate of the centroid of the rear half-section is greater than the ordinate of the centroid of the front half-section, the liquid surface is determined to be convex backward. When the ordinate of the centroid of the rear half-section is smaller than the ordinate of the centroid of the front half-section, the liquid surface is determined to be convex in the forward direction.

4. The method for coating insulation retrofitting of overhead lines as described in claim 1, characterized in that: The generation of adjustment instructions includes: The adjustment intensity level is determined based on the absolute value of the silhouette area deviation and the preset area threshold range; Based on the direction of the liquid surface bulge and the adjustment intensity level, the corresponding travel speed adjustment ratio and pumping rate adjustment ratio are retrieved from the pre-stored adjustment parameter table; Obtain the position coordinates of the conductor in the three most recent frames of the target device, and calculate the rate of change of conductor slope; The damping coefficient is obtained based on the rate of change of the conductor slope, and the travel speed adjustment ratio and the pumping rate adjustment ratio are smoothed based on the damping coefficient to generate smoothed speed control signal and pumping control signal. The speed control signal is sent to the drive motor, and the pumping control signal is sent to the paint pump controller.

5. The method for coating insulation retrofitting of overhead lines as described in claim 4, characterized in that: After initializing the paint pumping rate and the target equipment travel speed, the following steps are also included: Pre-scanning imaging is performed on the area to be coated on the surface of the conductor; Compare the pre-scanned image with the standard texture features of clean wires; When an abnormal surface adhesion is detected and its physical location is within a preset forward distance of the current coating path, a surface preprocessing instruction or a coating delay instruction is generated, and after execution, the process of acquiring the silhouette image of the meniscus begins.

6. The method for coating insulation retrofitting of overhead lines as described in claim 4, characterized in that: The area threshold of the corrected meniscus profile template includes: The thickness of the wet film in the coated section was continuously sampled to obtain the measured thickness sequence; The measured thickness sequence is compared with the target thickness, and the average thickness deviation is calculated; When the average thickness deviation is greater than the positive tolerance, the upper limit of the area threshold is lowered; when the average thickness deviation is less than the negative tolerance, the lower limit of the area threshold is raised. Among them, the upper limit and lower limit of the area threshold are maintained independently, and the adjustment amount is based on the coating spreading geometry model to convert the thickness deviation into an equivalent area correction amount.

7. The method for coating insulation retrofitting of overhead lines as described in claim 6, characterized in that: The calculation of the average thickness deviation includes: The most recent consecutive valid sampling points are extracted from the measured thickness sequence as the evaluation window; Remove outliers in the evaluation window that deviate from the window mean by more than three standard deviations; The arithmetic mean of the remaining sampling points is calculated and used as the mean thickness deviation.

8. The method for coating insulation retrofitting of overhead lines as described in claim 6, characterized in that: The process of converting thickness deviation into equivalent area correction includes: Based on the mapping relationship between wet film thickness and meniscus projection area established by calibration experiments, the average thickness deviation is substituted into the mapping relationship to obtain the equivalent area correction amount. A single maximum correction step size limit is imposed on the equivalent area correction.

9. The method for coating insulation retrofitting of overhead lines as described in claim 7, characterized in that: After completing the full-gap insulation coating operation, the following is also included: Based on the measured thickness sequence and adjustment command log of the complete range, construct a location-related process execution record and identify systematic thickness deviation sections. Based on the direction and magnitude of the thickness deviation in the systematic thickness deviation section, the initial area threshold compensation parameters are generated. Write the initial area threshold compensation parameters into the configuration file for the next job.

10. An overhead line insulation retrofit coating system, employing the overhead line insulation retrofit coating method as described in any one of claims 1 to 9, characterized in that, include: The vision initialization unit, in response to the starting signal of the target equipment traveling on the long-span overhead line, acquires the silhouette image of the meniscus and loads the meniscus contour template; The rheological discrimination unit obtains the silhouette area deviation and the direction of the surface bulge based on the geometric comparison results between the meniscus silhouette image and the meniscus contour template. The feedforward control unit generates corresponding adjustment commands for the coating pumping rate and travel speed based on the silhouette area deviation and the liquid surface bulge direction characteristics, and outputs the adjustment commands to the actuator. The reference calibration unit corrects the area threshold of the meniscus profile template based on the wet film thickness of the coated section, and updates the adjustment command based on the corrected area threshold to obtain the insulation coating result.

Images