Insulator creeping distance measurement method and device based on multi-vision

Abstract

The application discloses a kind of insulator creeping distance measurement method and device based on multi-vision, it is related to insulator measurement technical field, the method includes: to the insulator image that multiple cameras are collected is first image segmentation, generates insulator edge point cloud data, and insulator edge point cloud data is based on and calculates insulator edge perimeter, obtains initial creeping distance.Through first image segmentation, edge point cloud data is generated and initial creeping distance is calculated, establishes measurement datum.To target insulator image is second image segmentation, extract skirt edge point cloud data under view, and according to skirt edge point cloud data under view, calculate correction value.Combining elliptic fitting algorithm fitting recess feature, the local path deviation problem caused by umbrella skirt shielding in initial calculation is solved.According to multiple correction values, correct initial creeping distance, obtain target creeping distance.Reduce cumulative error, significantly improve the precision of the creeping distance measurement of complex structure insulator.

Description

Technical Field

[0001] This application relates to the field of insulator measurement technology, and in particular to a method and apparatus for measuring insulator creepage distance based on multi-view vision. Background Technology

[0002] As a crucial component of power facilities, the creepage distance of transmission and transformation insulators is an important parameter for assessing their insulation condition. The assessment of grid anti-pollution flashover and overvoltage insulation coordination requires actual measurement and verification, and the establishment of a standardized database of insulator creepage distances to provide a basis for insulator maintenance and replacement.

[0003] Currently, the main methods for measuring the creepage distance of insulators are contact measurement and total station measurement. However, improper operation during manual measurement can lead to deviations in the measurement results, and obstruction problems can occur when using instruments, which results in lower accuracy of the measurement results.

[0004] Therefore, improving the measurement accuracy of insulator creepage distance measurement is an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a method and device for measuring the creepage distance of insulators based on multi-view vision, which aims to solve the technical problem of low measurement accuracy in measuring the creepage distance of insulators.

[0006] To achieve the above objectives, this application proposes a method for measuring the creepage distance of insulators based on multi-view vision. The method includes:

[0007] The insulator images captured by the multi-view camera are first segmented to generate insulator edge point cloud data, and the insulator edge perimeter is calculated based on the insulator edge point cloud data to obtain the initial creepage distance;

[0008] The target insulator image is segmented twice to extract the skirt bottom edge point cloud data, and the correction value is calculated based on the skirt bottom edge point cloud data. The target insulator image is the insulator image currently captured by the target camera, and the skirt bottom edge point cloud data is used to represent the connection edge data between the insulator skirt and the column.

[0009] The initial creepage distance is corrected based on multiple correction values ​​to obtain the target creepage distance.

[0010] In one embodiment, the step of performing initial image segmentation on the insulator image captured by the multi-view camera to generate insulator edge point cloud data, and calculating the insulator edge perimeter based on the insulator edge point cloud data to obtain the initial creepage distance includes:

[0011] The insulator images captured by the multi-view camera are input into the deep learning model to obtain the edge point cloud data of the first insulator and the edge point cloud data of the second insulator.

[0012] The edge point cloud data of the first insulator and the edge point cloud data of the second insulator are smoothed to obtain the first edge perimeter and the second edge perimeter.

[0013] By splicing the perimeter of the first edge and the perimeter of the second edge, the Euclidean distance between adjacent skirts is calculated, resulting in multiple distance results.

[0014] The initial creepage distance is obtained by summing the multiple distance results.

[0015] In one embodiment, the step of splicing the first edge perimeter and the second edge perimeter, calculating the Euclidean distance between adjacent skirts, and obtaining multiple distance results includes:

[0016] The perimeters of the first edge and the second edge are spliced ​​together to obtain spliced ​​point cloud data;

[0017] The stitched point cloud data is sorted to obtain the sorting result;

[0018] Based on the sorting results, the Euclidean distance is calculated for every two points in sequence to obtain multiple distance results.

[0019] In one embodiment, the step of performing secondary image segmentation on the target insulator image, extracting the skirt bottom-view edge point cloud data, and calculating a correction value based on the skirt bottom-view edge point cloud data, wherein the target insulator image is an insulator image captured by a target camera, and the skirt bottom-view edge point cloud data is used to represent the connection edge data between the insulator skirt and the post, includes:

[0020] Select the target camera based on the current images captured by the multi-view camera, and acquire the image of the target insulator;

[0021] Secondary image segmentation is performed on the target insulator image to extract the point cloud data of the skirt's downward-looking edge;

[0022] Based on the point cloud data of the lower edge of the skirt, occlusion features are determined and correction values ​​are generated.

[0023] In one embodiment, the step of determining occlusion features and generating correction values ​​based on the skirt-viewing edge point cloud data includes:

[0024] The point cloud data of the lower edge of the skirt is smoothed to extract the connection line between the insulator skirt and the column.

[0025] Correction values ​​are generated by fitting the shading characteristics based on the connection line between the insulator skirt and the column.

[0026] In one embodiment, the step of generating a correction value by fitting shading features based on the connection edge line between the insulator skirt and the post includes:

[0027] The first and second ellipses are fitted by the connecting edge line between the insulator skirt and the column;

[0028] Calculate the inward distance of the skirt hem based on the first ellipse;

[0029] Calculate the distance from the outer edge to the cylinder based on the second ellipse;

[0030] The occlusion length is calculated based on the inward camber of the skirt and the distance from the outer edge to the column, and a correction value is obtained.

[0031] In one embodiment, the step of correcting the initial creepage distance based on a plurality of correction values ​​to obtain the target creepage distance includes:

[0032] The initial creepage distance is corrected according to the correction value to obtain the correction result of a single frame image;

[0033] Based on the correction results of the single-frame image, the correction results of multiple frames are weighted and averaged to output the target creepage distance.

[0034] In addition, to achieve the above objectives, this application also proposes an insulator creepage distance measuring device based on multi-view vision, which includes: an insulator detection unit, a transportation unit, a processing unit, and a sorting unit;

[0035] The insulator detection unit includes a vision component and a camera motion component. The vision component is slidably mounted on the camera motion component. The camera motion component controls the movement of the vision component and performs multi-view image acquisition on the insulator at the detection station.

[0036] The transport unit includes an insulator loading assembly and an electric conveying assembly. The insulator loading assembly is used to load insulators, and the electric conveying assembly is used to move the insulator loading assembly to the testing station. After the insulators in the insulator loading assembly are tested, the insulator loading assembly is transported out of the testing station.

[0037] The processing unit is connected to the insulator detection unit and is used to perform the insulator creepage distance measurement method based on multi-view vision as described above.

[0038] The sorting unit is used to sort out unqualified insulators after the inspection is completed.

[0039] In one embodiment, the vision component includes at least a first camera and a second camera, and the camera motion component includes a first stepper motor and a first optical guide rail;

[0040] The first camera and the second camera are respectively installed on both sides of the first optical rod guide rail. The first stepper motor drives the first camera and the second camera to move along the first optical rod guide rail to collect multi-view images of the insulator at the inspection station.

[0041] The vision component also includes an illumination element for providing a uniform lighting environment for the insulators at the inspection station.

[0042] In one embodiment, the insulator loading assembly includes a limiting support and a handrail. The limiting support is used to limit the insulator, and the handrail is provided on the limiting support for the first gripper to grasp.

[0043] The electric conveying assembly includes a first gripper, a second stepper motor, and a second guide rail. The first gripper is slidably mounted on the second guide rail. The second stepper motor drives the first gripper to move on the second guide rail, transporting the insulator loading assembly gripped by the first gripper to the testing station. After the insulator in the insulator loading assembly is tested, the second stepper motor drives the first gripper to continue moving on the second guide rail, transporting the insulator loading assembly gripped by the first gripper out of the testing station.

[0044] One or more technical solutions proposed in this application have at least the following technical effects:

[0045] The insulator images acquired by the multi-camera system were initially segmented to generate insulator edge point cloud data. Based on this data, the insulator edge perimeter was calculated, and the initial creepage distance was obtained. A binocular camera synchronous acquisition technique was employed, which solved the problem of missing image information due to occlusion in traditional monocular views, achieving complete coverage of the three-dimensional point cloud data of the insulator edges. By generating edge point cloud data through initial image segmentation and calculating the initial creepage distance, a measurement benchmark was established.

[0046] A secondary image segmentation process is performed on the target insulator image to extract the skirt's downward-looking edge point cloud data, and correction values ​​are calculated based on this data. The target insulator image is the insulator image currently captured by the target camera, and the skirt's downward-looking edge point cloud data represents the connection line data between the insulator skirt and the column. Further, by extracting the skirt's downward-looking edge point cloud data through secondary image segmentation of the skirt connection area, and combining this with an ellipse fitting algorithm to fit the concave features, the local path deviation problem caused by skirt occlusion in the initial calculation is resolved.

[0047] The initial creepage distance is corrected using multiple correction values ​​to obtain the target creepage distance. Based on the multi-correction value dynamic correction mechanism, compensation amounts are independently generated for each shed connection point and fused into the initial path, reducing cumulative errors and significantly improving the accuracy of creepage distance measurement for complex insulator structures. Attached Figure Description

[0048] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating the first embodiment of the insulator creepage distance measurement method based on multi-view vision of this application;

[0051] Figure 2 This is a flowchart illustrating the second embodiment of the insulator creepage distance measurement method based on multi-view vision in this application;

[0052] Figure 3 This is a schematic diagram of a two-dimensional scan image of the insulator edge in the second embodiment of this application;

[0053] Figure 4 This is a flowchart illustrating the third embodiment of the insulator creepage distance measurement method based on multi-view vision in this application;

[0054] Figure 5 This is a schematic diagram of the ellipse fitting effect in the third embodiment of this application;

[0055] Figure 6 This is a flowchart illustrating the fourth embodiment of the insulator creepage distance measurement method based on multi-view vision in this application;

[0056] Figure 7 This is a schematic diagram illustrating the effect of the EMA exponential moving average algorithm in the fourth embodiment of this application;

[0057] Figure 8 This is a schematic diagram of the method flow of a preferred embodiment of this application;

[0058] Figure 9 This is a schematic diagram of the module structure of the insulator creepage distance measuring device based on multi-view vision according to an embodiment of this application;

[0059] Figure 10This is a first schematic diagram of the hardware device structure involved in the multi-view vision-based insulator creepage distance measurement device in the embodiments of this application;

[0060] Figure 11 This is a second schematic diagram of the hardware device structure involved in the multi-view vision-based insulator creepage distance measurement device in the embodiments of this application;

[0061] Figure 12 This is a third schematic diagram of the hardware structure of the insulator creepage distance measuring device based on multi-view vision in this application embodiment.

[0062] Explanation of icon numbers:

[0063] 10. Insulator detection unit; 101. Vision component; 1011. First camera; 1012. Second camera; 1013. Illumination element; 102. Camera motion component; 1021. First stepper motor; 1022. First optical guide rail; 20. Transport unit; 201. Insulator loading component; 2011. Limiting support; 2012. Handrail; 202. Electric transmission component; 2021. First gripper; 2022. The... 2. Stepper motor; 2023. Second guide rail; 2024. Third stepper motor; 2025. Third lead screw guide rail; 2026. Limit switch; 2027. First servo motor; 30. Processing unit; 40. Sorting unit; 401. Fourth stepper motor; 402. Fourth guide rail; 403. Rotary platform; 404. Second gripper; 405. Second servo motor; 406. Third servo motor; 407. Fourth servo motor; 408. Robotic arm.

[0064] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0065] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0066] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0067] As a crucial component of power facilities, the creepage distance of transmission and transformation insulators is an important parameter for assessing their insulation condition. The assessment of grid anti-pollution flashover and overvoltage insulation coordination requires actual measurement and verification, and the establishment of a standardized database of insulator creepage distances to provide a basis for insulator maintenance and replacement.

[0068] Currently, the main methods for measuring insulator creepage distance are contact measurement and measurement using a total station. Contact measurement, which involves manual measurement using thin, non-elastic adhesive tape or wire along the semiconductor enamel layer on the insulator surface, suffers from significant deviations in measurement results due to limitations in equipment and methods, and is also costly.

[0069] Total stations can perform coordinate, distance, and angle measurements and are widely used in the surveying field. When using a total station to measure insulators, because the installation location is fixed, obstructions can prevent the acquisition of complete insulator creepage distance information. Furthermore, the actual measurement results are directly related to the operator's methods.

[0070] Based on this, embodiments of this application provide a method for measuring insulator creepage distance based on multi-view vision, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the insulator creepage distance measurement method based on multi-view vision of this application.

[0071] In this embodiment, the insulator creepage distance measurement method based on multi-view vision includes steps S10~S30:

[0072] Step S10: Perform initial image segmentation on the insulator image captured by the multi-view camera to generate insulator edge point cloud data, and calculate the insulator edge perimeter based on the insulator edge point cloud data to obtain the initial creepage distance.

[0073] It should be noted that a multi-view camera can be understood as multiple cameras in an insulator creepage distance measurement device based on multi-view vision. It can simultaneously acquire insulator images from multiple perspectives, generating insulator edge point cloud data. For example, an image semantic segmentation algorithm can be used to separate the insulator body and background from the insulator images acquired by the multi-view cameras, generating continuous insulator edge point cloud data. The insulator edge perimeter can be calculated based on the edge point cloud data using path integration to obtain the initial creepage distance.

[0074] Step S20: Perform secondary image segmentation on the target insulator image, extract the point cloud data of the skirt's downward view edge, and calculate the correction value based on the point cloud data of the skirt's downward view edge.

[0075] It should be noted that the target insulator image is the insulator image currently captured by the target camera, and the skirt bottom-view edge point cloud data is used to represent the connection line data between the insulator skirt and the post. The camera position will change according to the measurement requirements; the camera that can capture the area below the insulator skirt is defined as the target camera. For the area where the lower edge of the skirt connects to the post, an edge enhancement segmentation algorithm can be used to extract the skirt bottom-view edge point cloud data. For example, correction can be made by comparing the indentation distance and the outer edge distance of the connection line.

[0076] Step S30: Correct the initial creepage distance based on multiple correction values ​​to obtain the target creepage distance.

[0077] It is understandable that the target creepage distance can be understood as the result of fusing the corrected creepage distance data from multiple frames of images, corresponding to multiple correction values.

[0078] In this embodiment, a binocular camera synchronous acquisition technology is adopted to solve the problem of missing image information caused by occlusion in the traditional monocular view, and to achieve complete coverage of the three-dimensional point cloud data of the insulator edge. The edge point cloud data is generated by the first image segmentation and the initial creepage distance is calculated to establish a measurement benchmark. Furthermore, the downward edge point cloud data of the skirt edge is extracted by the secondary image segmentation of the skirt connection area, and the concave features are fitted by the ellipse fitting algorithm to solve the problem of local path deviation caused by the skirt occlusion in the initial calculation. Based on the multi-correction value dynamic correction mechanism, the compensation amount is generated independently for each skirt connection point and fused into the initial path, which reduces the cumulative error and significantly improves the accuracy of creepage distance measurement of complex structure insulators.

[0079] Reference Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the insulator creepage distance measurement method based on multi-view vision of this application. Figure 1 The first embodiment shown presents a second embodiment of the insulator creepage distance measurement method based on multi-view vision.

[0080] In the second embodiment, step S10 includes:

[0081] Step S101: Input the insulator images captured by the multi-view camera into the deep learning model to obtain the edge point cloud data of the first insulator and the edge point cloud data of the second insulator.

[0082] It should be noted that, for ease of explanation, this embodiment uses a binocular camera as an example. The binocular camera can be arranged vertically, with an upper and lower viewpoint. For example, the deep learning model can be a YOLOv8 model. The first insulator edge point cloud data corresponds to the insulator body contour data obtained from the upper viewpoint, and the second insulator edge point cloud data corresponds to the umbrella skirt connection area data obtained from the lower viewpoint.

[0083] For example, after the YOLOv8 model segments the insulator from the acquired image, it can perform Canny edge detection on the segmented image to obtain a two-dimensional image of the insulator containing only edge information. The two-dimensional image of the insulator after YOLOv8 recognition and Canny edge detection processing is shown below. Figure 3 As shown.

[0084] Step S102: Smooth the point cloud data of the first insulator edge and the point cloud data of the second insulator edge respectively to obtain the perimeter of the first edge and the perimeter of the second edge.

[0085] For example, the Savitzky-Golay filter can be used to smooth the point cloud data of the first and second insulator edges, respectively. This removes noise while preserving the edge features of the insulator, making the point cloud data smoother and thus obtaining the perimeters of the first and second edges. It can be understood that the first edge perimeter corresponds to the point cloud data of the first insulator edge, and the second edge perimeter corresponds to the point cloud data of the second insulator edge. The Savitzky-Golay filter is a digital filter based on local polynomial regression, which uses linear least squares to fit a low-order polynomial to a sliding window of adjacent data points. A fixed-size window slides across the signal sequence, and polynomials are fitted to the data points within the window. The window size and the polynomial order are two key parameters of this algorithm. The algorithm calculates the value of the polynomial at the center point of each window position and uses it as the filtered output for that point. By repeating this process for each data point, the complete filtered signal is finally obtained.

[0086] For example, suppose the data sequence is ,in ∈[1, N], the goal is to use p A polynomial of order 1 is used to fit the local data. The polynomial expression is shown in formula (1).

[0087]

[0088] For the center is located The window needs to determine the coefficient vector. , , ..., This allows the polynomial to best fit the data points within the window. This optimization problem can be solved by minimizing the mean square error, as expressed in formula (2).

[0089]

[0090] In formula (2), m represents the window size, and the center of the window is a point. .

[0091] Step S103: The perimeters of the first edge and the second edge are spliced ​​together, and the Euclidean distance between adjacent skirts is calculated to obtain multiple distance results.

[0092] It should be noted that step S103 includes: splicing the perimeter of the first edge and the perimeter of the second edge to obtain spliced ​​point cloud data; sorting the spliced ​​point cloud data to obtain sorting results; and calculating the Euclidean distance between each pair of points based on the sorting results to obtain multiple distance results.

[0093] For example, this can be achieved by retaining the relatively frontal view area captured by the binocular camera lens as the image stitching material, sorting the smoothed stitching point cloud data, and calculating the Euclidean distance between every two points in turn.

[0094] Step S104: Sum the multiple distance results to obtain the initial creepage distance.

[0095] This embodiment employs multi-view synchronous acquisition technology (top view to acquire the main outline, bottom view to focus on the skirt connection area), which solves the path breakage problem caused by skirt occlusion in traditional monocular imaging, and achieves complete modeling of the insulator's three-dimensional outline. By denoising the point cloud, optical noise is suppressed while preserving the sharpness of the skirt edges, overcoming the measurement deviation caused by image jitter. The dual-view data are fused into a continuous three-dimensional path, reducing the segmentation error caused by viewpoint differences. Finally, the complex surface creepage path is transformed into a linearly solvable problem through a discrete point Euclidean distance accumulation algorithm.

[0096] Reference Figure 4 , Figure 4 This is a flowchart illustrating the third embodiment of the insulator creepage distance measurement method based on multi-view vision of this application. Figure 2 The second embodiment shown presents a third embodiment of the insulator creepage distance measurement method based on multi-view vision.

[0097] In the third embodiment, step S20 includes:

[0098] Step S201: Select the target camera based on the current captured image of the multi-view camera and acquire the image of the target insulator.

[0099] For example, the target insulator diagram includes key geometric features at the connection between the skirt and the column.

[0100] Step S202: Perform secondary image segmentation on the target insulator image and extract the point cloud data of the skirt's downward-looking edge.

[0101] For example, the target insulator image is input into the YOLOv8 model for secondary image segmentation, and the skirt bottom edge point cloud data is extracted by Canny edge detection.

[0102] Step S203: Determine occlusion features based on the point cloud data of the skirt's downward view edge, and generate correction values.

[0103] It should be noted that the occlusion feature can be understood as a quantitative parameter used to describe the geometric concavity at the connection of the insulator skirt. Step S203 includes: smoothing the point cloud data of the lower edge of the skirt, extracting the connection edge line between the insulator skirt and the column; fitting the occlusion feature according to the connection edge line between the insulator skirt and the column, and generating a correction value.

[0104] Specifically, the steps for generating correction values ​​based on the shading characteristics fitted by the connection line between the insulator skirt and the post include: fitting a first ellipse and a second ellipse by the connection line between the insulator skirt and the post; calculating the skirt edge indentation distance based on the first ellipse; calculating the distance from the outer edge to the post based on the second ellipse; and calculating the shading length based on the skirt edge indentation distance and the distance from the outer edge to the post to obtain the correction value.

[0105] For example, such as Figure 5 As shown, the first and second ellipses are used to fit the indentation distance of the skirt and the distance from the outer edge to the cylinder, respectively. Simultaneously, the coordinates of the vertices of the major axes of the two ellipses in the two-dimensional image are obtained. Then, by calculating the distance between the vertices of the major axes of the two ellipses on the same side using the Pythagorean theorem, the length of the lower edge line of the insulator skirt, i.e., the occluded distance, can be obtained. The correction value is the compensation amount calculated based on the occlusion characteristics.

[0106] In this embodiment, a secondary image segmentation and edge enhancement algorithm (deep learning model and Canny detection) is used to accurately extract the point cloud data of the skirt's downward-looking edge from a complex background, overcoming the shortcomings of traditional methods in accurately identifying geometric features at the connection points. Furthermore, based on a double ellipse fitting model, the occlusion features are quantified, and the surface concavity in three-dimensional space is converted into calculable geometric parameters (inward concavity distance and outer edge distance). Finally, the Pythagorean theorem spatial projection algorithm is used to convert the occlusion features into the actual occlusion length and generate a correction value, reducing the systematic error caused by the skirt occlusion in the initial creepage distance.

[0107] Reference Figure 6 , Figure 6 This is a flowchart illustrating the fourth embodiment of the insulator creepage distance measurement method based on multi-view vision of this application. Figure 4 The second embodiment shown presents a fourth embodiment of the insulator creepage distance measurement method based on multi-view vision of this application.

[0108] In the fourth embodiment, step S30 includes:

[0109] Step S301: Correct the initial creepage distance according to the correction value to obtain the correction result of a single frame image.

[0110] It should be noted that the correction result of a single frame image is the correction of the skirt occlusion of the insulator image acquired in a single acquisition.

[0111] Step S302: Based on the correction results of a single frame image, perform a weighted average of the correction results of multiple frames images to output the target creepage distance.

[0112] For example, the EMA exponential moving average algorithm can be used to perform a weighted average of the correction results of multiple frames of images, and the target creepage distance can be output as formula (3) and formula (4).

[0113]

[0114] In formulas (3) and (4), let the window be N (fixed length), and the data points in the current queue be { , ,..., }( (≤N), each data point represents the raw distance result at a given time. It refers to The smoothed distance result at each moment. This refers to the smoothing factor. This represents the average distance within the window at the current moment, as shown in the image. Figure 7 As shown.

[0115] In this embodiment, a dynamic compensation mechanism for correction values ​​is adopted to incorporate the geometric concave features at the umbrella skirt connection into the initial creepage distance, thus solving the systematic deviation problem of path measurement caused by the occlusion of the three-dimensional structure of the umbrella skirt. Furthermore, through multi-frame time-series weighted fusion technology (EMA exponential moving average algorithm), the continuously acquired correction results are dynamically smoothed, which significantly improves the accuracy compared with the traditional single measurement or fixed window averaging method.

[0116] To make the above embodiments clearer, refer to Figure 8 , Figure 8 This is a complete flowchart of a preferred embodiment of the present application.

[0117] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the multi-view vision-based insulator creepage distance measurement method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0118] This application also provides an insulator creepage distance measuring device based on multi-view vision, please refer to... Figure 9 The insulator creepage distance measuring device based on multi-view vision includes: an insulator creepage distance measuring device: an insulator detection unit 10, a transportation unit 20, a processing unit 30, and a sorting unit 40.

[0119] The insulator inspection unit 10 includes a vision component 101 and a camera motion component 102. The vision component 101 is slidably mounted on the camera motion component 102. The camera motion component 102 controls the movement of the vision component 101 and performs multi-view image acquisition on the insulators at the inspection station. The transport unit 20 includes an insulator loading component 201 and an electric conveying component 202. The insulator loading component 201 is used to move the insulator loading component 201 to the inspection station, and after the insulators in the insulator loading component 201 are inspected, the insulator loading component 201 is transported out of the inspection station. The processing unit 30 is connected to the insulator inspection unit 10 and is used to execute the insulator creepage distance measurement method based on multi-view vision as described in the above embodiments and implementations. The sorting unit 40 is used to sort the unqualified insulators after the inspection is completed.

[0120] It should be noted that the processing unit 30 is also connected to the transportation unit 20 and the sorting unit 40, and is used to issue transportation instructions to the transportation unit 20 and sorting instructions to the sorting unit 40. For example, the processing unit 30 can be a host computer, including but not limited to mobile terminals such as laptops, digital broadcast receivers, and PDAs (Personal Digital Assistants), as well as fixed terminals such as digital TVs and desktop computers. The processing unit 30 implements human-machine interaction functions, providing operators with real-time device operation data and images, while also receiving operator instructions.

[0121] Based on this, this application proposes an implementation method, referring to... Figure 10 , Figure 10This is a 3D diagram of the insulator creepage distance measurement device based on multi-view vision according to this application. The vision component 101 includes a first camera 1011 and a second camera 1012. The camera motion component 102 includes a first stepper motor 1021 and a first optical guide rail 1022. The first camera 1011 and the second camera 1012 are respectively mounted on both sides of the first optical guide rail 1022. The first stepper motor 1021 drives the first camera 1011 and the second camera 1012 to move along the first optical guide rail 1022 to acquire multi-view images of the insulator at the inspection station. The vision component 101 also includes an illumination element 1013, which is used to provide a uniform lighting environment for the insulator at the inspection station. For example, the illumination element 1013 can be an LED light.

[0122] The insulator loading assembly 201 includes a limiting support 2011 and a handle 2012. The limiting support 2011 is used to limit the insulator, and the handle 2012 is provided on the limiting support 2011 for the first gripper 2021 to grip. The electric transmission assembly 202 includes a first gripper 2021, a second stepper motor 2022, and a second smooth rod guide rail 2023. The first gripper 2021 is slidably mounted on the second smooth rod guide rail 2023. The second stepper motor 2022 drives the first gripper 2021 to move on the second smooth rod guide rail 2023 to transport the insulator loading assembly 201 gripped by the first gripper 2021 to the testing station. After the insulator in the insulator loading assembly 201 is tested, the second stepper motor 2022 drives the first gripper 2021 to continue moving on the second smooth rod guide rail 2023 to transport the insulator loading assembly 201 gripped by the first gripper 2021 out of the testing station.

[0123] It should be noted that the second stepper motor 2022 drives the first gripper 2021 to move on the second guide rail 2023, and the first servo motor 2027 drives the first gripper 2021 to grasp the insulator loading assembly. (Reference) Figure 11 , Figure 11From another perspective of the three-dimensional diagram of the insulator creepage distance measuring device based on multi-view vision in this application, the electric transmission assembly 202 also includes a third stepper motor 2024, a third lead screw guide rail 2025, and a limit switch 2026. When the second stepper motor 2022 drives the first gripper 2021 to move on the second guide rail 2023 to the entrance of the detection station, the insulator loading assembly 201 gripped by the first gripper 2021 is placed at the entrance of the detection station, and the first gripper 2021 waits for insulator detection at the current position. The third stepper motor 2024 drives the third lead screw guide rail 2025 to move the detection station inward. When the detection station moves to the target position, the limit switch 2026 sends a stop signal. After the insulator inspection is completed, the third stepper motor 2024 drives the third lead screw guide rail 2025 to pull the inspection station outward to the entrance of the inspection station. The first servo motor 2027 drives the first gripper 2021 to grab the insulator loading assembly 201. The second stepper motor 2022 drives the first gripper 2021 to continue moving along the second guide rail 2023 and transport the insulator loading assembly 201 to the sorting area.

[0124] Specifically, the sorting unit 40 includes a fourth stepper motor 401, a fourth guide rail 402, a rotating platform 403, and a second gripper 404. The second gripper 404 is connected to the rotating platform 403 and is mounted above the rotating platform 403. The rotating platform 403 is slidably mounted on the fourth guide rail 402. After inspection, the fourth stepper motor 401 drives the rotating platform 403 to slide on the fourth guide rail 402. When the rotating platform 403 slides to the position of the defective insulator, the third servo motor 406 drives the robotic arm 408 to lift up, and the first servo motor 405 drives the second gripper 404 to grab the defective insulator. After the fourth servo motor 407 drives the rotating platform 403 to rotate to the target position (such as in front of the defective insulator rack), the third servo motor 406 drives the robotic arm 408 to lower towards the defective insulator rack, and the first servo motor 405 drives the second gripper 404 to release the defective insulator (see reference). Figure 12 The process involves placing unqualified insulators on a defective insulator rack to complete the sorting of unqualified insulators.

[0125] In this embodiment, the host computer sends a control signal to cause the first gripper to pick up the handrails on both sides of the insulator loading frame. Next, the second stepper motor drives the insulator loading frame (i.e., the insulator loading assembly) and transports it to the measurement station (before entering the measurement chamber). The third stepper motor drives the third lead screw guide rail to transport the insulator loading frame into the machine's measurement chamber, placing it directly below the binocular cameras (first and second cameras). The first stepper motor drives the binocular cameras to measure the insulator creep distance sequentially, and the measurement process and results are updated in real time on the host computer. After the measurement is completed, the binocular cameras return to the zero position, and the third stepper motor drives the third lead screw guide rail to move the insulator loading frame out of the chamber. Next, the first gripper picks up the insulator loading frame again, and the second stepper motor drives the belt again to transport the insulator loading frame to the sorting area. Immediately afterwards, the fourth stepper motor drives the belt to transport the horizontal rotating gimbal of the robotic arm to the front of the defective insulator, and then controls the robotic arm to tilt down towards the insulator to pick up the defective insulator. Next, the horizontal rotating gimbal rotates, and the second gripper places the unqualified insulator into the unqualified area. Finally, the second gripper returns to its original position, waiting for the next sorting. Through the cooperation of various components, multiple batches of insulators can be measured continuously and automatically, which greatly improves the measurement efficiency and realizes high-precision measurement of insulator creepage distance.

[0126] The electric conveyor assembly precisely transports the insulators under test to the testing station via the insulator loading assembly, and then transports the measured insulators to the sorting area, reducing manual handling and ensuring a stable and reliable transportation process that avoids damage to the insulators that may be caused by manual handling. The sorting unit automatically sorts out unqualified insulators based on signals from the host computer, eliminating the need for manual screening and further enhancing the automation of the overall measurement process, saving significant time. The fully automated control reduces human intervention and dependence on operators, resulting in measurement repeatability errors far lower than traditional methods. Compared to traditional measuring devices, this multi-view vision-based insulator creepage distance measuring device offers significant improvements in efficiency, accuracy, adaptability, and intelligence, enhancing its adaptability and versatility in complex scenarios and better meeting the power grid's needs for insulator creepage distance measurement.

[0127] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for measuring the creepage distance of insulators based on multi-view vision, characterized in that, The method includes: The insulator images captured by the multi-camera are first segmented to generate insulator edge point cloud data, and the insulator edge perimeter is calculated based on the insulator edge point cloud data to obtain the initial creepage distance; A secondary image segmentation process is performed on the target insulator image to extract the skirt's downward-looking edge point cloud data. A correction value is then calculated based on this data. The target insulator image is the insulator image currently captured by the target camera. The skirt's downward-looking edge point cloud data represents the connection line data between the insulator skirt and the post. The process includes: selecting a target camera based on the currently captured image from the multi-camera system and acquiring the target insulator image; performing secondary image segmentation on the target insulator image to extract the skirt's downward-looking edge point cloud data; smoothing the skirt's downward-looking edge point cloud data to extract the connection line between the insulator skirt and the post; fitting a first ellipse and a second ellipse using the connection line between the insulator skirt and the post; calculating the skirt's indentation distance based on the first ellipse; calculating the distance from the outer edge to the post based on the second ellipse; and calculating the occlusion length based on the skirt's indentation distance and the distance from the outer edge to the post to obtain the correction value. The initial creepage distance is corrected according to the correction value to obtain the correction result of a single frame image; Based on the correction results of the single-frame image, the correction results of multiple frames are weighted and averaged to output the target creepage distance.

2. The method as described in claim 1, characterized in that, The steps of performing initial image segmentation on the insulator images captured by the multi-view camera, generating insulator edge point cloud data, and calculating the insulator edge perimeter based on the insulator edge point cloud data to obtain the initial creepage distance include: The insulator images captured by the multi-view camera are input into the deep learning model to obtain the edge point cloud data of the first insulator and the edge point cloud data of the second insulator. The edge point cloud data of the first insulator and the edge point cloud data of the second insulator are smoothed to obtain the first edge perimeter and the second edge perimeter. By splicing the perimeter of the first edge and the perimeter of the second edge, the Euclidean distance between adjacent skirts is calculated, resulting in multiple distance results. The initial creepage distance is obtained by summing the multiple distance results.

3. The method as described in claim 2, characterized in that, The step of splicing the perimeters of the first and second edges, calculating the Euclidean distance between adjacent skirts, and obtaining multiple distance results includes: The perimeters of the first edge and the second edge are spliced ​​together to obtain spliced ​​point cloud data; The stitched point cloud data is sorted to obtain the sorting result; Based on the sorting results, the Euclidean distance is calculated for every two points in sequence to obtain multiple distance results.

4. A multi-view vision-based insulator creepage distance measuring device, characterized in that, The insulator creepage distance measuring device based on multi-view vision includes: an insulator detection unit, a transportation unit, a processing unit, and a sorting unit; The insulator detection unit includes a vision component and a camera motion component. The vision component is slidably mounted on the camera motion component. The camera motion component controls the movement of the vision component and performs multi-view image acquisition on the insulator at the detection station. The transport unit includes an insulator loading assembly and an electric conveying assembly. The insulator loading assembly is used to load insulators, and the electric conveying assembly is used to move the insulator loading assembly to the testing station. After the insulators in the insulator loading assembly are tested, the insulator loading assembly is transported out of the testing station. The processing unit is connected to the insulator detection unit and is used to perform the insulator creepage distance measurement method based on multi-view vision as described in any one of claims 1 to 3; The sorting unit is used to sort out unqualified insulators after the inspection is completed.

5. The insulator creepage distance measuring device based on multi-view vision as described in claim 4, characterized in that, The vision component includes at least a first camera and a second camera, and the camera motion component includes a first stepper motor and a first optical rod guide rail; The first camera and the second camera are respectively installed on both sides of the first optical rod guide rail. The first stepper motor drives the first camera and the second camera to move along the first optical rod guide rail to collect multi-view images of the insulator at the inspection station. The vision component also includes an illumination element for providing a uniform lighting environment for the insulators at the inspection station.

6. The insulator creepage distance measuring device based on multi-view vision as described in claim 4, characterized in that, The insulator loading assembly includes a limiting support and a handrail. The limiting support is used to limit the insulator, and the handrail is provided on the limiting support for the first gripper to grasp. The electric conveying assembly includes a first gripper, a second stepper motor, and a second guide rail. The first gripper is slidably mounted on the second guide rail. The second stepper motor drives the first gripper to move on the second guide rail, transporting the insulator loading assembly gripped by the first gripper to the testing station. After the insulator in the insulator loading assembly is tested, the second stepper motor drives the first gripper to continue moving on the second guide rail, transporting the insulator loading assembly gripped by the first gripper out of the testing station.

Images