Insulator creepage distance detection method, system and equipment based on visual recognition

Abstract

The present invention relates to a method, system and device for detecting the creepage distance of an insulator based on visual recognition. The detection method obtains three-dimensional point cloud data of the surface of the insulator to be tested, uses a random consistency check method to perform multiple calculations on the point cloud data, forms a number of virtual cylinders that preliminarily meet preset conditions, and further selects a target cylinder from among them, thereby performing cross-sectional calculations on the target cylinder to generate a corresponding creepage distance arc. Finally, after calculation, the length of the shortest creepage distance arc is selected as the creepage distance of the insulator. This detection method is suitable for insulators with complex outer contour shapes and can quickly feedback the creepage distance detection results. Compared with traditional distance measurement tools and methods, it can reduce measurement errors caused by human factors and improve the measurement accuracy of the creepage distance. In addition, the implementation of this method can rely on machines to collect point cloud data and analyze and calculate, effectively reducing the safety hazards of manual operation.

Description

Technical Field

[0001] The present invention relates to the technical field of disc-shaped suspension insulator detection, and in particular to a method, system and device for detecting the creepage distance of an insulator based on visual recognition. Background Art

[0002] High-voltage insulators, as electrical insulation components, are widely used in power transmission projects. Their sheds provide the primary creepage distance for these insulators. Due to long-term operation in harsh climates, exposed to factors such as high and low temperatures, snowstorms, and other factors, these sheds are susceptible to aging and deformation, which reduces creepage distance and can lead to arc bridging between the sheds. This reduces the flashover voltage of the high-voltage insulators, resulting in flashover accidents and threatening the safe operation of the power grid. Therefore, insulator quality testing is a key component of power grid material testing, and parameters such as insulator diameter, surface area, and creepage distance are key indicators of insulator quality.

[0003] Currently, creepage distance measurements on construction sites or already installed insulators primarily rely on manual measurement. For example, a flexible wire or measuring tape is wrapped around the insulator's surface, closely replicating the insulator's contour at the corresponding location. The length of the stretched wire or measuring tape is then measured. Due to the complex outer contours of insulators, manual measurement can lead to significant deviations in creepage distance results due to uncertainties such as tool accuracy, measurement method, and human error. Furthermore, manual measurement is labor-intensive and poses safety risks. Summary of the Invention

[0004] Based on this, it is necessary to address the technical problem in the prior art that the measurement accuracy of the insulator creepage distance needs to be improved. The present invention provides an insulator creepage distance detection method, system and device based on visual recognition.

[0005] The present invention discloses a method for detecting creepage distance of an insulator based on visual recognition, which comprises the following steps, namely step one to step three.

[0006] 1. Obtain 3D point cloud data of the surface of an insulator to be tested, and construct a 3D model of the insulator based on the point cloud data fitting. The insulator includes a column and several sheds.

[0007] 2. Generate the creepage distance arc of the insulator based on the point cloud data and the three-dimensional model of the insulator. The method for generating the creepage distance arc includes the following steps, namely steps S1 to S5.

[0008] S1. Use the random consistency check method to perform multiple calculations on the point cloud data to obtain multiple different cylinder parameters. The cylinder parameters include the cross-sectional diameter and centerline of the cylinder.

[0009] S2. Construct multiple corresponding virtual cylinders based on different cylinder parameters, and compare and analyze the coordinates of all sampling points in the point cloud data with the cylindrical coordinates of each virtual cylinder in turn, and then count the number of sampling points belonging to the cylindrical surface on different virtual cylinders.

[0010] S3. Select the virtual cylinder with the largest number of cylinder sampling points as the target cylinder.

[0011] S4. Slice the target cylinder along the radial direction to obtain multiple cylinder sections.

[0012] S5. Calculate and process multiple cylindrical sections to generate a creepage distance arc within each section.

[0013] 3. Calculate the length of the creepage distance arc in each section cyclically and select the shortest length as the insulator creepage distance.

[0014] As a further improvement of the above solution, in step S1 , each calculation of the point cloud data to obtain corresponding cylinder parameters includes the following steps, namely steps S11 to S17 .

[0015] S11. Randomly select two sampling points in the point cloud as the starting point and the end point, and obtain a spatial straight line passing through the two points based on the spatial coordinates of the starting point and the end point [x1, y1, z1] and [x2, y2, z2].

[0016] S12. Analyze the linear distances between each sampling point in the point cloud and the spatial line. If the linear distance d between any sampling point and the spatial line is less than a preset distance threshold δ, the sampling point is considered to be near the spatial line and included in the statistics. The spatial coordinates of the sampling point are [x0, y0, z0].

[0017] S13. When the statistical number N0 of sampling points near the spatial line is greater than a preset number threshold M0, the spatial line is marked as a candidate line belonging to the cylinder surface line.

[0018] S14. Establish a vertical plane 1 located at the starting point and a vertical plane 2 located at the end point. Both the vertical plane 1 and the vertical plane 2 are perpendicular to the candidate line.

[0019] S15. Count the number of first sampling points N1 and the number of second sampling points N2 located on the vertical plane 1 and the vertical plane 2 respectively.

[0020] S16. When both the number of first sampling points N1 and the number of second sampling points N2 are greater than a preset threshold number M1, fit all first sampling points and the second sampling points to a circle, thereby obtaining a first fitting circle and a second fitting circle. The ratio of the diameters of the first fitting circle to the second fitting circle satisfies a ratio less than a preset ratio σ, and the number of sampling points on the circumferences of both the first fitting circle and the second fitting circle is greater than a preset threshold number M2.

[0021] S17. Take the average of the diameters of the first fitting circle and the second fitting circle as the cross-sectional diameter d0 of the virtual cylinder, then extend the length of the cross-sectional diameter d0 from the starting point to the center of the first fitting circle to obtain the center point one C1, and extend the length of the cross-sectional diameter d0 from the midpoint to the center of the second fitting circle to obtain the center point two C2, and then take the line connecting the center point one C1 and the center point two C2 as the center line of the virtual cylinder.

[0022] As a further improvement to the above solution, the straight-line distance d is expressed as:

[0023]

[0024] As a further improvement of the above solution, the expression equation of the vertical plane 1 is:

[0025] (x-x1,y-y1,z-z1)·(x2-x1,y2-y1,z2-z1)=0

[0026] The expression equation of vertical plane 2 is:

[0027] (x-x2,y-y2,z-z2)·(x2-x1,y2-y1,z2-z1)=0.

[0028] As a further improvement of the above solution, in step S4, the specific process of slicing the cylinder is as follows, namely steps S41 to S42.

[0029] S41. Determine the slice parameters: set any point P[x3, y3, z3] on the surface of the cylinder as the starting calculation point, and set the number of slices to N3, the slice interval angle θ, and the distance threshold to ε.

[0030] The angle between the sections is θ = 360° / N3. The foot of the perpendicular to the center line of the column is P. n =[k(x2-x1)+x1,k(y2-y1)+y1,k(z2-z1)+z1]. The initial tangent v is: [x3-k(x2-x1)-x1,y3-k(y2-y1)-y1,z3-k(z2-z1)-z1]. k is expressed as:

[0031]

[0032] S42. Use the slicing parameters to divide the cylinder into N3 sections.

[0033] As a further improvement of the above solution, in step S5 , the method for generating the creepage distance arc in each section includes the following steps, namely steps S51 to S53 .

[0034] S51. Rotate the initial tangent around the center line of the cylinder according to the section interval angle θ to obtain a new tangent vector, and then obtain a target section perpendicular to the new tangent.

[0035] The new tangent vector is v' = v cosΦ + u × v sinΦ + (u·v)u(1cosΦ). u is the centerline vector of the cylinder. The expression equation of the target section is: [(x, y, z)-P n ]·v'=0.

[0036] S52. Calculate the distance between each sampling point on the cylinder surface and the target section in sequence, and if the distance is less than a distance threshold ε, consider the sampling point as a point within the target section.

[0037] S53. Calculate the projection coordinates of all sampling points considered to be within the target section, and sequentially connect each projection point along a preset direction to obtain a creepage distance arc within the target section.

[0038] As a further improvement of the above solution, in step three, the method for calculating the creepage distance arc length in each section includes the following steps, namely steps S6 to S9.

[0039] S6. Divide the generated creepage distance arc into a flat portion and a continuous curved portion.

[0040] S7. Perform straight line fitting on the continuous flat parts using the least squares method to calculate the length of the flat part of the connecting line.

[0041] S8. Fit a spline curve to the continuous curved portion using the Gauss-Newton method, and sequentially obtain the coordinates of multiple points on the spline curve at intervals of a preset measurement accuracy. Then, connect each adjacent point and add up the lengths to obtain the length of the continuous curved portion.

[0042] S9. The length of the creepage distance arc shall be the sum of the lengths of the continuous flat portion and the continuous curved portion.

[0043] As a further improvement of the above solution, in step one, the method for constructing the three-dimensional model of the insulator includes the following steps, namely step (1) to step (5).

[0044] (1) Establish an insulator geometry library to assist point cloud computing.

[0045] (2) Use a three-dimensional laser scanner to collect initial point cloud data of the insulator to be tested multiple times.

[0046] (3) Align and splice the initial point cloud data collected multiple times to form complete point cloud data.

[0047] (4) Locate the control points in the laser-marked image corresponding to the complete point cloud data, and perform template matching on the laser-marked image.

[0048] (5) Extract features from the complete point cloud data, perform component fitting modeling based on the extraction results, and then combine them into a three-dimensional model of the insulator.

[0049] The present invention also discloses an insulator creepage distance detection system based on visual recognition, which applies any of the above-mentioned insulator creepage distance detection methods based on visual recognition. The detection system includes:

[0050] The model building module is used to obtain three-dimensional point cloud data of the surface of an insulator to be tested, and to construct a three-dimensional model of the insulator based on the point cloud data fitting.

[0051] The arc generation module is used to generate the creepage distance arc of the insulator based on the point cloud data and the insulator 3D model.

[0052] The distance calculation module is used to calculate a shortest insulator creepage distance based on the generated creepage distance arc.

[0053] The present invention also discloses an insulator creepage distance detection device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the device implements any of the steps of the aforementioned visual recognition-based insulator creepage distance detection method, thereby completing the creepage distance detection of an insulator to be tested.

[0054] Compared with the prior art, the technical solution disclosed in the present invention has the following beneficial effects:

[0055] 1. This visual recognition-based insulator creepage distance detection method obtains three-dimensional point cloud data of the surface of the insulator to be tested, then uses a random consistency check method to perform multiple calculations on the point cloud data to form several virtual cylinders that initially meet the preset conditions. The target cylinder is further selected from these virtual cylinders, and a cross-sectional calculation is performed on the target cylinder to generate the corresponding creepage distance arc. Finally, the length of the shortest creepage distance arc is selected as the insulator creepage distance. This detection method is targeted at insulators with complex outer contours. By utilizing visual recognition technology and systematic calculation and analysis, it can quickly feedback creepage distance detection results. Compared with traditional distance measurement tools and methods, it can reduce measurement errors caused by human factors and improve the measurement accuracy of creepage distance. In addition, the implementation of this method relies to a greater extent on machines to collect point cloud data and analyze and calculate, effectively reducing the safety hazards of manual operation.

[0056] 2. The beneficial effects of the insulator creepage distance detection system and the insulator creepage distance detection device based on visual recognition are the same as those of the above-mentioned detection method, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a step diagram of the insulator creepage distance detection method based on visual recognition in Example 1 of the present invention;

[0058] Figure 2 for Figure 1 Flow chart of creepage distance detection method;

[0059] Figure 3 Schematic diagram of the structure of the insulator in Example 1 of the present invention;

[0060] Figure 4 This is a correlation coefficient matrix diagram after the template traverses the laser marked image in Example 1 of the present invention;

[0061] Figure 5 This is a partial enlarged view of the insulator three-dimensional model after component fitting based on point cloud data in Example 1 of the present invention;

[0062] Figure 6 A diagram showing the steps of the method for calculating the point cloud data each time in Example 1 of the present invention;

[0063] Figure 7 Schematic diagram of parameters for forming a virtual column in Example 1 of the present invention. DETAILED DESCRIPTION

[0064] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used herein in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "or / and" as used herein includes any and all combinations of one or more of the associated listed items.

[0066] Example 1

[0067] See also Figure 1 and Figure 2 This embodiment provides an insulator creepage distance detection method based on visual recognition, including the following steps, namely step one to step three.

[0068] 1. Obtain 3D point cloud data of the surface of an insulator to be tested, and construct a 3D model of the insulator based on the point cloud data fitting.

[0069] See also Figure 3 , wherein the insulator includes a column 1 and a plurality of sheds 2. In this embodiment, the three-dimensional point cloud data can be collected and pre-processed based on visual recognition technology, and the three-dimensional model of the insulator can be constructed. The method for constructing the three-dimensional model of the insulator can include the following steps, namely, steps (1) to (5).

[0070] (1) Establish an insulator geometry library to assist point cloud computing.

[0071] In this embodiment, a library of common geometric structures is built for point cloud computing assistance. By building an insulator graphic library and a geometric constraint library, automatic parametric modeling of insulators is performed on point cloud data using insulator fitting and combination methods.

[0072] (2) Use a three-dimensional laser scanner to collect initial point cloud data of the insulator to be tested multiple times.

[0073] In this embodiment, the scanner SDK can be used to acquire point cloud data. Image acquisition can be divided into two steps. First, the laser pointer is turned on to capture an image with the laser mark. Then, the laser pointer is turned off to capture an image without the laser mark. Finally, the image format, bit number, naming method, and storage path are determined for the acquired image.

[0074] (3) Align and splice the initial point cloud data collected multiple times to form complete point cloud data.

[0075] Since insulators are rigid devices with fixed geometric shapes, the point cloud data obtained by the 3D laser scanner can well reflect the shape characteristics of the scanned object. The point cloud data collected multiple times is matched and spliced ​​to form a complete point cloud data.

[0076] (4) Locate the control points in the laser-marked image corresponding to the complete point cloud data, and perform template matching on the laser-marked image.

[0077] In this embodiment, the laser crosshairs form two cross-sections in the laser-marked image. The center pixel of each cross-section serves as the image control point. Each image has two control points, one above and one below. These control points are extracted and used as the scale for the corresponding point matching process. The vertical range determined by the vertical coordinates of the two control points serves as the vertical range of the target area.

[0078] Template matching methods are frequently used in pattern recognition and search, enabling pre-selected objects to be visualized within the image to be matched and their coordinates to be determined. Template matching algorithms typically operate on grayscale images and include methods such as the mean absolute difference (MAD) and the sum of squared deviation (SSD). The primary difference lies in the similarity measurement formulas used. This project uses the SSD algorithm.

[0079] like Figure 4 As shown in the figure, after the template traverses the entire laser-marked image, the correlation coefficient corresponding to each pixel is obtained, forming a correlation coefficient matrix. The correlation coefficient matrix is ​​displayed in pseudo-color. Yellow (light color in the figure) indicates a high correlation coefficient value and a high degree of match, while blue (dark color in the figure) indicates a low correlation coefficient value and a low degree of match.

[0080] (5) Extract features from the complete point cloud data, perform component fitting modeling based on the extraction results, and then combine them into a three-dimensional model of the insulator.

[0081] See also Figure 5 In this embodiment, effective point cloud data preprocessing can be used to improve the accuracy and efficiency of point cloud recognition. Feature extraction is then performed based on the complete spliced ​​point cloud data. Component fitting modeling is then performed based on the extracted results to create a 3D insulator model.

[0082] 2. Generate the creepage distance arc of the insulator based on the point cloud data and the three-dimensional model of the insulator. The method for generating the creepage distance arc includes the following steps, namely steps S1 to S5.

[0083] S1. Use the random consistency check method to perform multiple calculations on the point cloud data to obtain multiple different cylinder parameters. The cylinder parameters include the cross-sectional diameter and centerline of the cylinder.

[0084] See also Figure 6 In this embodiment, the number of cycles of the random consistency check method can be set to 1000 times, and the method of calculating the point cloud data each time to obtain the corresponding cylinder parameters can include the following steps, namely steps S11 to S17.

[0085] S11. Randomly select two sampling points in the point cloud as the starting point and the end point, and obtain a spatial straight line passing through the two points based on the spatial coordinates of the starting point and the end point [x1, y1, z1] and [x2, y2, z2].

[0086] Among them, the difference between the starting point and the end point is the direction vector, that is, (x2-x1, y2-y1, z2-z1).

[0087] S12. Analyze the linear distances between each sampling point in the point cloud and the spatial line. When the linear distance d between any sampling point and the spatial line is less than a preset distance threshold δ, the sampling point is determined to be near the spatial line and included in the statistics. The spatial coordinates of the sampling point are [x0, y0, z0]. The linear distance d can be expressed as:

[0088]

[0089] S13. When the statistical number N0 of sampling points near the spatial line is greater than a preset number threshold M0, the spatial line is marked as a candidate line belonging to the cylinder surface line for further identification; otherwise, the processing ends and enters the next cycle.

[0090] S14. Establish a vertical plane 1 located at the starting point and a vertical plane 2 located at the end point. Both vertical plane 1 and vertical plane 2 are perpendicular to the candidate line. In this embodiment, the expression equation of vertical plane 1 can be:

[0091] (x-x1, y-y1, z-z1)·(x2-x1, y2-y1, z2-z1)=0 The expression equation of the vertical plane 2 can be:

[0092] (x-x2,y-y2,z-z2)·(x2-x1,y2-y1,z2-z1)=0

[0093] S15. Count the number of first sampling points N1 and the number of second sampling points N2 located on the vertical plane 1 and the vertical plane 2 respectively.

[0094] S16. When both the number of first sampling points N1 and the number of second sampling points N2 are greater than a preset threshold number M1, fit all first sampling points and the second sampling points to a circle, thereby obtaining a first fitting circle and a second fitting circle. The ratio of the diameters of the first fitting circle to the second fitting circle satisfies a ratio less than a preset ratio σ, and the number of sampling points on the circumferences of both the first fitting circle and the second fitting circle is greater than a preset threshold number M2.

[0095] In this embodiment, for vertical plane 1, if the number of first sampling points N1 does not exceed a preset threshold value M1, the current process ends and the next loop begins. Otherwise, a circle is fitted using these points. If the number of sampling points on the circle does not exceed a preset threshold value M2, the current process ends and the next loop begins. The same principle applies to vertical plane 2 and will not be further described here.

[0096] S17. Take the average of the diameters of the first fitting circle and the second fitting circle as the cross-sectional diameter d0 of the virtual cylinder, then extend the length of the cross-sectional diameter d0 from the starting point to the center of the first fitting circle to obtain the center point one C1, and extend the length of the cross-sectional diameter d0 from the midpoint to the center of the second fitting circle to obtain the center point two C2, and then take the line connecting the center point one C1 and the center point two C2 as the center line of the virtual cylinder.

[0097] See also Figure 7 ,In the figure, P1 and P2 represent vertical plane 1 and vertical plane 2 respectively; L represents the center line (central axis) of the virtual cylinder.

[0098] S2. Construct multiple corresponding virtual cylinders based on different cylinder parameters, and compare and analyze the coordinates of all sampling points in the point cloud data with the cylindrical coordinates of each virtual cylinder in turn, and then count the number of sampling points belonging to the cylindrical surface on different virtual cylinders.

[0099] S3. Select the virtual cylinder with the largest number of cylinder sampling points as the target cylinder.

[0100] In this embodiment, the centerline and cross-sectional diameter of the cylinder are used as parameters to count the number of points on the cylinder surface in the surface point cloud. This is used as the basis for multiple loop comparisons. After the loop is completed, the cylinder parameters with the largest number of cylinder points are taken as the final cylinder parameters.

[0101] S4. Slice the target cylinder along the radial direction to obtain a plurality of cylinder sections. In this embodiment, the specific process of slicing the cylinder is as follows, namely steps S41 to S42.

[0102] S41. Determine the slice parameters: set any point P[x3, y3, z3] on the surface of the cylinder as the starting calculation point, and set the number of slices to N3, the slice interval angle θ, and the distance threshold to ε.

[0103] The angle between the sections is θ = 360° / N3. The foot of the perpendicular to the center line of the column is P. n =[k(x2-x1)+x1,k(y2-y1)+y1,k(z2-z1)+z1]. The initial tangent v is: [x3-k(x2-x1)-x1,y3-k(y2-y10-y1,z3-k(z2-z1)-z1]. k is expressed as:

[0104]

[0105] S42. Use the slicing parameters to divide the cylinder into N3 sections.

[0106] S5. Calculate and process multiple cylinder sections to generate a creepage distance arc within each section. The method for generating a creepage distance arc within each section includes the following steps, namely steps S51 to S53.

[0107] S51. Rotate the initial tangent around the center line of the cylinder according to the section interval angle θ to obtain a new tangent vector, and then obtain a target section perpendicular to the new tangent.

[0108] The new tangent vector is v' = v cosΦ + u × v sinΦ + (u·v)u(1cosΦ). u is the centerline vector of the cylinder. The expression equation of the target section is: [(x, y, z)-P n ]·v'=0.

[0109] S52. Calculate the distance between each sampling point on the cylinder surface and the target section in sequence, and if the distance is less than a distance threshold ε, consider the sampling point as a point within the target section.

[0110] S53. Calculate the projection coordinates of all sampling points considered to be within the target section, and sequentially connect each projection point along a preset direction to obtain a creepage distance arc within the target section.

[0111] 3. Calculate the length of the creepage distance arc in each section in a loop and select the shortest length as the insulator creepage distance. In this embodiment, the method for calculating the length of the creepage distance arc in each section includes the following steps, namely steps S6 to S9.

[0112] S6. Divide the generated creepage distance arc into a flat portion and a continuous curved portion.

[0113] S7. Perform straight line fitting on the continuous flat parts using the least squares method to calculate the length of the flat part of the connecting line.

[0114] S8. Fit a spline curve to the continuous curved portion using the Gauss-Newton method, and sequentially obtain the coordinates of multiple points on the spline curve at intervals of a preset measurement accuracy. Then, connect each adjacent point and add up the lengths to obtain the length of the continuous curved portion.

[0115] S9. The length of the creepage distance arc shall be the sum of the lengths of the continuous flat portion and the continuous curved portion.

[0116] Example 2

[0117] This embodiment provides an insulator creepage distance detection system based on visual recognition, which can apply the insulator creepage distance detection method based on visual recognition in Example 1. The detection system includes: a model building module, an arc generation module, and a distance calculation module.

[0118] The model building module is used to obtain three-dimensional point cloud data of the surface of an insulator to be tested, and to construct a three-dimensional model of the insulator based on the point cloud data fitting.

[0119] The arc generation module is used to generate the creepage distance arc of the insulator based on point cloud data and the insulator three-dimensional model.

[0120] The distance calculation module is used to calculate a shortest insulator creepage distance based on the generated creepage distance arc.

[0121] Example 3

[0122] This embodiment provides an insulator creepage distance testing device. This testing device is a computer device. The testing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements any of the steps of the aforementioned visual recognition-based insulator creepage distance testing method, thereby completing the creepage distance test of an insulator under test.

[0123] The computer device may be a smartphone, tablet computer, laptop computer, desktop computer, rack server, blade server, tower server, or cabinet server (including a standalone server or a server cluster consisting of multiple servers) capable of executing programs. The computer device of this embodiment includes at least, but is not limited to, a memory and a processor that can be interconnected via a system bus.

[0124] In this embodiment, the memory (i.e., readable storage medium) includes flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, magnetic disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of a computer device, such as the hard disk or internal memory of the computer device. In other embodiments, the memory may also be an external storage device of the computer device, such as a plug-in hard disk equipped with the computer device, a smart media card (SMC), a secure digital (SD) card, a flash memory card, etc. Of course, the memory may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory is generally used to store the operating system and various application software installed on the computer device. In addition, the memory may also be used to temporarily store various types of data that have been output or are about to be output.

[0125] In some embodiments, the processor can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. Such a processor is typically used to control the overall operation of a computer device. In this embodiment, the processor is used to execute program code stored in a memory or process data to implement the process of the insulator creepage distance detection method described in Example 1.

[0126] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned 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 specification.

[0127] The above-described embodiments merely illustrate several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that a person skilled in the art would be able to make numerous modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for detecting creepage distance of insulators based on visual recognition, characterized in that: Including steps:

1. Obtaining three-dimensional point cloud data of the surface of an insulator to be tested, and constructing a three-dimensional model of the insulator based on the point cloud data; wherein the insulator includes a column and a plurality of sheds; 2. Generating a creepage distance arc of the insulator based on the point cloud data and the three-dimensional model of the insulator; wherein the method for generating the creepage distance arc comprises the following steps: S1. The point cloud data is calculated multiple times using a random consistency check method to obtain multiple different cylinder parameters; the cylinder parameters include the cross-sectional diameter and center line of the cylinder; S2. Constructing multiple virtual cylinders according to different cylinder parameters, and comparing the coordinates of all sampling points in the point cloud data with the cylindrical coordinates of each virtual cylinder in turn, and then counting the number of sampling points belonging to the cylindrical surface on different virtual cylinders; S3. Select the virtual cylinder with the largest number of cylinder sampling points as the target cylinder; S4. The target cylinder is sliced ​​radially to obtain multiple cylinder sections; S5. By calculating and processing multiple cylindrical sections, a creepage distance arc is generated within each section; 3. Calculate the length of the creepage distance arc in each section cyclically, and select the shortest length as the insulator creepage distance; the calculation method of the creepage distance arc length in each section includes the following steps: S6. The generated creepage distance arc is divided into a continuous flat portion and a continuous curved portion; S7. A linear fit is performed on the continuous flat portion by the least squares method to calculate the length of the continuous flat portion; S8. Fitting a spline curve to the continuous curved portion using the Gauss-Newton method, and sequentially obtaining the coordinates of multiple points on the spline curve at intervals of a preset measurement accuracy, then connecting each adjacent point and adding up the lengths to obtain the length of the continuous curved portion; S9. Taking the sum of the lengths of the continuous flat portion and the continuous curved portion as the length of the creepage distance arc.

2. The insulator creepage distance detection method based on visual recognition according to claim 1, characterized in that: In step S1, the method of calculating the point cloud data each time to obtain the corresponding cylinder parameters includes the following steps: S11. Randomly select two sampling points in the point cloud as the starting point and the end point, and calculate the spatial coordinates of the starting point and the end point. and Get a straight line in space passing through two points; S12. Analyze the linear distance between each sampling point in the point cloud and the spatial straight line in sequence. When the linear distance between any sampling point and the spatial straight line is d Less than a preset distance threshold δ When , the sampling point is determined to be a point near the spatial straight line and included in the statistics; the spatial coordinates of the sampling point are ; S13. When the statistical number of sampling points near the spatial line N 0 is greater than a preset number threshold M When 0, the spatial line is marked as a candidate line belonging to the cylinder surface line; S14. Establishing a vertical plane 1 located at the starting point and a vertical plane 2 located at the end point; both the vertical plane 1 and the vertical plane 2 are perpendicular to the candidate line; S15. Count the number of first sampling points located on the vertical plane 1 and the vertical plane 2 respectively. N 1 and the number of second sampling points N 2; S16. When the number of first sampling points N 1 and the number of second sampling points N 2 are both greater than a preset threshold M 1, all the first sampling points and the second sampling points are fitted into circles respectively, thereby obtaining the first fitting circle and the second fitting circle; wherein the diameter ratio of the first fitting circle to the second fitting circle satisfies a preset ratio. σ , and the number of sampling points on the circumference of the first fitting circle and the second fitting circle is greater than a preset number threshold M 2; S17. Taking the average of the diameters of the first fitting circle and the second fitting circle as the cross-sectional diameter of the virtual cylinder d 0, and then extend the cross-sectional diameter from the starting point to the center of the first fitting circle d The length of 0 gets the center point one C 1. Extend the cross-sectional diameter from the end point to the center of the second fitting circle d The length of 0 gets the center point 2 C 2, and then the center point C 1 and center point 2 C 2 is used as the center line of the virtual cylinder.

3. The insulator creepage distance detection method based on visual recognition according to claim 2, characterized in that: The straight-line distance d The expression formula is: 。 4. The insulator creepage distance detection method based on visual recognition according to claim 3, characterized in that: The expression equation of the vertical plane 1 is: The expression equation of the vertical plane 2 is: 。 5. The insulator creepage distance detection method based on visual recognition according to claim 4, characterized in that: In step S4, the specific process of slicing the cylinder is as follows: S41. Determine slice parameters: Set any point on the surface of the cylinder P As the starting point for calculation, and set the number of slices to N 3. Section interval angle θ And the distance threshold is ε ; Among them, the section interval angle θ =360° / N 3; P The foot of the perpendicular point to the center line of the cylinder is ; Initial tangent v for: ; Expressed as: S42. Use the slice parameters to divide the cylinder into N 3 sections.

6. The insulator creepage distance detection method based on visual recognition according to claim 5, characterized in that: In step S5, the method for generating the creepage distance arc in each section includes the following steps: S51. According to the section interval angle θ Rotating the initial tangent line around the center line of the cylinder to obtain a new tangent line vector, and then obtaining a target section perpendicular to the new tangent line; The new tangent vector is ; is the centerline vector of the cylinder; the expression equation of the target section is: ; S52. Calculate the distance between each sampling point on the cylinder surface and the target section in sequence, and determine whether the distance is less than the distance threshold. ε , then the sampling point is regarded as a point within the target section; S53. Calculate the projection coordinates of all sampling points considered to be within the target section, and sequentially connect each projection point along a preset direction to obtain a creepage distance arc within the target section.

7. The insulator creepage distance detection method based on visual recognition according to claim 1, characterized in that: In step 1, the method for constructing the three-dimensional model of the insulator includes the following steps: (1) Establishing an insulator geometry library to assist in point cloud computing; (2) using a three-dimensional laser scanner to collect initial point cloud data of the insulator to be tested multiple times; (3) Aligning and splicing the initial point cloud data collected multiple times to form complete point cloud data; (4) locating control points in the laser-marked image corresponding to the complete point cloud data, and performing template matching on the laser-marked image; (5) Extracting features from the complete point cloud data, and performing component fitting modeling based on the extraction results, and then combining them into the three-dimensional model of the insulator.

8. An insulator creepage distance detection system based on visual recognition, characterized in that: The application is the insulator creepage distance detection method based on visual recognition according to any one of claims 1 to 7; the detection system comprises: A model building module is used to obtain three-dimensional point cloud data of the surface of an insulator to be tested, and to fit and build a three-dimensional model of the insulator based on the point cloud data; an arc generation module, configured to generate a creepage distance arc of the insulator based on the point cloud data and the three-dimensional model of the insulator; and The distance calculation module is used to calculate a shortest insulator creepage distance based on the generated creepage distance arc.

9. An insulator creepage distance detection device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; characterized in that: When the processor executes the program, the steps of the insulator creepage distance detection method based on visual recognition as described in any one of claims 1 to 7 are implemented, thereby completing the work of detecting the creepage distance of an insulator to be tested.

Images