Power transmission line porcelain insulator positioning method and system based on main shaft detection

Abstract

The application provides a power transmission line porcelain insulator positioning method and system based on a main shaft detection, which comprises the following steps: acquiring an infrared detection image of a power transmission line porcelain insulator; performing gray scale processing and edge detection on the infrared detection image; performing Hough transform straight line detection based on the result of the edge detection, wherein, in the Hough transform detection, a long line segment representing the main shaft direction of the insulator string is acquired, and a main shaft line segment is selected from the long line segment; the whole infrared detection image is rotated by using the rotation angle of the main shaft line segment to eliminate the rotation of the insulator string itself; the threshold value of the binaryzation of the rotated image is determined to obtain an adaptive segmentation image, the segmentation image is further marked to obtain a marked image, and the marked image is equivalent to the insulator region positioned. The method can realize fast and accurate positioning of the whole insulator string, and greatly improves the efficiency and reliability of the infrared image porcelain insulator zero value diagnosis.

Description

Technical Field

[0001] This invention belongs to the field of insulator positioning technology, and particularly relates to a method and system for positioning porcelain insulators of transmission lines based on spindle detection. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Insulators, as the most widely used components in power systems, play a vital role in connecting, supporting, and insulating transmission lines. A complete and healthy string of insulators ensures the safety of line workers and the safe and stable operation of transmission and transformation equipment. Statistical studies on insulation accidents in power systems have found that insulator failure, especially porcelain insulator failure, is one of the most serious faults in overhead transmission lines, causing extremely adverse effects on the safety of power transmission lines and the surrounding environment.

[0004] Currently, the main methods for detecting deteriorated insulators include ultrasonic measurement, insulation resistance testing, distributed voltage testing, infrared imaging, and ultraviolet imaging. Frontline workers on power lines commonly use contact-based methods such as distributed voltage testing, which requires manual pole climbing and zero-point checking. This is a massive workload, posing a serious threat to personal safety. Furthermore, limitations imposed by the length and location of the insulator strings can lead to missed or false detections due to operational errors. Infrared detection for determining low zero values ​​in insulators has become increasingly widely used in recent years due to its advantages such as long-distance operation, ease of use, and safety.

[0005] However, reliable intelligent diagnostic methods are still lacking for zero-detection diagnosis based on infrared images of porcelain insulators. The key to judging temperature changes lies in the accurate positioning of the porcelain insulators in the image. In actual live-line testing, to achieve good detection results, the detection angles of porcelain insulators in infrared images vary, and the proportion of the entire string of porcelain insulators also varies, which makes accurate positioning of the porcelain insulators difficult. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention provides a method for locating porcelain insulators in transmission lines based on spindle detection, which can achieve rapid and accurate positioning of the entire string of insulators, greatly improving the efficiency and reliability of zero-value diagnosis of porcelain insulators using infrared images.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0008] Firstly, a method for locating porcelain insulators in transmission lines based on spindle detection is disclosed, including:

[0009] Acquire infrared images of porcelain insulators in power transmission lines;

[0010] Perform grayscale processing and edge detection on infrared detection images;

[0011] Based on the edge detection results, Hough transform line detection is performed. In the Hough transform detection, long line segments representing the main axis direction of the insulator string are obtained, and the main axis line segment is selected from the long line segments.

[0012] The rotation angle of the main axis segment is used to rotate the entire infrared detection image to eliminate the rotation of the insulator string itself;

[0013] The binarization segmentation threshold of the rotated image is determined to obtain an adaptive segmented image. Further labeling processing is performed on the segmented image to obtain a labeled image, which can be regarded as the insulator region obtained by localization.

[0014] As a further technical solution, if the porcelain insulator of the transmission line is a double-string structure, the number of horizontal white pixels in the located insulator area is counted, and the two strings are distinguished based on the statistical results.

[0015] As a further technical solution, the infrared detection image of the porcelain insulator of the transmission line is acquired by an infrared thermal imager. The temperature matrix output by the infrared thermal imager is stored as a pseudo-color image in the form of an iron red color palette. The resolution of the image is W×H, where W is the number of columns and H is the number of rows.

[0016] As a further technical solution, when using an infrared thermal imager to collect data, the object being measured, i.e., the porcelain insulator, should be located at the center of the entire infrared image layout, and at the same time, the position of the iron cap of each porcelain insulator should be visible.

[0017] As a further technical solution, among all the line segment results of the Hough transform straight line detection, the longest line segment is retained as the main axis segment retained in the final detection. This line segment basically connects most of the vertices of the entire string of insulators.

[0018] As a further technical solution, for the main axis segment, the segment [x1,y1,x2,y2,θ] is characterized by five parameters, where x1,y1 are the coordinates of the starting point X1 of the segment, x2,y2 are the coordinates of the ending point X2 of the segment, and θ is the rotation angle of the segment, defined as the angle between the segment and the positive horizontal direction in the pixel coordinate system.

[0019] As a further technical solution, the rotation angle of the main axis segment is used to rotate the entire infrared detection image, converting the entire insulator string into a horizontal position, so as to locate the porcelain insulator in the vicinity of the main axis position.

[0020] As a further technical solution, the segmented image is further labeled to obtain a labeled image, specifically as follows:

[0021] Extract the region of interest from the segmented image;

[0022] Extract the largest binarized region from the region of interest and take its smallest bounding quadrilateral to obtain the labeled image. The labeled image location is the gray quadrilateral labeled region.

[0023] As a further technical solution, the number of horizontal white pixels in the located insulator area is counted, specifically as follows:

[0024] Count the number of white pixels in the marked image in each row;

[0025] Based on statistical results, the boundary position between the two strings of insulators is obtained in the vertical direction.

[0026] As a further technical solution, it also includes: counting the number of white pixels in the horizontal direction for each segmented porcelain insulator string, specifically by column, counting the number of white pixels in each column;

[0027] Based on the statistical calculation results S H The width of each column of white area is obtained, the area near the peak is the position of the disk, and the area near the trough is the position of the iron cap;

[0028] For S H The smoothed curve is obtained by filtering using the moving average method, and then the same smoothing filter is applied again to obtain S'. H ;

[0029] For S' H Perform valley detection to obtain valley locations;

[0030] Record the x-axis coordinates and corresponding y-axis coordinates at the trough locations;

[0031] Cluster the y-axis coordinates into two classes. The group with the smaller mean y-axis coordinate is the x-axis coordinate of the location of the iron cap. The y-axis coordinate of these points in the column number is H. k / 2,H k For L k number of rows;

[0032] The detected point coordinates are remapped and re-mapped into the grayscale image to obtain the final position of the iron cap.

[0033] Secondly, a transmission line porcelain insulator positioning system based on spindle detection is disclosed, including:

[0034] The image acquisition module is configured to acquire infrared detection images of porcelain insulators of transmission lines;

[0035] The detection module is configured to perform grayscale processing and edge detection on the infrared detection image;

[0036] The main axis segment acquisition module is configured to: perform Hough transform line detection based on the edge detection results, wherein, in the Hough transform detection, long line segments representing the main axis direction of the insulator string are acquired, and the main axis segment is selected from the long line segments;

[0037] The rotation module is configured to rotate the entire infrared detection image using the rotation angle of the main axis segment to eliminate the rotation of the insulator string itself;

[0038] The localization module is configured to: determine the binarization segmentation threshold of the rotated image, obtain an adaptive segmented image, and further label the segmented image to obtain a labeled image, which is equivalent to the localized insulator region.

[0039] The above one or more technical solutions have the following beneficial effects:

[0040] This invention discloses a method for locating porcelain insulators in transmission lines based on main axis detection using infrared images. Combining the requirements of live-line testing standards, this method detects the main direction in the infrared image, rotates the entire image, and searches within the vicinity of the main axis to achieve rapid and accurate location of the entire string of insulators, greatly improving the efficiency and reliability of zero-value diagnosis of porcelain insulators using infrared images.

[0041] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0042] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0043] Figure 1 This is a typical infrared detection image of a porcelain insulator for a power transmission line according to an embodiment of the present invention;

[0044] Figure 2 This is a grayscale image of an embodiment of the present invention;

[0045] Figure 3 This is an example of edge detection results for an image converted to grayscale, as described in this invention.

[0046] Figure 4 This is a schematic diagram of the insulator vertex according to an embodiment of the present invention;

[0047] Figure 5 This is a diagram showing the detection results of the main shaft long segment in an embodiment of the present invention;

[0048] Figure 6 This is a de-rotation result image from an embodiment of the present invention;

[0049] Figure 7 This is an image binarization result from an embodiment of the present invention;

[0050] Figure 8 This is a region of interest image extracted based on the main axis position according to an embodiment of the present invention;

[0051] Figure 9 This is a schematic diagram showing the distribution of white pixels in the insulator region according to an embodiment of the present invention;

[0052] Figures 10(a)-10(b) This is a schematic diagram showing the separation result of two strings of insulators in an embodiment of the present invention;

[0053] Figures 11(a)-11(b) This is a schematic diagram comparing the horizontal statistical curves and smoothing results of an embodiment of the present invention;

[0054] Figure 12 This is a schematic diagram of the smooth curve trough extraction results according to an embodiment of the present invention;

[0055] Figure 13 This is a schematic diagram of the detection results in an embodiment of the present invention. Detailed Implementation

[0056] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0057] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0058] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0059] Example 1

[0060] This embodiment discloses a method for locating porcelain insulators in transmission lines based on spindle detection. By detecting the main direction in the infrared image and rotating the entire image, the method can quickly and accurately locate the entire string of insulators in the measurement image by searching in the vicinity of the spindle, which greatly improves the efficiency and reliability of zero-value diagnosis of porcelain insulators in infrared images.

[0061] To more clearly illustrate this embodiment, the specific positioning implementation process can be described in detail as follows:

[0062] In an implementation example, a typical infrared detection image of a porcelain insulator in a transmission line is as follows: Figure 1 As shown, the temperature matrix output by the infrared thermal imager is stored as a pseudo-color image in the form of an iron red palette. The image resolution is W×H, where W is the number of columns and H is the number of rows.

[0063] According to the requirements for on-site live-line testing, the object being tested, namely the porcelain insulator, should be located in the center of the entire infrared image layout, with clear edges, and the position of the iron cap of each porcelain insulator should be clearly visible.

[0064] Meanwhile, due to the location of porcelain insulators in transmission lines, a portion of the transmission line can be seen. Given the connection properties of porcelain insulators, for a 220kV porcelain insulator string, it can be considered that the entire string of insulators and connecting conductors are approximately straight lines with the same direction.

[0065] Therefore, in this embodiment, the principal axis of the porcelain insulator in the infrared image is first detected using Hough transform. Then, the detection result is used to rotate the entire image. Finally, the entire string and each insulator are located in the vicinity of the principal axis.

[0066] More specifically, the infrared image is first processed into grayscale. Based on the color distribution pattern of the iron oxide red color palette, the background in the image is dominated by the blue component. To better highlight the difference between the tested insulator and the background, only the red component in the RGB color space of the infrared image is selected, denoted as I. g ,like Figure 2 As shown.

[0067] Then for I g For edge detection, operators such as Sobel and Canny can be selected. Here, the Canny operator is used with a Gaussian filter size of 5×5, and the edge detection results are as follows. Figure 3 As shown.

[0068] Based on the edge detection results, Hough transform line detection is performed. From the edge detection map, it can be seen that the vertices of each insulator (like...) Figure 4 There is a certain distance between the circular markers, but the inclination angle of the line connecting the vertices is consistent with the direction of the conductor. Therefore, in the Hough transform detection, increasing the spacing threshold for merging different line segments allows these short lines to be merged into a longer line segment representing the main axis direction of the insulator string. Of all the detected line segment results, only the longest line segment is retained. The final detection result is as follows: Figure 5 As shown. Figure 5The white line segment is the main axis segment retained for final inspection. It can be seen that this line segment connects most of the vertices of the insulators in the entire string.

[0069] The test result is recorded as L. m The line segment [x1, y1, x2, y2, θ] can be characterized by five parameters, where x1, y1 are the coordinates of the starting point X1 of the line segment, x2, y2 are the coordinates of the ending point X2 of the line segment, and θ is the rotation angle of the line segment, defined as the angle between the line segment and the positive horizontal direction in the pixel coordinate system.

[0070] The rotation of the insulator string can be eliminated by rotating the entire image using θ. The rotated image is denoted as I. r .

[0071] Remember I g The coordinates of each point in the array are [x g ,y g ], I r The coordinates of this point are [x r ,y r ].

[0072] The transformation relationship between the two coordinates is:

[0073]

[0074]

[0075] Where [x] r,0 ,y r,0 [x] represents the original coordinates of the image after rotation. In actual rotation, there may be cases where the transformed coordinates are negative. Therefore, the original coordinates of the four vertices after rotation should be calculated, and negative coordinates should be translated to display all information from the original image in the rotated image. In the above formula, x... bound y bound This represents the minimum value of the original coordinates of the three vertices [0,H-1], [W-1,0], and [W-1,H-1] in the original image after rotation. The coordinate [0,0] remains [0,0] after rotation, so no comparison is needed.

[0076]

[0077] After rotation L m The starting and ending coordinates are marked as [x'1,y',x'2,y']. Since the insulator is placed in a horizontal direction, the y-coordinates are the same, both being y'.

[0078] like Figure 6 As shown, by converting the entire insulator string to a horizontal orientation, the insulators can be positioned more quickly and accurately within the vicinity of the main shaft. Figure 1 For example, in the image, the resolution is 640×480, the X1 coordinate is (90, 346), the X2 coordinate is (497, 201), and θ is -19°. After rotation, the X1 coordinate is transformed to (130, 357), and the X2 coordinate is transformed to (562, 357).

[0079] It should be noted that, due to the high degree of structural complexity of the entire insulator string, the overall method of this embodiment extracts the centerline of the main axis direction after detecting the main axis to compare and extract the temperature curve. The above conversion process can be equivalent to the fact that extracting relevant information of pixels on the diagonal line is more cumbersome. After converting to the horizontal direction, it is simply extracted by column within a fixed row.

[0080] Determining image I using Otsu's method r The binary segmentation threshold, denoted as I, can yield an adaptive segmentation result. b The segmentation results are as follows Figure 7 As shown.

[0081] Using x'1 and x'2 as the left and right edges, For the left and right edges, cut off I. b Region of Interest I R I R The number of rows is The number of columns is x'2-x'1+1, such as Figure 8 As shown. Further, extract I. R The image obtained by taking the largest binarized region in the image and its smallest bounding quadrilateral is denoted as I. c I c Location such as Figure 8 The medium gray quadrilateral markings can be considered as equivalent to the insulator areas located.

[0082] Generally, 220kV transmission line porcelain insulators are double-string structures, meaning there are usually two independent insulator strings within the insulator area. By counting the number of white pixels horizontally within the insulator area, the two strings can be effectively distinguished.

[0083] S v (i) = length(I) c =255), i=1,2,…H1

[0084] In the above formula, S v To perform statistics row by row, count the I values ​​in each row. c The number of pixels with a grayscale value of 255 (i.e., white dots), where H1 is I. c number of rows.

[0085] In this example, S v The result is as follows Figure 9As shown. Due to the presence of the double insulator string, the boundary between the two insulator strings is clearly visible in the vertical direction. Similarly, S can be obtained using Otsu's method. v The segmentation threshold t is the position of the dividing line separating two insulator strings.

[0086] The result after separation is as follows Figures 10(a)-10(b) As shown, the upper and lower strings of insulators are denoted as L. k , k = 1, 2.

[0087] By examining the structure of each porcelain insulator, it can be observed that the disc area is relatively wide, while the iron cap area, which is the connection point between each insulator, is relatively narrow. The widths of the disc and iron cap are approximately equal between different insulators in the image.

[0088] The number of white pixels in the horizontal direction is counted for each segmented porcelain insulator string.

[0089] S H (i) = length(L) k =255), i=1,2,…W1

[0090] In the above formula, S H To count the number of pixels with a grayscale value of 255 (i.e., white dots) in each column, W1 is L k The number of columns. S H The value is the width of the white area in each column of the graph. The area near the peak is the position of the disk, and the area near the trough is the position of the iron cap.

[0091] To reiterate, Sv refers to the vertical direction, counted by row, used to distinguish multiple strings, while SH refers to the horizontal direction, counted by column, used to distinguish different insulator plates.

[0092] For S H Filtering is performed using the moving average method, with a smoothing window length of w, which is usually an odd number. The smoothed curve is S. H1 The calculation process is as follows:

[0093]

[0094] For S H,1 Perform the same smoothing filtering process again to obtain S' H .

[0095] Because there are many burrs in Figure 10(a), and considering the structure of a single insulator, the disc surface is wide, while the iron cap, located in the middle of the crossbar, is relatively narrow. Extracting the troughs is to locate the iron cap. If the burrs are not smoothed out, multiple small troughs will appear, therefore large-scale smoothing is required.

[0096] Set the smoothing window to 5, and statistically analyze S in the horizontal direction of L1. H With S' H The comparison results are shown in Figure 11.

[0097] For S' H The valley detection process is as follows:

[0098] a. Calculate S' H First-order difference diff1

[0099]

[0100] b. Calculate S' H The second-order difference diff2

[0101] diff2(i)=diff1(i)-diff1(i-1)

[0102] c. Extracting the points where the second difference equals 2 is the trough. idx obtains the x-coordinate values ​​of all points where the second difference is 2.

[0103] idx = find(diff2 = 2)

[0104] Simultaneously, maximum suppression is applied to the troughs within the neighborhood: traverse idx and calculate the difference between the x-coordinates of two adjacent points. If the difference is less than 3, the two troughs can be considered as troughs within the neighborhood, representing a small perturbation of a trough. In this case, only S' is retained. H Points with smaller values. Extraction results are as follows: Figure 12 As shown, the '+' sign marks the extracted trough locations.

[0105] The above-mentioned trough detection step is a fixed step using second-order difference. The reason for performing maximum suppression is that the iron cap itself has a certain length, and the ultimate goal of this application's technical solution is simply to find a single location point within the iron cap on each insulator. The purpose of maximum suppression is to delete multiple troughs within the neighborhood of an iron cap, ensuring that only one point is extracted from each iron cap. In the final extraction result, the distance between each trough point should be relatively consistent.

[0106] The extracted x-axis coordinate value of the trough is denoted as X. v ={x v,i i = 1, 2, ..., n v},n v The number of troughs extracted; the corresponding y-axis coordinate is Y. v ={S' H (x v,i ), i = 1, 2, ..., n v}

[0107] At this point, the remaining troughs fall into two categories:

[0108] (1) The fluctuations on the chart;

[0109] (2) The location of the iron cap.

[0110] The two have significantly different widths, so these troughs are located in two relatively evenly distributed sequences.

[0111] Therefore, Y v The clusters were divided into two classes using the k-means clustering method, with Y being the most significant value. v The set of points with the smaller mean represents the x-axis coordinates of the location of the iron cap; these points are located on L. k The y-axis coordinate value is H. k / 2,H k For L k number of rows.

[0112] Finally, the coordinates of these detected points are remapped to image I following the steps described above. g In this way, the final position of the iron cap was obtained. The final test results are as follows: Figure 13 As shown in the figure, all detected locations of the iron caps are marked with blue dots.

[0113] In actual infrared detection images of porcelain insulators in transmission lines, defect analysis of porcelain insulators mainly relies on the temperature of the iron cap to characterize the temperature of the entire insulator. Therefore, based on detecting the entire string of insulators and locating the position of the iron cap, the temperature value of the corresponding pixel position can be obtained from the bottom temperature matrix of the infrared detection image according to the corresponding iron cap pixel coordinate value. This allows for the rapid acquisition of the temperature distribution of the entire string of porcelain insulators, thereby enabling defect analysis of the porcelain insulators.

[0114] Each infrared image stores a temperature matrix at its core, and the infrared image is a pseudo-color image drawn from the temperature matrix using a selected color palette. Once the location of each cap is achieved, the temperature value in the corresponding temperature matrix can be extracted using the location coordinates.

[0115] Example 2

[0116] The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method.

[0117] Example 3

[0118] The purpose of this embodiment is to provide a computer-readable storage medium.

[0119] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the above-described method.

[0120] Example 4

[0121] The purpose of this embodiment is to provide a transmission line porcelain insulator positioning system based on spindle detection, including:

[0122] The image acquisition module is configured to acquire infrared detection images of porcelain insulators of transmission lines;

[0123] The detection module is configured to perform grayscale processing and edge detection on the infrared detection image;

[0124] The main axis segment acquisition module is configured to: perform Hough transform line detection based on the edge detection results, wherein, in the Hough transform detection, long line segments representing the main axis direction of the insulator string are acquired, and the main axis segment is selected from the long line segments;

[0125] The rotation module is configured to rotate the entire infrared detection image using the rotation angle of the main axis segment to eliminate the rotation of the insulator string itself;

[0126] The localization module is configured to: determine the binarization segmentation threshold of the rotated image, obtain an adaptive segmented image, and further label the segmented image to obtain a labeled image, which is equivalent to the localized insulator region.

[0127] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0128] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0129] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for locating porcelain insulators in transmission lines based on spindle detection, characterized in that, include: Acquire infrared images of porcelain insulators in power transmission lines; Perform grayscale processing and edge detection on infrared detection images; Based on the edge detection results, Hough transform line detection is performed. In the Hough transform detection, long line segments representing the main axis direction of the insulator string are obtained, and the main axis line segment is selected from the long line segments. The rotation angle of the main axis segment is used to rotate the entire infrared detection image to eliminate the rotation of the insulator string itself; Determining image I using Otsu's method r The binary segmentation threshold, denoted as I, can yield an adaptive segmentation result. b Extract I b Region of Interest I R Extract I R The image obtained by taking the largest binarized region in the image and its smallest bounding quadrilateral is denoted as I. c I c It can be equivalently viewed as the insulator region obtained from the location; The number of horizontal white pixels in the insulator region is counted to distinguish the two strings: In the above formula, S v To perform statistics row by row, count the I values ​​in each row. c The number of pixels with a grayscale value of 255, H1 is I c number of rows; The two strings of insulators are denoted as L. k k=1,2; The number of white pixels in the horizontal direction is counted for each segmented string of porcelain insulators. In the above formula, S H To count the number of pixels with a grayscale value of 255 in each column, W1 is L k The number of columns; S H The value is the width of the white area in each column of the graph. The area near the peak is the position of the board, and the area near the trough is the position of the iron cap. For S H Filtering is performed using the moving average method, with a smoothing window length of [value missing]. w The smoothed curve is ,right Perform the same smoothing filtering process again to obtain ; right The valley detection process is as follows: a. Calculation First-order difference diff 1 b. Calculation Second-order difference diff 2 c. Extracting the points where the second difference equals 2 is the trough. idx obtains the x-coordinate values ​​of all points where the second difference is 2; Simultaneously, maximum suppression is applied to the troughs within the neighborhood: traverse idx, calculate the difference between the x-coordinates of two adjacent points; if the difference is less than 3, the two troughs can be considered as troughs within the neighborhood, representing a small perturbation of a trough, and in this case, only the maximum value of the trough is retained. Points with smaller values; The extracted trough x-axis coordinate value is denoted as , The number of extracted troughs; the corresponding y-axis coordinate is Will The classes were divided into two groups using the k-means clustering method. The set of points with the smaller mean represents the x-axis coordinates of the location of the iron cap; these points are located on L. k The y-axis coordinate value is H k / 2, H k For L k number of rows; The detected point coordinates are remapped and re-mapped into the grayscale image to obtain the final position of the iron cap.

2. The method for locating porcelain insulators of transmission lines based on spindle detection as described in claim 1, characterized in that, The acquired infrared images of porcelain insulators on transmission lines were obtained using an infrared thermal imager. The temperature matrix output by the infrared thermal imager was stored as a pseudo-color image in the form of an iron red color palette. The resolution of this image is [resolution missing]. , where W is the number of columns in the image and H is the number of rows in the image.

3. The method for locating porcelain insulators of transmission lines based on spindle detection as described in claim 1, characterized in that, When using an infrared thermal imager, the object being measured, i.e., the porcelain insulator, should be located in the center of the entire infrared image layout, and the position of the iron cap should be visible for each porcelain insulator.

4. The method for locating porcelain insulators of transmission lines based on spindle detection as described in claim 1, characterized in that, In the Hough transform line detection, the line segment with the longest retention length is the main axis segment retained in the final detection.

5. The method for locating porcelain insulators of transmission lines based on spindle detection as described in claim 1, characterized in that, For the main axis segment, five parameters characterize the segment. ,in Let X1 be the coordinate of the starting point of the line segment. Let X2 be the coordinate of the endpoint of the line segment. The rotation angle of the line segment is defined as the angle between the line segment and the positive horizontal direction in the pixel coordinate system. The entire infrared detection image is rotated by the rotation angle of the main axis segment, and the entire insulator string is converted to a horizontal orientation so that the porcelain insulator can be positioned in the vicinity of the main axis position.

6. A transmission line porcelain insulator positioning system based on spindle detection, characterized in that, include: The image acquisition module is configured to acquire infrared detection images of porcelain insulators of transmission lines; The detection module is configured to perform grayscale processing and edge detection on the infrared detection image; The main axis segment acquisition module is configured to: perform Hough transform line detection based on the edge detection results, wherein, in the Hough transform detection, long line segments representing the main axis direction of the insulator string are acquired, and the main axis segment is selected from the long line segments; The rotation module is configured to rotate the entire infrared detection image using the rotation angle of the main axis segment to eliminate the rotation of the insulator string itself; The positioning module is configured as follows: Determining image I using Otsu's method r The binary segmentation threshold, denoted as I, can yield an adaptive segmentation result. b Extract I b Region of Interest I R Extract I R The image obtained by taking the largest binarized region in the image and its smallest bounding quadrilateral is denoted as I. c I c It can be equivalently viewed as the insulator region obtained from the location; The number of horizontal white pixels in the insulator region is counted to distinguish the two strings: In the above formula, S v To perform statistics row by row, count the I values ​​in each row. c The number of pixels with a grayscale value of 255, H1 is I c number of rows; The two strings of insulators are denoted as L. k k=1,2; The number of white pixels in the horizontal direction is counted for each segmented string of porcelain insulators. In the above formula, S H To count the number of pixels with a grayscale value of 255 in each column, W1 is L k The number of columns; S H The value is the width of the white area in each column of the graph. The area near the peak is the position of the board, and the area near the trough is the position of the iron cap. For S H Filtering is performed using the moving average method, with a smoothing window length of [value missing]. w The smoothed curve is ,right Perform the same smoothing filtering process again to obtain ; right The valley detection process is as follows: a. Calculation First-order difference diff 1 b. Calculation Second-order difference diff 2 c. Extracting the points where the second difference equals 2 is the trough. idx obtains the x-coordinate values ​​of all points where the second difference is 2; Simultaneously, maximum suppression is applied to the troughs within the neighborhood: traverse idx, calculate the difference between the x-coordinates of two adjacent points; if the difference is less than 3, the two troughs can be considered as troughs within the neighborhood, representing a small perturbation of a trough, and in this case, only the maximum value of the trough is retained. Points with smaller values; The extracted trough x-axis coordinate value is denoted as , The number of extracted troughs; the corresponding y-axis coordinate is Will The classes were divided into two groups using the k-means clustering method. The set of points with the smaller mean represents the x-axis coordinates of the location of the iron cap; these points are located on L. k The y-axis coordinate value is H k / 2, H k For L k number of rows; The detected point coordinates are remapped and re-mapped into the grayscale image to obtain the final position of the iron cap.

7. The transmission line porcelain insulator positioning system based on spindle detection as described in claim 6, characterized in that, The acquired infrared images of porcelain insulators on transmission lines were obtained using an infrared thermal imager. The temperature matrix output by the infrared thermal imager was stored as a pseudo-color image in the form of an iron red color palette. The resolution of this image is [resolution missing]. , where W is the number of columns in the image and H is the number of rows in the image.

8. The transmission line porcelain insulator positioning system based on spindle detection as described in claim 6, characterized in that, When using an infrared thermal imager, the object being measured, i.e., the porcelain insulator, should be located in the center of the entire infrared image layout, and the position of the iron cap should be visible for each porcelain insulator.

9. A computer 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, it implements the steps of the method described in any one of claims 1-5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it performs the steps of the method described in any one of claims 1-5 above.

Images