An overhead high-voltage wire high-temperature anomaly diagnosis algorithm

Abstract

The application provides an overhead high-voltage wire high-temperature anomaly diagnosis algorithm, acquires an infrared image of an overhead high-voltage wire, separates channels of the infrared image, and takes a G channel most sensitive to temperature as a gray-scale image; an adaptive gray-scale threshold processing method is used to obtain a gray-scale threshold; a watershed algorithm is used to process the infrared image and extract an ROI region; the position of a maximum connected domain is determined through the largest region of the contour in the image, and whether the image can be detected is judged through the position and size of the maximum connected domain, wherein the maximum connected domain in the image is an insulator; the conductor is detected through the position of the insulator; after the position of the conductor is determined, the current-carrying wire is detected according to the position of the conductor; the temperature of the conductor is obtained through a temperature measurement SDK provided by a UAV manufacturer, and whether the temperature at this time is abnormal is judged according to a temperature anomaly judgment standard of the State Grid. The application has the beneficial effects of strong anti-interference ability, good robustness, and the ability to realize overhead high-voltage conductor high-temperature anomaly detection in a complex environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image processing and pattern recognition, and particularly relates to an overhead high-voltage line high-temperature anomaly diagnosis algorithm. BACKGROUND

[0002] Overhead transmission lines usually have the characteristics of complex structure, complex installation, large scale and repairable. The reason for the complex structure and installation is that there are many elements constituting the line, mainly including overhead conductors, towers and stay wires, fittings, insulators, etc., and the change of the state of any one element will have a corresponding influence on the normal operation of the line. Its large scale means that the line spans a wide range of areas, and the climate along the line is complex and changeable, so the geographical factors in different places are difficult to be uniformly processed. As an indispensable part of power transmission, overhead transmission lines are particularly susceptible to internal and external disturbances and abnormal operation or even failure, which may even lead to shutdown and seriously interfere with the normal life order of the people. Therefore, early prevention, monitoring and system troubleshooting of overhead line faults play a key role in improving the quality, reliability and economy of power transmission.

[0003] The research on intelligent line patrol technology based on unmanned aerial vehicles started relatively late in China, and currently mainly focuses on the research and development of hardware devices. There are many interference items in the inspection process, and the images differ greatly due to the influence of the shooting habits of the unmanned aerial vehicle photographers, resulting in inaccurate detection results. SUMMARY

[0004] In view of the high difficulty and low detection efficiency of overhead high-voltage line inspection, the present application provides an overhead high-voltage line high-temperature anomaly diagnosis algorithm. In the daily inspection task of high-voltage transmission lines, the traditional inspection method is mainly manual inspection, and researchers at home and abroad have shifted their research direction from the initial robot inspection to unmanned aerial vehicle inspection technology. The use of unmanned aerial vehicles for inspection not only changes the defects of traditional manual line inspection in the power system, but also improves the inspection efficiency, reduces the cost and improves the attendance rate. The application of this method in the modern smart grid also indirectly indicates that the technology used in the field of power inspection is constantly improving. The overhead high-voltage line high-temperature anomaly diagnosis algorithm mainly includes the detection of conductors and current-carrying wires, and the processing flow only needs the on-site staff to send the images collected by the unmanned aerial vehicle aerial photography to the data center, and the data center returns the signal of whether the temperature is abnormal. This greatly saves the labor cost and accurately and effectively realizes the high-temperature anomaly diagnosis of overhead conductors. The algorithm mainly includes:

[0005] S1: Obtain the infrared image of the overhead high-voltage line by unmanned aerial vehicle aerial photography. It is found through experiments that the G channel in the RGB three channels is most sensitive to temperature, and the G channel image is obtained by channel separation and taken as a gray image;

[0006] S2: obtaining a gray threshold value of the G channel separation image obtained in step S1 by using an adaptive gray threshold processing method;

[0007] S3: processing the infrared image by using a watershed algorithm to extract an ROI region;

[0008] S4: after performing morphological processing on the ROI region, a horizontal direction edge detection template is proposed according to the characteristics of the conductor image, inspired by the traditional edge detection algorithm. It can be analyzed from the conductor image that the conductor is often in the horizontal direction and has a certain inclination. The width of the conductor is generally 2-3 pixels. The higher the gray value is, the more in the middle of the conductor.

[0009] The size of the template is 5*5. The template is convolved with the image to extract the boundary gradient of the image to improve the contrast of the straight line and the background. Finally, the image after the above operation is subjected to noise removal processing by median filtering;

[0010] S5: converting the ROI region obtained in step S3 into a binary image by using the gray threshold value obtained in step S2, and filling the noise points inside the insulator by using the closing operation in image morphology. All pixel points in the image are traversed, and the gray values of the surrounding pixel points are counted. For the case that there are many high pixel points around, it is considered as the inside of the image. Considering that the largest region in the image at this time is the insulator, the insulator is connected with the conductor and the drain wire, and the image is a binary image, the width of the conductor is generally 2-3 pixels; therefore, by traversing each pixel in the image, the number of high pixel points within a range of 3 pixels around the pixel is counted. When the number of high pixel points is small, the gray value of the point is set to 0. At this time, the conductor in the image will be corroded because the width is not enough, and only the insulator part with a larger area in the image is left. Then, the profile is screened to detect the insulator with the largest connected domain.

[0011] S6: positioning the conductor in the processing result image in step S4 by using the position of the insulator obtained in step S5, and positioning the processing result twice;

[0012] S7: after the conductor is positioned, the points on the conductor are traversed to determine the starting point of the drain wire and detect the drain wire according to the depth-first search algorithm;

[0013] S8: obtaining the temperature of the conductor by using the temperature measurement SDK provided by the unmanned aerial vehicle manufacturer, and determining whether the temperature is abnormal according to the temperature abnormality judgment standard of the State Grid.

[0014] Further, step S2 uses the gray feature to segment the highlight area and the dark area in the image to obtain the gray threshold value of the infrared image.

[0015] Further, step S3 uses the gray information of the infrared image to project in the horizontal direction, and uses the watershed algorithm and position information to distinguish the tower and insulator in the image.

[0016] Further, in step S4, according to the characteristics of the high gray of the wire center, a horizontal edge detection template is used:

[0017]

[0018] The horizontal line in the image is enhanced, and the broken wire condition that may occur during the gray threshold processing is repaired, so as to improve the contrast between the straight line and the background.

[0019] Further, in step S5, by traversing each pixel in the image, the number of high pixel points in the range of 3 pixels around the pixel is counted, and when the number of high pixel points is small, the gray of the point is set to 0, the wire in the image is completely corroded, only the insulator is left, and the insulator position coordinates are obtained by using contour detection.

[0020] Further, in step S6, the boundary of the insulator on the side of the wire is taken as the starting point, and the intersection of the wire and the image boundary is taken as the end point, and the position of the wire is determined according to the number of wire points on the connecting line as the criterion.

[0021] Further, in the final temperature anomaly detection, the temperature field data is needed to be obtained according to the wire coordinates to judge the temperature on the wire, but the convolution operation of morphological expansion, corrosion and filtering is used in the process of detecting the wire, so that the detected wire may deviate from the position of the original wire, and finally the detected wire needs to be positioned again. The method of secondary positioning is to determine the position of the detected wire according to the fact that the wire does not exist low temperature anomaly, so when the detected wire is positioned in the background, the highest gray coordinate in the 8-connected domain around the coordinate is taken as the new coordinate of the point.

[0022] Further, in step S7, after the position of the wire is determined, a 3*20 rectangular slider similar to a sliding window is used under the wire detected in the wire detection, all points on the wire are traversed in turn, and the current point is highly coincident with the slider (the coincidence rate is at least 50%) in the place where there is a current-carrying line, so as to determine the position of the starting point of the current-carrying line. The current-carrying line is obtained by traversing in a depth-first manner, and the slope is used as a constraint condition for traversal.

[0023] Further, in step S8, if all the detected wire temperatures are within the judgment range, the wire temperature is normal, otherwise the temperature is abnormal.

[0024] The technical scheme provided by the present application has the beneficial effects that the present application is suitable for high-temperature anomaly detection of overhead high-voltage conductors in a complex background, coordinates of the overhead conductors are accurately obtained through image enhancement, gradient edge extraction, horizontal projection and other methods, and corresponding temperatures are obtained through a temperature SDK provided by a UAV manufacturer, so that it is judged whether the overhead conductors are abnormal according to high-temperature anomaly criteria of the overhead conductors provided by the State Grid. The present application has strong anti-interference ability, good robustness, and can realize high-temperature anomaly detection of overhead high-voltage conductors in a complex environment. BRIEF DESCRIPTION OF DRAWINGS

[0025] The present application will be further described below in combination with the drawings and embodiments, and the drawings are as follows:

[0026] Figure 1 is a flowchart of an overhead high-voltage wire high-temperature anomaly diagnosis algorithm in an embodiment of the present application.

[0027] Figure 2 is a boundary graph after filtering and optimizing the conductors in an embodiment of the present application.

[0028] Figure 3 is an insulator detection result graph in an embodiment of the present application.

[0029] Figure 4 is a conductor detection result graph in an embodiment of the present application.

[0030] Figure 5 is an initial state of a hypothetical streamer in an embodiment of the present application.

[0031] Figure 6 is a second-order sub-pixel edge detection principle diagram in an embodiment of the present application.

[0032] Figure 7 is a final detection result graph in an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to have a clearer understanding of the technical features, objects and effects of the present application, the specific embodiments of the present application will be described in detail with reference to the drawings.

[0034] The embodiment of the present application provides an overhead high-voltage line high-temperature anomaly diagnosis algorithm. First, the infrared image of the high-voltage line is obtained by the DJI / KeYi unmanned aerial vehicle aerial photography and is transmitted to the data center. For any image, according to the processing result of the algorithm, it is divided into a conductor type and an unidentifiable type. The conductor type returns the fault detection result, and the unidentifiable type skips the detection step and directly enters the detection of the next picture. The present application mainly reflects in the detection of the conductor type. Specifically, the image is preprocessed, including channel separation, adaptive threshold processing, extraction of the region of interest (ROI), conductor image enhancement and detection of the maximum connected domain. After the preprocessing of the image, the conductor and the drain wire are detected. After the preprocessing is completed, the end points of the conductor in the boundary are judged according to the position and inclination of the insulator, the high gray value pixels between the end points in the boundary and the end points at the insulator are counted, which are the positions of the conductor. The drain wire is obtained by the distribution of the pixels below the coordinate points of the conductor. The starting point of the drain wire is obtained by the downward depth traversal, and the complete drain wire is obtained. According to the length of the obtained drain wire, the starting point of the drain wire is determined again. The temperature of all position points in the overhead conductor is extracted by the SDK of DJI / KeYi, and whether the overhead conductor has high-temperature anomaly is judged according to the criterion for high-temperature anomaly in the national grid standard. If the region of interest or the maximum connected domain has problems such as too large or too small region, it is considered that an unidentifiable image appears. The present application can accurately detect the position of the overhead conductor and judge whether there is high-temperature anomaly, which meets the industrial requirements.

[0035] Please refer to Figure 1 , Figure 1 is a flow chart of an overhead high-voltage line high-temperature anomaly diagnosis algorithm in the embodiment of the present application, specifically comprising:

[0036] S1: collecting the infrared image of the overhead conductor by unmanned aerial vehicle aerial photography;

[0037] S2: using the gray scale feature to segment the highlight region and the dark region in the image, calculating the gray scale threshold of the image to enhance the robustness of the algorithm;

[0038] S3: in order to obtain a better extraction of the insulator contour, the position information and the gray scale information of the image are used. The gray scale information of the image is projected in the horizontal direction to obtain one-dimensional data. Whether there is a tower and other interference items in the vertical direction is judged according to the number of high gray scale pixels in the data. Then, the watershed algorithm and the position information are used to distinguish the tower, the insulator, the ground and the sky in the image. The ROI region is obtained by cutting the image, and the interference of the tower and the ground in the background is eliminated.

[0039] S4: Inspired by traditional edge detection algorithm, according to the characteristics of overhead conductor image, a horizontal and vertical edge detection template is proposed. From the overhead conductor image, it can be analyzed that the conductor is often in the horizontal direction, with a certain degree of inclination. And the conductor width is generally 2-3 pixels. The higher the gray value is, the more in the middle of the conductor.

[0040] The detection template is used to enhance the horizontal line in the image, repair the possible conductor or drainage line fracture caused by gray threshold processing. The contrast between straight lines is improved, the median filter is used for denoising the processing result, and according to the characteristics of high gray value in the center of the conductor, the horizontal and vertical edge detection template is as follows:

[0041]

[0042] The size of the template is 5*5, each convolution kernel can be regarded as a combination of median filter and horizontal filter. The template is convolved with the image to extract the boundary gradient of the image, repair the possible conductor or drainage line fracture caused by gray threshold processing, and improve the contrast between straight lines and background. The extracted boundary graph is shown in Figure 2 .

[0043] S5: Since the traditional edge detection algorithm will produce many false edges when detecting edges, affecting the detection accuracy, and the insulator is just a whole in the image. After the image morphological closing operation is done, all pixel points in the image are traversed, and the gray values of the surrounding pixel points are counted. For the case of more high pixel points, it is considered as the inside of the image. Considering that the largest region in the image at this time is the insulator, the insulator is connected with the conductor and the drainage line, and the image is a binary image, and the conductor width is generally 2-3 pixels. Therefore, by traversing each pixel in the image, the number of high pixel points in the range of 3 pixels around the pixel is counted. When the number of high pixel points is small, the gray value of the point is set to 0, and at this time the conductor in the image will be completely eroded. The processing result is shown in the left image of Figure 3 . Then through the method of extracting the contour, the contour of the largest connected domain is extracted after the contour is constrained, and the largest connected domain is the insulator. The insulator detection result is shown in the right image of Figure 3 . The white box is the marked position of the insulator;

[0044] S6: After the edges of the candidate image are detected, the algorithm traces the adjacent edge pixels to form lines and converts the edge image to a line image using the 8-connected scan algorithm. The line image can be represented by a two-dimensional linked list. Each node in the outer linked list corresponds to a line in the image. Each node in the inner linked list corresponds to an edge pixel that makes up the line. image(n) represents the two-dimensional linked list containing n lines, and linei represents the ith line. The algorithm calculates the coordinates of the vertices of the minimum bounding rectangle of each line and extracts the following four features: the height (h) of the line, which is the height of the minimum bounding rectangle; the area (a) of the line, which is the area of the minimum bounding rectangle; the center position (p) of the line, which is the coordinates of the center of the minimum bounding rectangle; and the edge density (d) of the line, which is the ratio of the number of edge pixels to the area of the minimum bounding rectangle. Thus, each line can be represented by a four-dimensional vector (L).

[0045] After obtaining the line vector image, the algorithm performs preliminary filtering on the lines based on their features. By observing the edge image, it is noted that lines are generally parallel and the background is clean sky. Therefore, the following three rules can be used to filter out lines: 1. Some lines with a large number of pixels are removed, which are likely to be disturbed by other objects in the image, such as clouds in the sky, towers, and ground. 2. Non-parallel lines are removed, which are less likely to be conductors. 3. Since a conductor may be detected multiple times, lines that are very close to each other are removed, which are less likely to be conductors.

[0046] After preliminary filtering of the lines, adjacent lines need to be clustered to obtain candidate lines. The process of generating candidate lines uses an improved 8-connected region scanning algorithm, but unlike the previous section where the connected element is a pixel, the connected element here is a line. The improved 8-connected algorithm is described as follows: First, according to the scale factor γ (γ > 1), all lines in the original image with a width of W and a height of H are mapped to a line distribution thumbnail S with a width of M and a height of N. M = W / γ, N = H / γ. If the center position of the original line L is (x0, y0), and the cell at position (x, y) in S is mapped, the mapping rule is (x, y) = f(x0, y0) = (x0 / γ, y0 / γ). Each cell in the thumbnail S contains all lines in the original image that are mapped to this cell. It can be represented as 0)

[0047] S(x, y) = {L(x0, y0) | x = x0 / γ, y = y0 / γ}

[0048] ​Then, the cell S(x, y) is regarded as a scanning element, and the thumbnail S is divided into several connected regions, i.e. candidate conductor regions, by using an eight-connected scanning algorithm. Each region contains several adjacent lines. In order to avoid the interference of some noise lines near the conductor region, a higher value is set for the horizontal distance when generating the conductor connected region. After judging the number of conductors, the boundary of the insulator on the conductor side is taken as the starting point, and the intersection of the conductor and the image boundary is taken as the end point. The number of pixel points with higher gray values on these lines is counted. The conductor with the most pixels is the detected conductor, and the detection result is as shown in Figure 4 .

[0049] S7: Before the identification of the drainage line, the identification of the conductor needs to be completed first. For the detection of the drainage line, a depth-first search traversal method is designed. The depth-first search algorithm belongs to a blind search algorithm, and thus the search speed is relatively slow, and the search result is often not the optimal solution. Therefore, a pruning process is needed to optimize the algorithm. Pruning is to judge which nodes should be accessed and which nodes should be discarded in the search process, so as to avoid unnecessary search process and improve the search efficiency. The pruning process is used in most search algorithms.

[0050] Referring to the idea of pruning optimization design, the depth-first search algorithm is improved according to the pruning optimization principle, and the process is as follows:

[0051] In order to obtain the edge information of the drainage line, a sub-pixel edge detection method is used to obtain the two-dimensional contour point features of the line. Compared with the pixel-level detection method, the edge detected by the sub-pixel detection method is smoother, and the specific algorithm is as follows:

[0052] Since the edge of the drainage line cannot be a straight line, it is assumed that the initial condition is as shown in Figure 5 (a). A second-order curve is assumed to divide the image into two regions, i.e. a background and a drainage line, and E and F represent the pixel values of the background and the drainage line respectively. The middle part of the image is an uncertain pixel, as shown in Figure 5 (b). SE and SF are two intensity covering pixel regions, and h is the length of a unit pixel. Therefore, h 2 = SE + SF.

[0053] A 5-pixel x 3-pixel neighborhood centered on pixel D(i, j) is selected to ensure that the edge of the drainage line passes through the window from left to right. According to the initial assumption, the pixel size meets

[0054]

[0055] In the formula: P i,j (0≤Pi,j≤h 2 ) is the area under the edge line of pixel (i, j).

[0056] Let SL, SM and SR be the left, middle and right columns of pixels respectively. Figure 6

[0057]

[0058] where i, j are the horizontal and vertical coordinates of pixel D respectively; n is the number of pixels; L, M and R represent the pixel area in each column under the edge line. The sub-pixel position calculated by the known coefficients L, M and R on the vertical center line of the pixel (x = 0), the change of intensity on both sides and the normal vector are the same as the linear case. The curvature at x = 0 is calculated according to the above algorithm to measure the yarn edge and the curvature. Then these two-dimensional edge point sets are extracted.

[0059] According to the number of wires, find the starting point of the drainage line corresponding to each wire. Under the wire extracted in the wire detection, use a 3*20 rectangular slider similar to a sliding window to traverse all points on the wire in turn, and the drainage line will coincide with the height of the slider in the place with the drainage line, so the starting point of the drainage line will be obtained;

[0060] After obtaining the starting point, use the depth-first traversal algorithm to traverse the gray points downward, and use the constraint condition to make the traversal always on the drainage line;

[0061] In the process of traversing the drainage line, if the curvatures of the front and rear two points tend to 0, the retrieval of the drainage line will be ended;

[0062] Repeat the above steps and count the length of each drainage line, and the longest one is the position of the detected drainage line.

[0063] If multiple drainage lines need to be detected, the relative positions of each wire and the wire below will be counted to determine who should be on the left and who should be on the right. Then the drainage line measured by the insulator is fine-tuned from the previous starting point and the drainage line is searched again. In this way, multiple drainage lines can be avoided. Finally, the found drainage line is positioned again. The final detection result is shown in

[0064] Figure 7 As shown in the left figure, the white box represents the detected insulator, and the wires and drainage lines are also marked in the figure, but they coincide with the original figure, so in the right figure, the detected wires and drainage lines are marked with a black image as a background.

[0065] S8: Through the temperature measurement SDK provided by the unmanned aerial vehicle manufacturer, the temperature of the wire is obtained, and according to the temperature abnormality judgment standard of the State Grid, it is judged whether the temperature is abnormal or not, and the detection result is given. After the detection is finished, the algorithm will start to read the next infrared overhead high-voltage wire picture and start the high-temperature abnormality diagnosis of the wire again.

[0066] ​The application is suitable for high-voltage overhead conductor high-temperature anomaly detection in a complex background, coordinates of the overhead conductor are accurately obtained through image enhancement, gradient edge extraction, horizontal projection and other methods, and corresponding temperatures are obtained through a temperature SDK provided by a UAV manufacturer, so that it is judged whether the overhead conductor is abnormal according to overhead conductor high-temperature anomaly criteria provided by the State Grid. The application has strong anti-interference ability, good robustness and can realize high-voltage overhead conductor high-temperature anomaly detection in a complex environment.

[0067] The above description is only the preferred embodiment of the application, and is not intended to limit the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. An overhead high-voltage line high-temperature anomaly diagnosis algorithm, characterized in that: The method comprises the following steps: S1: Obtain an infrared image of an overhead high-voltage wire by aerial photography of a UAV, and obtain a G channel image by channel separation and take the G channel image as a gray image, because the G channel is most sensitive to temperature; S2: Obtain a gray threshold value of the G channel image obtained in step S1 by using an adaptive gray threshold processing method; S3: Process the infrared image by using a watershed algorithm to extract an ROI region; S4: After morphological processing of the ROI region, according to the characteristics of the wire image, a horizontal edge detection template is used to extract the boundary gradient of the image, and then the image after the above operation is subjected to noise removal processing by median filtering; S5: The ROI region obtained in step S3 is converted into a binary image by using the gray threshold value obtained in step S2, and the noise points in the insulator are filled by using a closing operation in image morphology, and all the pixel points in the image are traversed to count the gray values of the surrounding pixel points, and at this time, the largest region in the image is the insulator; S6: The position of the insulator obtained in step S5 is used to locate the wire in the processing result image obtained in step S4, and the processing result is located again; S7: After the wire is located, the points on the wire are traversed to determine the starting point of the current-carrying wire and detect the current-carrying wire according to a depth-first search algorithm; S8: The temperature of the wire is obtained by using a temperature measurement SDK provided by the UAV manufacturer, and it is judged whether the temperature is abnormal according to the abnormal temperature judgment standard of the State Grid; In step S5, by traversing each pixel in the image, the number of high pixel points in the range of 3 pixels around the pixel is counted, and when the number of high pixel points is less than a preset value, the gray value of the pixel is set to 0, the wire in the image is completely eroded, and only the insulator is left, and the position coordinates of the insulator are obtained by using contour detection; In step S6, the boundary of the insulator on the side of the wire is taken as the starting point, the intersection of the wire and the image boundary is taken as the end point, the 8-connected scanning algorithm is used to track adjacent edge pixels to form a line, and the position of the wire is determined according to the number of wire points on the line as a criterion; In step S6, the method of secondary positioning is that according to the fact that the wire does not have low temperature abnormity, so when the detected wire positioning fails, the coordinates of the highest gray value in the 8-connected domain around the coordinates are taken as the new coordinates of the point; The 8-connected scanning algorithm is as follows: First, according to the scale factor γ, γ>1, all lines in the original image with width W and height H are mapped into a line distribution thumbnail S with width M and height N; wherein M=W / γ, N=H / γ, if the center position of the original line L is (x0, y0), the position of the cell in S is (x, y), the mapping rule is (x, y)=f(x0, y0)=(x0 / γ, y0 / γ), each cell of the thumbnail S contains all lines in the original image which are mapped into this cell, denoted as: 0) ​ S(x, y) = {L(x0, y0) | x = x0 / γ, y = y0 / γ} Then, the cell S(x, y) is regarded as a scanning element, and the thumbnail S is divided into a plurality of connected regions, i.e. candidate wire regions, by using the 8-connected scanning algorithm, and each region contains a plurality of adjacent lines.

2. The overhead high-voltage line high-temperature anomaly diagnosis algorithm according to claim 1, characterized in that: In step S2, the gray threshold value of the infrared image is obtained by dividing the highlight area and the dark area in the image according to the gray feature.

3. The overhead high-voltage line high-temperature anomaly diagnosis algorithm according to claim 1, characterized in that: In step S3, the gray information of the infrared image is projected in the horizontal direction, and the position information is used to distinguish the tower and the insulator in the image by using the watershed algorithm.

4. The overhead high-voltage line high-temperature anomaly diagnosis algorithm of claim 1, characterized in that: In step S4, according to the characteristics that the center of the wire has a high gray value, a horizontal edge detection template shown in the following formula is used: The horizontal line in the image is enhanced, and the wire breakage that may occur during the gray threshold processing is repaired to improve the contrast between the straight line and the background.

5. The overhead high-voltage line high-temperature anomaly diagnosis algorithm according to claim 1, characterized in that: After the step S7 determines the position of the wire, a rectangular slider is used to traverse all the points on the wire in sequence under the wire extracted in the wire detection. Where there is a drainage line, the coincidence rate with the slider will be more than 50%, so as to determine the position of the starting point of the drainage line. The drainage line is obtained through the traversal in a depth-first manner, and the slope is taken as the constraint condition of the traversal.

6. The overhead high-voltage line high-temperature anomaly diagnostic algorithm of claim 1, characterized in that: In the step S8, if all the wire temperatures detected are within the judgment range, the wire temperature is normal, otherwise the temperature is abnormal.

Images