A method for detecting wire breakage

Abstract

The application discloses a wire broken strand detection method, comprising the following steps: a front-end device collects images in real time; target detection of the wire is carried out on the images, the wire region is highlighted, and processing is carried out through color model conversion and graying; the wire edge in the wire region is extracted, the wire region in the image is obtained, the wire region is cropped to obtain a contour map and an image filled by graying of the contour; the image of the wire body is cut into several cut parts; the area of each cut part and the area ratio of the wire region area to the area of the corresponding cut part are calculated; a threshold interval of the area ratio is set, and whether the area ratio of each cut part of the wire and the area ratio of the adjacent cut part exceed the threshold interval of the area ratio is judged. The application avoids that the wire image is affected by backlight, the wire surface color is dark, the texture feature is blurred, and the gray scale change is not obvious, so that the wire broken strand detection is affected.

Description

Technical Field

[0001] This invention relates to the field of conductor strand breakage detection technology, and more specifically, to a method for detecting conductor strand breakage. Background Technology

[0002] To ensure a reliable and stable power supply to the power grid, it is essential to monitor distribution lines in real time and take timely measures to resolve any abnormalities. Conductors are the main equipment in the entire power grid, responsible for transmitting electrical energy; their safe operation is crucial to the stability of the entire grid. One of the major defects in transmission line conductors is broken strands.

[0003] Traditional image detection of conductor defects in power transmission lines relies on deep learning models. However, due to limitations in training conditions, these models are prone to false detections. For images containing conductor targets, color model conversion and image enhancement are necessary to facilitate effective detection of broken strands. However, images of conductors taken from a side angle are affected by sunlight, resulting in higher brightness at the top and lower clarity at the bottom, hindering conductor segmentation and extraction. Images taken from a low angle are backlit, leading to darker conductor surfaces, blurred textures, and indistinct grayscale variations. Images taken along the conductor's direction or from a slightly downward angle show uniform brightness distribution and clear individual strand outlines, but the complex background, including trees, traffic, and buildings, presents challenges for conductor edge extraction.

[0004] like Figure 1 The image shown is of the conductor taken from three different angles. Figure 1 In the middle left image, it is clear that the conductor is divided into two parts, one light and one dark, and there is a more complex background below the conductor than above it. Figure 1 The image on the right is a bottom view of the conductor. The background is simple, mostly the sky, but the texture of the conductor surface is not obvious due to the backlight. Figure 1 The lower image shows a top-down view of the conductor. In transmission lines, the front-end camera is usually positioned slightly above one end of the conductor. From this angle, the background is complex, but the surface texture of the conductor is uniform and clear. However, due to the camera's focal length, the background at this angle is relatively blurry, making it difficult to extract the conductor area and affecting strand breakage detection. Summary of the Invention

[0005] The purpose of this invention is to provide a method for detecting broken strands in a conductor, avoiding the impact of backlighting, dark surface color, blurred texture features, and insignificant grayscale changes on conductor images that affect the detection of broken strands.

[0006] To achieve the above objectives, a method for detecting broken strands in a conductor is provided, comprising the following steps:

[0007] (1) The front-end device acquires images in real time;

[0008] (2) Target detection of wires is performed on the image to highlight the wire area, and the image is processed by color model conversion and grayscale conversion;

[0009] (3) Extract the edges of the conductor within the conductor region to obtain the conductor region in the image, and crop the conductor region to obtain the contour map and the grayscale contour-filled image:

[0010] (4) Divide the image of the main body of the conductor into several segments;

[0011] (5) Calculate the area of ​​each segment and the ratio of the area of ​​the conductor region to the area of ​​the corresponding segment;

[0012] (6) Set a percentage threshold range and determine whether the percentage of each cut part of the conductor exceeds the percentage threshold range. If it is not within the percentage threshold range, the conductor in the cut part is considered to have a broken strand abnormality; otherwise, the conductor in the cut part is considered to have no abnormality.

[0013] Specifically, the front-end device is a pan-tilt camera.

[0014] Specifically, in step (2), the method for detecting the target wire in the image is as follows: the region of the target wire is identified by the YOLOv5 algorithm, and the region is taken as the region of interest.

[0015] Specifically, in step (3), extracting the wire edge within the wire region involves extracting the wire contour in the region of interest using the Sobel edge detection algorithm.

[0016] In particular, step (3) uses an edge detection algorithm based on the Sobel operator to achieve accurate segmentation and extraction of the wire region in the image.

[0017] Specifically, in step (4), each of the several sections of the conductor body is a section of equal size.

[0018] Specifically, in step (2), the color model conversion uses the HSV model conversion, and the specific method is to apply the conversion formula:

[0019]

[0020]

[0021] V = MAX (1-3);

[0022] Where H is the hue parameter value, S is the saturation parameter value, and V is the brightness parameter value. Among the three RGB components, the maximum value is represented by MAX and the minimum value is represented by MIN.

[0023] Specifically, in step (2), the grayscale conversion method is as follows:

[0024] Set a traversal template to traverse the entire image. During the traversal, calculate the mean grayscale value, variance, and normalized value of each pixel within the traversal template, using the following formula:

[0025]

[0026] Where Q is the average grayscale value of each pixel within the template, Q i This involves iterating through the grayscale value of the i-th pixel within the template.

[0027]

[0028] Where q5 is the gray value variance of a heavily shaded pixel. By traversing the entire image using the above method, the complete gray value variance map corresponding to the original image can be obtained. The variance of gray values ​​within the obtained template is then used to replace the gray value of the center pixel of the heavily shaded part in the image.

[0029] The beneficial effects of the present invention are as follows:

[0030] This invention proposes a method for detecting broken strands in conductors. A front-end device acquires images in real time; the image is processed to detect the conductor, highlighting the conductor region, and then processed through color model conversion and grayscale conversion. HSV model conversion further emphasizes the conductor, and grayscale processing ensures that surface details are preserved while effectively resisting interference from similar colors and distant complex backgrounds during conductor segmentation. The conductor edges within the conductor region are extracted to obtain the conductor region in the image. The conductor region is then cropped to obtain a contour map and a grayscale contour-filled image. Contour filling makes subsequent segmentation calculations more standardized. The image of the conductor body is divided into several segments. The area of ​​each segment and the ratio of the conductor region area to the corresponding segment area are calculated. A ratio threshold range is set, and it is determined whether the ratio of each segment of the conductor exceeds the ratio of adjacent segments. If it does not exceed the ratio threshold range, the conductor in that segment is considered to have broken strands; otherwise, the conductor in that segment is considered to be normal.

[0031] This invention avoids the problem of backlighting, which can cause the conductor surface to appear dark, the texture features to be blurred, and the grayscale changes to be indistinct, thus affecting the detection of broken strands in conductors. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 The images shown in the background are of a conductor taken from three different angles. The left image is of the conductor from a normal perspective, the right image is a bottom view of the conductor, and the bottom image is a top view of the conductor.

[0034] Figure 2 The images shown are before and after grayscale conversion in an embodiment of the present invention. The upper image is the original image, and the lower image is the image after grayscale conversion.

[0035] Figure 3 This is a sample image from an embodiment of the present invention.

[0036] Figure 4 The images shown are transformation diagrams of sample images after the method of this embodiment of the invention; the upper image is the contour map of the original sample image after target detection of the wire, the middle image is the grayscale contour filling image, and the lower image is the image of the contour map divided into 7 segments. Detailed Implementation

[0037] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby providing a clearer and more explicit definition of the scope of protection of the present invention.

[0038] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0039] It should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use. They are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," and "third," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0040] Furthermore, terms such as "horizontal," "vertical," and "sag" do not imply that components must be absolutely horizontal or suspended, but rather that they can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal relative to "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0041] like Figure 1 As shown, a method for detecting broken strands in a conductor according to this embodiment includes the following steps:

[0042] (1) The front-end device acquires images in real time;

[0043] (2) Target detection of wires is performed on the image to highlight the wire area, and the image is processed by color model conversion and grayscale conversion;

[0044] (3) Extract the edges of the conductor within the conductor region to obtain the conductor region in the image, and crop the conductor region to obtain the contour map and the grayscale contour-filled image:

[0045] (4) Divide the image of the main body of the conductor into several segments;

[0046] (5) Calculate the area of ​​each segment and the ratio of the area of ​​the conductor region to the area of ​​the corresponding segment.

[0047] (6) Set a percentage threshold range and determine whether the percentage of each cut part of the conductor exceeds the percentage threshold range. If it is not within the percentage threshold range, the conductor in the cut part is considered to have a broken strand abnormality; otherwise, the conductor in the cut part is considered to have no abnormality.

[0048] Specifically, the front-end device is a pan-tilt camera.

[0049] Specifically, in step (2), the method for detecting the target wire in the image is as follows: the region of the target wire is identified by the YOLOv5 algorithm, and the region is taken as the region of interest.

[0050] Specifically, in step (3), extracting the wire edge within the wire region involves extracting the wire contour in the region of interest using the Sobel edge detection algorithm.

[0051] In particular, step (3) uses an edge detection algorithm based on the Sobel operator to achieve accurate segmentation and extraction of the wire region in the image.

[0052] Specifically, in step (4), each of the several sections of the conductor body is a section of equal size.

[0053] Specifically, in step (2), the color model conversion uses the HSV model conversion, and the specific method is to apply the conversion formula:

[0054]

[0055]

[0056] V = MAX (1-3);

[0057] Where H is the hue parameter value, S is the saturation parameter value, and V is the brightness parameter value. Among the three RGB components, the maximum value is represented by MAX and the minimum value is represented by MIN.

[0058] Specifically, in step (2), the grayscale conversion method is as follows:

[0059] Set a traversal template to traverse the entire image. During the traversal, calculate the mean grayscale value, variance, and normalized value of each pixel within the traversal template, using the following formula:

[0060]

[0061] Where Q is the average grayscale value of each pixel within the template, Q i This involves iterating through the grayscale value of the i-th pixel within the template.

[0062]

[0063] Where q5 is the gray value variance of a heavily shaded pixel. By traversing the entire image using the above method, the complete gray value variance map corresponding to the original image can be obtained. The variance of gray values ​​within the obtained template is then used to replace the gray value of the center pixel of the heavily shaded part in the image.

[0064] The detection of broken strands in wires is verified using sample images containing broken strands. (Sample images are shown.) Figure 3 As shown.

[0065] Edge detection is performed using an edge detection algorithm based on the Sobel operator, resulting in a contour image as follows: Figure 4 The above image and the grayscale outline-filled image are as follows: Figure 4 As shown in the Chinese image:

[0066] Divide the outline into multiple segments of equal area using a 70*100 grid, as shown below. Figure 4 The image below.

[0067] Statistically analyze the grayscale distribution within each segment contour and the area of ​​pixels (white region pixels) with a BGR value of (255, 255, 255) in each segmented image:

[0068] The pixel area is calculated segment by segment from left to right as follows: 1543, 1498, 1631, 1710, 2344, 1693, 1368;

[0070] The percentages of the area of ​​each individual segment are as follows:

[0071] 27.34%, 26.54%, 28.90%, 30.30%, 41.53%, 30.00%, 24.24%;

[0072] Calculate the area ratio of adjacent segments (the area ratio of the right segment divided by the area ratio of the left segment).

[0073] From left to right: 0.97, 1.09, 1.05, 1.37, 0.72, 0.81;

[0075] Using 1.2 and 0.8 as the upper and lower limits for judgment, the slope shows that positions below 0.8 and above 1.2 are precisely where the stock breakage occurs.

[0076] Although embodiments of the present invention have been described in conjunction with the accompanying drawings, the patent owner may make various modifications or alterations within the scope of the appended claims, as long as they do not exceed the protection scope described in the claims of the present invention, they shall be within the protection scope of the present invention.

Claims

1. A method of detecting a broken wire, the method comprising: The method comprises the following steps: (1) The front-end device collects images in real time; (2) The image is subjected to target detection of the conductor, the conductor area is highlighted, and the conductor area is processed through color model conversion and grayscale processing; (3) The conductor edge in the conductor area is extracted, the conductor area in the image is obtained, the conductor area is cropped to obtain a contour map and an image filled by grayscale contour filling; (4) The image of the conductor body is cut into several cut parts; (5) The area of each cut part and the proportion of the area of the conductor area to the area of the corresponding cut part are calculated; (6) The adjacent segment ratio is calculated, the adjacent segment ratio is the ratio of the right segment area ratio to the left segment area ratio, a threshold interval is set, and it is judged whether the adjacent segment ratio exceeds the threshold interval, if not, it is considered that the conductor of the adjacent segment has the abnormality of broken strands, otherwise, it is considered that the conductor of the adjacent segment has no abnormality. The front-end device adopts a pan-tilt camera.

2. The method of claim 1, wherein: In the step (2), the method for detecting the target conductor of the image is specifically that: the region of the target conductor is recognized by using a YOLOv5 algorithm, and the region is taken as a region of interest.

3. The method of claim 1, wherein: In the step (3), the conductor edge in the conductor area is extracted specifically by using a Sobel edge detection algorithm to extract the conductor contour in the region of interest.

4. The method of claim 1, wherein: In the step (3), the Sobel operator-based edge detection algorithm is used to realize accurate segmentation and extraction of the conductor area in the image.

5. The method of claim 1, wherein: In the step (4), each cut part of the conductor body is a cut part of equal size.

6. The method of claim 1, wherein: In the step (2), the color model conversion uses an HSV model conversion, and the specific method is to use a conversion formula:

7. The method of claim 1, wherein: Wherein, H is a parameter value of the hue, S is a parameter value of the saturation, and V is a parameter value of the brightness, in the RGB three components, the maximum value is represented by MAX, and the minimum value is represented by MIN. In the step (2), the specific method of the grayscale processing is as follows:

8. The method of claim 1, wherein: The whole image is traversed by setting a traversal template, in the traversal process, the mean value of the gray values of the pixels in the traversal template, the variance of each traversal point and the normalized value are calculated, and the formula is as follows: ​ Wherein, Q is the mean value of the gray scale of each pixel in the template, Q i is the gray scale of the i-th pixel in the template. wherein, is the variance of the gray value of a heavy shadow pixel point, and the complete gray variance map corresponding to the original image can be obtained by traversing the complete image by the above method, so as to replace the gray value of the center pixel point of the heavy shadow part in the image with the variance of the gray value obtained in the template.

Images