Disconnecting link state diagnosis method and system based on image data
Through the diagnostic method based on image data, the Canny algorithm and Hough linear detection determine the status of the isolating switch blade, which solves the problems of low accuracy and poor reliability of traditional monitoring methods, and achieves higher monitoring accuracy and stronger reliability.
Patent Information
- Application Number
- CN202411852071.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-16
- Publication Date
- 2025-05-13
AI Technical Summary
The traditional method of monitoring the status of the isolating switch knife switch has low accuracy, poor reliability and is susceptible to external influences.
The diagnostic method based on image data is used, and the edge feature extraction is performed using the Canny algorithm, combined with Hough linear detection, the upper and lower edges of the knife switch arm are determined, and the state of the knife switch is determined by comparing the slope calculation with the set threshold.
It improves the accuracy and reliability of the status monitoring of the isolating switch knife switch, and reduces the impact on external interference.
Smart Images
Figure CN119991550A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a switch state diagnosis method and system based on image data, belonging to the field of image recognition. Background Art
[0002] As the most used primary switch equipment in the power transmission and transformation system, the improvement of the intelligence level of the isolating switch is of great significance. Since the switch knife is exposed to the outdoor environment and the long-term opening and closing, the knife may rust and deform, the transmission components may fail, and other problems may occur, resulting in typical faults such as heating, inadequate opening and closing, and damage to insulators. At present, the judgment of the opening and closing state of the isolating switch knife in the power system depends on the observation of the main inspection personnel. On the one hand, it is impossible to identify the state of the isolating knife after the motor drive action in real time. On the other hand, judging the state only by naked eye observation is too subjective and lacks accuracy. Online status monitoring technology monitors the status of key components in real time through sensors, and uploads data to the monitoring unit or network for processing. It can warn of faults, reduce inspection pressure, reduce costs and improve the intelligence level of the power grid.
[0003] By monitoring the opening and closing status of the disconnector, and mastering the real-time status of the disconnector during operation, it is of great significance to ensure the normal and stable operation of the disconnector, reduce the equipment failure rate, and improve the maintenance efficiency. It can effectively avoid power system safety accidents caused by disconnector failures, which is in line with the development trend of strong smart grids. The indirect measurement method using angle sensors has low accuracy and poor reliability; direct measurement methods such as posture sensors are affected by electromagnetic interference, and have poor practicality and safety. The image recognition method does not require close contact with the switch equipment, and is less affected by electromagnetic interference, so it has good application prospects. Summary of the invention
[0004] The purpose of the present invention is to provide a switch status diagnosis method and system based on image data, so as to solve the problems of low accuracy, poor reliability and susceptibility to external influences of traditional monitoring devices for disconnect switches.
[0005] To achieve the above object, the solution of the present invention includes: The present invention discloses a knife switch state diagnosis method based on image data, comprising: using the Canny algorithm to extract edge features of an image containing a knife switch, using the Hough line detection method to detect straight lines in the edge features to determine the upper and lower edges of a knife switch arm, calculating the slope of the straight line where the upper and lower edges of the knife switch arm are located and comparing it with a set slope threshold to determine the state of the knife switch.
[0006] Furthermore, the Canny algorithm is an improved Canny algorithm, which replaces Gaussian filtering with bilateral filtering to improve edge preservation effect, and replaces dual threshold edge screening with maximum inter-class variance method to achieve adaptive threshold selection to enhance applicable scenarios.
[0007] Furthermore, the image is preprocessed before edge feature extraction, and the preprocessing includes image grayscale processing to reduce redundant information of the image and improving image contrast through histogram equalization.
[0008] The beneficial effects of the present invention are as follows: using the Canny algorithm to extract edge features of an image containing a knife switch, using the Hough line detection method to detect straight lines in the edge features to determine the upper and lower edges of the knife switch arm, calculating the slopes of the straight lines where the upper and lower edges of the knife switch arm are located and comparing them with a set slope threshold to determine the state of the knife switch, so that the monitoring accuracy of the isolating switch is higher, the reliability is stronger and it is not affected by the outside world.
[0009] A knife switch status diagnosis system based on image data includes a processor, wherein the processor executes the computer program to implement the following steps.
[0010] The Canny algorithm is further used to extract edge features from the image containing the knife switch, and the Hough line detection method is used to detect the straight lines in the edge features to determine the upper and lower edges of the knife switch arm. The slope of the straight line where the upper and lower edges of the knife switch arm are located is calculated and compared with the set slope threshold to determine the state of the knife switch.
[0011] Furthermore, the Canny algorithm is an improved Canny algorithm, which replaces Gaussian filtering with bilateral filtering to improve edge preservation effect, and replaces dual threshold edge screening with maximum inter-class variance method to achieve adaptive threshold selection to enhance applicable scenarios.
[0012] Furthermore, the image is preprocessed before edge feature extraction, and the preprocessing includes image grayscale processing to reduce redundant information of the image and improving image contrast through histogram equalization.
[0013] The present invention can achieve the same beneficial effects as the knife switch state diagnosis method based on image data. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 This is a Hough line detection effect diagram provided by an embodiment of the present invention; Figure 2 The embodiment of the present invention provides an improved Canny algorithm flow chart; Figure 3This is a diagram showing the effect of modifying the filtering mode to bilateral filtering provided by an embodiment of the present invention; Figure 4 This is an adaptive threshold edge extraction effect diagram provided by an embodiment of the present invention; Figure 5 This is an effect diagram after adding grayscale processing and histogram equalization provided by an embodiment of the present invention; Figure 6 The present invention provides a working flow chart of a knife switch status diagnosis system based on image data. DETAILED DESCRIPTION
[0015] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail in a clear and complete manner in conjunction with the accompanying drawings and embodiments.
[0016] Method Example: The knife switch state diagnosis method based on image data of this embodiment first uses the Canny algorithm to extract edge features of the image containing the knife switch, and then uses the Hough line detection method to detect the straight lines in the edge features to determine the upper and lower edges of the knife switch arm. Figure 1 , the white lines are the upper and lower outlines of the knife gate arm. Figure 1 (a) is the determination of the upper and lower edges of the knife gate arm using the Hough line detection method in state 1; Figure 1 (b) is the determination of the upper and lower edges of the knife gate arm using the Hough line detection method in state 2; Figure 1 (c) is the determination of the upper and lower edges of the knife gate arm using the Hough line detection method in state 3. The slope of the straight line where the upper and lower edges of the knife gate arm are located is calculated and compared with the set slope threshold to determine the state of the knife gate. The edge feature extraction is performed using the Canny algorithm. The steps of the Canny algorithm to extract edge features are as follows: (1) Gaussian filtering: Gaussian filtering is performed on the preprocessed image to smooth the image and remove noise. Gaussian filtering is a linear smoothing filter whose weight coefficient is determined by a Gaussian function. Through Gaussian filtering, high-frequency noise in the image can be reduced while retaining the edge information of the image.
[0017] (2) Calculate the gradient magnitude and direction: Calculate the gradient magnitude and direction of each pixel in the image. The gradient magnitude indicates the intensity of the change in the grayscale value of the pixel in the image, while the gradient direction indicates the direction of the change in the grayscale value. This step is the key to edge detection because the edge is usually located at the location where the grayscale value changes most dramatically.
[0018] (3) Non-maximum suppression: Only the local maximum value in the gradient direction is retained, and other values are suppressed to 0. This step can further reduce the noise points in the image and make the edges clearer.
[0019] (4) Double threshold edge screening: Set two thresholds minVal and maxVal. For each pixel in the image, if its gradient amplitude is greater than maxVal, it is marked as a strong edge pixel; if the gradient amplitude is less than minVal, it is marked as a non-edge pixel; if the gradient amplitude is between the two, it is marked as a weak edge pixel. Next, the weak edge pixels are further processed. If the weak edge pixel is connected to the strong edge pixel (that is, there is a strong edge pixel in the 8-connected region), it is retained as an edge pixel; otherwise, it is suppressed to 0. This step can connect broken edges and remove isolated noise points.
[0020] In order to achieve better edge preservation and adapt to more scenarios, an implementable approach is to improve the Canny algorithm. Figure 2 The improved Canny algorithm flow chart is shown below. The Gaussian filter is changed to a bilateral filter, which makes the edges clearer when reducing the noise of the image. Figure 3 The left part is the original image. Figure 3 The right part in the middle is the effect of the original image after bilateral filtering; the dual threshold edge screening method is changed to the maximum inter-class variance method to realize adaptive threshold edge screening, which can improve flexibility and adapt to more application scenarios. Figure 4 (a) is the original image. Figure 4 (b) shows the effect of adaptive threshold edge screening.
[0021] In order to improve the effect of feature extraction, as an implementable method, grayscale processing and histogram equalization steps are added before feature extraction. Grayscale processing can reduce the redundant information of the image; histogram equalization can improve the contrast. Figure 5 (a) is the original image. Figure 5 (b) is the effect of the original image after histogram equalization. Figure 5 (c) is the grayscale image of the original image after grayscale processing. Figure 5 Middle (d) is the effect of grayscale image after histogram equalization.
[0022] System Example: The knife switch status diagnosis system based on image data of this embodiment, Figure 6 This is a workflow diagram of the knife switch state diagnosis system based on image data. The visible light knife switch image is collected by a visible light camera, and after the knife switch image is received by the processor, the knife switch state can be judged according to the method in Example 1, which will not be described in detail here.
Claims
1. A switch status diagnosis method based on image data, characterized in that: include: The Canny algorithm is used to extract edge features of the image containing the knife switch. The Hough line detection method is used to detect the straight lines in the edge features to determine the upper and lower edges of the knife switch arm. The slope of the straight line where the upper and lower edges of the knife switch arm are located is calculated and compared with the set slope threshold to determine the state of the knife switch.
2. The method for diagnosing knife switch status based on image data according to claim 1, characterized in that: The Canny algorithm is an improved Canny algorithm. The improved Canny algorithm replaces Gaussian filtering with bilateral filtering to improve edge preservation effect, and replaces dual threshold edge screening with maximum inter-class variance method to achieve adaptive threshold selection to enhance applicable scenarios.
3. The switch status diagnosis method based on image data according to claim 1 is characterized in that: The image is also preprocessed before edge feature extraction, and the preprocessing includes image grayscale processing to reduce redundant information of the image and improving image contrast through histogram equalization.
4. A switch status diagnosis system based on image data, comprising a processor, characterized in that: The processor executes the computer program to implement the following steps.
5. The knife switch status diagnosis system based on image data according to claim 4 is characterized in that: include: The Canny algorithm is used to extract edge features of the image containing the knife switch. The Hough line detection method is used to detect the straight lines in the edge features to determine the upper and lower edges of the knife switch arm. The slope of the straight line where the upper and lower edges of the knife switch arm are located is calculated and compared with the set slope threshold to determine the state of the knife switch.
6. The knife switch status diagnosis system based on image data according to claim 4 is characterized in that: The Canny algorithm is an improved Canny algorithm. The improved Canny algorithm replaces Gaussian filtering with bilateral filtering to improve edge preservation effect, and replaces dual threshold edge screening with maximum inter-class variance method to achieve adaptive threshold selection to enhance applicable scenarios.
7. The knife switch status diagnosis system based on image data according to claim 4 is characterized in that: The image is also preprocessed before edge feature extraction, and the preprocessing includes image grayscale processing to reduce redundant information of the image and improving image contrast through histogram equalization.