A switch opening and closing state recognition method and system

Abstract

The present application belongs to the technical field of intelligent monitoring of high-voltage switchgear, and relates to a switch opening and closing state recognition method, which comprehensively judges the results of opening and closing recognition, can accurately judge whether the switch is closed in place, opened in place, or in an intermediate state, and can overall judge the accurate state of the switch through one-time recognition and identification. In the closing recognition, an auxiliary feature device on the switch body side is designed, including a position indication plate with color features and an observation hole opened on the moving side support, which facilitates calculation and recognition during later image processing. In the opening recognition, the edge of the moving contact is taken as a key feature, a standard image edge curve data matrix is established, an envelope region is constructed, and edge data matrix proportion calculation is performed on the real-time image, and then it is judged whether the switch is opened in place. The present application solves the problems of existing switch state recognition methods, such as difficulty in determining the threshold range and poor adaptability of machine learning models.

Description

Technical Field

[0001] This invention belongs to the field of intelligent monitoring technology for high-voltage switchgear, specifically relating to a method and system for identifying the open and closed status of a switch. Background Technology

[0002] With the comprehensive construction of smart substations, the demand for intelligent monitoring technology for high-voltage switchgear in power systems is increasing. High-voltage disconnect switches play a crucial role in the power station's switching system, assisting in circuit breaking and protection; therefore, comprehensive and accurate monitoring of their opening and closing status is essential. Currently, there are visual video monitoring technologies for the internal status of high-voltage disconnect switches available on the market. These technologies use video sensors installed on the switch body to remotely monitor the internal status in real time, reducing the difficulty and danger of traditional on-site observation. However, this method still requires personnel to observe remotely and manually judge the switch status, resulting in weak real-time performance and significant subjective influence. Currently, there is no mature automatic identification and judgment method applied in this field.

[0003] A search revealed two image processing algorithms in the field of high-voltage switchgear that are closest to this solution for identifying the status of switchgear:

[0004] One approach is based on traditional image recognition technology. This method mainly involves basic techniques such as feature matching degree calculation, line segment fitting, and projection calculation. It relies on standard open and closed state images and determines the final result by setting a matching degree threshold range. This method places high demands on the stability of the switch's internal state, and determining the threshold range is also very difficult.

[0005] One approach is a classification and recognition algorithm based on machine learning and deep learning. This method collects a large number of images of the internal state of switches under different conditions, trains a network model, and then identifies the state of the switch images. Initially, a large number of samples under different conditions need to be collected for model training. Because the internal structure of switching devices is simple and the image features are basically consistent, it is difficult to train a highly adaptable machine learning model. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for identifying the opening and closing status of a switch, which solves the problems of difficulty in determining the threshold range and poor adaptability of machine learning models in existing switch status identification methods.

[0007] This invention is achieved through the following technical solution:

[0008] This invention discloses a method for identifying the open / closed state of a switch, comprising the following steps:

[0009] Acquire real-time images of the switch's internal components;

[0010] Perform a closing identification and obtain the result R1;

[0011] Perform circuit breaker identification to obtain result R2;

[0012] The identification results of closing and opening circuit breakers are comprehensively analyzed to obtain the final identification result, as follows:

[0013] When R1 is in the closed position and R2 is not in the open position, the switch position status is determined to be in the closed position.

[0014] When R1 is not in the closed position and R1 is in the open position, the switch position status is determined to be in the open position.

[0015] In other cases, the switch position is determined to be either an intermediate state or a fault state.

[0016] Furthermore, the remaining situations include R1 being in the closed position and R2 being in the open position, or R1 not being in the closed position and R2 not being in the open position.

[0017] Furthermore, an auxiliary feature device for indicating the position information of the moving contact is installed on the switch body side. The auxiliary feature device includes an observation hole pre-made on the moving side support and a position indicator plate set at the end of the moving contact. The position indicator plate has a color feature. When the switch is closed, the moving contact drives the position indicator plate to the observation hole.

[0018] The closing identification process includes the following steps:

[0019] S11. Obtain a real-time image of the switch, and after image processing, obtain the position and size information of the observation hole;

[0020] S12. Based on the position and size information of the observation hole, extract the image of the observation hole area, and extract the color area image of the position indicator plate from the image of the observation hole area;

[0021] S13. Process and calculate the color area image of the position indicator to determine whether the switch is closed in place.

[0022] Furthermore, S11 specifically includes the following steps:

[0023] S111. The acquired real-time switch image is simplified by performing image feature matrix simplification through grayscale conversion to obtain a simplified image;

[0024] S112. Redundant information in the simplified image is removed by smoothing filtering to obtain an intermediate image;

[0025] S113. Locate the observation hole on the intermediate image using shape detection;

[0026] S114. Obtain the position and size information of the observation hole.

[0027] Furthermore, S12 specifically includes the following steps:

[0028] S121. Extract the image of the observation hole area based on the position and size information of the observation hole;

[0029] S122. Perform HSV channel conversion on the image of the observation hole area, converting the image data matrix into an HSV channel information matrix;

[0030] S123. Based on the color of the position indicator, determine the threshold interval matrix in the HSV color space, and calculate the color area masking layer by combining the HSV channel information matrix and the threshold interval matrix.

[0031] S124. Perform bitwise operations on the color area mask layer and the observation hole area image extracted in step S121 to extract the color area image of the position indicator plate.

[0032] Furthermore, S13 specifically includes the following steps:

[0033] S131. Perform grayscale transformation, Gaussian filtering and thresholding calculation on the color area image of the position indicator in sequence to obtain a binarized image;

[0034] S132. Perform morphological processing on the binarized image to remove edge noise and interference;

[0035] S133. The template matching algorithm is used to perform a global matching calculation on the binary image of the image processed in S132 and the standard closing position image in the color area of ​​the position indicator. If the matching result is higher than the matching threshold, it is judged as closing position; if it is lower than the matching threshold, it is judged as closing position not in place.

[0036] The matching threshold was obtained by statistically summarizing a large number of actual calculation results in the early stage;

[0037] The process of obtaining the binarized image of the standard closing position image in the color area of ​​the position indicator is as follows: the standard closing position image is processed sequentially through S11, S12, S131 and S132 to obtain the binarized image of the standard closing position image in the color area of ​​the position indicator.

[0038] Furthermore, the circuit breaker identification process includes the following steps:

[0039] S21. Obtain the real-time image of the switch, detect the position of the moving contact edge line, extract the local area image of the moving contact edge, and obtain the moving contact edge curve data matrix;

[0040] S22. Calculate the proportion of the moving contact edge curve data matrix in the envelope region;

[0041] The process of constructing the envelope region is as follows: using the edge region of the moving contact as a key feature, construct a standard edge curve data matrix; and establish the envelope region based on the standard edge curve data matrix.

[0042] S23. Perform threshold analysis on the percentage data to determine whether the switch is fully open.

[0043] Furthermore, in S21, the specific process of obtaining the moving contact edge curve data matrix is ​​as follows:

[0044] The local area image at the edge of the moving contact is sequentially subjected to grayscale transformation and image feature enhancement to obtain the enhanced image.

[0045] The enhanced image is smoothed using Gaussian blur, and the smoothing result is thresholded to obtain a binarized image; then morphological processing is performed to remove excess noise and burrs at the edges.

[0046] The image edge detection method is used to obtain the edge curve data matrix of the moving contact.

[0047] Furthermore, in S22, a standard edge curve data matrix is ​​constructed using the moving contact edge region as a key feature, specifically as follows:

[0048] Collect a large number of images of the circuit breaker in position under different operating conditions to construct a standard image dataset;

[0049] Batch processing of standard image datasets is performed to extract standard image edge curve data matrices, which are then used to construct the data matrix of the standard image dataset.

[0050] The data matrix of the standard image dataset is serialized to construct an edge curve dataset;

[0051] We use mathematical statistical methods to fit and construct a standard edge curve data matrix.

[0052] This invention also discloses a switch opening / closing status identification system, comprising:

[0053] The data acquisition module is used to acquire real-time images of the switch's internal components.

[0054] The closing identification module is used to identify the closing operation and obtain the result R1.

[0055] The circuit breaker tripping identification module is used to identify circuit breaker tripping and obtain the result R2.

[0056] The comprehensive judgment module is used to comprehensively analyze the identification results of closing and opening circuit breakers to obtain the final identification result, specifically:

[0057] When R1 is in the closed position and R2 is not in the open position, the switch position status is determined to be in the closed position.

[0058] When R1 is not in the closed position and R1 is in the open position, the switch position status is determined to be in the open position.

[0059] In other cases, the switch position is determined to be either an intermediate state or a fault state.

[0060] Compared with the prior art, the present invention has the following beneficial technical effects:

[0061] This invention discloses a method for identifying the open and closed states of a switch. By comprehensively judging the results of the open and closed identification, it can accurately determine whether the switch is in the closed, open, or intermediate state. After one judgment and identification, the accurate state of the switch can be judged as a whole.

[0062] Furthermore, during the closing identification process, this invention incorporates an auxiliary feature device on the switch body side, namely a position indicator plate with color characteristics, which is connected to the moving contact of the switch. An opening is made in the moving-side support for observing the position indicator plate. The position indicator plate is colored, facilitating calculation and identification during subsequent image processing. An observation hole is designed to observe the actual operation of the position indicator plate; the opening position is determined based on the conductor position when the switch is fully closed, ensuring that the position indicator plate can be observed in the fully closed state.

[0063] This invention designs an image processing algorithm. During closing identification, based on the characteristics of the switch image, it sequentially locates the observation holes, identifies the color areas of the position indicator plate, performs calculations and analysis, and obtains the closing status identification result of the switchgear. Specifically, the original image is first processed, including grayscale conversion to simplify the image feature matrix, smoothing filtering to blur redundant information, and special shape detection to locate the observation holes, ultimately obtaining the position and size information of the observation holes on the moving side support of the switch. Next, based on the above-mentioned location information of the observation holes, the regional image is segmented, a mask layer for identifying the color features of the indicator plate is constructed, and the position indicator plate region image is calculated and extracted. A recognition and judgment algorithm is designed to process and calculate the position indicator plate region image, obtain a binarized matrix, eliminate edge noise and interference, and analyze and judge the matrix data within a specific region to obtain the closing identification result.

[0064] Furthermore, this invention designs an image processing algorithm. During circuit breaker opening identification, the edge of the moving contact is used as a key feature. A standard image edge curve data matrix is ​​established as a reference to construct an envelope region and calculate the proportion of the edge data matrix in the real-time image to determine whether the switch is fully open. Finally, the image processing algorithm is used to process the real-time image to extract the edge curve matrix, calculate its proportion in the constructed envelope region, and perform threshold analysis to obtain the circuit breaker opening identification result.

[0065] Whether it's closing or opening recognition, it doesn't rely on accumulated knowledge or experience, nor does it require a large number of standard images to train the data model. This method only requires configuring the percentage statistical threshold, constructing the standard image edge curve data matrix, and calculating and setting parameters such as the upper and lower thresholds of the envelope region in the initial stage of use. Attached Figure Description

[0066] Figure 1 This is a flowchart of a switch opening / closing status identification method according to the present invention;

[0067] Figure 2 This is a schematic diagram of the auxiliary feature device in the closed state of the switch;

[0068] In the diagram, 1. Moving side support; 2. Observation hole; 3. Moving contact; 4. Stationary side support; 5. Position indicator plate;

[0069] Figure 3 Here is a flowchart for the closing identification process;

[0070] Figure 4 A flowchart illustrating the location of the observation hole area during the closing identification process;

[0071] Figure 5 This is a flowchart illustrating the process of extracting the color region image during the power-on recognition process.

[0072] Figure 6 A flowchart for processing and calculating the color area image of the position indicator panel to determine whether the switch is fully closed;

[0073] Figure 7 A schematic diagram showing the location of the edge curve of the moving contact when the switch is in the open state;

[0074] In the diagram, 1. Moving side support; 2. Observation hole; 3. Moving contact; 4. Stationary side support; 5. Position indicator plate; 6. Edge area of ​​moving contact;

[0075] Figure 8 Here is a flowchart for the circuit breaker tripping identification process;

[0076] Figure 9 A flowchart illustrating the construction of a standard edge curve data matrix during the circuit breaker identification process;

[0077] Figure 10 A flowchart for constructing the envelope region;

[0078] Figure 11 A schematic diagram of the established envelope region; where A represents the moving contact edge curve extracted from the standard image, and B represents the envelope region established by combining the upper and lower thresholds;

[0079] Figure 12This is a schematic diagram of the acquired image photographs in Example 1;

[0080] Figure 13 This is a schematic diagram of the acquired image photographs in Example 2. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of the present invention clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the present invention, and not all of them.

[0082] The components described and illustrated in the accompanying drawings and embodiments of this invention can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the invention provided in the following drawings is not intended to limit the scope of the claimed invention, but merely to illustrate one selected embodiment of the invention. All other embodiments obtained by those skilled in the art based on the accompanying drawings and embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0083] It should be noted that the terms “comprising,” “including,” or any other variations are intended to cover non-exclusive inclusion, such that a process, element, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to the process, element, method, article, or apparatus.

[0084] This invention proposes a method for identifying the open / closed state of a switch, the main technical solution and process of which are as follows: Figure 1 As shown, the process includes the following:

[0085] Acquire real-time images of the switch's internal components;

[0086] Perform a closing identification and obtain the result R1;

[0087] Perform circuit breaker identification to obtain result R2;

[0088] The identification results of closing and opening circuit breakers are comprehensively analyzed to obtain the final identification result, as follows:

[0089] When R1 is in the closed position and R2 is not in the open position, the switch position status is determined to be in the closed position.

[0090] When R1 is not in the closed position and R1 is in the open position, the switch position status is determined to be in the open position.

[0091] In other cases, the switch position is determined to be either an intermediate state or a fault state.

[0092] For closing identification, this invention designs an auxiliary feature device for indicating the position information of the moving contact of the switch. The structure of this device is as follows: Figure 2 As shown, a position indicator plate 5 with color features is designed and connected to the moving contact 3 of the switch. It moves synchronously with the moving contact 3. An opening is made on the moving side support 1 for observing the position indicator plate 5. When the switch is closed, the moving contact 3 will drive the position indicator plate 5 to the center position of the observation hole 2. The closing position can be identified based on this feature.

[0093] For closing identification, the positioning observation hole 2 is sequentially located, the color area of ​​the position indicator plate 5 is identified, calculations and analyses are performed, and the closing status identification result of the switchgear is obtained. The specific process is as follows: Figure 3 As shown:

[0094] Step S11: Locate the observation hole area

[0095] In this step, the specific process is as follows: Figure 4 As shown, an observation hole positioning and recognition algorithm is designed to process the original image acquired by the video sensor, including grayscale conversion to simplify the image feature matrix, smoothing filtering to blur redundant information, and special shape detection to locate the observation hole 2, and finally obtain the position and size information of the observation hole 2 on the switch moving side support 1.

[0096] Step S12: Extract the color region image from position indicator 5

[0097] In this step, a location indicator detection algorithm is designed. Based on the positioning information of observation hole 2, the region image is segmented, a mask layer for recognizing the color features of the indicator is constructed, and the location indicator region image is calculated and extracted. Specific steps are as follows: Figure 5 As shown:

[0098] S121. First, based on the position and size information of the observation hole 2 obtained in the previous step, locate the area and extract the image of the observation hole area from the original image.

[0099] S122. Perform HSV channel conversion on the image of the observation hole area, converting the image data matrix into an HSV channel information matrix;

[0100] S123. Based on the color of the position indicator (5), determine the threshold interval matrix in the HSV color space. Combine the image HSV channel information matrix and the threshold interval matrix to calculate the color region mask layer. Perform bitwise operations on the mask layer and the observation hole area image extracted in step S121 to extract the color region image.

[0101] Step S13: Image processing and calculation to obtain the closing recognition result.

[0102] In this step, an identification and judgment algorithm is designed to process and calculate the color area image of the position indicator 5, obtain a binarized matrix, eliminate edge noise and interference, and perform statistical analysis on the matrix data within a specific area to obtain the closing identification result. The calculation process is as follows: Figure 6 As shown:

[0103] The specific steps for implementation are as follows:

[0104] S131. First, perform grayscale transformation, Gaussian filtering and thresholding calculation on the color region image in sequence to obtain a binarized image.

[0105] S132. Perform morphological processing on the binarized image to remove edge noise and reduce interference from other irrelevant information in the image.

[0106] S133. A template matching algorithm is used to perform a global matching calculation on the binarized image of the above image and the standard closing position image in the color area of ​​the position indicator 5. A matching threshold is set. If the matching result is higher than the matching threshold, it is judged as closing position. If it is lower than the matching threshold, it is judged as closing position not being in place.

[0107] The process of obtaining the binarized image of the standard closing position image in the color area of ​​position indicator 5 is as follows: after processing by S11 and S12, it can be obtained by processing by S131 and S132.

[0108] The above matching thresholds are derived from statistical summarization of a large number of actual calculation results, and are generally greater than 0.8.

[0109] The technical solution for circuit breaker trip identification is described in detail below:

[0110] like Figure 7 The diagram shown illustrates the internal structure of the switch in the tripped state. The core of tripped detection is to use the edge region 6 of the moving contact as a key feature, employing an image recognition algorithm to detect the position of the moving contact's edge line.

[0111] For circuit breaker tripping identification, a standard image edge curve data matrix is ​​established as a reference. An envelope region is constructed, and the proportion of the edge data matrix is ​​calculated on the real-time image to determine whether the circuit breaker has tripped.

[0112] The specific process is as follows: Figure 8 As shown, specifically:

[0113] S21. Use image processing algorithms to process and analyze real-time images, extract local area images of the moving contact edge, and obtain the moving contact edge curve data matrix.

[0114] S22. Calculate the proportion of the real-time image edge curve data matrix in the envelope region;

[0115] S23. Perform threshold analysis on the percentage data to determine whether the circuit breaker has been properly tripped. Specifically, a critical threshold needs to be designed first. If the percentage data is less than this critical threshold, it is determined that the circuit breaker has not been properly tripped.

[0116] In step S21, the specific process of obtaining the moving contact edge curve data matrix is ​​as follows:

[0117] First, the local area image of the moving contact edge is transformed into a grayscale image. Then, global histogram equalization is used to enhance the image features. Next, Gaussian blur is used to smooth the enhanced image. The smoothing result is then thresholded to obtain a binary image. Morphological processing is then performed to remove excess noise and burrs from the edges. Finally, image edge detection technology is used to obtain the moving contact edge curve data matrix.

[0118] The envelope region needs to be constructed in advance. The specific construction process is as follows: using the edge of the moving contact to be detected as the key feature, construct a standard edge curve data matrix; and establish the envelope region based on the standard edge curve data matrix.

[0119] The specific process for constructing a standard edge curve data matrix is ​​as follows: Figure 9 As shown, firstly, a large number of tripping state images under different working conditions are collected to construct a standard image dataset; then, an image processing algorithm is designed to extract the standard image edge curve data matrix; the standard image dataset data matrix calculated by the above image processing algorithm is serialized to construct an edge curve dataset; finally, a standard edge curve data matrix is ​​fitted and constructed using mathematical statistical methods, and saved as a reference for subsequent calculations.

[0120] The process of extracting the standard image edge curve data matrix is ​​the same as the specific process of obtaining the moving contact edge curve data matrix in S21.

[0121] The specific process of constructing the envelope region is as follows: Figure 10 As shown, the specific steps are as follows: Read the standard image edge curve data matrix, obtain the upper and lower thresholds, and establish the envelope region based on envelope theory. The upper and lower thresholds are calculated based on the obtained standard edge curve dataset using mathematical statistics and probability distribution.

[0122] like Figure 11 The diagram shows the envelope region established based on the standard edge curve data matrix. In the diagram, A represents the moving contact edge curve extracted from the standard image, and B represents the envelope region established by combining upper and lower thresholds.

[0123] This invention also discloses a switch opening / closing status identification system, comprising:

[0124] The data acquisition module is used to acquire real-time images of the switch's internal components.

[0125] The closing identification module is used to identify the closing operation and obtain the result R1.

[0126] The circuit breaker tripping identification module is used to identify circuit breaker tripping and obtain the result R2.

[0127] The comprehensive judgment module is used to comprehensively analyze the recognition results of closing and opening identification to obtain the final recognition result.

[0128] In the comprehensive judgment module, the recognition results of the closing recognition module and the opening recognition module are comprehensively analyzed to obtain the final recognition result. For the acquired real-time images inside the switch, after calculation by the closing recognition and opening recognition modules, results R1 and R2 are obtained. Based on the results of R1 and R2, the switch status is judged as follows: when R1 is closed and R2 is not open, the switch position is judged as closed; when R1 is not closed and R1 is open, the switch position is judged as open; in other cases, including R1 being closed and R2 being open, or R1 being not closed and R2 not open, the switch position is judged as an intermediate state, which is usually regarded as a fault state.

[0129] This invention is based on image processing algorithms, requiring neither the accumulation of knowledge and experience nor a large number of standard images to train the learning data model. The method only requires initial configuration of parameters such as the proportion statistical threshold, construction of a standard image edge curve data matrix, and calculation and setting of upper and lower thresholds for the envelope region before it can be deployed online.

[0130] The auxiliary feature device and image recognition method proposed in this invention can be applied to switching equipment of other voltage levels. When the method proposed in this invention is applied to switches of other voltage levels, it can adaptively identify and judge the opening and closing status of the switching equipment after the parameters are configured during initialization.

[0131] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0132] Example 1

[0133] like Figure 12 As shown, first, the original image a is obtained, and after image processing, the position and size information of the observation hole 2 is obtained; the image of the observation hole area is extracted; the color area image b of the position indicator plate 5 is extracted from the image of the observation hole area; the binarized image c is calculated, and template matching calculation is performed to determine that the switch is in the closed position.

[0134] Example 2

[0135] like Figure 13As shown, a real-time image of the switch is acquired, and the local area image of the moving contact edge is extracted as shown in Figure a. The image is processed to obtain a binarized image as shown in Figure b. The moving contact edge curve data matrix is ​​obtained as shown in Figure c. The proportion of the moving contact edge curve data matrix in the envelope region (the red area in Figure d) is calculated. Finally, a threshold analysis is performed on the proportion data to determine whether the switch is in the open position.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for identifying the open / closed state of a switch, characterized in that, The process includes the following: Acquire real-time images of the inside of the switch; Perform a closing identification and obtain the result R1; Perform circuit breaker identification to obtain result R2; The identification results of closing and opening circuit breakers are comprehensively analyzed to obtain the final identification result, as follows: When R1 is in the closed position and R2 is not in the open position, the switch position status is determined to be in the closed position. When R1 is not in the closed position and R1 is in the open position, the switch position status is determined to be in the open position. In other cases, the switch position is determined to be either an intermediate state or a fault state. The circuit breaker identification includes the following steps: S21. Obtain the real-time image of the switch, detect the position of the moving contact edge line, extract the local area image of the moving contact edge, and obtain the moving contact edge curve data matrix; S22. Calculate the proportion of the moving contact edge curve data matrix in the envelope region; The process of constructing the envelope region is as follows: using the moving contact edge region (6) as the key feature, construct a standard edge curve data matrix; establish the envelope region based on the standard edge curve data matrix; S23. Perform threshold analysis on the percentage data to determine whether the switch is fully open.

2. The method for identifying the open / closed state of a switch according to claim 1, characterized in that, Other cases include R1 being closed and R2 being open, or R1 not being closed and R2 not being open.

3. The method for identifying the open / closed state of a switch according to claim 1, characterized in that, An auxiliary feature device for indicating the position information of the moving contact is installed on the side of the switch body. The auxiliary feature device includes an observation hole (2) prefabricated on the moving side support (1) and a position indicator plate (5) set at the end of the moving contact (3). The position indicator plate (5) has a color feature. When the switch is closed, the moving contact (3) drives the position indicator plate (5) to the observation hole (2). The closing identification includes the following steps: S11. Obtain the real-time image of the switch, and after image processing, obtain the position and size information of the observation hole (2); S12. Based on the position and size information of the observation hole (2), extract the image of the observation hole (2) area and extract the color area image of the position indicator plate (5) from the image of the observation hole (2) area. S13. Process and calculate the color area image of the position indicator plate (5) to determine whether the switch is closed.

4. The method for identifying the open / closed state of a switch according to claim 3, characterized in that, S11 specifically includes the following steps: S111. The acquired real-time switch image is simplified by performing image feature matrix simplification through grayscale conversion to obtain a simplified image; S112. Redundant information in the simplified image is removed by smoothing filtering to obtain an intermediate image; S113. Locate the observation hole (2) on the intermediate image by shape detection. S114. Obtain the position and size information of the observation hole (2).

5. The method for identifying the open / closed state of a switch according to claim 3, characterized in that, S12 specifically includes the following steps: S121. Extract the image of the observation hole area based on the position and size information of the observation hole (2); S122. Perform HSV channel conversion on the image of the observation hole area, converting the image data matrix into an HSV channel information matrix; S123. Based on the color of the position indicator (5), determine the threshold interval matrix in the HSV color space, and calculate the color area masking layer by combining the HSV channel information matrix and the threshold interval matrix. S124. Perform bitwise operations on the color area mask layer and the observation hole area image extracted in step S121 to extract the color area image of the position indicator plate (5).

6. The method for identifying the open / closed state of a switch according to claim 3, characterized in that, S13 specifically includes the following steps: S131. Perform grayscale transformation, Gaussian filtering and thresholding calculation on the color area image of the position indicator (5) in sequence to obtain a binarized image; S132. Perform morphological processing on the binarized image to remove edge noise and interference; S133. The template matching algorithm is used to perform global matching calculation on the binary image of the image processed by S132 and the standard closing position image in the color area of ​​the position indicator plate (5). If the matching result is higher than the matching threshold, it is judged as closing position; if it is lower than the matching threshold, it is judged as closing position not in place. The matching threshold was obtained by statistically summarizing a large number of actual calculation results from the previous period; The process of obtaining the binarized image of the standard closing position image in the color area of ​​the position indicator plate (5) is as follows: the standard closing position image is processed sequentially through S11, S12, S131 and S132 to obtain the binarized image of the standard closing position image in the color area of ​​the position indicator plate (5).

7. The method for identifying the open / closed state of a switch according to claim 1, characterized in that, In S21, the specific process of obtaining the moving contact edge curve data matrix is ​​as follows: The local area image at the edge of the moving contact is sequentially subjected to grayscale transformation and image feature enhancement to obtain the enhanced image. The enhanced image is smoothed using Gaussian blur, and the smoothing result is thresholded to obtain a binarized image; then morphological processing is performed to remove excess noise and burrs at the edges. The image edge detection method is used to obtain the edge curve data matrix of the moving contact.

8. The method for identifying the open / closed state of a switch according to claim 1, characterized in that, In S22, the moving contact edge region (6) is used as the key feature to construct a standard edge curve data matrix, specifically: Collect a large number of images of the circuit breaker in position under different operating conditions to construct a standard image dataset; Batch processing of standard image datasets is performed to extract standard image edge curve data matrices, which are then used to construct the data matrix of the standard image dataset. The data matrix of the standard image dataset is serialized to construct an edge curve dataset; We use mathematical statistical methods to fit and construct a standard edge curve data matrix.

9. A switch opening / closing status identification system for implementing the switch opening / closing status identification method according to any one of claims 1-8, characterized in that, include: The data acquisition module is used to acquire real-time images of the switch's internal components. The closing identification module is used to identify the closing operation and obtain the result R1. The circuit breaker tripping identification module is used to identify circuit breaker tripping and obtain the result R2. The comprehensive judgment module is used to comprehensively analyze the identification results of closing and opening circuit breakers to obtain the final identification result, specifically: When R1 is in the closed position and R2 is not in the open position, the switch position status is determined to be in the closed position. When R1 is not in the closed position and R1 is in the open position, the switch position status is determined to be in the open position. In other cases, the switch position is determined to be either an intermediate state or a fault state.

Images