A method and system for identifying the status of the secondary panel of a substation based on image recognition

Abstract

The present invention discloses a method for identifying the state of the secondary switchgear panel based on image recognition. The method is as follows: S01, collect the images of the secondary switchgear pressure plates; S02, compare the switch states in the pressure plate switching-on / off list and the images of the secondary switchgear pressure plates to obtain the positions of the pressure plate switches with switching-on / off errors; S03, send out the positions of the pressure plate switches with switching-on / off errors and give an early warning. This is to solve the problems in the prior art that the inspection of the switch states of the secondary switchgear panel has high labor intensity, high error rate and untimely early warning.

Description

Technical Field

[0001] The present invention relates to a method and system for identifying the status of the secondary switchgear panel based on image recognition, belonging to the technical field of secondary switchgear panel inspection. Background Art

[0002] With the continuous increase in social power consumption, the structure of the power system has become increasingly complex. To meet the safe and stable operation of the power system, the number of intelligent substations under construction is gradually increasing. To ensure the high stability and safe operation of intelligent substations and improve the operation and maintenance level of secondary equipment, the structure of the equipment switchgear in substations has become increasingly complex, and the number of protection pressure plates has gradually increased. However, the protection pressure plates are vulnerable to human operations. For example, during manual inspections, due to various reasons such as visual fatigue and memory confusion, missed inspections and misinspections may occur. At the same time, it may also be affected by factors such as environmental pollution, resulting in incorrect withdrawal and incorrect input. In particular, incorrect operations of the protection pressure plates can only be discovered and restored after a power accident occurs, seriously affecting the safety and stability of the substation.

[0003] In a substation, the number of secondary switchgears is usually dozens or hundreds, and the number of pressure plate switches in each switchgear is as many as fifty-four. The inspection work needs to check these protection pressure plate switches, and it is required to inspect the operation status of the protection pressure plates at least once a month. However, in the current manual inspection process, there are problems such as heavy workload, poor timeliness, and paper-based work records. In addition, due to factors such as physical and mental qualities, sense of responsibility, technical level, and work experience of the inspection personnel, mistakes may occur, leaving potential safety hazards and even possibly causing major safety accidents. Therefore, we need to study a method for identifying the status of the secondary switchgear panel in a substation to timely discover and eliminate potential safety hazards that do not conform to the status of the secondary switchgear panel in the substation, reduce the labor intensity of the staff, and improve the power supply reliability of the power enterprise. Summary of the Invention

[0004] The technical problem to be solved by the present invention is: to provide a method and system for identifying the status of the secondary switchgear panel based on image recognition to overcome the deficiencies of the prior art.

[0005] The technical solution of the present invention is:

[0006] A method for identifying the status of the secondary switchgear panel based on image recognition, the method is as follows:

[0007] S01. Collect images of the secondary switchgear pressure plates;

[0008] S02. Compare the switch status in the pressure plate input / output table with the secondary switchgear pressure plate image to obtain the positions of the pressure plate switches with incorrect input / output;

[0009] S03. Send out the positions of the pressure plate switches with incorrect input / output and give an alarm.

[0010] Specifically, the step S02 is specifically as follows:

[0011] Identify the switch status in the image of the secondary switch cabinet pressure plate, compare the switch status with the pressure plate withdrawal and insertion table, mark the pressure plate switches with status different from the records in the pressure plate withdrawal and insertion table as withdrawal and insertion errors, and obtain the positions of the pressure plate switches with withdrawal and insertion errors.

[0012] Specifically, identifying the switch status in the image of the secondary switch cabinet pressure plate in the step S02 is specifically as follows:

[0013] Step 1: Divide each row of pressure plates using red lines;

[0014] Step 2: Perform image smoothing using the Gaussian filtering method while retaining the overall gray-scale distribution characteristics of the image;

[0015] Step 3: Convert the image from RGB to HSV through color region conversion, compare the effects of the three channels, and select the best channel for processing;

[0016] Step 4: Perform binarization and Otsu operation on the image;

[0017] Step 5: Determine the position of the red line by extracting the pixel distribution with a threshold of 255 for each row of pixels;

[0018] Step 6: Segment each row of pressure plates according to the determined red line;

[0019] Step 7: Re-binarize the segmented image to obtain a black-and-white image;

[0020] Step 8: Remove smaller noise points by dilating and eroding the image using morphological matrices of different sizes;

[0021] Step 9: Scan each processed single-row pressure plate image column by column and count the number of black pixels in each column;

[0022] Step 10: In the counted black pixel array, distinguish between peaks and valleys, and search for and record the edges of individual pressure plates;

[0023] Step 11: Determine the switch status of the pressure plate based on the shape of the corresponding pixel points in the original image according to the edge information;

[0024] Step 12: Outline the pressure plate switches with incorrect switch status in the original image.

[0025] Further, the morphological matrices selected in the step 8 are 3x3 and 5x5 matrices, which are used to process the image once respectively.

[0026] Further, the judgment of the switch status of the pressure plate is identified through Hough transform operation.

[0027] Specifically, step S02 is specifically as follows:

[0028] Generate corresponding standard switch status diagrams for each switch status in the pressure plate withdrawal list;

[0029] Cut the secondary switch cabinet pressure plate image according to different switches into corresponding actual switch status diagrams;

[0030] Calculate the similarity A between the standard switch status diagram and the actual switch status diagram;

[0031] If the similarity A is greater than the set threshold B, it is determined that the withdrawal is correct; otherwise, it is determined that the withdrawal is incorrect.

[0032] Preferably, the similarity A is obtained by the following method:

[0033] S02_1: Extract the features of the standard switch status diagram and the actual switch status diagram respectively;

[0034] S02_2: Map the features extracted in step S02_1 to an n-dimensional hyperspace to obtain the corresponding embedding;

[0035] S03_3: Calculate the distance between the two embeddings in the hyperspace as the similarity A.

[0036] Preferably, the distance in the hyperspace in step S03_3 is the cosine distance.

[0037] A method for identifying the status of the secondary switch cabinet panel based on image recognition, the system includes:

[0038] An image acquisition module for acquiring the secondary switch cabinet pressure plate image;

[0039] A data storage module, which can be used to store the pressure plate withdrawal list recording the switch status of the pressure plate;

[0040] A control module, which is used to compare the switch status of the acquired secondary switch cabinet pressure plate image with the pressure plate withdrawal list, obtain the switch position of the pressure plate with incorrect withdrawal, the control module is electrically connected to the image acquisition module, and the control module is electrically connected to the data storage module;

[0041] An early warning module, which is used to give an early warning of the switch position of the pressure plate with incorrect withdrawal, and the early warning module is electrically connected to the control module.

[0042] The beneficial effects of the present invention are: by acquiring the secondary switch cabinet pressure plate image, comparing the switch status in the pressure plate withdrawal list and the secondary switch cabinet pressure plate image, obtaining the switch position of the pressure plate with incorrect withdrawal, eliminating the need for manual inspection, greatly reducing the labor intensity, reducing human errors, and realizing real-time early warning. Description of the Drawings

[0043] Figure 1 It is the flowchart of the method according to the embodiment of the present invention;

[0044] Figure 2 It is the structural framework diagram according to the embodiment of the present invention;

[0045] Figure 3 It is the pressure plate state diagram when the connecting piece is in the inserted state in the normal mode of the present invention;

[0046] Figure 4 It is the pressure plate state diagram when the connecting piece is in the withdrawn state in the normal mode of the present invention;

[0047] Figure 5 It is the pressure plate state diagram when the connecting piece is in the removed state in the normal mode of the present invention;

[0048] Figure 6 It is the original image of the secondary cabinet pressure plate of the present invention;

[0049] Figure 7 It is the image of the secondary cabinet pressure plate of the present invention after binarization and Otsu operation;

[0050] Figure 8 It is the black and white image obtained by re-binarizing the image of the secondary cabinet pressure plate of the present invention;

[0051] Figure 9 It is the image in which the pressure plate switch with incorrect switch state is framed in the original image of the secondary cabinet pressure plate of the present invention. Detailed Embodiment

[0052] The technical solution in the embodiment of the present invention has the following general idea:

[0053] By collecting the image of the secondary cabinet pressure plate, comparing the switch states in the pressure plate withdrawal and insertion table and the image of the secondary cabinet pressure plate, the position of the pressure plate switch with incorrect withdrawal and insertion is obtained.

[0054] Embodiment 1

[0055] In order to solve the problems of high labor intensity, high error rate and untimely warning in the inspection of the switch state of the secondary cabinet panel in the prior art, referring to Figure 1 , the embodiment of the present invention adopts a method for identifying the state of the secondary cabinet panel of a substation based on image recognition, and the method is as follows:

[0056] S01. Collect the image of the secondary cabinet pressure plate;

[0057] S02. Compare the switch states in the pressure plate withdrawal and insertion table and the image of the secondary cabinet pressure plate, and obtain the position of the pressure plate switch with incorrect withdrawal and insertion;

[0058] S03. Send out the position of the pressure plate switch with a retraction error outward and give an early warning.

[0059] Embodiment 2

[0060] In this embodiment, in addition to the technical solution of Embodiment 1, it further includes a method for specifically obtaining the position of the pressure plate switch with a retraction error. The specific step S02 is as follows: Identify the switch state in the image of the secondary screen cabinet pressure plate, compare the switch state with the pressure plate retraction table, record the pressure plate switch with a state different from the recorded state in the pressure plate retraction table as a retraction error, and obtain the position of the pressure plate switch with a retraction error.

[0061] Here, first identify the switch state in the image of the secondary screen cabinet pressure plate, and then compare the switch state with the pressure plate retraction table.

[0062] In order to improve the recognition accuracy of the switch state in the image of the secondary screen cabinet pressure plate, specifically in this embodiment, the step of identifying the switch state in the image of the secondary screen cabinet pressure plate in step S02 is specifically as follows:

[0063] Step 1. Divide each row of pressure plates with a red line; as Figure 6 shown, there are 6 rows of pressure plates in the collected image of the secondary screen cabinet pressure plate. In order to reduce the influence and recognition of noise, here it is divided by a red line, and each row of pressure plates is divided and processed separately;

[0064] Step 2. Adopt the Gaussian filtering method for image smoothing processing, and at the same time retain the overall gray distribution characteristics of the image to facilitate the distinction and recognition of each row of pressure plate switches;

[0065] Step 3. Convert the image from RGB to HSV through color region conversion, compare the effects of the three channels, and select the best channel for processing. Here, the S channel is selected;

[0066] Step 4. Perform binarization and Otsu operation on the image to obtain a better image, as shown in Figure 7 shown;

[0067] Step 5. Determine the position of the red line by extracting the pixel distribution with a threshold of 255 for each row of pixels;

[0068] Step 6. Segment each row of pressure plates according to the judged red line;

[0069] Step 7. Re-binarize the already segmented image to obtain a black and white image, as shown in Figure 8 shown;

[0070] Step 8: Dilate and erode the image using morphological matrices of different sizes, and then remove the smaller noise points. Here, the morphological matrices selected are 3x3 and 5x5 matrices, which are used to process the image once respectively;

[0071] Step 9: Scan the processed single-line pressure plate image column by column, and count the number of black pixels in each column;

[0072] Step 10: In the counted black pixel array, distinguish the peaks and valleys, and search for and record the edges of a single pressure plate;

[0073] Step 11: Based on the edge information, judge the switch state of the pressure plate at the corresponding pixel points in the original image. Refer to Figures 3 - 5 , and the recognized output pressure plate states are: pressure plate on (operating), pressure plate off (maintenance), removed (hot standby);

[0074] Step 12: In the original image, frame out the pressure plate switches with incorrect switch states, as shown in Figure 9 shown.

[0075] In order to more effectively obtain the switch states of the pressure plates in the entire secondary switch cabinet, in this embodiment, further, the judgment of the switch state of the pressure plate is identified through Hough transform operation.

[0076] Embodiment 3

[0077] This embodiment provides another method for obtaining the position of the pressure plate switch with incorrect withdrawal and insertion. Specifically, the step S02 is specifically as follows:

[0078] Generate corresponding standard switch state diagrams for each switch state in the pressure plate withdrawal and insertion table;

[0079] Cut the secondary switch cabinet pressure plate image into corresponding actual switch state diagrams according to different switches;

[0080] Calculate the similarity A between the standard switch state diagram and the actual switch state diagram;

[0081] If the similarity A is greater than the set threshold B, it is determined that the withdrawal and insertion are correct, otherwise it is determined that the withdrawal and insertion are incorrect.

[0082] Here, by calculating the similarity between the corresponding standard switch state diagrams generated for each switch state in the pressure plate withdrawal and insertion table and the actual switch state diagrams cut from the secondary switch cabinet pressure plate image according to different switches, it is determined whether the switch state of the secondary switch cabinet pressure plate is the same as the recorded state in the withdrawal and insertion table. If the similarity is low, it can be judged that the withdrawal and insertion are incorrect and a warning is issued.

[0083] In this embodiment, in order to obtain the similarity between the standard switch state diagram and the actual switch state diagram, preferably in this embodiment, the similarity A is obtained by the following method:

[0084] S02_1. Extract the features of the standard switch status diagram and the actual switch status diagram respectively;

[0085] S02_2. Map the features extracted in step S02_1 to an n-dimensional hyperspace to obtain the corresponding embedding;

[0086] S03_3. Calculate the distance between the two embeddings in the hyperspace as the similarity A.

[0087] Here, the distance between the embeddings of the two switch status diagrams in the n-dimensional hyperspace is used as the similarity.

[0088] Preferably, the distance in the hyperspace in step S03_3 is the cosine distance.

[0089] Embodiment 4

[0090] In order to solve the problems of high labor intensity, high error rate and untimely warning in the inspection of the switch status of the secondary cabinet panel in the prior art, referring to Figure 2 , a method for identifying the status of the secondary cabinet panel of a substation based on image recognition, the system includes:

[0091] An image acquisition module for acquiring the image of the secondary cabinet pressure plate;

[0092] A data storage module, which can be used to store the pressure plate withdrawal and investment table recording the switch status of the pressure plate;

[0093] A control module for comparing the image of the secondary cabinet pressure plate acquired with the switch status of the pressure plate withdrawal and investment table, obtaining the position of the pressure plate switch with withdrawal and investment errors, the control module is electrically connected to the image acquisition module, and the control module is electrically connected to the data storage module;

[0094] An early warning module for warning the position of the pressure plate switch with withdrawal and investment errors, the early warning module is electrically connected to the control module.

[0095] The above content is a further detailed description of the present invention in combination with specific preferred embodiments. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, which should all be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for identifying the status of the secondary panel of a substation based on image recognition, the method is as follows: S01. Collect the images of the secondary panel pressure plates; S02. Compare the switch status in the pressure plate withdrawal / insertion table with that in the images of the secondary panel pressure plates to obtain the positions of the pressure plate switches with incorrect withdrawal / insertion; S03. Send out the positions of the pressure plate switches with incorrect withdrawal / insertion and give an early warning; The specific step S02 is as follows: Identify the switch status in the images of the secondary panel pressure plates, compare the switch status with the pressure plate withdrawal / insertion table, mark the pressure plate switches with status different from the records in the pressure plate withdrawal / insertion table as those with incorrect withdrawal / insertion, and obtain the positions of the pressure plate switches with incorrect withdrawal / insertion; The method for identifying the switch status in the images of the secondary panel pressure plates in step S02 is specifically as follows: Step 1. Divide each row of pressure plates with red lines; Step 2. Use the Gaussian filtering method for image smoothing while retaining the overall gray-scale distribution characteristics of the image; Step 3. Convert the image from RGB to HSV through color area conversion, compare the effects of the three channels, and select the best channel for processing; Step 4. Perform binarization and Otsu operation on the image; Step 5. Determine the position of the red line by extracting the pixel distribution with a threshold of 255 for each row of pixels; Step 6. Segment each row of pressure plates according to the determined red line; Step 7. Re-binarize the segmented image to obtain a black-and-white image; Step 8. Remove noise by dilating and eroding the image with morphological matrices of different sizes; Step 9. Scan each processed single-row pressure plate image column by column and count the number of black pixels in each column; Step 10. In the counted black pixel array, distinguish the peaks and valleys, and find and record the edges of individual pressure plates; Step 11. Determine the switch status of the pressure plates according to the pixel point shapes corresponding to the edges in the original image; Step 12. Outline the pressure plate switches with incorrect switch status in the original image.

2. The method for identifying the status of the secondary panel of a substation based on image recognition according to claim 1, characterized in that, in step 8, the selected morphological matrices are 3x3 and 5x5 matrices, and the image is processed once with each matrix.

3. The method for identifying the status of the secondary panel of a substation based on image recognition according to claim 1, characterized in that, the judgment of the switch status of the pressure plates is recognized through Hough transform operation.

4. The method for identifying the status of the secondary panel of a substation based on image recognition according to claim 1, characterized in that, the specific step S02 is as follows: Generate corresponding standard switch status diagrams for each switch status in the pressure plate withdrawal / insertion table; Cut the images of the secondary panel pressure plates according to different switches into corresponding actual switch status diagrams; Calculate the similarity A between the standard switch status diagram and the actual switch status diagram; If the similarity A is greater than the set threshold B, it is determined that the withdrawal / insertion is correct, otherwise it is determined that the withdrawal / insertion is incorrect.

5. The method for identifying the status of the secondary panel of a substation based on image recognition according to claim 4, characterized in that, the similarity A is obtained by the following method: S02_1. Extract the features of the standard switch status diagram and the actual switch status diagram respectively; S02_2. Map the features extracted in step S02_1 to an n-dimensional hyperspace to obtain the corresponding embedding; S03_3. Calculate the distance between the two embeddings in the hyperspace as the similarity A.

6. The method for identifying the state of the secondary switchgear panel based on image recognition according to claim 5, wherein, the distance in the hyperspace in step S03_3 is the cosine distance.

7. An image recognition-based system for identifying the state of the secondary switchgear panel, wherein, the system is used for the method described in any one of claims 1-6, and the system includes: an image acquisition module for acquiring the images of the secondary switchgear pressure plates; a data storage module, which can be used to store the pressure plate withdrawal and insertion table for recording the switch states of the pressure plates; a control module for comparing the acquired images of the secondary switchgear pressure plates with the switch states in the pressure plate withdrawal and insertion table to obtain the positions of the pressure plate switches with incorrect withdrawal and insertion, the control module is electrically connected to the image acquisition module, and the control module is electrically connected to the data storage module; a warning module for warning the positions of the pressure plate switches with incorrect withdrawal and insertion, the warning module is electrically connected to the control module.

Images