Switch cabinet operation debugging monitoring method and system based on image recognition

Abstract

The invention discloses a switch cabinet operation debugging monitoring method and system based on image recognition, and relates to the technical field of image recognition, an image preprocessing module improves image quality through denoising, enhancement and feature extraction, calculates an image index Txzs and eliminates redundant images according to a threshold value; a recognition model is established by using a convolutional neural network through an image recognition analysis module, and a breaking prediction unit calculates a breaking stability difference value Kdcz by analyzing historical and real-time data and evaluates the breaking stability of the circuit breaker through a preset difference value threshold value; the time management module dynamically adjusts the monitoring time according to the actual on-off condition of the circuit breaker, guarantees the timeliness and accuracy of the monitoring operation, effectively deals with the influence of various unpredictable factors, and avoids the problems of inaccuracy and non-timeliness caused by fixed time management. The anomaly detection unit calculates an on-off index Kdzs through real-time data, and the multi-level threshold value strategy unit sends out early warning signals of different levels according to a preset threshold value.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a switch cabinet operation debugging monitoring method and system based on image recognition. Background Art

[0002] The switchgear operation monitoring system uses advanced sensing technology and data analysis algorithms to monitor the operating status and environmental parameters of electrical equipment in real time, such as current, voltage, temperature, etc. Through remote monitoring and data analysis, the system can achieve equipment failure early warning and operation optimization, improve equipment reliability and safety, reduce downtime and maintenance costs, thereby improving production efficiency and the overall operation quality of the equipment.

[0003] In the invention patent with application number 202410276345.3, a switch cabinet operation monitoring system and monitoring method are disclosed, which relate to the field of switch cabinet monitoring technology, including an operation data acquisition module, a monitoring and analysis modeling module, a monitoring time management module and a switch cabinet operation monitoring module. The operation data acquisition module is used to collect the circuit breaker opening and closing historical data during the operation of the switch cabinet, and pre-process the opening and closing historical data; the monitoring and analysis modeling module is used to predict the circuit breaker opening and closing stable cycle based on the pre-processed opening and closing historical data; the monitoring time management module is used to analyze the prediction results of the opening and closing stable cycle and manage the circuit breaker opening and closing monitoring time; the switch cabinet operation monitoring module is used to monitor the circuit breaker of the switch cabinet according to the management time, and issue an abnormal warning when a opening and closing fault is detected, thereby improving the probability of successful acquisition of monitoring data and improving the difficulties caused to the monitoring work by the differences in the working conditions of the circuit breakers in the switch cabinet.

[0004] However, when the above invention is actually applied to the operation of the switch cabinet of a power plant, the monitoring time management module manages the circuit breaker disconnection monitoring time according to the prediction result of the stable disconnection cycle. However, in the actual operation of the power plant, the disconnection frequency and time of the circuit breaker will be affected by a variety of unpredictable factors, which leads to inaccurate and untimely monitoring time management; secondly, the circuit breaker disconnection monitoring is carried out according to the management time. If the monitoring time management is not accurate, the actual operating status of the circuit breaker cannot be reflected in real time, which increases the probability of false alarm. Summary of the invention

[0005] Technical issues solved

[0006] In view of the deficiencies in the prior art, the present invention provides a switch cabinet operation debugging monitoring method and system based on image recognition, which solves the technical problems mentioned in the above background technology.

[0007] Technical Solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a switch cabinet operation debugging monitoring system based on image recognition, including an image acquisition module, an image preprocessing module, an image recognition and analysis module, a time management module and an abnormal warning module;

[0009] The image acquisition module is used to collect image data of the switch cabinet and circuit breaker in real time, and ensures the clarity and stability of the image data through a high-resolution camera and anti-interference design;

[0010] The image preprocessing module is used to preprocess the collected image data, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, the image data is calculated, the image index Txzs is obtained and evaluated, so as to eliminate redundant images;

[0011] The image recognition and analysis module identifies the working state of the circuit breaker based on the preprocessed image data through a deep learning algorithm and predicts the circuit breaker's breaking stability cycle to provide more accurate monitoring and analysis;

[0012] The time management module dynamically adjusts the monitoring time of the circuit breaker according to the prediction results of the image recognition and analysis module, and then adjusts the monitoring plan again based on the monitoring time;

[0013] The abnormal warning module is based on real-time data of image recognition. When abnormal circuit breaker breaking is detected, it sends out a warning signal and generates different response strategies according to preset multi-level thresholds.

[0014] Preferably, the image acquisition module includes an acquisition unit and an image storage unit, the acquisition unit acquires real-time data through a high-resolution camera and an anti-interference device; the high-resolution camera is installed at a key position of the switch cabinet, including images of the circuit breaker and its operating area, and is used to acquire image data of the switch cabinet and the circuit breaker in real time; the anti-interference device includes an electromagnetic shielding cover and an anti-vibration bracket, which are used to eliminate the influence of electromagnetic interference and mechanical vibration on the camera;

[0015] The image storage unit stores the collected image data for maintaining data integrity and traceability; the image data storage unit is communicatively connected with the image preprocessing module and transmits the image data via an API.

[0016] Preferably, the image preprocessing module includes a preprocessing unit and an image screening unit; the preprocessing unit is used to receive the image data collected by the image acquisition module, and then eliminate the noise in the image data through a denoising algorithm, and then improve the contrast and brightness of the image by adjusting the grayscale distribution and local contrast of the image, and then extract key feature information from the image data through a feature extraction algorithm;

[0017] Finally, the overlap Dch, clarity Dqx, contrast Ddb and brightness Dl of all switchgear and circuit breaker image data are extracted and dimensionlessly processed to construct an image quality data set.

[0018] Preferably, the image screening unit is used to calculate the preprocessed switch cabinet and circuit breaker image data, and obtain the image index Txzs by extracting the image quality data set and merging the data for calculation. The specific calculation formula is as follows:

[0019]

[0020] After obtaining the image index Txzs, the image screening unit removes redundant images by using the preset image quality threshold Q and the image index Txzs. The specific contents are as follows:

[0021] If the image quality threshold Q ≥ the image index Txzs, it means that the current image is qualified and put into use;

[0022] If the image quality threshold Q is less than the image index Txzs, it means that the current image is unqualified and the image is discarded.

[0023] Preferably, the image recognition and analysis module includes a recognition model building unit and a disconnection prediction unit; the recognition model building unit establishes a deep learning model by selecting a convolutional neural network, and then trains and recognizes the screened switch cabinet and circuit breaker image data, inputs the image data of the switch cabinet and circuit breaker into the deep learning model, and then outputs the current working status of the circuit breaker through the model, including no abnormality, abnormality and warning status.

[0024] Preferably, the breaking prediction unit collects historical breaking image data and current real-time image data for analysis and comparison, and calculates the breaking stability difference Kdcz. The specific formula is as follows:

[0025]

[0026] In the formula, Kqd represents the electrical value before breaking, Khd represents the electrical value after breaking, Hzt represents the closing time, Fzt represents the opening time, Kqw represents the contact temperature before breaking, and Khw represents the contact temperature after breaking;

[0027] The disconnection prediction unit is configured to preset a difference threshold value W 1 Sum difference threshold W 2 , and the breaking stability difference Kdcz is evaluated, and the difference threshold W 1 >Difference threshold W 2 , the specific contents are as follows:

[0028] If the breaking stability difference Kdcz ≥ difference threshold W 1, indicating that the breaking stability of the circuit breaker is abnormal, in this case, the circuit breaker should be maintained or replaced;

[0029] If the difference threshold W 1 > Breaking stability difference Kdcz > Difference threshold W 2 , it means that the breaking stability of the circuit breaker is abnormal, but it does not exceed the preset difference threshold W 2 ,At this time, the result is transmitted to the time management module, and the monitoring time of the circuit breaker is adjusted twice;

[0030] If the difference threshold W 2 ≥ breaking stability difference Kdcz, it means that there is no abnormality in the breaking stability of the circuit breaker and the operating status of the circuit breaker.

[0031] Preferably, the time management module includes a time adjustment unit and a monitoring plan generation unit; the time adjustment unit adjusts the time point of the next monitoring according to the predicted breaking stability cycle; the monitoring plan generation unit generates a monitoring plan for the circuit breaker according to the time point calculated by the time adjustment unit, including the monitoring time, monitoring frequency and required monitoring parameters.

[0032] Preferably, the abnormal warning module includes an abnormal detection unit and a multi-level threshold strategy unit; the abnormal detection unit is used to extract the real-time image data in the image quality data set, and fit it with the breaking stability difference Kdcz of the breaking prediction unit, and calculate the breaking index Kdzs by the following formula:

[0033]

[0034] Where, Ykd represents the average monthly number of interruptions, and Ygz represents the average monthly number of failures.

[0035] Preferably, the multi-level threshold strategy unit presets a first-level abnormal threshold E 1 and the secondary anomaly threshold E 2 , analyze and evaluate the breaking index Kdzs, and the first-level abnormal threshold E 1 <Second level abnormal threshold E 2 <Level 3 abnormal threshold E 3 The specific evaluation contents are as follows:

[0036] If the secondary abnormal threshold E 2 >Breaking index Kdzs ≥ first level abnormal threshold E 1 , indicating that the breaking index Kdzs is the first abnormality, at this time it is determined as the first warning level, and a warning signal is issued, and the prompt information is sent to the maintenance personnel's patrol list;

[0037] If the third level abnormal threshold E 3 >Breaking index Kdzs ≥ secondary abnormal threshold E2 , indicating that the breaking index Kdzs is the second abnormality, at this time it is determined as the second warning level, and a warning signal is issued, and the prompt information is sent and placed on the maintenance personnel's waiting patrol list;

[0038] If the breaking index Kdzs ≥ the third level abnormal threshold E 3 , indicating that the breaking index Kdzs is the third abnormal, it is judged as the third warning level, and an alarm signal is issued, and the prompt information is sent to the communication equipment of the maintenance personnel.

[0039] A switch cabinet operation debugging monitoring method based on image recognition comprises the following steps:

[0040] Step 1: Use the image acquisition module to collect image data of the switch cabinet and circuit breaker in real time, and ensure the clarity and stability of the image data through high-resolution cameras and anti-interference design;

[0041] Step 2: Preprocess the collected image data through the image preprocessing module, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, calculate the image data, obtain the image index Txzs and evaluate it to eliminate redundant images;

[0042] Step 3: Through the image recognition and analysis module, based on the pre-processed image data, the working state of the circuit breaker is identified through a deep learning algorithm, and the circuit breaker's breaking stability cycle is predicted to provide more accurate monitoring and analysis;

[0043] Step 4: Through the time management module, according to the prediction results of the image recognition and analysis module, the monitoring time of the circuit breaker is dynamically adjusted, and then based on the monitoring time, the monitoring plan is adjusted again;

[0044] Step 5: Through the abnormal warning module, based on the real-time data of image recognition, when the circuit breaker is detected to be abnormal, a warning signal is issued, and different response strategies are generated according to the preset multi-level thresholds.

[0045] Beneficial Effects

[0046] The present invention provides a communication network fault early warning method and system based on artificial intelligence, which has the following beneficial effects:

[0047] (1) The image acquisition module uses a high-resolution camera and anti-interference device to ensure the clarity and stability of image acquisition, and the image storage unit ensures the integrity and traceability of the data; the image preprocessing module improves the image quality through denoising, enhancement and feature extraction, and calculates the image index Txzs and eliminates redundant images according to the threshold; the image recognition and analysis module uses a convolutional neural network to establish a recognition model, and the breaking prediction unit calculates the breaking stability difference Kdcz by analyzing historical and real-time data, and evaluates the breaking stability of the circuit breaker through a preset difference threshold.

[0048] (2) The time management module dynamically adjusts the monitoring time according to the actual opening and closing conditions of the circuit breaker to ensure the timeliness and accuracy of the monitoring operation, effectively cope with the influence of various unpredictable factors, and avoid the inaccuracy and untimeliness caused by fixed time management; the anomaly detection unit calculates the opening and closing index Kdzs through real-time data, and the multi-level threshold strategy unit issues different levels of warning signals according to the preset threshold. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the structural framework of a switch cabinet operation debugging monitoring system based on image recognition according to the present invention;

[0050] Figure 2 The present invention is a schematic flow chart of the steps of a switch cabinet operation debugging monitoring method based on image recognition. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Example 1

[0053] A switch cabinet operation and debugging monitoring system based on image recognition, comprising an image acquisition module, an image preprocessing module, an image recognition and analysis module, a time management module and an abnormal warning module;

[0054] The image acquisition module is used to collect image data of the switch cabinet and circuit breaker in real time, and ensures the clarity and stability of the image data through a high-resolution camera and anti-interference design;

[0055] The image preprocessing module is used to preprocess the collected image data, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, the image data is calculated, the image index Txzs is obtained and evaluated, so as to eliminate redundant images;

[0056] The image recognition and analysis module identifies the working state of the circuit breaker based on the preprocessed image data through a deep learning algorithm and predicts the circuit breaker's breaking stability cycle to provide more accurate monitoring and analysis;

[0057] The time management module dynamically adjusts the monitoring time of the circuit breaker according to the prediction results of the image recognition and analysis module, and then adjusts the monitoring plan again based on the monitoring time;

[0058] The abnormal warning module is based on real-time data of image recognition. When abnormal circuit breaker breaking is detected, it sends out a warning signal and generates different response strategies according to preset multi-level thresholds.

[0059] In this embodiment, the image acquisition module uses a high-resolution camera and anti-interference design to ensure the clarity and stability of the image data, thereby improving the accuracy of subsequent processing and analysis; the image preprocessing module improves the accuracy of image recognition through denoising, enhancement and feature extraction. The image index Txzs is calculated and evaluated, and redundant images are eliminated to ensure the data quality of processing and analysis; the image recognition and analysis module uses a deep learning algorithm to identify the working state of the circuit breaker and predict the stable opening cycle based on the preprocessed image data, providing more accurate monitoring and analysis; the time management module dynamically adjusts the monitoring time of the circuit breaker according to the prediction results of the image recognition and analysis module, ensuring the timeliness and accuracy of the monitoring operation and improving the monitoring efficiency; the abnormal warning module is based on real-time image recognition data, and when the circuit breaker is detected to be abnormal, a warning signal is issued. Through a multi-level threshold strategy, false alarms and missed alarms are reduced, and the accuracy and reliability of the warning are improved.

[0060] Example 2

[0061] The image acquisition module includes an acquisition unit and an image storage unit. The acquisition unit acquires real-time data through a high-resolution camera and an anti-interference device. The high-resolution camera is installed at a key position of the switch cabinet, including images of the circuit breaker and its operating area, and is used to acquire image data of the switch cabinet and the circuit breaker in real time. The anti-interference device includes an electromagnetic shielding cover and an anti-vibration bracket, which are used to eliminate the influence of electromagnetic interference and mechanical vibration on the camera.

[0062] The image storage unit stores the collected image data for maintaining data integrity and traceability; the image data storage unit is communicatively connected with the image preprocessing module and transmits the image data via an API.

[0063] In this embodiment, the high-resolution camera has high definition and wide-angle lens, can capture details in complex environments, and ensure the accuracy of image data. Specifically, the resolution of the high-resolution camera is not less than 1080p, has low-light performance and autofocus function, and can keep the image clear under changing light conditions. The electromagnetic shielding cover of the anti-interference device is made of highly conductive material, which can effectively shield the interference of external electromagnetic waves on the camera, and the anti-vibration bracket is made of elastic shock-absorbing material, which can absorb mechanical vibration and prevent the camera from shifting during operation.

[0064] Example 3

[0065] The image preprocessing module includes a preprocessing unit and an image screening unit; the preprocessing unit is used to receive the image data collected by the image acquisition module, and then eliminate the noise in the image data through a denoising algorithm, and then improve the contrast and brightness of the image by adjusting the grayscale distribution and local contrast of the image, and then extract key feature information from the image data through a feature extraction algorithm;

[0066] Finally, the overlap Dch, clarity Dqx, contrast Ddb and brightness Dl of all switchgear and circuit breaker image data are extracted and dimensionlessly processed to construct an image quality data set.

[0067] The image screening unit is used to calculate the pre-processed switch cabinet and circuit breaker image data, and obtain the image index Txzs by extracting the image quality data set and merging the data. The specific calculation formula is as follows:

[0068]

[0069] After obtaining the image index Txzs, the image screening unit removes redundant images by using the preset image quality threshold Q and the image index Txzs. The specific contents are as follows:

[0070] If the image quality threshold Q ≥ the image index Txzs, it means that the current image is qualified and put into use;

[0071] If the image quality threshold Q is less than the image index Txzs, it means that the current image is unqualified and the image is discarded.

[0072] In this embodiment, the acquisition overlap Dch ensures the consistency of the content and perspective of the images acquired multiple times. A high overlap means that the changes between the images are small, which helps to reduce the interference caused by different perspectives or positions and improve the consistency and comparability of the image data; the clarity Dqx ensures that the details of the image are clearly visible. High-definition images can provide more feature information, facilitate subsequent feature extraction and recognition, and improve the accuracy of image recognition; the contrast Ddb enhances the distinction between different areas in the image, making the key features more obvious. Appropriate contrast can highlight important details in the image, facilitate algorithm recognition and analysis, and improve recognition accuracy; the brightness Dl ensures that the overall brightness of the image is moderate, and the visibility and detail performance of the image are not affected by being too bright or too dark. Moderate brightness can avoid overexposure or underexposure of the image, so that all important details can be clearly captured and recognized.

[0073] Example 4

[0074] The image recognition and analysis module includes a recognition model building unit and a disconnection prediction unit; the recognition model building unit establishes a deep learning model by selecting a convolutional neural network, and then trains and recognizes the screened switch cabinet and circuit breaker image data, inputs the switch cabinet and circuit breaker image data into the deep learning model, and then outputs the current working status of the circuit breaker through the model, including no abnormality, abnormality and warning status.

[0075] The breaking prediction unit collects historical breaking image data and analyzes and compares the current real-time image data, and calculates the breaking stability difference Kdcz. The specific formula is as follows:

[0076]

[0077] In the formula, Kqd represents the electrical value before breaking, Khd represents the electrical value after breaking, Hzt represents the closing time, Fzt represents the opening time, Kqw represents the contact temperature before breaking, and Khw represents the contact temperature after breaking;

[0078] The disconnection prediction unit is configured to preset a difference threshold value W 1 Sum difference threshold W 2 , and the breaking stability difference Kdcz is evaluated, and the difference threshold W 1 >Difference threshold W 2 , the specific contents are as follows:

[0079] If the breaking stability difference Kdcz ≥ difference threshold W 1 , indicating that the breaking stability of the circuit breaker is abnormal, in this case, the circuit breaker should be maintained or replaced;

[0080] If the difference threshold W 1 > Breaking stability difference Kdcz > Difference threshold W 2, it means that the breaking stability of the circuit breaker is abnormal, but it does not exceed the preset difference threshold W 2 ,At this time, the result is transmitted to the time management module, and the monitoring time of the circuit breaker is adjusted twice;

[0081] If the difference threshold W 2 ≥ breaking stability difference Kdcz, it means that there is no abnormality in the breaking stability of the circuit breaker and the operating status of the circuit breaker.

[0082] In this embodiment, the collected electrical value Kqd before breaking represents the electrical state of the circuit breaker before the breaking operation, reflecting the current, voltage and other parameters under normal operation; the collected electrical value Khd after breaking represents the electrical state of the circuit breaker after the breaking operation, helping to determine whether the breaking operation is successful and the recovery of the electrical system after breaking. By comparing with the electrical value Kqd, the effect of the breaking operation and the performance of the circuit breaker are evaluated; the closing time Hzt refers to the time required for the circuit breaker from the start of the closing operation to the completion of the closing operation. This time reflects the response speed and mechanical performance of the closing operation of the circuit breaker. Too long or too short a closing time may indicate potential mechanical or electrical problems; the opening time Fzt It refers to the time required for the circuit breaker to start the opening operation and complete the opening. The opening time also reflects the response speed and mechanical performance of the circuit breaker opening operation, and is an important indicator for evaluating the reliability of the circuit breaker; the contact temperature Kqw before breaking is collected, which indicates the temperature of the contacts of the circuit breaker before the breaking operation. High temperature may indicate problems such as poor contact or overload of the contacts. By monitoring this parameter, potential faults can be discovered and prevented in advance; the contact temperature Khw after breaking indicates the temperature of the contacts of the circuit breaker after the breaking operation. The change in contact temperature reflects the impact of the breaking operation on the contacts. Excessive temperature changes may indicate contact wear or other problems. By monitoring this parameter, the health status of the circuit breaker can be evaluated.

[0083] Example 5

[0084] The time management module includes a time adjustment unit and a monitoring plan generation unit; the time adjustment unit adjusts the time point of the next monitoring according to the predicted breaking stability cycle; the monitoring plan generation unit generates a monitoring plan for the circuit breaker according to the time point calculated by the time adjustment unit, including the monitoring time, monitoring frequency and required monitoring parameters.

[0085] In this embodiment, the time adjustment unit dynamically adjusts the specific time point of the next monitoring according to the real-time monitoring results and the output of the prediction model to ensure that detection and maintenance are performed before the circuit breaker may become abnormal; the monitoring plan generation unit sets a reasonable monitoring frequency according to the working status and historical data of the circuit breaker; including appropriately reducing the monitoring frequency during the stable period of the circuit breaker and increasing the monitoring frequency during the unstable period.

[0086] Example 6

[0087] The abnormal warning module includes an abnormal detection unit and a multi-level threshold strategy unit; the abnormal detection unit is used to extract the real-time image data in the image quality data set, and fit it with the breaking stability difference Kdcz of the breaking prediction unit, and calculate the breaking index Kdzs by the following formula:

[0088]

[0089] Where, Ykd represents the average monthly number of interruptions, and Ygz represents the average monthly number of failures.

[0090] The multi-level threshold strategy unit presets a first-level abnormal threshold E 1 and the secondary anomaly threshold E 2 , analyze and evaluate the breaking index Kdzs, and the first-level abnormal threshold E 1 <Second level abnormal threshold E 2 <Level 3 abnormal threshold E 3 The specific evaluation contents are as follows:

[0091] If the secondary abnormal threshold E 2 >Breaking index Kdzs ≥ first level abnormal threshold E 1 , indicating that the breaking index Kdzs is the first abnormality, at this time it is determined as the first warning level, and a warning signal is issued, and the prompt information is sent to the maintenance personnel's patrol list;

[0092] If the third level abnormal threshold E 3 >Breaking index Kdzs ≥ secondary abnormal threshold E 2 , indicating that the breaking index Kdzs is the second abnormality, at this time it is determined as the second warning level, and a warning signal is issued, and the prompt information is sent and placed on the maintenance personnel's waiting patrol list;

[0093] If the breaking index Kdzs ≥ the third level abnormal threshold E 3 , indicating that the breaking index Kdzs is the third abnormal, it is judged as the third warning level, and an alarm signal is issued, and the prompt information is sent to the communication equipment of the maintenance personnel.

[0094] In this embodiment, the average monthly number of breaking times Ykd represents the average number of breaking operations performed by the circuit breaker in one month. This parameter reflects the use frequency and load conditions of the circuit breaker. A higher average monthly number of breaking times may mean that the circuit breaker bears a higher load and requires more frequent monitoring and maintenance. The average monthly number of failures Ygz represents the average number of failures of the circuit breaker in one month. This parameter is a key indicator for evaluating the reliability of the circuit breaker. A higher average monthly number of failures indicates that the circuit breaker may have potential problems or is aging and needs to be repaired or replaced. The breaking stability difference Kdcz represents the performance difference of the circuit breaker before and after the breaking operation, including changes in electrical values ​​and contact temperature, reflecting the stability and consistency of the circuit breaker in the breaking operation. A normal breaking stability difference Kdcz indicates that the breaking operation of the circuit breaker is stable, and an abnormal breaking stability difference Kdcz indicates that there are mechanical or electrical problems.

[0095] Example 7

[0096] A switch cabinet operation debugging monitoring method based on image recognition comprises the following steps:

[0097] Step 1: Use the image acquisition module to collect image data of the switch cabinet and circuit breaker in real time, and ensure the clarity and stability of the image data through high-resolution cameras and anti-interference design;

[0098] Step 2: Preprocess the collected image data through the image preprocessing module, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, calculate the image data, obtain the image index Txzs and evaluate it to eliminate redundant images;

[0099] Step 3: Through the image recognition and analysis module, based on the pre-processed image data, the working state of the circuit breaker is identified through a deep learning algorithm, and the circuit breaker's breaking stability cycle is predicted to provide more accurate monitoring and analysis;

[0100] Step 4: Through the time management module, according to the prediction results of the image recognition and analysis module, the monitoring time of the circuit breaker is dynamically adjusted, and then based on the monitoring time, the monitoring plan is adjusted again;

[0101] Step 5: Through the abnormal warning module, based on the real-time data of image recognition, when the circuit breaker is detected to be abnormal, a warning signal is issued, and different response strategies are generated according to the preset multi-level thresholds.

[0102] In this embodiment, step one ensures that the collected image data is clear, stable and free of interference, providing a high-quality data basis for subsequent image processing and analysis; step two improves the accuracy of image recognition, eliminates redundant images, and ensures that only high-quality image data is used in the subsequent analysis process, thereby improving the overall efficiency and accuracy of the system; step three uses a deep learning algorithm to accurately identify the working state of the circuit breaker and predict its stable breaking cycle, providing data support for dynamically adjusting the monitoring time and improving the accuracy and effectiveness of monitoring; step four dynamically adjusts the monitoring time according to the prediction results to ensure the timeliness and accuracy of the monitoring operation, avoids the lag and inflexibility of traditional fixed-time monitoring, and improves the efficiency of monitoring and maintenance; step five monitors the breaking state of the circuit breaker in real time, issues early warning signals in a timely manner, reduces false alarms and missed alarms based on a multi-level threshold strategy, and ensures the safety and reliability of the system.

[0103] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A switch cabinet operation debugging monitoring system based on image recognition, characterized in that: It includes image acquisition module, image preprocessing module, image recognition and analysis module, time management module and abnormal warning module; The image acquisition module is used to collect image data of the switch cabinet and circuit breaker in real time, and ensures the clarity and stability of the image data through a high-resolution camera and anti-interference design; The image preprocessing module is used to preprocess the collected image data, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, the image data is calculated, the image index Txzs is obtained and evaluated, so as to eliminate redundant images; The image recognition and analysis module identifies the working state of the circuit breaker based on the preprocessed image data through a deep learning algorithm and predicts the circuit breaker's breaking stability cycle to provide more accurate monitoring and analysis; The time management module dynamically adjusts the monitoring time of the circuit breaker according to the prediction results of the image recognition and analysis module, and then adjusts the monitoring plan based on the monitoring time; The abnormal warning module is based on real-time data of image recognition. When abnormal circuit breaker breaking is detected, it sends out a warning signal and generates different response strategies according to preset multi-level thresholds.

2. The switch cabinet operation debugging monitoring system based on image recognition according to claim 1 is characterized in that: The image acquisition module includes an acquisition unit and an image storage unit. The acquisition unit acquires real-time data through a high-resolution camera and an anti-interference device. The high-resolution camera is installed at a key position of the switch cabinet, including images of the circuit breaker and its operating area, and is used to acquire image data of the switch cabinet and the circuit breaker in real time. The anti-interference device includes an electromagnetic shielding cover and an anti-vibration bracket, which are used to eliminate the influence of electromagnetic interference and mechanical vibration on the camera. The image storage unit stores the collected image data for maintaining data integrity and traceability; the image data storage unit is communicatively connected with the image preprocessing module and transmits the image data via an API.

3. The switch cabinet operation debugging monitoring system based on image recognition according to claim 2 is characterized in that: The image preprocessing module includes a preprocessing unit and an image screening unit; the preprocessing unit is used to receive the image data collected by the image acquisition module, and then eliminate the noise in the image data through a denoising algorithm, and then improve the contrast and brightness of the image by adjusting the grayscale distribution and local contrast of the image, and then extract key feature information from the image data through a feature extraction algorithm; Finally, the overlap Dch, clarity Dqx, contrast Ddb and brightness Dl of all switchgear and circuit breaker image data are extracted and dimensionlessly processed to construct an image quality data set.

4. The switch cabinet operation debugging monitoring system based on image recognition according to claim 3 is characterized in that: The image screening unit is used to calculate the pre-processed switch cabinet and circuit breaker image data, and obtain the image index Txzs by extracting the image quality data set and merging the data. The specific calculation formula is as follows: After obtaining the image index Txzs, the image screening unit removes redundant images by using the preset image quality threshold Q and the image index Txzs. The specific contents are as follows: If the image quality threshold Q ≥ the image index Txzs, it means that the current image is qualified and put into use; If the image quality threshold Q is less than the image index Txzs, it means that the current image is unqualified and the image is discarded.

5. The switch cabinet operation debugging monitoring system based on image recognition according to claim 4 is characterized in that: The image recognition and analysis module includes a recognition model building unit and a disconnection prediction unit; the recognition model building unit establishes a deep learning model by selecting a convolutional neural network, and then trains and recognizes the screened switch cabinet and circuit breaker image data, inputs the switch cabinet and circuit breaker image data into the deep learning model, and then outputs the current working status of the circuit breaker through the model, including no abnormality, abnormality and warning status.

6. The switch cabinet operation debugging monitoring system based on image recognition according to claim 5 is characterized in that: The breaking prediction unit collects historical breaking image data and analyzes and compares the current real-time image data, and calculates the breaking stability difference Kdcz. The specific formula is as follows: In the formula, Kqd represents the electrical value before breaking, Khd represents the electrical value after breaking, Hzt represents the closing time, Fzt represents the opening time, Kqw represents the contact temperature before breaking, and Khw represents the contact temperature after breaking; The disconnection prediction unit is configured to preset a difference threshold value W 1 Sum difference threshold W 2 , and the breaking stability difference Kdcz is evaluated, and the difference threshold W 1 >Difference threshold W 2 , the specific contents are as follows: If the breaking stability difference Kdcz ≥ difference threshold W 1 , indicating that the breaking stability of the circuit breaker is abnormal, in this case, the circuit breaker should be maintained or replaced; If the difference threshold W 1 > Breaking stability difference Kdcz > Difference threshold W 2 , it means that the breaking stability of the circuit breaker is abnormal, but it does not exceed the preset difference threshold W 2 ,At this time, the result is transmitted to the time management module, and the monitoring time of the circuit breaker is adjusted twice; If the difference threshold W 2 ≥ breaking stability difference Kdcz, it means that there is no abnormality in the breaking stability of the circuit breaker and the operating status of the circuit breaker.

7. The switch cabinet operation debugging monitoring system based on image recognition according to claim 6 is characterized in that: The time management module includes a time adjustment unit and a monitoring plan generation unit; the time adjustment unit adjusts the time point of the next monitoring according to the predicted breaking stability cycle; the monitoring plan generation unit generates a monitoring plan for the circuit breaker according to the time point calculated by the time adjustment unit, including the monitoring time, monitoring frequency and required monitoring parameters.

8. The switch cabinet operation debugging monitoring system based on image recognition according to claim 7 is characterized in that: The abnormal warning module includes an abnormal detection unit and a multi-level threshold strategy unit; the abnormal detection unit is used to extract the real-time image data in the image quality data set, and fit it with the breaking stability difference Kdcz of the breaking prediction unit, and calculate the breaking index Kdzs by the following formula: Where, Ykd represents the average monthly number of interruptions, and Ygz represents the average monthly number of failures.

9. The switch cabinet operation debugging monitoring system based on image recognition according to claim 8 is characterized in that: The multi-level threshold strategy unit presets a first-level abnormal threshold E 1 and the secondary anomaly threshold E 2 , analyze and evaluate the breaking index Kdzs, and the first-level abnormal threshold E 1 <Second level abnormal threshold E 2 <Level 3 abnormal threshold E 3 The specific evaluation contents are as follows: If the secondary abnormal threshold E 2 >Breaking index Kdzs ≥ first level abnormal threshold E 1 , indicating that the breaking index Kdzs is the first abnormality, at this time it is determined as the first warning level, and a warning signal is issued, and the prompt information is sent to the maintenance personnel's patrol list; If the third level abnormal threshold E 3 >Breaking index Kdzs ≥ secondary abnormal threshold E 2 , indicating that the breaking index Kdzs is the second abnormality, at this time it is determined as the second warning level, and a warning signal is issued, and the prompt information is sent and placed on the maintenance personnel's waiting patrol list; If the breaking index Kdzs ≥ the third level abnormal threshold E 3 , indicating that the breaking index Kdzs is the third abnormal, it is judged as the third warning level, and an alarm signal is issued, and the prompt information is sent to the communication equipment of the maintenance personnel.

10. A switch cabinet operation debugging monitoring method based on image recognition, according to a switch cabinet operation debugging monitoring system based on image recognition according to claims 1 to 9, characterized in that: The following steps are involved: Step 1: Use the image acquisition module to collect image data of the switch cabinet and circuit breaker in real time, and ensure the clarity and stability of the image data through high-resolution cameras and anti-interference design; Step 2: Preprocess the collected image data through the image preprocessing module, including image denoising, enhancement and feature extraction, to improve the accuracy of image recognition; finally, calculate the image data, obtain the image index Txzs and evaluate it to eliminate redundant images; Step 3: Through the image recognition and analysis module, based on the pre-processed image data, the working state of the circuit breaker is identified through a deep learning algorithm, and the circuit breaker's breaking stability cycle is predicted to provide more accurate monitoring and analysis; Step 4: Through the time management module, dynamically adjust the monitoring time of the circuit breaker according to the prediction results of the image recognition and analysis module, and then adjust the monitoring plan again based on the monitoring time; Step 5: Through the abnormal warning module, based on the real-time data of image recognition, when the circuit breaker is detected to be abnormal, a warning signal is issued, and different response strategies are generated according to the preset multi-level thresholds.

Images