Automatic detection method and system for electric power equipment detection instruments based on machine vision

Abstract

The present invention discloses an automatic detection method and system for electric power equipment detection instruments based on machine vision, and relates to the technical field of automatic instrument detection. The automatic detection method for electric power equipment detection instruments based on machine vision continuously obtains display panel image data of the electric power equipment detection instrument, performs preprocessing and regional division, extracts measurement data of the detection instrument, such as voltage, current, frequency, etc., and performs comprehensive analysis, calculates the detection evaluation index, and compares it with a preset threshold. The present invention continuously obtains and processes the display panel image data of the electric power equipment detection instrument, and can monitor the working status of the equipment in real time at multiple time points, and timely discover the abnormality of the data during the detection process. This improvement in real-time performance and accuracy helps to realize data collection for electric power equipment detection instruments that do not have data transmission functions, and can directly generate experimental records and reports, which can realize unmanned operation and ensure personal safety in high-risk operation sites.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic instrument detection, and in particular to an automatic detection method and system for electric power equipment detection instruments based on machine vision. Background Art

[0002] As the power industry's requirements for measurement accuracy, equipment stability and reliability continue to increase, the detection and testing of power equipment has become increasingly important. Power equipment testing instruments are usually used to measure electrical parameters such as voltage, current, frequency, power factor and power of power equipment, and provide key information for the metering performance of various components of the power system (such as transformers, power distribution equipment, etc.), ensuring the safe and stable operation of the power system. In order to ensure the accuracy and reliability of the metering data of power equipment, regular testing and testing has become an indispensable task.

[0003] The inspection and testing of power equipment are usually carried out in special indoor laboratories or testing sites. These environments are relatively closed and safe, which helps to avoid interference from the external environment on the measurement results. Testing power equipment indoors can not only effectively avoid the impact of external weather and temperature fluctuations on equipment performance, but also provide inspectors with a safer and controllable working environment. However, traditional testing methods still face some challenges, especially in the inspection process of high-voltage power equipment, where inspectors often need to operate close to the power equipment, which poses potential safety risks.

[0004] Traditional methods of testing and inspecting power equipment usually rely on manual operation, and the power equipment is inspected by manually adjusting the testing instruments. This method not only has low work efficiency, but also has a complex and error-prone testing process. Especially in the indoor testing process of high-voltage power equipment, although the environment is relatively controlled, the inspectors still need to come into contact with dangerous power equipment such as high voltage and high current, which greatly increases the safety risks. In this environment, operators are often exposed to dangerous conditions such as high voltage electricity and strong electromagnetic interference, which makes it very easy to make operational errors. In serious cases, it may even cause electrical accidents, fires or casualties.

[0005] In addition, traditional detection methods usually make it difficult to achieve real-time data uploading and remote monitoring. It is often difficult to obtain test results and make adjustments in real time during the detection process, making it difficult to quickly discover potential problems and deal with them. Therefore, traditional detection and detection methods not only have safety hazards brought about by manual operation, but also have much room for improvement in efficiency and accuracy.

[0006] The limitations of the existing technology include at least the following problems. First, the detection methods of existing power equipment detection instruments usually rely on manual operation and manual recording, and have problems such as low efficiency, large errors and poor real-time performance. Especially in continuous monitoring and multi-time point data analysis, traditional methods often find it difficult to process large amounts of data in a timely manner, and the possibility of human intervention is high, which can easily cause errors. Secondly, traditional detection methods often find it difficult to achieve real-time and dynamic equipment status evaluation, and it is difficult to monitor the health status of the equipment for a long time and continuously, which can easily lead to potential abnormalities of the equipment being difficult to detect and handle in a timely manner, thereby affecting the operating stability and safety of the equipment. In addition, the existing technology does not fully utilize machine vision and automated image analysis technology, and it is difficult to efficiently extract and process real-time data from the display panel, which not only increases the complexity of the operation, but also leads to an extension of the detection cycle and excessive reliance on personnel. Therefore, the existing technology has obvious deficiencies in efficient, real-time and accurate power equipment detection, and it is difficult to meet the high requirements of modern power systems for detection accuracy and automation. Summary of the invention

[0007] In view of the deficiencies of the prior art, the present invention provides an automatic detection method and system for electric power equipment detection instruments based on machine vision, which solves the problems that the detection methods of the prior electric power equipment detection instruments usually rely on manual operation and manual recording, and have low efficiency, large errors and poor real-time performance. In particular, in continuous monitoring and multi-time point data analysis, traditional methods are often difficult to process a large amount of data in a timely manner, and the possibility of human intervention is high, which is easy to cause errors. Secondly, traditional detection methods are often difficult to achieve real-time and dynamic equipment status evaluation, and it is difficult to monitor the health status of the equipment for a long time and continuously, which easily leads to potential abnormalities of the equipment being difficult to discover and handle in time, thereby affecting the operational stability and safety of the equipment. In addition, the prior art does not make full use of machine vision and automated image analysis technology, and it is difficult to efficiently extract and process real-time data of the display panel, which not only increases the complexity of the operation, but also leads to the extension of the detection cycle and excessive reliance on personnel. Therefore, the prior art has obvious deficiencies in efficient, real-time and accurate detection of electric power equipment, and it is difficult to meet the high requirements of modern power systems for detection accuracy and automation.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: an automatic detection method for electric power equipment detection instruments based on machine vision, comprising the following steps: continuously acquiring display panel image data of the electric power equipment detection instrument at several time points, and performing preprocessing, wherein the display panel image data includes the pixel value of each pixel in the display panel image; performing regional division on the display panel image data of the electric power equipment detection instrument at each time point after preprocessing, and obtaining several display areas of the display panel image of the electric power equipment detection instrument at each time point; performing information processing on each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, respectively. Extract and summarize the measurement data of the power equipment detection instrument at each time point, and perform storage and backup processing, wherein the measurement data includes voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value; perform comprehensive analysis on the measurement data of the power equipment detection instrument at each time point, obtain the detection evaluation index of the power equipment detection instrument, and perform storage and backup processing; determine whether the detection evaluation index of the power equipment detection instrument is within a preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, issue a detection abnormality alarm to relevant staff.

[0009] The automatic detection system of electric power equipment detection instrument based on machine vision includes: an image acquisition and processing module, a region division module, an information extraction module, an evaluation and analysis module, and a detection and judgment module; the image acquisition and processing module is used to continuously acquire the display panel image data of the electric power equipment detection instrument at several time points and perform preprocessing, and the display panel image data includes the pixel value of each pixel in the display panel image; the region division module is used to perform region division on the display panel image data of the electric power equipment detection instrument at each time point after preprocessing, and obtain several display areas of the display panel image of the electric power equipment detection instrument at each time point; the information extraction module is used to perform region division on each display panel image of the electric power equipment detection instrument at each time point The information is extracted from each display area respectively, and summarized and processed to obtain the measurement data of the power equipment detection instrument at each time point, and the measurement data includes voltage measurement value, current measurement value, frequency measurement value, power factor measurement value and power measurement value; the evaluation and analysis module is used to perform comprehensive analysis on the measurement data of the power equipment detection instrument at each time point respectively, obtain the detection evaluation index of the power equipment detection instrument, and perform storage and backup processing; the detection judgment module is used to judge whether the detection evaluation index of the power equipment detection instrument is within a preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, send a detection abnormality alarm to relevant staff.

[0010] The present invention has the following beneficial effects:

[0011] (1) The automatic detection method of power equipment detection instruments based on machine vision can monitor the working status of the equipment in real time at multiple time points and promptly discover data anomalies during the detection process by continuously acquiring and processing the display panel image data of the power equipment detection instruments. In the image preprocessing step, the image clarity and contrast are effectively improved through meticulous grayscale processing and gradient analysis, providing an accurate basis for subsequent character segmentation and information extraction. This improvement in real-time performance and accuracy helps to realize data collection for power equipment detection instruments that do not have data transmission functions, and can directly generate experimental records and reports. At the same time, unmanned operation can be realized to ensure personal safety in high-risk operating locations.

[0012] (2) The automatic detection method of the power equipment detection instrument based on machine vision automatically detects the power equipment without human intervention. It analyzes the display panel image through machine vision and intelligent algorithms, extracts the measurement data of the detection instrument, calculates the detection evaluation index, and automatically triggers an abnormal alarm according to the set threshold. This automated process not only greatly reduces the complexity and error of manual operation, but also can promptly issue an abnormal alarm when the power equipment deviates from the standard, so as to ensure the stable operation of the power equipment.

[0013] (3) The automatic detection method for power equipment detection instruments based on machine vision can evaluate the working status of the instrument from multiple angles by comprehensively analyzing multiple dimensions in the measurement data of the detection instrument, such as voltage, current, frequency, etc., as well as their rate of change and deviation. This multi-dimensional evaluation method is more comprehensive and accurate than the traditional single-indicator detection method. It can effectively identify potential problems in power equipment and avoid misjudgment or omission caused by a single indicator. At the same time, it combines reference data for evaluation to ensure that the detection process is more scientific and objective, thereby enhancing the long-term stability and accuracy of power equipment.

[0014] (4) The automatic detection system for power equipment detection instruments based on machine vision can continuously capture and analyze images of the display panel of the power equipment detection instrument at multiple time points, thereby monitoring the actual operating data of the power equipment in real time and performing automatic detection. Compared with the traditional manual detection method, this method has higher detection accuracy and efficiency. Through automated information extraction and comprehensive evaluation and analysis, it can immediately issue abnormal alarms when deviations occur in the power equipment, thus avoiding performance degradation caused by equipment errors. In addition, the system can also dynamically evaluate the measurement data of the detection instrument to ensure that the power equipment is always within the preset normal working range, thereby improving the stability and service life of the power equipment and reducing equipment maintenance costs and the need for manual intervention.

[0015] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The present invention is a flow chart of the automatic detection method of the electric power equipment detection instrument based on machine vision.

[0017] Figure 2 The present invention is a block diagram of the automatic detection system of the electric power equipment detection instrument based on machine vision. DETAILED DESCRIPTION

[0018] The embodiment of the present application solves the problem that the detection method of the existing power equipment detection instrument usually relies on manual operation and manual recording, and has low efficiency, large errors and poor real-time performance, especially in continuous monitoring and multi-time point data analysis. Traditional methods are often difficult to process a large amount of data in a timely manner, and the possibility of human intervention is high, which is easy to cause errors. Secondly, traditional detection methods are often difficult to achieve real-time and dynamic equipment status evaluation, and it is difficult to monitor the health status of the equipment for a long time and continuously, which easily leads to potential abnormalities of the equipment being difficult to detect and handle in time, thereby affecting the operation stability and safety of the equipment. In addition, the existing technology does not make full use of machine vision and automated image analysis technology, and it is difficult to efficiently extract and process the real-time data of the display panel, which not only increases the complexity of the operation, but also leads to the extension of the detection cycle and excessive reliance on personnel. Therefore, the existing technology has obvious deficiencies in efficient, real-time and accurate power equipment detection, and it is difficult to meet the high requirements of modern power systems for detection accuracy and automation.

[0019] The overall idea of ​​the problem in the embodiment of this application is as follows:

[0020] Firstly, by continuously acquiring the display panel image data of the power equipment detection instrument, a series of preprocessing and analysis steps are performed to realize automatic detection and evaluation. Firstly, after the image data is preprocessed by grayscale processing, gradient analysis and area division, the feature information of each display area is extracted, and enhancement and filtering are performed. Then, the image of each area is marked with connected domains and character segmentation, and the recognized characters are input into the pre-trained neural network model for classification, and the measurement data (such as voltage, current, power, etc.) are extracted. In the data analysis stage, the various detection and evaluation indexes of the power equipment are obtained through change rate analysis and deviation calculation. Finally, according to the set evaluation threshold, it is judged whether there is any abnormality in the power equipment, and an alarm is issued when an abnormality is found. This method not only has a high degree of automation, but also improves the detection accuracy and efficiency through the combination of image analysis and deep learning. It is suitable for remote monitoring and status evaluation of power equipment.

[0021] See also Figure 1 The embodiment of the present invention provides a technical solution: an automatic detection method for electric power equipment detection instruments based on machine vision, comprising the following steps: continuously acquiring display panel image data of the electric power equipment detection instrument at several time points through a camera, and performing preprocessing, the display panel image data including the pixel value of each pixel in the display panel image; performing area division on the display panel image data of the electric power equipment detection instrument at each time point after preprocessing, respectively, to obtain several display areas of the display panel image of the electric power equipment detection instrument at each time point; performing information extraction on each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, and performing summary processing, to obtain the display area of ​​each time point. The measurement data of the power equipment detection instrument (this data is the measurement data of the power equipment detected by the power equipment detection instrument, such as the transformer), and storage backup processing is performed. The measurement data includes voltage measurement values, current measurement values, frequency measurement values, power factor measurement values, and power measurement values; the measurement data of the power equipment detection instrument at each time point are comprehensively analyzed to obtain the detection evaluation index of the power equipment detection instrument, and storage backup processing is performed; it is determined whether the detection evaluation index of the power equipment detection instrument is within the preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, a detection abnormality alarm is issued to relevant staff.

[0022] Specifically, the power equipment testing instrument can be placed on the quick positioning trolley, and different pads can be selected according to different instruments, so that the testing instrument is placed on the pad to carry out the testing work.

[0023] Among them, the storage backup processing is to store the measured data in a preset data storage library, and the staff can download and view the measured data from the data storage library.

[0024] Specifically, the specific steps of preprocessing the display panel image data of the power equipment detection instrument at each time point are as follows: grayscale processing is performed on the display panel image data of the power equipment detection instrument at each time point to obtain the display panel grayscale image data of the power equipment detection instrument at each time point, and the display panel grayscale image data includes the grayscale pixel value of each grayscale pixel point in the display panel grayscale image; based on a preset gradient operator (such as a Sobel operator), the gradient value of each grayscale pixel point in the display panel grayscale image of the power equipment detection instrument at each time point is analyzed; the display panel grayscale image data of the power equipment detection instrument at each time point is analyzed. The image processing area data of the power equipment detection instrument at each time point are divided and processed separately to obtain a number of image processing area data of the power equipment detection instrument at each time point, and each image processing area data includes a number of grayscale pixels; each image processing area data of the power equipment detection instrument at each time point is compared and analyzed (i.e., the grayscale pixel values ​​are compared) to obtain the maximum grayscale pixel value and the minimum grayscale pixel value of each image processing area of ​​the power equipment detection instrument at each time point, and the regional contrast value of each image processing area of ​​the power equipment detection instrument at each time point is analyzed; the grayscale pixel value of each grayscale pixel point in each image processing area data of the power equipment detection instrument at each time point is compared and analyzed. The pixel values ​​of the image processing area of ​​the power equipment detection instrument are comprehensively analyzed to obtain the regional texture complexity (i.e., standard deviation) of each image processing area of ​​the power equipment detection instrument at each time point; the regional contrast value and regional texture complexity of each image processing area of ​​the power equipment detection instrument at each time point are comprehensively analyzed to obtain the pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point; the pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point is comprehensively analyzed in combination with the grayscale pixel value and gradient value of each grayscale pixel point in the data of each image processing area of ​​the power equipment detection instrument at each time point to obtain the grayscale pixel enhancement value of each grayscale pixel point in the data of each image processing area of ​​the power equipment detection instrument at each time point, and the grayscale pixel enhancement data of each grayscale pixel point in the data of each image processing area of ​​the power equipment detection instrument at each time point are summarized to obtain the display panel image enhancement processing data of the power equipment detection instrument at each time point, and the display panel image enhancement processing data includes the grayscale pixel enhancement value of each grayscale pixel point in the display panel image; the display panel image enhancement processing data of the power equipment detection instrument at each time point are filtered to obtain the display panel image filtering processing data of the power equipment detection instrument at each time point, and the display panel image filtering processing data includes the grayscale pixel filtering value of each grayscale pixel point in the display panel image.

[0025] The specific formulas for calculating the regional contrast value, regional texture complexity, pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point, and the grayscale pixel enhancement value of each grayscale pixel point in each image processing area data of the power equipment detection instrument at each time point are as follows: ;in, For the The first time point of the power equipment testing instrument The regional contrast value of the image processing area, For the The first time point of the power equipment testing instrument The maximum grayscale pixel value of the image processing area, For the The first time point of the power equipment testing instrument The minimum grayscale pixel value of the image processing area, For the The first time point of the power equipment testing instrument The regional texture complexity of the image processing area, For the The first time point of the power equipment testing instrument The image processing area data The grayscale pixel value of the grayscale pixel point, For the The first time point of the power equipment testing instrument The pixel adjustment coefficient of the image processing area, For the The first time point of the power equipment testing instrument The image processing area data The grayscale pixel enhancement value of grayscale pixels, For the The first time point of the power equipment testing instrument The image processing area data The gradient value of a grayscale pixel, =1, 2, 3, ..., , is the number of time points, =1, 2, 3, ..., , is the number of image processing regions, =1, 2, 3, …, , It is the number of grayscale pixels in the image processing area data.

[0026] It needs to be explained that is a smoothing term to prevent the pixel adjustment coefficient from being too large when the contrast is extremely small (low contrast area), where For the The first time point of the power equipment testing instrument The regional contrast value of the image processing area.

[0027] In this implementation scheme, by performing grayscale processing and gradient analysis on the display panel image data of the power equipment detection instrument at each time point, the details in the image can be effectively enhanced, especially in the low-contrast area, and the contrast and texture complexity of the image can be enhanced. This can ensure that important measurement data information is not ignored in the subsequent area division and information extraction, thereby improving the accuracy of data extraction. By dividing the display panel image at each time point into multiple image processing areas and analyzing the maximum and minimum grayscale values ​​of each area, the contrast and texture complexity of different areas can be evaluated more carefully, thereby helping the system to identify and separate key electrical parameter areas, reducing errors and interference. After performing contrast and texture complexity analysis on each image processing area, the pixel adjustment coefficient obtained comprehensively can dynamically adjust the pixel value of the image according to the image characteristics. This adaptive adjustment mechanism can optimize the image. The processing process avoids the processing limitations brought by fixed parameters, especially in some areas with large grayscale changes. By adding contrast smoothing items, excessive pixel adjustment in low-contrast areas is further avoided, maintaining the stability and accuracy of the processing. By comprehensively analyzing the enhancement value of each grayscale pixel point and combining the gradient value of each pixel for further optimization, noise can be effectively eliminated and the overall quality of image data can be improved. The processed image data is more stable and has higher recognition accuracy, providing a more accurate basis for subsequent measurement data extraction and detection judgment. The final filtering processing step makes the image clearer by eliminating noise and smoothing irregularities in the image, avoiding over-enhancement or over-smoothing of details in the image, and ensuring the authenticity and availability of the image, which is particularly important for subsequent measurement data analysis and equipment detection, and can further improve the accuracy of detection.

[0028] Specifically, the specific steps for obtaining the filtered processing data of the display panel image of the electric power equipment detection instrument at each time point are as follows: for each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point, a number of adjacent grayscale pixels in a set neighborhood are identified respectively; the grayscale pixel enhancement value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point is comprehensively analyzed in combination with the grayscale pixel enhancement value of each adjacent grayscale pixel in the set neighborhood to obtain the filtering adjustment coefficient of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point; the grayscale pixel enhancement value and filtering adjustment coefficient of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point are comprehensively analyzed respectively to obtain the grayscale pixel filtering value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point.

[0029] The specific formula for calculating the filter adjustment coefficient and the grayscale pixel filter value of each grayscale pixel in the display panel image of the power equipment detection instrument at each time point is as follows: ;in, For the The display panel image of the power equipment detection instrument at a time point The filter adjustment coefficient of grayscale pixels, For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel in the set neighborhood The grayscale pixel enhancement value of adjacent grayscale pixels, is the filter adjustment factor stored in the database, which is a constant that controls the sensitivity of noise removal. For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel filter value of grayscale pixels, for The display panel image of the power equipment detection instrument at a time point The grayscale pixel enhancement value of grayscale pixels, =1, 2, 3, …, , is the number of time points, =1, 2, 3, …, , is the number of grayscale pixels in the display panel image, =1, 2, 3, …, , The number of adjacent grayscale pixels within the set neighborhood of the grayscale pixel.

[0030] In this embodiment, by comprehensively analyzing each grayscale pixel point in combination with the enhanced values ​​of the adjacent pixels in its neighborhood, the noise in the image can be effectively removed, especially in areas with complex image details. The neighborhood analysis makes the filtering result of each pixel point not only depend on its own grayscale value, but also takes into account the influence of the surrounding pixels, so as to more accurately smooth the noise in the image and maintain the overall structure and details of the image. The analysis combined with the enhanced values ​​of the adjacent grayscale pixels in the neighborhood can avoid excessive blurring of the image details by simple mean filtering or Gaussian filtering. By comprehensively analyzing the enhanced values ​​of each grayscale pixel point and its surrounding pixels, the filtered image can retain more detail information while denoising, ensuring that subsequent measurement data extraction is not affected by image quality degradation. The introduction of the filter adjustment coefficient provides the ability of adaptive adjustment, and can dynamically adjust the denoising intensity according to the grayscale value of each pixel point and the changes of the pixels in its neighborhood. This adaptive filtering method allows the image processing to be adjusted according to the noise characteristics of different regions, avoiding processing distortion or detail that may be caused by unified filtering parameters. Loss, thereby enhancing the flexibility of image processing. By setting the grayscale enhancement value of the pixel points in the neighborhood for filtering processing, not only the noise in the image is removed, but also the edges and details are better preserved, which is crucial for subsequent steps such as area division and information extraction, because edges and details are the identifiers of key electrical parameter areas in the display panel data of the power equipment detection instrument, ensuring the accuracy and precision of subsequent analysis. Through the filtering adjustment and comprehensive analysis performed on each grayscale pixel point, the filtering processing can not only remove noise, but also improve image quality and processing efficiency. Compared with the traditional simple filtering method, this processing method based on neighborhood analysis and adaptive adjustment can reduce unnecessary calculation while retaining key features, making image processing more efficient and meeting the needs of real-time detection and analysis. The introduction of the filtering adjustment factor can adjust the sensitivity of noise removal to ensure that the filtering processing can be flexibly adapted under different environmental or equipment conditions to avoid too weak filtering effect when the noise is strong, or excessive filtering when the noise is less. This control mechanism can improve the applicability of filtering and make it more stable and reliable.

[0031] Specifically, the specific steps for obtaining several display areas of the display panel image of the electric power equipment detection instrument at each time point are as follows: gradient analysis is performed on the grayscale pixel filter values ​​of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point to obtain the horizontal gradient value and vertical gradient value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point; the horizontal gradient value and vertical gradient value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point are comprehensively analyzed to obtain the gradient intensity value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point; based on a preset gradient intensity threshold, the gradient intensity value of each grayscale pixel in the display panel image of the electric power equipment detection instrument at each time point is compared and analyzed respectively, and the grayscale pixel points in the display panel image of the electric power equipment detection instrument at each time point whose gradient intensity value is higher than the preset gradient intensity threshold are marked as grayscale edge pixels, and connected domain marking processing is performed to obtain several display areas of the display panel image of the electric power equipment detection instrument at each time point.

[0032] The specific formula for calculating the horizontal gradient value, vertical gradient value, and gradient intensity value of each grayscale pixel in the display panel image of the power equipment detection instrument at each time point is as follows: ;in, For the The display panel image of the power equipment detection instrument at a time point The horizontal gradient value of a grayscale pixel, is the horizontal convolution kernel stored in the database, For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel filter value of grayscale pixels, For the The display panel image of the power equipment detection instrument at a time point The vertical gradient value of a grayscale pixel, is the vertical convolution kernel stored in the database, For the The display panel image of the power equipment detection instrument at a time point The gradient intensity value of the grayscale pixel, =1, 2, 3, …, , is the number of time points, =1, 2, 3, …, , It is the number of grayscale pixels in the display panel image.

[0033] It needs to be explained that the horizontal convolution kernel stored in the database is .

[0034] The vertical convolution kernel stored in the database is .

[0035] In this example, 8-connectivity (each pixel is connected to 8 neighboring pixels above, below, left, right, and diagonally opposite to it) is selected for connected domain labeling. The specific process is as follows:

[0036] Create a marker image of the same size as the display panel image and initialize it to zero.

[0037] In this labeled image, each connected region is gradually assigned a unique label.

[0038] Scan each grayscale pixel separately. If the grayscale pixel is a grayscale edge pixel and has not been marked, start marking the connected domain from the grayscale pixel. This can be processed by depth-first search (DFS) or breadth-first search (BFS):

[0039] Depth-first search: Starting from the current grayscale pixel, recursively check all its neighboring grayscale pixels and mark all grayscale pixels connected to the grayscale pixel.

[0040] Breadth-first search: Use a queue, starting from the current grayscale pixel, scan its neighboring grayscale pixels layer by layer, and mark all connected grayscale pixels.

[0041] Labeling process: If the current grayscale pixel is an edge pixel and has not been labeled, then create a new region label and attach the same label to this grayscale pixel and all its connected neighboring grayscale pixels. This process will be recursively performed until all connected pixels are labeled.

[0042] In this embodiment, by performing gradient analysis on each grayscale pixel, especially the combined analysis of horizontal and vertical gradients, the edge information in the image can be effectively identified. In the display panel image of the power equipment detection instrument, the electrical measurement area usually has obvious gradient changes with the background area. The processing method based on the gradient intensity value can accurately identify these edge pixels, thereby dividing different display areas. This method is more accurate than the traditional edge detection method and helps to improve the accuracy of subsequent information extraction. The connected domain marking processing method is used to mark the edge pixels, which can effectively classify adjacent edge pixels in the image into the same display area. This display area division method based on gradient intensity can automatically identify and separate different electrical measurement areas, avoiding errors caused by manually setting thresholds or manual intervention, thereby improving the automation and processing efficiency of the system. By setting the gradient intensity threshold, the sensitivity to edge pixels can be flexibly adjusted to adapt to the characteristics of the display panel images of different power equipment detection instruments. For example, for areas with large grayscale differences in the display panel, gradient analysis can effectively identify their boundaries;For areas with small grayscale differences, flexible control can be achieved by adjusting the threshold, thereby avoiding over-detection or missing edges, and improving the applicability and stability of the method. In the process of connected domain labeling, the 8-connectivity rule is used to connect neighborhood pixels to ensure that all adjacent pixels in the image are correctly marked as the same area. This method avoids the omission of low-contrast areas or parts with unclear edges, thereby improving the accuracy of regional division. Whether it is processed by depth-first search (DFS) or breadth-first search (BFS), it can ensure that all interconnected edge pixels are uniformly labeled, thereby better dividing the display area in the image. By combining gradient analysis with connected domain labeling, the processing process is not only accurate but also highly robust. When faced with complex or noisy images, it can still effectively identify the electrical measurement area and maintain a good processing effect. By adopting depth-first search or breadth-first search, it can flexibly handle The method can process regions of different sizes and complexities, further improving the efficiency of image processing. The gradient intensity threshold can be dynamically adjusted according to the characteristics of the display panel image of the power equipment detection instrument at different time points, making the method have strong adaptability. In addition, the regional division method can be extended to other types of equipment detection and image processing tasks, and has good versatility. Through precise display area division, the subsequent information extraction module can extract more accurate measurement data (such as voltage, current, power, etc.) from each area, thereby reducing the interference of erroneous data and improving the accuracy and reliability of detection evaluation and abnormal detection of power equipment detection instruments. Through automated gradient analysis and regional division, manual intervention is reduced and the automation of the system is improved. Users do not need to manually set the area. The system can automatically identify and divide the display area according to the gradient characteristics of the image, greatly improving the intelligent processing capability, and is suitable for batch and real-time power equipment detection scenarios. ;

[0043] Specifically, the specific steps for obtaining the measurement data of the electric power equipment detection instrument at each time point are as follows: perform connected domain labeling processing on each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, obtain a number of connected areas of each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, and perform character segmentation processing to obtain a number of characters of the display panel image of the electric power equipment detection instrument at each time point; input the several characters of the display panel image of the electric power equipment detection instrument at each time point into the pre-trained neural network model for classification analysis to obtain the category information of each character of the display panel image of the electric power equipment detection instrument at each time point; determine and summarize the category information of each character of the display panel image of the electric power equipment detection instrument at each time point, and obtain the measurement data of the electric power equipment detection instrument at each time point.

[0044] The specific process of performing connected domain marking processing on each display area of ​​the display panel image of the power equipment detection instrument at each time point is as follows:

[0045] According to a preset grayscale pixel threshold, the grayscale pixel filter value of each grayscale pixel point in each display area of ​​the display panel image of the electric power equipment detection instrument at each time point is binarized to obtain a binarized image of each display area of ​​the display panel image of the electric power equipment detection instrument at each time point.

[0046] A connected domain analysis is performed on each display area binary image of the display panel image of the power equipment detection instrument at each time point, and each connected area in the image is marked (using the 8-connected strategy, that is, each pixel is connected to its 8 surrounding neighboring pixels). The depth-first search (DFS) or breadth-first search (BFS) method can be used to mark the connected areas. Whenever an unmarked edge pixel is found, a new round of marking starts, and the pixel and all its connected pixels are marked as the same area.

[0047] After the connected domains are labeled, a number of connected regions in each display region of the display panel image of the electric power equipment detection instrument at each time point are obtained, and each connected region will be assigned a unique label.

[0048] The specific process of performing character segmentation processing on a plurality of connected areas of each display area of ​​the display panel image of the power equipment detection instrument at each time point is as follows:

[0049] For each connected region, its shape and structure are further analyzed to determine whether it contains multiple characters or irregular character boundaries.

[0050] Whether the connected area contains multiple characters is determined based on geometric features such as the aspect ratio of the connected area, the pixel density within the area, and the spacing between characters. Gabor wavelet transform is used to extract texture features (to help determine the segmentation boundaries between characters. Gabor wavelet transform can conduct in-depth analysis of the texture of characters to identify the boundaries of characters). Based on the extracted texture features, the characters are segmented and each character is extracted separately. All the characters in each connected area are obtained and marked and segmented separately.

[0051] In this embodiment, the pre-trained neural network is a convolutional neural network, which is good at extracting local features (such as edges, shapes, etc.) and is suitable for character recognition tasks.

[0052] The convolutional neural network includes an input layer for inputting a segmented single character image, a convolution layer for extracting local features of the character using multiple convolution kernels, a pooling layer for reducing dimensionality through a pooling layer, reducing computational complexity, and enhancing the robustness of the features, a fully connected layer for mapping the extracted features to different character categories through a fully connected layer, and an output layer for outputting the category label of each character, such as the number "5" or the unit "V".

[0053] The pre-training process of the convolutional neural network is as follows:

[0054] Collect image samples containing numbers (such as 0-9) and units (such as V, A, Hz, etc.). Each character should be input into the neural network as an image instance.

[0055] Randomly rotate the character image to simulate the possible angle changes of the characters on the display panel.

[0056] The character image is scaled and cropped to simulate the deformation of the characters on the actual measuring instrument.

[0057] Simulate the noise in the sensor or image acquisition process (such as Gaussian noise, salt and pepper noise, etc.).

[0058] By adjusting image brightness, contrast, etc., the image differences in different display environments can be simulated.

[0059] Each character image should have a corresponding annotation label, indicating the character (number or unit) corresponding to the image. These labels will be used as the target output of training.

[0060] The convolutional neural network is trained based on the cross entropy loss function and the Adam optimization algorithm, including:

[0061] Choose an appropriate batch size (such as 32 or 64) to balance training efficiency and memory usage.

[0062] Select 0.001 as the initial learning rate. If the model training is unstable, you can adjust it appropriately.

[0063] Select 10-50 epochs for training. The specific number of epochs can be adjusted through cross-validation.

[0064] The validation set is used during the training process to evaluate the generalization ability of the model and avoid overfitting.

[0065] After training is completed, a test set is used to evaluate the final performance of the model. The test set should be different from the training set and validation set to ensure that the model has good generalization ability.

[0066] Use evaluation indicators such as accuracy, recall, and F1 value to estimate model performance.

[0067] A specific example of determining and summarizing the category information of each character in the display panel image of the power equipment testing instrument at each time point is as follows: based on the previous character classification results, the characters in the same display area are summarized. For example, a display area contains the characters "1", "2", "3" and the unit "V", then the information is combined to obtain the measurement result "123V". After summarizing the information of all display areas, the complete measurement data for each time point is obtained.

[0068] In this implementation, through the connected domain labeling process, each character area in the display panel image can be accurately identified and separated. The connected domain analysis can not only identify independent characters, but also effectively handle the relationship between different display areas to avoid character loss or misrecognition. For each display area, character segmentation is further performed to ensure that each character can be extracted separately. By combining Gabor wavelet transform to extract texture features, the separation boundary between characters can be accurately identified, especially for the case where the character spacing is small and the shape is irregular, the integrity and independence of the characters can be ensured. The pre-trained convolutional neural network (CNN) is used to classify and analyze each extracted character, and the character type (such as numbers, units, etc.) can be accurately identified. , CNN is particularly good at extracting local features (such as edges, textures, etc.) when processing image classification tasks. It has extremely high accuracy and robustness in recognizing various types of characters (including numbers, symbols, units, etc.) in power equipment detection instruments. Through the deep learning structure of convolutional layers and pooling layers, CNN can adapt to character images in different angles, different forms and different noise environments, improving the accuracy of recognition. In the pre-training process, data enhancement operations such as random rotation, scaling, cropping and noise simulation are performed on character images to ensure that the neural network model can adapt to character deformation and environmental interference that may occur in different display environments. This process not only improves the generalization ability of the model, but also enhances its robustness in different actual scenarios, making it more accurate. It can accurately identify the characters displayed in the power equipment detection instrument, regardless of the display angle, character deformation, or image noise and other factors that affect the recognition. This method can summarize the character information extracted from the display panel image at each time point to ensure the integrity and accuracy of the measurement data. For example, by combining the numbers and unit characters (such as "123V") in the same area, it is ensured that the final measurement data can accurately reflect the real-time measurement results of the instrument. After the data of all display areas are summarized, complete measurement data is obtained, which improves the accuracy of data extraction and avoids human errors. The use of a neural network model can dynamically adapt to various changes in the display panel image of the power equipment, including the rotation, tilt, and light of the characters. In traditional image processing methods, changes in the shape of characters and display environment often lead to recognition errors, while this method can effectively cope with these challenges through the training of deep learning models, and ensure high accuracy in complex and uncertain environments. The entire process from image preprocessing, character segmentation to neural network classification is highly automated without human intervention or manual adjustment of parameters. The system can automatically process image data at different time points, extract and classify each character, and summarize the measurement data, thereby greatly improving work efficiency. It is especially suitable for batch and real-time power equipment detection tasks. Although the present invention focuses on character recognition and data extraction of power equipment detection instruments, the method has strong versatility.It can be applied to image character recognition tasks in other fields. For example, similar technologies can be applied to electronic equipment, instrument panels, meters and other fields to provide efficient and accurate automatic data collection and analysis capabilities. Through automated recognition and data extraction, the workload of manual detection, manual reading and data input is greatly reduced, while also reducing the errors of manual operations, saving a lot of manpower and time costs. This is especially important for large-scale and long-term power equipment detection work. By accurately extracting measurement data and combining it with real-time image analysis, it can provide accurate basic data for subsequent abnormal detection intelligent decision-making, which will enhance the overall intelligence level of the system and make the management and maintenance of power equipment more efficient and scientific.

[0069] Specifically, the specific steps for obtaining the detection evaluation index of the power equipment detection instrument are as follows: obtain the electrical measurement reference data of the power equipment detection instrument (the reference data is the electrical measurement reference data of the power equipment detection instrument for the power equipment, such as the measurement reference value of the transformer), the electrical measurement reference data includes voltage measurement reference value, current measurement reference value, frequency measurement reference value, power factor measurement reference value, power measurement reference value; comprehensively analyze the voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value of the power equipment detection instrument at each time point, respectively, in combination with the voltage measurement reference value, current measurement reference value, frequency measurement reference value, power factor measurement reference value, and power measurement reference value of the power equipment detection instrument, to obtain the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of several groups of power equipment detection instruments; comprehensively analyze the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of several groups of power equipment detection instruments to obtain the detection evaluation index of the power equipment detection instrument.

[0070] Among them, the specific formula for calculating the detection evaluation index of power equipment detection instruments is as follows: ;in, It is the test evaluation index of power equipment testing instruments. For the The voltage detection evaluation index of the power equipment detection instrument is is the voltage detection evaluation coefficient, For the The current detection evaluation index of the group of power equipment detection instruments, is the current detection evaluation coefficient, For the Frequency detection evaluation index of power equipment detection instruments, is the frequency detection evaluation coefficient, For the The power factor detection evaluation index of the group of power equipment detection instruments, is the power factor detection evaluation coefficient, For the The power detection evaluation index of the group of power equipment detection instruments, is the power detection evaluation coefficient, , is a natural constant, and in this embodiment, its value is 2.718. =1, 2, 3, ..., , is the total number of groups.

[0071] It should be explained that the voltage measurement reference value of the power equipment detection instrument can be obtained by obtaining the historical voltage measurement values ​​of the power equipment by the power equipment detection instrument at several historical time points and performing mean analysis.

[0072] The current measurement reference value of the power equipment detection instrument can be obtained by obtaining historical current measurement values ​​of the power equipment by the power equipment detection instrument at several historical time points and performing mean analysis.

[0073] The frequency measurement reference value of the power equipment detection instrument can be obtained by obtaining the historical frequency measurement values ​​of the power equipment by the power equipment detection instrument at several historical time points and performing mean analysis.

[0074] The power factor measurement reference value of the power equipment detection instrument can be obtained by obtaining the historical power factor measurement values ​​of the power equipment by the power equipment detection instrument at several historical time points and performing mean analysis.

[0075] The power measurement reference value of the electric power equipment detection instrument can be obtained by obtaining the historical power measurement values ​​of the electric power equipment by the electric power equipment detection instrument at several historical time points and performing mean analysis.

[0076] , , , , The specific calculation process is: standardize the voltage measurement reference value, current measurement reference value, frequency measurement reference value, power factor measurement reference value, and power measurement reference value of the electric power equipment testing instrument (i.e., remove the unit), and perform sum analysis after standardization to obtain the measurement reference and value of the electric power equipment testing instrument; then perform proportion analysis on the voltage measurement reference value, current measurement reference value, frequency measurement reference value, power factor measurement reference value, and power measurement reference value of the standardized electric power equipment testing instrument and the measurement reference and value of the electric power equipment testing instrument, and the proportion analysis result is the corresponding evaluation coefficient.

[0077] In the formula The function is the hyperbolic tangent function, which is a nonlinear function. Its mathematical expression is: , where e is a natural constant, which is 2.718 in this embodiment, and The result is always between (-1, 1) and is symmetric about the origin. As the independent variable x increases, the value of tanh(x) gradually increases.

[0078] In this implementation scheme, by comparing and analyzing the various measurement values ​​of the power equipment detection instruments with historical reference data, the working status of the power equipment can be accurately evaluated. The historical data provides a reference for the performance of the equipment under normal operating conditions, so that the current test results can be more objectively compared with the preset standards, thereby avoiding the deviation that may be caused by a single measurement value. By weighted analysis of multiple parameters such as voltage, current, frequency, power factor, power, etc., and comprehensive consideration of the influence of various factors, a comprehensive evaluation index can be obtained. This can not only reflect the abnormality of a single parameter, but also make comprehensive judgments in multiple dimensions, thereby enhancing the comprehensiveness and accuracy of the test results. Standardizing (de-unitizing) historical data helps to eliminate the differences between different measurement units, so that various parameters can be compared and analyzed under the same benchmark. The standardized data facilitates the establishment of a unified Evaluation coefficient, thereby improving the comparability and consistency between various detection indicators. By analyzing the proportion of measurement reference values ​​and historical data of equipment, the evaluation coefficient of each parameter is obtained. These coefficients can be dynamically adjusted according to changes in the historical status of the equipment. This flexibility ensures that the system can adapt to the operation of power equipment in different time periods and under different load conditions, and improves the adaptability of the evaluation. This method can monitor the operating status of power equipment in real time, detect anomalies and issue warnings in time. Through the change trend of the comprehensive evaluation index, it can provide strong data support for equipment maintenance, identify potential faults in advance, reduce equipment failure rate, and improve system reliability and safety. By aggregating multiple independent detection data into a comprehensive evaluation index, it not only simplifies the data processing process, but also provides concise and clear evaluation results for power equipment maintenance personnel, avoiding tedious single indicator analysis.

[0079] Specifically, the specific steps for obtaining the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of each group of power equipment detection instruments are as follows: adjacent time points are regarded as a group of time series data, and the voltage measurement values, current measurement values, frequency measurement values, power factor measurement values, and power measurement values ​​at adjacent time points in each group of time series data are analyzed for change rates to obtain the voltage change index, current change index, frequency change index, power factor change index, and power change index of several groups of power equipment detection instruments; the voltage measurement values, current measurement values, frequency measurement values, power factor measurement values, and power measurement values ​​at adjacent time points in each group of time series data are combined with the voltage measurement reference value, The current measurement reference value, frequency measurement reference value, power factor measurement reference value and power measurement reference value are comprehensively analyzed to obtain the voltage deviation index, current deviation index, frequency deviation index, power factor deviation index and power deviation index of several groups of power equipment detection instruments; the voltage change index, current change index, frequency change index, power factor change index and power change index of several groups of power equipment detection instruments are respectively combined with the voltage deviation index, current deviation index, frequency deviation index, power factor deviation index and power deviation index of several groups of power equipment detection instruments for comprehensive analysis to obtain the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index and power detection evaluation index of several groups of power equipment detection instruments.

[0080] The specific formulas for calculating the voltage variation index, voltage deviation index, and voltage detection evaluation index of each group of power equipment detection instruments are as follows: ;in, For the The voltage variation index of the group power equipment detection instrument, For the The voltage measurement value of the first adjacent time point in the group time series data, For the The voltage measurement value at the second adjacent time point in the group time series data, For the The voltage deviation index of the power equipment detection instrument is It is the voltage measurement reference value of the power equipment detection instrument. is the first deviation coefficient stored in the database, is the second deviation coefficient stored in the database, For the The voltage detection evaluation index of the power equipment detection instrument is is a natural constant, and in this embodiment, its value is 2.718. is the voltage variation coefficient stored in the database, is the voltage deviation coefficient stored in the database, , =1, 2, 3, ..., , is the total number of groups.

[0081] It needs to be explained that the specific logic for calculating the current change index, current deviation index, current detection evaluation index, frequency change index, frequency deviation index, frequency detection evaluation index, power factor change index, power factor deviation index, power factor detection evaluation index, power change index, power deviation index and power detection evaluation index of each group of power equipment detection instruments is consistent with the specific logic for calculating the voltage change index, voltage deviation index and voltage detection evaluation index of each group of power equipment detection instruments.

[0082] In this implementation scheme, by analyzing the rate of change of parameters such as voltage, current, frequency, power factor and power at adjacent time points, the dynamic changes in the operating status of the equipment can be captured and the fluctuation trend of the equipment in a short period of time can be identified. This dynamic monitoring can more accurately reflect the actual working status of the power equipment and provide more timely information than static data. The measured data and historical reference values ​​are comprehensively analyzed. By calculating the deviation index, the difference between the equipment and the preset standard can be revealed, which helps to timely detect deviations during the operation of the power equipment and discover potential problems at an early stage, thereby reducing the risk of equipment failure and providing accurate data support for equipment maintenance. This method not only considers the change of a single parameter, but also combines the change rate and deviation analysis, and obtains the evaluation index through comprehensive analysis of multiple dimensions. This multi-dimensional evaluation can comprehensively reflect the operation of the power equipment and avoid the misjudgment that may be caused by judging the performance of the equipment by only a single indicator. In the comprehensive analysis, the change index and the deviation index are combined, and the deviation coefficient and the change coefficient stored in the database are used to calculate the deviation index. The weighted data are weighted. This weighted method can dynamically adjust the evaluation weight according to the different states of the equipment, so that different situations and equipment can be properly weighted during the evaluation, ensuring the rationality and accuracy of the evaluation results. Through the comprehensive analysis of the change index and the deviation index, the abnormal changes of the power equipment can be reflected in time, and an early warning function can be provided. When the change of a certain parameter exceeds the set threshold, the system can quickly identify and issue an alarm, thereby helping maintenance personnel to intervene in time to debug or repair the equipment and reduce the probability of equipment failure. Through this meticulous detection and evaluation method, it is possible to effectively discover possible deviations or anomalies in the operation of the equipment, avoid the chain reaction of equipment failure to the system, and improve the stability and safety of the power system. By subdividing multiple parameters such as voltage, current, frequency, power factor, and analyzing the changes and deviations of each parameter, the specific parameters or time points where the equipment problems are located can be accurately located, which is of great significance for fault location and the formulation of maintenance plans, and helps to improve the efficiency of fault repair.

[0083] See also Figure 2The embodiment of the present invention provides a technical solution: an automatic detection system for electric power equipment detection instruments based on machine vision, comprising: an image acquisition processing module, a region division module, an information extraction module, an evaluation and analysis module, and a detection and judgment module; the image acquisition processing module is used to continuously acquire the display panel image data of the electric power equipment detection instrument at several time points and perform preprocessing, the display panel image data includes the pixel value of each pixel in the display panel image; the region division module is used to perform region division on the display panel image data of the electric power equipment detection instrument at each time point after preprocessing, and obtain several display areas of the display panel image of the electric power equipment detection instrument at each time point; the information extraction module is used to perform preprocessing on the display panel image data of the electric power equipment detection instrument at each time point Information is extracted from each display area of ​​the display panel image respectively, and summarized and processed to obtain the measurement data of the power equipment detection instrument at each time point, and the measurement data includes voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value; the evaluation and analysis module is used to perform comprehensive analysis on the measurement data of the power equipment detection instrument at each time point, obtain the detection evaluation index of the power equipment detection instrument, and perform storage and backup processing; the detection judgment module is used to judge whether the detection evaluation index of the power equipment detection instrument is within the preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, a detection abnormality alarm is issued to relevant staff.

[0084] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An automatic detection method for electric power equipment detection instruments based on machine vision, characterized in that: The following steps are involved: Continuously acquiring display panel image data of the power equipment detection instrument at several time points and performing preprocessing, wherein the display panel image data includes a pixel value of each pixel point in the display panel image; Performing area division on the preprocessed display panel image data of the electric power equipment detection instrument at each time point, respectively, to obtain a plurality of display areas of the display panel image of the electric power equipment detection instrument at each time point; Extract information from each display area of ​​the display panel image of the power equipment detection instrument at each time point, perform summary processing, obtain measurement data of the power equipment detection instrument at each time point, and perform storage backup processing, wherein the measurement data includes voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value; Comprehensively analyze the measurement data of the power equipment detection instrument at each time point, obtain the detection evaluation index of the power equipment detection instrument, and perform storage and backup processing; Determine whether the detection evaluation index of the power equipment detection instrument is within the preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, issue a detection abnormality alarm to relevant staff; The specific steps of preprocessing the display panel image data of the power equipment detection instrument at each time point are as follows: Performing grayscale processing on the display panel image data of the electric power equipment detection instrument at each time point, respectively, to obtain the display panel grayscale image data of the electric power equipment detection instrument at each time point, wherein the display panel grayscale image data includes the grayscale pixel value of each grayscale pixel point in the display panel grayscale image; Analyze the gradient value of each grayscale pixel in the grayscale image of the display panel of the power equipment detection instrument at each time point based on a preset gradient operator; The grayscale image data of the display panel of the electric power equipment detection instrument at each time point are divided and processed respectively to obtain a plurality of image processing area data of the electric power equipment detection instrument at each time point, each image processing area data including a plurality of grayscale pixel points; Compare and analyze the data of each image processing area of ​​the power equipment detection instrument at each time point, obtain the maximum grayscale pixel value and the minimum grayscale pixel value of each image processing area of ​​the power equipment detection instrument at each time point, and analyze the regional contrast value of each image processing area of ​​the power equipment detection instrument at each time point; Comprehensively analyzing the grayscale pixel value of each grayscale pixel point in each image processing area data of the electric power equipment detection instrument at each time point, and obtaining the regional texture complexity of each image processing area of ​​the electric power equipment detection instrument at each time point; Comprehensively analyze the regional contrast value and regional texture complexity of each image processing area of ​​the power equipment detection instrument at each time point to obtain the pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point; The pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point is combined with the grayscale pixel value and gradient value of each grayscale pixel point in the data of each image processing area of ​​the power equipment detection instrument at each time point to perform a comprehensive analysis, so as to obtain the grayscale pixel enhancement value of each grayscale pixel point in the data of each image processing area of ​​the power equipment detection instrument at each time point, and the grayscale pixel enhancement value is obtained by summarizing the data to obtain the display panel image enhancement processing data of the power equipment detection instrument at each time point, wherein the display panel image enhancement processing data includes the grayscale pixel enhancement value of each grayscale pixel point in the display panel image; The display panel image enhancement processing data of the power equipment detection instrument at each time point is filtered respectively to obtain the display panel image filtering processing data of the power equipment detection instrument at each time point, and the display panel image filtering processing data includes the grayscale pixel filtering value of each grayscale pixel point in the display panel image.

2. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 1 is characterized in that: The specific formulas for calculating the regional contrast value, regional texture complexity, pixel adjustment coefficient of each image processing area of ​​the power equipment detection instrument at each time point, and the grayscale pixel enhancement value of each grayscale pixel point in each image processing area data of the power equipment detection instrument at each time point are as follows: ； in, For the The first time point of the power equipment testing instrument The regional contrast value of the image processing area, For the The first time point of the power equipment testing instrument The maximum grayscale pixel value of the image processing area, For the The first time point of the power equipment testing instrument The minimum grayscale pixel value of the image processing area, For the The first time point of the power equipment testing instrument The regional texture complexity of the image processing area, For the The first time point of the power equipment testing instrument The image processing area data The grayscale pixel value of the grayscale pixel point, For the The first time point of the power equipment testing instrument The pixel adjustment coefficient of the image processing area, For the The first time point of the power equipment testing instrument The image processing area data The grayscale pixel enhancement value of grayscale pixels, For the The first time point of the power equipment testing instrument The image processing area data The gradient value of a grayscale pixel, =1, 2, 3, ..., , is the number of time points, =1, 2, 3, ..., , is the number of image processing regions, =1, 2, 3, ..., , It is the number of grayscale pixels in the image processing area data.

3. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 2 is characterized in that: The specific steps for obtaining the display panel image filtering processing data of the power equipment detection instrument at each time point are as follows: For each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point, a number of adjacent grayscale pixel points in a set neighborhood are identified respectively; The grayscale pixel enhancement value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point is combined with the grayscale pixel enhancement value of each adjacent grayscale pixel point in the set neighborhood to obtain the filter adjustment coefficient of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point; Comprehensively analyzing the grayscale pixel enhancement value and the filter adjustment coefficient of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point, respectively, to obtain the grayscale pixel filter value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point; The specific formula for calculating the filter adjustment coefficient and the grayscale pixel filter value of each grayscale pixel in the display panel image of the power equipment detection instrument at each time point is as follows: ； in, For the The display panel image of the power equipment detection instrument at a time point The filter adjustment coefficient of grayscale pixels, For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel in the set neighborhood The grayscale pixel enhancement value of adjacent grayscale pixels, is the filter adjustment factor stored in the database, For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel filter value of grayscale pixels, for The display panel image of the power equipment detection instrument at a time point The grayscale pixel enhancement value of grayscale pixels, =1, 2, 3, ..., , is the number of time points, =1, 2, 3, ..., , is the number of grayscale pixels in the display panel image, =1, 2, 3, ..., , The number of adjacent grayscale pixels within the set neighborhood of the grayscale pixel.

4. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 3 is characterized in that: The specific steps of obtaining several display areas of the display panel image of the power equipment detection instrument at each time point are as follows: Performing gradient analysis on the grayscale pixel filter value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point, respectively, to obtain the horizontal gradient value and the vertical gradient value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point; Comprehensively analyzing the horizontal gradient value and the vertical gradient value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point, and obtaining the gradient intensity value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point; Based on a preset gradient intensity threshold, the gradient intensity value of each grayscale pixel point in the display panel image of the power equipment detection instrument at each time point is compared and analyzed respectively, and the grayscale pixel points whose gradient intensity value in the display panel image of the power equipment detection instrument at each time point is higher than the preset gradient intensity threshold are marked as grayscale edge pixels, and connected domain marking processing is performed to obtain a plurality of display areas of the display panel image of the power equipment detection instrument at each time point; The specific formula for calculating the horizontal gradient value, vertical gradient value, and gradient intensity value of each grayscale pixel in the display panel image of the power equipment detection instrument at each time point is as follows: ； in, For the The display panel image of the power equipment detection instrument at a time point The horizontal gradient value of a grayscale pixel, is the horizontal convolution kernel stored in the database, For the The display panel image of the power equipment detection instrument at a time point The grayscale pixel filter value of grayscale pixels, For the The display panel image of the power equipment detection instrument at a time point The vertical gradient value of a grayscale pixel, is the vertical convolution kernel stored in the database, For the The display panel image of the power equipment detection instrument at a time point The gradient intensity value of the grayscale pixel, =1, 2, 3, ..., , is the number of time points, =1, 2, 3, ..., , It is the number of grayscale pixels in the display panel image.

5. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 1 is characterized in that: The specific steps to obtain the measurement data of the power equipment detection instrument at each time point are as follows: Performing connected domain labeling processing on each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, respectively, to obtain a plurality of connected areas of each display area of ​​the display panel image of the electric power equipment detection instrument at each time point, and performing character segmentation processing, to obtain a plurality of characters of the display panel image of the electric power equipment detection instrument at each time point; Inputting several characters of the display panel image of the power equipment detection instrument at each time point into the pre-trained neural network model for classification analysis, and obtaining the category information of each character of the display panel image of the power equipment detection instrument at each time point; The category information of each character of the display panel image of the electric power equipment detection instrument at each time point is determined and summarized to obtain the measurement data of the electric power equipment detection instrument at each time point.

6. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 1 is characterized in that: The specific steps to obtain the detection evaluation index of power equipment detection instruments are as follows: Acquire electrical measurement reference data of an electric power equipment detection instrument, wherein the electrical measurement reference data includes a voltage measurement reference value, a current measurement reference value, a frequency measurement reference value, a power factor measurement reference value, and a power measurement reference value; The voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value of the power equipment detection instrument at each time point are respectively combined with the voltage measurement reference value, current measurement reference value, frequency measurement reference value, power factor measurement reference value, and power measurement reference value of the power equipment detection instrument to perform a comprehensive analysis to obtain voltage detection evaluation indexes, current detection evaluation indexes, frequency detection evaluation indexes, power factor detection evaluation indexes, and power detection evaluation indexes of several groups of power equipment detection instruments; Comprehensively analyze the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of several groups of power equipment detection instruments to obtain the detection evaluation index of the power equipment detection instruments; Among them, the specific formula for calculating the detection evaluation index of power equipment detection instruments is as follows: ； in, It is the test evaluation index of power equipment testing instruments. For the The voltage detection evaluation index of the power equipment detection instrument is is the voltage detection evaluation coefficient, For the The current detection evaluation index of the group of power equipment detection instruments, is the current detection evaluation coefficient, For the Frequency detection evaluation index of power equipment detection instruments, is the frequency detection evaluation coefficient, For the The power factor detection evaluation index of the group of power equipment detection instruments, is the power factor detection evaluation coefficient, For the The power detection evaluation index of the group of power equipment detection instruments, is the power detection evaluation coefficient, , is a natural constant, =1, 2, 3, ..., , is the total number of groups.

7. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 6 is characterized in that: The specific steps for obtaining the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of each group of power equipment detection instruments are as follows: Adjacent time points are regarded as a group of time series data, and the voltage measurement values, current measurement values, frequency measurement values, power factor measurement values, and power measurement values ​​at adjacent time points in each group of time series data are analyzed for change rates, so as to obtain voltage change indexes, current change indexes, frequency change indexes, power factor change indexes, and power change indexes of several groups of power equipment detection instruments; The voltage measurement values, current measurement values, frequency measurement values, power factor measurement values, and power measurement values ​​at adjacent time points in each group of time series data are respectively combined with the voltage measurement reference values, current measurement reference values, frequency measurement reference values, power factor measurement reference values, and power measurement reference values ​​of the power equipment detection instrument to perform a comprehensive analysis, and obtain voltage deviation indexes, current deviation indexes, frequency deviation indexes, power factor deviation indexes, and power deviation indexes of several groups of power equipment detection instruments; The voltage change index, current change index, frequency change index, power factor change index, and power change index of several groups of power equipment detection instruments are combined with the voltage deviation index, current deviation index, frequency deviation index, power factor deviation index, and power deviation index of several groups of power equipment detection instruments for a comprehensive analysis to obtain the voltage detection evaluation index, current detection evaluation index, frequency detection evaluation index, power factor detection evaluation index, and power detection evaluation index of several groups of power equipment detection instruments.

8. The automatic detection method of electric power equipment detection instrument based on machine vision according to claim 7 is characterized in that: The specific formulas for calculating the voltage variation index, voltage deviation index, and voltage detection evaluation index of each group of power equipment detection instruments are as follows: ； in, For the The voltage variation index of the group power equipment detection instrument, For the The voltage measurement value of the first adjacent time point in the group time series data, For the The voltage measurement value at the second adjacent time point in the group time series data, For the The voltage deviation index of the power equipment detection instrument is It is the voltage measurement reference value of the power equipment detection instrument. is the first deviation coefficient stored in the database, is the second deviation coefficient stored in the database, For the The voltage detection evaluation index of the power equipment detection instrument is is a natural constant, is the voltage variation coefficient stored in the database, is the voltage deviation coefficient stored in the database, , =1, 2, 3, ..., , is the total number of groups.

9. An automatic detection system for electric power equipment detection instruments based on machine vision, using the automatic detection method for electric power equipment detection instruments based on machine vision according to any one of claims 1 to 8, characterized in that: include: Image acquisition and processing module, area division module, information extraction module, evaluation and analysis module, detection and judgment module; The image acquisition processing module is used to continuously acquire the display panel image data of the power equipment detection instrument at several time points and perform preprocessing, wherein the display panel image data includes the pixel value of each pixel point in the display panel image; The region division module is used to divide the display panel image data of the power equipment detection instrument at each time point after preprocessing into regions, so as to obtain a plurality of display regions of the display panel image of the power equipment detection instrument at each time point; The information extraction module is used to extract information from each display area of ​​the display panel image of the power equipment detection instrument at each time point, and perform summary processing to obtain measurement data of the power equipment detection instrument at each time point, wherein the measurement data includes voltage measurement value, current measurement value, frequency measurement value, power factor measurement value, and power measurement value; The evaluation and analysis module is used to comprehensively analyze the measurement data of the power equipment detection instrument at each time point to obtain the detection evaluation index of the power equipment detection instrument; The detection judgment module is used to determine whether the detection evaluation index of the power equipment detection instrument is within a preset detection evaluation threshold range, and when the detection evaluation index of the power equipment detection instrument is outside the preset detection evaluation threshold range, issue a detection abnormality alarm to relevant staff.

Images