Ultraviolet imaging live detection method and system for power equipment

Abstract

The present invention discloses a method and system for ultraviolet imaging live detection of electric equipment, which relates to the technical field of ultraviolet imaging live detection and aims to solve the problem of inaccurate imaging live detection data. The present invention can display the results of live detection from different angles through different data types, which helps detection personnel to understand the status of the equipment more comprehensively. During the report generation process, when the report is displayed on the display terminal, an abnormal warning is synchronously performed according to the warning information, which helps to achieve real-time monitoring and early warning, and prevent unexpected power outages and damage caused by equipment failure. Different intensities of early warning processing are performed according to different degrees of abnormality, which can ensure that maintenance personnel are aware of the severity of the discharge problem in a timely manner, take corresponding maintenance measures, avoid damage to power equipment and the occurrence of safety accidents, and use historical data to train the deep learning model to learn the complex patterns and laws of discharge characteristics, thereby improving the analysis accuracy of live identification data.

Description

Technical Field

[0001] The present invention relates to the technical field of ultraviolet imaging charging detection, and in particular to an ultraviolet imaging charging detection method and system for electric power equipment. Background Art

[0002] UV imaging live testing refers to the use of UV cameras to capture UV light signals emitted from the surface of electrical equipment and perform image analysis to identify and evaluate potential defects and faults.

[0003] Chinese patent publication number CN114034995A discloses an ultraviolet imager and its use method. The invention mainly collects existing ultraviolet typical defect maps and establishes a typical defect fault database of external insulation equipment by setting up a database management module. The original ultraviolet features are analyzed by the image intelligent processing module and matched and compared with the original features of the typical library, thereby realizing intelligent identification and detection. The detection monitoring module detects, records and analyzes the detection operations of discharge detection personnel according to the standardized process and standardized operation template of ultraviolet imaging live detection operation. The detection report generation module combines the detection operation requirements of the detection personnel by the detection monitoring module and the processing results of the ultraviolet image by the image intelligent processing module to regularly generate periodic reports of the monitoring equipment, which facilitates the standardization and standardization of inspection operations. Although the above patent solves the problems of using ultraviolet imagers, the following problems still exist in actual operation:

[0004] 1. The collected data is not effectively processed and image fused, resulting in the inability to extract key features more accurately.

[0005] 2. The processed image data is not effectively confirmed for data features, and the feature data is not subjected to more accurate data analysis and abnormality judgment, resulting in the inability to judge the detection results with the highest efficiency.

[0006] 3. The final test results were not generated and presented in an inspection report, which resulted in staff being unable to make faster abnormal decisions based on the test results. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and system for ultraviolet imaging live detection of power equipment. Different data types can be used to display the results of live detection from different angles, which helps detection personnel to understand the status of the equipment more comprehensively. During the report generation process, when the report is displayed on the display terminal, an abnormal warning is synchronously performed according to the warning information, which helps to achieve real-time monitoring and warning, and prevent unexpected power outages and damage caused by equipment failure. Different intensities of warning processing are performed according to different degrees of abnormality, which can ensure that maintenance personnel are aware of the severity of the discharge problem in a timely manner, take corresponding maintenance measures, avoid damage to power equipment and the occurrence of safety accidents, and solve the problems in the existing technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] The method for detecting live power equipment by ultraviolet imaging includes the following steps:

[0010] S1: Equipment detection and optical signal acquisition: Before optical signal acquisition, the acquisition equipment is first tested. After all the equipment passes the test, optical signal acquisition is performed. After the acquisition is completed, the signal data to be processed is obtained.

[0011] S2: Image generation and processing: Generate an image from the signal data to be processed, perform image preprocessing on the generated image, and obtain the target processed image after image preprocessing;

[0012] S3: Image fusion: performing image fusion on the target processed image, which includes an ultraviolet image and a visible light image, and obtaining a target fused image after image fusion;

[0013] S4: Fusion image recognition: The target fusion image is used to identify the discharge target, and the discharge target in the identified target fusion image is marked. After marking, the discharge target is integrated to obtain the charged identification data;

[0014] S5: Identification data analysis and early warning: Analyze the live identification data, screen abnormal data based on the analysis results, judge the abnormality of the screened abnormal data, and issue an early warning based on the judgment results.

[0015] Preferably, in S1, the acquisition equipment is tested, and optical signal acquisition is performed after all the equipment passes the test, including:

[0016] The acquisition equipment includes an ultraviolet imager, which collects ultraviolet light signals and visible light signals;

[0017] Before collecting UV and visible light signals, the UV imager should be tested first;

[0018] Among them, the equipment inspection of UV imagers includes appearance inspection, function inspection, performance inspection and safety inspection;

[0019] After all the equipment are tested and qualified, signal collection will be carried out; if the equipment fails to pass the test, repair or replacement will be carried out;

[0020] The ultraviolet imager converts the collected ultraviolet light signals and visible light signals into digital signals, and then performs digital signal processing after the conversion is completed;

[0021] Digital signal processing includes signal amplification and filtering;

[0022] After digital signal processing, the signal data to be processed is obtained.

[0023] Preferably, in step S2, the image is generated from the signal data to be processed, and the generated image is subjected to image preprocessing, including:

[0024] Reconstructing the signal data to be processed into digital signals using digital signal processing technology, and forming preliminary image data of the signal data to be processed after the digital signal reconstruction;

[0025] The preliminary image data of the signal data to be processed includes a preliminary ultraviolet light image and a preliminary visible light image;

[0026] performing image preprocessing on the preliminary ultraviolet light image and the preliminary visible light image;

[0027] Image preprocessing includes denoising, enhancing, rectifying, cropping and scaling the image;

[0028] Unifying the formats of the preliminary ultraviolet image and the preliminary visible light image after image preprocessing;

[0029] After the format is unified, the target processed image is obtained.

[0030] Preferably, performing image enhancement on the preliminary ultraviolet image and the preliminary visible light image comprises:

[0031] Extracting grayscale values ​​of pixels contained in the preliminary ultraviolet image as first grayscale value data;

[0032] Extracting the central grayscale value of the preliminary visible light image;

[0033] Obtaining a difference between the central grayscale values ​​of the preliminary visible light image and the preliminary ultraviolet light image as an observed difference;

[0034] Obtaining a preliminary ultraviolet image enhancement coefficient using the first grayscale value data and the central grayscale value of the preliminary visible light image in combination with an observed difference;

[0035] The preliminary ultraviolet image enhancement coefficient is obtained by the following formula:

[0036]

[0037] Among them, S z represents the preliminary ultraviolet image enhancement coefficient; n represents the number of gray values ​​contained in the first gray value data; H 01i represents the grayscale value corresponding to the i-th first grayscale value data; H 01i+1 Indicates the grayscale value corresponding to the i+1th first grayscale value data; H 01z Represents the central gray value of the preliminary UV image; H 02z Represents the central grayscale value of the preliminary visible light image; H 01b represents the standard deviation of the grayscale value corresponding to the first grayscale value data; H 02b represents the grayscale value standard deviation corresponding to the second grayscale value data;

[0038] The contrast of the preliminary ultraviolet light image and the preliminary visible light image is adjusted using the preliminary ultraviolet light image enhancement coefficient.

[0039] Preferably, using the preliminary ultraviolet image enhancement coefficient to adjust the contrast of the preliminary ultraviolet image and the preliminary visible light image includes:

[0040] Extracting grayscale values ​​of pixels contained in the preliminary visible light image as second grayscale value data;

[0041] extracting a central grayscale value and an observation difference value of the preliminary ultraviolet image;

[0042] Obtaining a preliminary visible light image enhancement coefficient by using the central grayscale value and the observed difference of the preliminary ultraviolet image in combination with each grayscale value data in the second grayscale value data;

[0043] The preliminary visible light image enhancement coefficient is obtained by the following formula:

[0044]

[0045] Among them, S k represents the preliminary visible light image enhancement coefficient; m represents the number of grayscale data contained in the second grayscale value data; H 02i represents the grayscale value corresponding to the i-th second grayscale value data; H 02i+1 Indicates the grayscale value corresponding to the i+1th second grayscale value data; H 01z Represents the central gray value of the preliminary UV image; H 02z represents the central grayscale value of the preliminary visible light image; f represents the adjustment factor, and the adjustment factor is obtained by the following formula:

[0046]

[0047] Where, f represents the adjustment factor; H 01b represents the standard deviation of the grayscale value corresponding to the first grayscale value data; H 02b represents the grayscale value standard deviation corresponding to the second grayscale value data;

[0048] The contrast between the preliminary ultraviolet light image and the preliminary visible light image is adjusted using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient.

[0049] Preferably, adjusting the contrast of the preliminary ultraviolet image and the preliminary visible light image using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet image enhancement coefficient includes:

[0050] extracting preliminary visible light image enhancement coefficients and preliminary ultraviolet light image enhancement coefficients;

[0051] Obtaining a comprehensive contrast adjustment coefficient using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0052] The comprehensive contrast adjustment coefficient is obtained by the following formula:

[0053]

[0054] Among them, S x Comprehensive contrast adjustment coefficient; S k represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 Respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0055] performing contrast adjustment on the preliminary visible light image using the comprehensive contrast adjustment coefficient to obtain a preliminary visible light image after contrast adjustment;

[0056] The contrast value corresponding to the preliminary visible light image after contrast adjustment is obtained by the following formula:

[0057]

[0058] Among them, D k represents the contrast value corresponding to the preliminary visible light image after contrast adjustment; D k0 Indicates the contrast value of the preliminary visible light image before adjustment; S x Comprehensive contrast adjustment coefficient; Sk represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 Respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0059] performing contrast adjustment on the preliminary ultraviolet image using the comprehensive contrast adjustment coefficient to obtain a contrast-adjusted preliminary ultraviolet image;

[0060] The contrast value corresponding to the preliminary ultraviolet image after contrast adjustment is obtained by the following formula:

[0061]

[0062] Among them, D z represents the contrast value corresponding to the preliminary ultraviolet image after contrast adjustment; D z0 Indicates the contrast value of the preliminary UV image before adjustment; S x Comprehensive contrast adjustment coefficient; S k represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 They respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient.

[0063] Preferably, performing image fusion on the target processed image in S3 includes:

[0064] The preliminary ultraviolet image and the preliminary visible light image in the target processing image are aligned using an image registration technology, and in the image alignment process, a feature point matching model is used to perform alignment processing, and the feature point matching model is retrieved from a database;

[0065] After the preliminary ultraviolet image and the preliminary visible light image are aligned, the preliminary ultraviolet image and the preliminary visible light image are merged using a pixel-level fusion method;

[0066] After the images are merged, image fusion processing is performed. The image fusion processing is to calculate the weighted average value of each corresponding pixel of the preliminary ultraviolet image and the preliminary visible light image. After the calculation is completed, the pixel value of the preliminary ultraviolet image and the preliminary visible light image after fusion is obtained;

[0067] performing contrast enhancement and brightness adjustment on the preliminary ultraviolet image and the preliminary visible light image after image fusion processing;

[0068] After contrast enhancement and brightness adjustment, the target fused image is obtained.

[0069] Preferably, the target fusion image is subjected to discharge target identification in S4, and the discharge targets in the identified target fusion image are marked, and integration is performed after marking is completed, including:

[0070] The target fusion image is separated from the foreground and background by edge detection, wherein the foreground in the target fusion image is the discharge target recognition area, and the background in the target fusion image is the remaining area except the foreground;

[0071] Extract features from the foreground of the target fusion image. The extracted features include shape features, texture features, spectral features, color features, brightness features, time data, intensity data, position data and spatial relationship features.

[0072] The extracted features are used to identify discharge targets using convolutional neural networks;

[0073] Marking and integrating the identified discharge targets;

[0074] Finally, the charged identification data is obtained.

[0075] Preferably, the charged identification data in S5 is analyzed, abnormal data is screened according to the analysis results, the abnormal degree of the screened abnormal data is judged, and an early warning is issued according to the judgment result, including:

[0076] The charged identification data is preprocessed, which includes removing noise, processing missing data and normalizing data;

[0077] Use feature selection algorithm to extract key indicator data of discharge characteristics from charged identification data;

[0078] Using historical data to train a deep learning model, where the historical data and the deep learning model are retrieved from a database;

[0079] After the deep learning model training is completed, the key indicator data of the discharge characteristics in the charged identification data are analyzed;

[0080] The analysis results are checked for abnormal data, wherein the analysis results are compared with the standard data for data thresholds, and based on the threshold comparison results, the analysis results that are not within the standard data threshold range are marked as abnormal data;

[0081] Determine the degree of abnormality based on the abnormal threshold range of abnormal data;

[0082] The degree of abnormality is divided into mild abnormality, moderate abnormality and severe abnormality;

[0083] Provide early warning processing of different intensities according to the degree of abnormality;

[0084] Abnormal data and non-abnormal data, as well as the abnormality degree and warning information of abnormal data are uniformly marked as standard live detection data.

[0085] The UV imaging live detection system for power equipment includes:

[0086] Live detection report generation unit, used for:

[0087] Generate a report using a visualization tool using standard live-test data, wherein the report layout is designed before generating the report;

[0088] The report layout includes text data, image data, table data and chart data. The chart data includes pie charts, curve charts, heat maps and bar charts.

[0089] After the report layout design is completed, use the visualization tool to generate a report for abnormal data and non-abnormal data in the standard live detection data;

[0090] The abnormal degree and warning information in the abnormal data are marked separately;

[0091] Finally, the generated report is transmitted to the display terminal for display. When the report is displayed, an abnormal warning is synchronously issued according to the warning information.

[0092] Compared with the prior art, the present invention has the following beneficial effects:

[0093] 1. The UV imaging method and system for detecting live power equipment provided by this invention can avoid prolonged downtime or damage caused by equipment failures through timely equipment inspection and repair and replacement. This helps reduce maintenance costs and extend the service life of power equipment. Image fusion technology enables rapid detection of power equipment status. A single capture and processing step yields a fused image containing both discharge and morphological information, eliminating the need for multiple captures and separate analyses, thus improving detection efficiency.

[0094] 2. The ultraviolet imaging live detection method and system for electric power equipment provided by the present invention performs early warning processing of different intensities according to different degrees of abnormality, which can ensure that maintenance personnel are aware of the severity of the discharge problem in a timely manner and take corresponding maintenance measures to avoid damage to electric equipment and safety accidents. The deep learning model is trained using historical data to learn the complex patterns and laws of discharge characteristics, improve the analysis accuracy of live identification data, and achieve a comprehensive description and accurate identification of discharge targets by fusing multiple feature information. The convolutional neural network has a strong generalization ability and can adapt to the discharge target identification needs in different scenarios.

[0095] 3. The UV imaging live detection method and system for electric power equipment provided by the present invention can display the results of live detection from different angles for different data types, which helps the inspectors to understand the status of the equipment more comprehensively. During the report generation process, abnormal data, its abnormality degree and warning information are separately marked, which helps the inspectors to quickly identify and locate the problem, so as to take timely and effective measures. When the report is displayed on the display terminal, abnormal warnings are synchronously issued according to the warning information, which helps to achieve real-time monitoring and warnings, and prevent accidental power outages and damage caused by equipment failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 Schematic diagram of the ultraviolet imaging charge detection steps of the present invention;

[0097] Figure 2 It is a schematic diagram of the ultraviolet imaging charge detection process of the present invention. DETAILED DESCRIPTION

[0098] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0099] In order to solve the problem in the prior art that after collecting UV and visible light signals, the collected data are not effectively processed and image fused, resulting in the inability to extract key features more accurately, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0100] The method for detecting live power equipment by ultraviolet imaging includes the following steps:

[0101] S1: Equipment detection and optical signal acquisition: Before optical signal acquisition, the acquisition equipment is first tested. After all the equipment passes the test, optical signal acquisition is performed. After the acquisition is completed, the signal data to be processed is obtained.

[0102] Among them, through automated and standardized equipment detection and signal acquisition processes, detection efficiency can be significantly improved;

[0103] S2: Image generation and processing: Generate an image from the signal data to be processed, perform image preprocessing on the generated image, and obtain the target processed image after image preprocessing;

[0104] Among them, the preliminary ultraviolet image and the preliminary visible light image are processed in a unified format to ensure that they have the same parameters such as resolution and color space;

[0105] S3: Image fusion: performing image fusion on the target processed image, which includes an ultraviolet image and a visible light image, and obtaining a target fused image after image fusion;

[0106] Among them, by training the machine learning model, the discharge phenomenon and morphological information in the fused image can be automatically identified and analyzed, further improving the automation and accuracy of detection;

[0107] S4: Fusion image recognition: The target fusion image is used to identify the discharge target, and the discharge target in the identified target fusion image is marked. After marking, the discharge target is integrated to obtain the charged identification data;

[0108] Among them, the flexibility and robustness of image processing are further enhanced by integrating information from different sources;

[0109] S5: Identification data analysis and early warning: Analyze the live identification data, screen abnormal data based on the analysis results, determine the abnormality level of the screened abnormal data, and issue an early warning based on the judgment results;

[0110] Among them, carrying out early warning processing of different intensities according to different degrees of abnormality can ensure that maintenance personnel are aware of the severity of the discharge problem in a timely manner and take corresponding maintenance measures.

[0111] The acquisition equipment in S1 is tested. After all the equipment passes the test, optical signal acquisition is carried out, including:

[0112] The acquisition equipment includes an ultraviolet imager, which collects ultraviolet light signals and visible light signals;

[0113] Before collecting UV and visible light signals, the UV imager should be tested first;

[0114] Among them, the equipment inspection of UV imagers includes appearance inspection, function inspection, performance inspection and safety inspection;

[0115] After all the equipment are tested and qualified, signal collection will be carried out; if the equipment fails to pass the test, repair or replacement will be carried out;

[0116] The ultraviolet imager converts the collected ultraviolet light signals and visible light signals into digital signals, and then performs digital signal processing after the conversion is completed;

[0117] Digital signal processing includes signal amplification and filtering;

[0118] After digital signal processing, the signal data to be processed is obtained.

[0119] Specifically, by subjecting the UV imager to rigorous equipment testing (including appearance inspection, functional testing, performance testing, and safety testing), its accuracy and reliability in collecting UV and visible light signals can be ensured. This helps improve the accuracy of subsequent digital signal processing and data analysis, thereby more accurately judging the operating status of power equipment. The equipment testing process not only includes basic functional inspections of the UV imager, but also covers performance and safety assessments. This helps to promptly discover and resolve potential equipment problems, avoid failures or errors during the acquisition process, and thus enhance the overall reliability of the equipment. The UV imager can simultaneously collect UV and visible light signals, which provides rich data support for a comprehensive assessment of the status of power equipment. In addition, through digital signal conversion and processing (including signal amplification and filtering), the signal quality can be further optimized, and the signal readability and analyzability can be improved. Through automated and standardized equipment testing and signal acquisition processes, the detection efficiency can be significantly improved. Once the equipment passes the inspection, signal acquisition and processing can be carried out immediately without additional waiting time. This helps shorten the inspection cycle and improve the timeliness of live inspections of power equipment. Through timely equipment inspection and repair and replacement, prolonged downtime or damage caused by equipment failure can be avoided. This helps reduce the maintenance cost of power equipment and extend the service life of the equipment.

[0120] In step S2, the signal data to be processed is image-generated, and the generated image is image-preprocessed, including:

[0121] Reconstructing the signal data to be processed into digital signals using digital signal processing technology, and forming preliminary image data of the signal data to be processed after the digital signal reconstruction;

[0122] The preliminary image data of the signal data to be processed includes a preliminary ultraviolet light image and a preliminary visible light image;

[0123] performing image preprocessing on the preliminary ultraviolet light image and the preliminary visible light image;

[0124] Image preprocessing includes denoising, enhancing, rectifying, cropping and scaling the image;

[0125] Unifying the formats of the preliminary ultraviolet image and the preliminary visible light image after image preprocessing;

[0126] After the format is unified, the target processed image is obtained.

[0127] Specifically, the preliminary UV and visible light images may contain various noise artifacts, such as granular, temporally and spatially random flickering bright spots. Denoising during image preprocessing can effectively reduce this noise interference and improve image clarity. However, the image features of the discharge region may be relatively faint and difficult to directly observe. Image enhancement techniques can highlight these features, making them more distinct and facilitating subsequent analysis and judgment. Image correction can eliminate these distortions and ensure image accuracy. Cropping can remove irrelevant image portions and highlight the discharge region. Scaling can adjust the image size for better suitability for subsequent analysis and processing. Unifying the format of the preliminary UV and visible light images ensures that they have the same resolution, color space, and other parameters. This facilitates subsequent image fusion, feature extraction, and pattern recognition. The unified format allows for easier fusion processing, where the UV imaging channel image is superimposed on the visible light channel image. This facilitates more intuitive observation of the discharge phenomenon and improves detection efficiency and accuracy.

[0128] Specifically, performing image enhancement on the preliminary ultraviolet image and the preliminary visible light image includes:

[0129] Extracting grayscale values ​​of pixels contained in the preliminary ultraviolet image as first grayscale value data;

[0130] Extracting the central grayscale value of the preliminary visible light image;

[0131] Obtaining a difference between the central grayscale values ​​of the preliminary visible light image and the preliminary ultraviolet light image as an observed difference;

[0132] Obtaining a preliminary ultraviolet image enhancement coefficient using the first grayscale value data and the central grayscale value of the preliminary visible light image in combination with an observed difference;

[0133] The preliminary ultraviolet image enhancement coefficient is obtained by the following formula:

[0134]

[0135] Among them, S z represents the preliminary ultraviolet image enhancement coefficient; n represents the number of gray values ​​contained in the first gray value data; H 01i represents the grayscale value corresponding to the i-th first grayscale value data; H 01i+1 Indicates the grayscale value corresponding to the i+1th first grayscale value data; H 01z Represents the central gray value of the preliminary UV image; H 02z Represents the central grayscale value of the preliminary visible light image; H 01brepresents the standard deviation of the grayscale value corresponding to the first grayscale value data; H 02b represents the grayscale value standard deviation corresponding to the second grayscale value data;

[0136] The contrast of the preliminary ultraviolet light image and the preliminary visible light image is adjusted using the preliminary ultraviolet light image enhancement coefficient.

[0137] The technical effect of the above-mentioned technical solution is to calculate the difference between the central grayscale values ​​of the preliminary UV image and the preliminary visible light image as the observed difference, and use this difference and grayscale data to calculate the preliminary UV image enhancement coefficient, thereby adjusting the image contrast. This process can significantly improve image contrast, making details in the image clearer. Image enhancement not only improves contrast but also enriches image details by adjusting the grayscale value distribution. This is of great significance for subsequent tasks such as image analysis and object detection. This technical solution can calculate the enhancement coefficient based on the actual grayscale value distribution of the preliminary UV image and the preliminary visible light image, thus having strong adaptability. Different images can achieve different enhancement effects, thus meeting different application requirements. By calculating statistical information such as the grayscale value standard deviation and combining it with the central grayscale value for adjustment, this technical solution can maintain the overall image contrast while enhancing local details. This helps to avoid local overexposure or underexposure during the image enhancement process. In the field of UV image detection, this technical solution can significantly improve image quality, thereby enhancing the accuracy and reliability of detection. This is of great significance in fields such as environmental monitoring and materials analysis. In addition to ultraviolet images, this technical solution is also applicable to the enhancement and optimization of visible light images. By adjusting contrast and enriching details, the visualization and application value of visible light images can be significantly improved. This technical solution also provides technical support for multispectral image fusion. By calculating the enhancement coefficients of images with different spectra and fusing them, richer and more comprehensive image information can be obtained, providing better data support for subsequent image analysis, target recognition, and other tasks.

[0138] In summary, the above technical solution, which calculates the enhancement coefficient by calculating the difference between the central grayscale values ​​of the preliminary UV image and the preliminary visible light image, and adjusts the image contrast, has significant technical effects and application prospects. It not only improves image quality and detail, but also has strong adaptability and broad application value.

[0139] Specifically, using the preliminary ultraviolet image enhancement coefficient to adjust the contrast of the preliminary ultraviolet image and the preliminary visible light image includes:

[0140] Extracting grayscale values ​​of pixels contained in the preliminary visible light image as second grayscale value data;

[0141] extracting a central grayscale value and an observation difference value of the preliminary ultraviolet image;

[0142] Obtaining a preliminary visible light image enhancement coefficient by using the central grayscale value and the observed difference of the preliminary ultraviolet image in combination with each grayscale value data in the second grayscale value data;

[0143] The preliminary visible light image enhancement coefficient is obtained by the following formula:

[0144]

[0145] Among them, S k represents the preliminary visible light image enhancement coefficient; m represents the number of grayscale data contained in the second grayscale value data; H 02i represents the grayscale value corresponding to the i-th second grayscale value data; H 02i+1 Indicates the grayscale value corresponding to the i+1th second grayscale value data; H 01z Represents the central gray value of the preliminary UV image; H 02z represents the central grayscale value of the preliminary visible light image; f represents the adjustment factor, and the adjustment factor is obtained by the following formula:

[0146]

[0147] Where, f represents the adjustment factor; H 01b represents the standard deviation of the grayscale value corresponding to the first grayscale value data; H 02b represents the grayscale value standard deviation corresponding to the second grayscale value data;

[0148] The contrast between the preliminary ultraviolet light image and the preliminary visible light image is adjusted using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient.

[0149] The technical solution described above significantly expands the image's dynamic range by calculating enhancement coefficients for the preliminary UV image and the preliminary visible light image and performing contrast adjustment on the two images. This means that details in both bright and dark areas of the image are better preserved and displayed, thereby improving the overall visual quality of the image. This solution utilizes information such as the central grayscale value, the observed difference, and the grayscale standard deviation to calculate the enhancement coefficient, helping to achieve balanced image contrast. This not only avoids overexposure or underexposure but also creates a more visually balanced and harmonious appearance across the image's regions. During the contrast adjustment process, this solution prioritizes preserving and enhancing image details. By finely adjusting the grayscale value of each pixel, details such as textures and edges in the image are made clearer and more prominent. While enhancing image details, this solution also suppresses noise to a certain extent. This is because statistical information such as the grayscale standard deviation is taken into account when calculating the enhancement coefficient. This information helps distinguish noise from true signal, effectively suppressing noise. This solution exhibits strong adaptability. It can calculate the enhancement coefficient based on the actual grayscale value distribution of the preliminary ultraviolet image and the preliminary visible light image, so it is suitable for image contrast adjustment under different scenes and lighting conditions. By adjusting the parameters in the formula (such as the adjustment factor f), the intensity of the contrast adjustment can be flexibly controlled. This provides users with more options and allows them to optimize the image effect according to actual needs. When calculating the enhancement coefficient, this technical solution mainly relies on statistical information such as the standard deviation of the grayscale value and the center grayscale value. The calculation of this information is relatively simple and efficient, so it can achieve a faster processing speed while ensuring image quality. Due to its high computational efficiency, this technical solution can be applied to real-time image processing systems. For example, in fields such as video surveillance and autonomous driving, real-time contrast adjustment of images can be achieved, thereby improving the performance and reliability of the system.

[0150] In summary, the technical benefits of the above-mentioned technical solution are primarily reflected in significantly improved contrast adjustment, image detail preservation and enhancement, adaptability and flexibility, and computational efficiency and real-time performance. These benefits give this technical solution broad application prospects and potential value in the field of image processing.

[0151] Specifically, adjusting the contrast of the preliminary ultraviolet image and the preliminary visible light image using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet image enhancement coefficient includes:

[0152] extracting preliminary visible light image enhancement coefficients and preliminary ultraviolet light image enhancement coefficients;

[0153] Obtaining a comprehensive contrast adjustment coefficient using the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0154] The comprehensive contrast adjustment coefficient is obtained by the following formula:

[0155]

[0156] Among them, S x Comprehensive contrast adjustment coefficient; S k represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 Respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0157] performing contrast adjustment on the preliminary visible light image using the comprehensive contrast adjustment coefficient to obtain a preliminary visible light image after contrast adjustment;

[0158] The contrast value corresponding to the preliminary visible light image after contrast adjustment is obtained by the following formula:

[0159]

[0160] Among them, D k represents the contrast value corresponding to the preliminary visible light image after contrast adjustment; D k0 Indicates the contrast value of the preliminary visible light image before adjustment; S x Comprehensive contrast adjustment coefficient; S k represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 Respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient;

[0161] performing contrast adjustment on the preliminary ultraviolet image using the comprehensive contrast adjustment coefficient to obtain a contrast-adjusted preliminary ultraviolet image;

[0162] The contrast value corresponding to the preliminary ultraviolet image after contrast adjustment is obtained by the following formula:

[0163]

[0164] Among them, D z represents the contrast value corresponding to the preliminary ultraviolet image after contrast adjustment; D z0 Indicates the contrast value of the preliminary UV image before adjustment; S x Comprehensive contrast adjustment coefficient; Sk represents the preliminary visible light image enhancement coefficient; S z represents the preliminary UV image enhancement coefficient; k 01 and k 02 They respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient.

[0165] The technical effect of the above technical solution is: by calculating the comprehensive contrast adjustment coefficient, the technical solution can comprehensively consider the contrast characteristics of the preliminary visible light image and the preliminary ultraviolet light image, and achieve more accurate and comprehensive contrast adjustment. The comprehensive contrast adjustment coefficient combines the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient, and introduces a preset weight coefficient, making the adjustment process more flexible and controllable. The technical solution provides a quantitative calculation method for the contrast value, which can accurately reflect the changes before and after the contrast adjustment, helping users to understand the adjustment effect more intuitively. Through quantitative control of the contrast value, users can more accurately adjust the contrast of the image to meet different application requirements. In the process of contrast adjustment, the technical solution focuses on preserving and enhancing the details of the image. By finely adjusting the grayscale value of each pixel, the texture, edge and other detail information in the image can be made clearer and more prominent. Detail preservation and enhancement help improve the readability and recognition of the image, and provide better data support for subsequent image analysis, target recognition and other tasks.

[0166] Images after contrast adjustment will experience significant visual improvements, including more vivid colors, clearer outlines, and richer details. This improved visual experience enhances the user's viewing experience, making the image more consistent with their visual needs and aesthetic standards. Contrast adjustment not only improves the visual quality of the image but also increases the amount of information contained in the image. By enhancing image contrast, more image information can be retained and displayed. This increased image information helps improve image analysis and processing capabilities, providing a better data foundation for subsequent image processing tasks. This technical solution is highly adaptable. It calculates a comprehensive contrast adjustment coefficient based on the actual contrast characteristics of the preliminary visible light image and preliminary ultraviolet image, making it suitable for image contrast adjustment in various scenes and lighting conditions. By adjusting the preset weighting coefficients k01 and k02, users can flexibly control the contribution of the preliminary visible light image enhancement coefficient and the preliminary ultraviolet image enhancement coefficient to the comprehensive contrast adjustment coefficient. This parameter adjustability provides users with more options to optimize the image contrast adjustment effect according to their actual needs.

[0167] In summary, the technical benefits of the above-mentioned technical solution are primarily reflected in significantly improved contrast adjustment, image detail preservation and enhancement, overall image quality improvements, and adaptability and flexibility. These benefits give this technical solution broad application prospects and potential value in the field of image processing.

[0168] Perform image fusion on the target processed image in S3, including:

[0169] The preliminary ultraviolet image and the preliminary visible light image in the target processing image are aligned using an image registration technology, and in the image alignment process, a feature point matching model is used to perform alignment processing, and the feature point matching model is retrieved from a database;

[0170] After the preliminary ultraviolet image and the preliminary visible light image are aligned, the preliminary ultraviolet image and the preliminary visible light image are merged using a pixel-level fusion method;

[0171] After the images are merged, image fusion processing is performed. The image fusion processing is to calculate the weighted average value of each corresponding pixel of the preliminary ultraviolet image and the preliminary visible light image. After the calculation is completed, the pixel value of the preliminary ultraviolet image and the preliminary visible light image after fusion is obtained;

[0172] performing contrast enhancement and brightness adjustment on the preliminary ultraviolet image and the preliminary visible light image after image fusion processing;

[0173] After contrast enhancement and brightness adjustment, the target fused image is obtained.

[0174] Specifically, UV imaging technology can detect partial discharge on the surface of electrical equipment, while visible light images provide intuitive morphological information about the equipment. By combining the two through image fusion, the discharge point can be more accurately located, improving detection accuracy. The application of a feature point matching model further improves the accuracy of image alignment, ensuring that the fused image truly reflects the equipment status. The pixel-level fusion method allows the information from the preliminary UV image and preliminary visible light image to be fully integrated. The weighted average calculation of each corresponding pixel retains the discharge information in the UV image while incorporating the morphological information from the visible light image, enhancing the richness and completeness of the image information. Contrast enhancement and brightness adjustment further improve the visual effect of the image, making the discharge phenomenon more obvious and easier to observe and analyze. Image fusion technology enables rapid detection of the status of electrical equipment. A fused image containing both discharge and morphological information can be obtained through a single capture and processing, eliminating the need for multiple captures and separate analysis, thereby improving detection efficiency. In addition, image fusion technology can also reduce the impact of human factors on detection results and improve the objectivity and accuracy of detection. By training machine learning models, it can automatically identify and analyze discharge phenomena and morphological information in fused images, further improving the degree of automation and accuracy of detection.

[0175] In order to solve the problem in the existing technology that the processed image data is not effectively confirmed for data features, and the feature data is not analyzed and abnormality judged more accurately, which leads to the inability to judge the detection results with the highest efficiency, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0176] In S4, the target fusion image is used to identify the discharge target, and the discharge target in the identified target fusion image is marked. After the marking is completed, the discharge target is integrated, including:

[0177] The target fusion image is separated from the foreground and background by edge detection, wherein the foreground in the target fusion image is the discharge target recognition area, and the background in the target fusion image is the remaining area except the foreground;

[0178] Extract features from the foreground of the target fusion image. The extracted features include shape features, texture features, spectral features, color features, brightness features, time data, intensity data, position data and spatial relationship features.

[0179] The extracted features are used to identify discharge targets using convolutional neural networks;

[0180] Marking and integrating the identified discharge targets;

[0181] Finally, the charged identification data is obtained.

[0182] Specifically, comprehensive feature extraction is performed on the foreground (discharge target identification area) in the target fusion image, including shape features, texture features, spectral features, color features, brightness features, time data, intensity data, location data, and spatial relationship features. These features together constitute a complete description of the discharge target, helping to improve recognition accuracy. Using a convolutional neural network (CNN) to perform discharge target identification on the extracted features, a more abstract and advanced feature representation is learned, enabling accurate recognition of the discharge target. Edge detection is used to separate the foreground and background in the target fusion image, making the discharge target identification area more distinct. This separation method not only helps reduce background interference but also improves image processing flexibility. In the feature extraction stage, the scheme integrates multiple feature information, fully leveraging the diversity of image data. In the target fusion stage, the flexibility and robustness of image processing are further enhanced by integrating information from different sources. By labeling and integrating the identified discharge targets, the scheme can generate charged identification data. These data provide strong support for live detection of power equipment, helping to timely discover and deal with potential discharge problems. By integrating multiple feature information, a comprehensive description and accurate identification of discharge targets are achieved. The convolutional neural network has strong generalization capabilities and can adapt to the discharge target identification needs in different scenarios.

[0183] Analyze the live identification data in S5, filter abnormal data based on the analysis results, determine the abnormality level of the filtered abnormal data, and issue an early warning based on the judgment results, including:

[0184] The charged identification data is preprocessed, which includes removing noise, processing missing data and normalizing data;

[0185] Use feature selection algorithm to extract key indicator data of discharge characteristics from charged identification data;

[0186] Using historical data to train a deep learning model, where the historical data and the deep learning model are retrieved from a database;

[0187] After the deep learning model training is completed, the key indicator data of the discharge characteristics in the charged identification data are analyzed.

[0188] The analysis results are checked for abnormal data, wherein the analysis results are compared with the standard data for data thresholds, and based on the threshold comparison results, the analysis results that are not within the standard data threshold range are marked as abnormal data;

[0189] Determine the degree of abnormality based on the abnormal threshold range of abnormal data;

[0190] The degree of abnormality is divided into mild abnormality, moderate abnormality and severe abnormality;

[0191] Provide early warning processing of different intensities according to the degree of abnormality;

[0192] Abnormal data and non-abnormal data, as well as the abnormality degree and warning information of abnormal data are uniformly marked as standard live detection data.

[0193] Specifically, by removing noise from the data, the accuracy and reliability of the data can be improved, providing a high-quality data foundation for subsequent analysis. Missing data may affect the accuracy of the analysis results. By processing missing data through appropriate methods (such as interpolation, filling, etc.), the integrity of the data can be ensured. Normalizing the data to the same order of magnitude helps to eliminate the dimensional differences between different features and improve the accuracy and generalization ability of the model. Using feature selection algorithms to extract key indicator data of discharge characteristics can reduce the dimension of the data and improve analysis efficiency while retaining the data that has the greatest impact on discharge characteristics. Using historical data to train deep learning models can learn the complex patterns and laws of discharge characteristics and improve the analysis accuracy of live identification data. Deep learning models can process large-scale data and have strong adaptability. They can cope with different power equipment and different discharge types. By comparing thresholds with standard data, abnormal data can be accurately confirmed to avoid false alarms and omissions. Judging the degree of abnormality based on the abnormal threshold range of abnormal data helps to take different treatment measures for discharge problems of different severity, improve maintenance efficiency, and carry out early warning treatment of different intensities according to different abnormal degrees, which can ensure that maintenance personnel are aware of the severity of the discharge problem in a timely manner and take corresponding maintenance measures to avoid damage to power equipment and safety accidents. By timely discovering and handling discharge problems, the normal operation of power equipment can be ensured and the stability and safety of the power grid can be improved.

[0194] In order to solve the problem in the existing technology that the final test results are not generated and presented in the inspection report, which makes it impossible for staff to make faster abnormal decisions based on the test results, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0195] The UV imaging live detection system for power equipment includes:

[0196] Live detection report generation unit, used for:

[0197] Generate a report using a visualization tool using standard live-test data, wherein the report layout is designed before generating the report;

[0198] The report layout includes text data, image data, table data and chart data. The chart data includes pie charts, curve charts, heat maps and bar charts.

[0199] After the report layout design is completed, use the visualization tool to generate a report for abnormal data and non-abnormal data in the standard live detection data;

[0200] The abnormal degree and warning information in the abnormal data are marked separately;

[0201] Finally, the generated report is transmitted to the display terminal for display. When the report is displayed, an abnormal warning is synchronously issued according to the warning information.

[0202] Specifically, by integrating text data, image data, table data, and chart data (including pie charts, curve charts, heat maps, and bar charts) into the report, the test results are made more intuitive and easy to understand. Different data types can display the results of live detection from different angles, which helps testers to have a more comprehensive understanding of the status of the equipment. During the report generation process, abnormal data and its abnormality level and warning information are individually marked, which helps testers quickly identify and locate problems, so as to take timely and effective measures. Visualization tools can automatically convert live detection data into reports, reducing the time of manual analysis and data collation, and improving detection efficiency. Rich data display forms (such as pie charts, curve charts, etc.) can help testers analyze data more deeply, discover potential trends and problems, and provide decision support for equipment maintenance and repair. When the report is displayed on the display terminal, abnormal warnings are synchronously issued according to the warning information, which helps to achieve real-time monitoring and warnings, and prevent unexpected power outages and damage caused by equipment failures. By designing a unified report layout and data display method, the standardization and normalization of live detection reports can be ensured, facilitating communication and collaboration between different personnel.

[0203] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0204] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. The method for detecting electric power equipment by ultraviolet imaging is characterized in that: The steps include: S1: Equipment detection and optical signal acquisition: Before optical signal acquisition, the acquisition equipment is first tested. After all the equipment passes the test, optical signal acquisition is performed. After the acquisition is completed, the signal data to be processed is obtained. S2: Image generation and processing: Generate an image from the signal data to be processed, perform image preprocessing on the generated image, and obtain the target processed image after image preprocessing; S3: Image fusion: performing image fusion on the target processed image, which includes an ultraviolet image and a visible light image, and obtaining a target fused image after image fusion; S4: Fusion image recognition: The target fusion image is used to identify the discharge target, and the discharge target in the identified target fusion image is marked. After marking, the discharge target is integrated to obtain the charged identification data; S5: Identification data analysis and early warning: Analyze the live identification data, screen abnormal data based on the analysis results, determine the abnormality level of the screened abnormal data, and issue an early warning based on the judgment results; In step S2, the generated image is subjected to image preprocessing, including: image enhancement of the preliminary ultraviolet image and preliminary visible light image formed after the digital signal is reconstructed to form the signal data to be processed, as follows: Extracting grayscale values ​​of pixels contained in the preliminary ultraviolet image as first grayscale value data; Extracting grayscale values ​​of pixels contained in the preliminary visible light image as second grayscale value data; extracting the central grayscale value of the preliminary ultraviolet image and the central grayscale value of the preliminary visible light image respectively; The preliminary UV image enhancement coefficient is obtained by the following formula: ； Among them, S z represents the preliminary ultraviolet image enhancement coefficient; n represents the number of gray values ​​contained in the first gray value data; H 01i represents the grayscale value corresponding to the i-th first grayscale value data; H 01i+1 Indicates the grayscale value corresponding to the i+1th first grayscale value data; H 01z Represents the central gray value of the preliminary UV image; H 02z Represents the central grayscale value of the preliminary visible light image; H 01b represents the standard deviation of the grayscale value corresponding to the first grayscale value data; H 02b represents the grayscale value standard deviation corresponding to the second grayscale value data; The preliminary visible light image enhancement coefficient is obtained by the following formula: ； Among them, S k represents the preliminary visible light image enhancement coefficient; m represents the number of grayscale data contained in the second grayscale value data; H 02i represents the grayscale value corresponding to the i-th second grayscale value data; H 02i+1 represents the grayscale value corresponding to the i+1th second grayscale value data; f represents the adjustment factor, which is obtained by the following formula: ; The preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient are used to obtain a comprehensive contrast adjustment coefficient, and the comprehensive contrast adjustment coefficient is obtained by the following formula: ； Among them, S x Comprehensive contrast adjustment coefficient; k 01 and k 02 Respectively represent the preset weight coefficients corresponding to the preliminary visible light image enhancement coefficient and the preliminary ultraviolet light image enhancement coefficient; The comprehensive contrast adjustment coefficient is used to perform contrast adjustment on the preliminary visible light image and the preliminary ultraviolet light image, respectively, to obtain a preliminary visible light image and a preliminary ultraviolet light image after contrast adjustment.

2. The method for detecting charged power equipment by ultraviolet imaging according to claim 1, characterized in that: In S1, the acquisition equipment is tested. After all the equipment passes the test, optical signal acquisition is performed, including: The acquisition equipment includes an ultraviolet imager, which collects ultraviolet light signals and visible light signals; Before collecting UV and visible light signals, the UV imager should be tested first; Among them, the equipment inspection of UV imagers includes appearance inspection, function inspection, performance inspection and safety inspection; After all the equipment are tested and qualified, signal collection will be carried out; if the equipment fails to pass the test, repair or replacement will be carried out; The ultraviolet imager converts the collected ultraviolet light signals and visible light signals into digital signals, and then performs digital signal processing after the conversion is completed; Digital signal processing includes signal amplification and filtering; After digital signal processing, the signal data to be processed is obtained.

3. The method for detecting charged power equipment by ultraviolet imaging according to claim 2, characterized in that: In step S2, the signal data to be processed is image-generated, and the generated image is image-preprocessed, including: Reconstructing the signal data to be processed into digital signals using digital signal processing technology, and forming preliminary image data of the signal data to be processed after the digital signal reconstruction; The preliminary image data of the signal data to be processed includes a preliminary ultraviolet light image and a preliminary visible light image; performing image preprocessing on the preliminary ultraviolet light image and the preliminary visible light image; Image preprocessing also includes denoising, rectification, cropping and scaling of images; Unifying the formats of the preliminary ultraviolet image and the preliminary visible light image after image preprocessing; After the format is unified, the target processed image is obtained.

4. The method for detecting charged power equipment by ultraviolet imaging according to claim 3, characterized in that: Using the comprehensive contrast adjustment coefficient to perform contrast adjustment on the preliminary visible light image and the preliminary ultraviolet light image respectively to obtain the contrast-adjusted preliminary visible light image and preliminary ultraviolet light image, comprising: The contrast value corresponding to the preliminary visible light image after contrast adjustment is obtained by the following formula: ； Among them, D k represents the contrast value corresponding to the preliminary visible light image after contrast adjustment; D k0 Indicates the contrast value of the preliminary visible light image before adjustment; The contrast value corresponding to the preliminary ultraviolet image after contrast adjustment is obtained by the following formula: ； Among them, D z represents the contrast value corresponding to the preliminary ultraviolet image after contrast adjustment; D z0 Indicates the contrast value of the preliminary UV image before adjustment.

5. The method for detecting charged power equipment by ultraviolet imaging according to claim 4, characterized in that: Perform image fusion on the target processed image in S3, including: The preliminary ultraviolet image and the preliminary visible light image in the target processing image are aligned using an image registration technology, and in the image alignment process, a feature point matching model is used to perform alignment processing, and the feature point matching model is retrieved from a database; After the preliminary ultraviolet image and the preliminary visible light image are aligned, the preliminary ultraviolet image and the preliminary visible light image are merged using a pixel-level fusion method; After the images are merged, image fusion processing is performed. The image fusion processing is to calculate the weighted average value of each corresponding pixel of the preliminary ultraviolet image and the preliminary visible light image. After the calculation is completed, the pixel value of the preliminary ultraviolet image and the preliminary visible light image after fusion is obtained; performing contrast enhancement and brightness adjustment on the preliminary ultraviolet image and the preliminary visible light image after image fusion processing; After contrast enhancement and brightness adjustment, the target fused image is obtained.

6. The method for detecting charged power equipment by ultraviolet imaging according to claim 5, characterized in that: In S4, the target fusion image is used to identify the discharge target, and the discharge target in the identified target fusion image is marked. After the marking is completed, the discharge target is integrated, including: The target fusion image is separated from the foreground and background by edge detection, wherein the foreground in the target fusion image is the discharge target recognition area, and the background in the target fusion image is the remaining area except the foreground; Extract features from the foreground of the target fusion image. The extracted features include shape features, texture features, spectral features, color features, brightness features, time data, intensity data, position data and spatial relationship features. The extracted features are used to identify discharge targets using convolutional neural networks; Marking and integrating the identified discharge targets; Finally, the charged identification data is obtained.

7. The method for detecting charged power equipment by ultraviolet imaging according to claim 6, characterized in that: Analyze the live identification data in S5, filter abnormal data based on the analysis results, determine the abnormality level of the filtered abnormal data, and issue an early warning based on the judgment results, including: The charged identification data is preprocessed, which includes removing noise, processing missing data and normalizing data; Use feature selection algorithm to extract key indicator data of discharge characteristics from charged identification data; Using historical data to train a deep learning model, where the historical data and the deep learning model are retrieved from a database; After the deep learning model training is completed, the key indicator data of the discharge characteristics in the charged identification data are analyzed; The analysis results are checked for abnormal data, wherein the analysis results are compared with the standard data for data thresholds, and based on the threshold comparison results, the analysis results that are not within the standard data threshold range are marked as abnormal data; Determine the degree of abnormality based on the abnormal threshold range of abnormal data; The degree of abnormality is divided into mild abnormality, moderate abnormality and severe abnormality; Provide early warning processing of different intensities according to the degree of abnormality; Abnormal data and non-abnormal data, as well as the abnormality degree and warning information of abnormal data are uniformly marked as standard live detection data.

8. A UV imaging charged detection system for electric power equipment, used in the UV imaging charged detection method for electric power equipment according to claim 7, characterized in that: include: Live detection report generation unit, used for: Generate a report using a visualization tool using standard live-test data, wherein the report layout is designed before generating the report; The report layout includes text data, image data, table data and chart data. The chart data includes pie charts, curve charts, heat maps and bar charts. After the report layout design is completed, use the visualization tool to generate a report for abnormal data and non-abnormal data in the standard live detection data; The abnormal degree and warning information in the abnormal data are marked separately; Finally, the generated report is transmitted to the display terminal for display. When the report is displayed, an abnormal warning is synchronously issued according to the warning information.

Images