A method and device for detecting deterioration of an energized non-contact insulator

Abstract

The present invention provides a live, non-contact insulator deterioration detection method and device. The method receives infrared and ultraviolet images of an insulator string to be inspected, along with the ambient humidity; pre-processes the infrared image to obtain a target grayscale image; extracts all grayscale image pixels from the image and determines a target threshold using the maximum inter-class variance method; determines the characteristic temperature of the steel leg area and each insulator based on the target threshold; constructs a grayscale-temperature relationship curve and then exports an axial temperature data file. A temperature difference threshold is determined based on all axial temperature data files to identify deteriorated insulators; and the contamination level of the deteriorated insulator is estimated based on the ultraviolet image and ambient humidity, generating and outputting a detection report. By capturing infrared and ultraviolet images and establishing a corresponding relationship between the contamination levels, the method uses the maximum inter-class variance method to effectively enhance the segmentation effect and improve image segmentation accuracy. Deteriorated insulators are identified based on statistical analysis, significantly reducing workload and improving detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of insulator detection, and in particular to a method and device for detecting degradation of an energized non-contact insulator. Background Art

[0002] In high-voltage transmission lines, insulators not only serve as anchors and suspension systems but also provide crucial electrical insulation, making them crucial to the transmission and distribution of electricity. Because insulators are constantly exposed to high-voltage electric fields, bearing mechanical loads, and facing harsh environments such as high temperatures, intense sunlight, high humidity, and dirt, degradation is inevitable. When insulator degradation reaches a certain level, its insulation performance deteriorates, potentially leading to flashovers, string drop, and conductor drop, compromising transmission line reliability. Therefore, timely detection of deteriorated insulators is crucial.

[0003] However, the sheer number of line insulators, coupled with the random and irregular location and distribution of deteriorated insulators, makes insulator inspection extremely laborious. Reducing the inspection workload not only reduces misjudgments due to fatigue among operators, but also reduces time costs, allowing for more frequent inspections. This allows for more timely detection of potential insulator failures, reducing the potential for line outages and equipment damage caused by these failures.

[0004] Therefore, how to optimize the detection judgment process and logic to reduce the detection workload and improve detection accuracy is an urgent problem that needs to be solved. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a live non-contact insulator degradation detection method and device to solve the problems of low detection efficiency and unreliability.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] A first aspect of the present invention discloses a method for detecting degradation of a live non-contact insulator, the method comprising:

[0008] receiving multiple infrared images, multiple ultraviolet images and environmental humidity of the insulator string to be inspected sent by the inspection equipment;

[0009] For each of the infrared images, preprocessing the infrared image to obtain a target grayscale image;

[0010] Based on all grayscale pixels of the target grayscale image, calculating the inter-class variance corresponding to each threshold within the image grayscale threshold range, and marking the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold;

[0011] Converting the target grayscale image into a binary image according to the target threshold value to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determining the average temperature as the characteristic temperature of each insulator;

[0012] Constructing a grayscale-temperature relationship curve based on the maximum temperature in the target grayscale image and the maximum grayscale in the binary image;

[0013] Deriving an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the gray-scale temperature relationship curve;

[0014] Calculating the temperature difference between each insulator and its adjacent insulator in the insulator string to be inspected based on all axial temperature data files;

[0015] determining a temperature difference threshold based on the temperature difference, and determining the insulator having a temperature difference greater than the temperature difference threshold as a degraded insulator in the insulator string to be detected;

[0016] The contamination level of the degraded insulator is estimated according to the ultraviolet image corresponding to the degraded insulator and the ambient humidity, and a detection report of the degraded insulator is generated and output.

[0017] Preferably, for each infrared image, preprocessing the infrared image to obtain a target grayscale image includes:

[0018] For each infrared image, obtaining the RGB value of each pixel of the infrared image, and converting the infrared image into a grayscale image based on the RGB value;

[0019] Performing median filtering on the grayscale image to reduce noise, thereby obtaining a grayscale image after noise reduction;

[0020] The denoised grayscale image is subjected to tilt correction processing based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

[0021] Preferably, the step of calculating the inter-class variance corresponding to each threshold within the image grayscale threshold range based on all grayscale image pixels of the target grayscale image, and marking the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold, includes:

[0022] Obtain the number of all grayscale pixels of the target grayscale image;

[0023] Calculating the grayscale mean of the target grayscale image according to the quantity;

[0024] For each threshold value in the image grayscale threshold range, calculate a first number of all grayscale image pixel points that are less than or equal to the threshold value, and a second number of all grayscale image pixel points that are greater than the threshold value;

[0025] Calculating a first between-class variance corresponding to the first quantity, a first category weight corresponding to the first quantity, a second between-class variance corresponding to the second quantity, and a second category weight corresponding to the second quantity;

[0026] Calculating the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range;

[0027] The maximum inter-class variance is extracted from all inter-class variances, and the threshold corresponding to the maximum inter-class variance is marked as the target threshold.

[0028] Preferably, converting the target grayscale image into a binary image according to the target threshold value to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determining the average temperature as the characteristic temperature of each insulator includes:

[0029] Converting the target grayscale image into a binary image according to the target threshold;

[0030] Finding the first n connected regions in descending order of area in the binary image, and marking the connected regions as the steel foot regions of each insulator in the infrared image;

[0031] Extracting the image grayscale value corresponding to each steel foot area according to the target grayscale image;

[0032] For each of the steel foot regions, calculating an average grayscale value of the steel foot region based on the area of ​​the steel foot region and the grayscale value of the image;

[0033] An average temperature corresponding to each of the steel leg regions is determined based on the average grayscale value, and the average temperature is determined as a characteristic temperature of each insulator.

[0034] Preferably, determining the temperature difference threshold based on the temperature difference includes:

[0035] The temperature difference mean and standard deviation corresponding to all axial temperature data files are calculated based on the temperature difference, and the temperature difference threshold is determined according to the temperature difference mean and the standard deviation.

[0036] A second aspect of the present invention discloses a live non-contact insulator degradation detection device, the device comprising:

[0037] A receiving unit, configured to receive multiple infrared images, multiple ultraviolet images, and ambient humidity of the insulator string to be inspected sent by the inspection equipment;

[0038] a preprocessing unit, configured to preprocess each infrared image to obtain a target grayscale image;

[0039] A first calculation unit is configured to calculate, based on all grayscale image pixels of the target grayscale image, the inter-class variance corresponding to each threshold within the image grayscale threshold range, and mark the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold;

[0040] a first determining unit, configured to convert the target grayscale image into a binary image according to the target threshold value, so as to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determine the average temperature as the characteristic temperature of each insulator;

[0041] A construction unit, configured to construct a grayscale-temperature relationship curve based on the maximum temperature in the target grayscale image and the maximum grayscale in the binary image;

[0042] an exporting unit, configured to export an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the grayscale-temperature relationship curve;

[0043] A second calculation unit is used to calculate the temperature difference between each insulator in the insulator string to be tested and its adjacent insulators according to all axial temperature data files;

[0044] a second determining unit, configured to determine a temperature difference threshold based on the temperature difference, and determine the insulator having a temperature difference greater than the temperature difference threshold as a degraded insulator in the insulator string to be detected;

[0045] The output unit is configured to estimate the contamination level of the deteriorated insulator according to the ultraviolet image corresponding to the deteriorated insulator and the ambient humidity, and generate and output a detection report of the deteriorated insulator.

[0046] Preferably, the pre-processing unit includes:

[0047] an acquisition module, configured to acquire, for each infrared image, an RGB value of each pixel of the infrared image, and convert the infrared image into a grayscale image based on the RGB value;

[0048] A noise reduction module, configured to perform median filtering on the grayscale image to reduce noise, thereby obtaining a grayscale image after noise reduction;

[0049] The correction module is used to perform tilt correction processing on the denoised grayscale image based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

[0050] Preferably, the first computing unit includes:

[0051] A quantity acquisition module is used to obtain the quantity of all grayscale pixels of the target grayscale image;

[0052] A first calculation module, configured to calculate a grayscale mean of the target grayscale image according to the quantity;

[0053] A second calculation module is configured to calculate, for each threshold value in the grayscale threshold range of the image, a first number of pixels in all grayscale images that are less than or equal to the threshold value, and a second number of pixels in all grayscale images that are greater than the threshold value;

[0054] a third calculation module, configured to calculate a first between-class variance corresponding to the first quantity, a first category weight corresponding to the first quantity, a second between-class variance corresponding to the second quantity, and a second category weight corresponding to the second quantity;

[0055] a fourth calculation module, configured to calculate the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range;

[0056] The marking module is used to extract the maximum inter-class variance from all inter-class variances, and mark the threshold corresponding to the maximum inter-class variance as the target threshold.

[0057] Preferably, the first determining unit includes:

[0058] A conversion module, configured to convert the target grayscale image into a binary image according to the target threshold;

[0059] A search module is used to search for the first n connected regions in the binary image in descending order of area, and mark the connected regions as the steel foot regions of each insulator in the infrared image;

[0060] An extraction module, configured to extract an image grayscale value corresponding to each steel foot region according to the target grayscale image;

[0061] a fifth calculation module, configured to calculate, for each of the steel foot regions, an average grayscale value of the steel foot region based on the area of ​​the steel foot region and the grayscale value of the image;

[0062] A determination module is configured to determine an average temperature corresponding to each of the steel leg regions based on the average grayscale value and determine the average temperature as a characteristic temperature of each insulator.

[0063] Preferably, the second determining unit is specifically configured to calculate a temperature difference mean and a standard deviation corresponding to all axial temperature data files based on the temperature difference, and determine a temperature difference threshold according to the temperature difference mean and the standard deviation.

[0064] Based on the above-mentioned embodiments of the present invention, a live, non-contact insulator degradation detection method and device are provided. The method receives infrared images, ultraviolet images, and ambient humidity of the insulator string to be inspected; preprocesses the infrared image to obtain a target grayscale image; obtains all grayscale image pixels from the image and determines a target threshold using the maximum inter-class variance method; determines the characteristic temperature of the steel leg area and each insulator based on the target threshold; constructs a grayscale-temperature relationship curve and then exports an axial temperature data file. A temperature difference threshold is determined based on all axial temperature data files to identify degraded insulators; the contamination level of the degraded insulator is estimated based on the ultraviolet image and ambient humidity, and a detection report is generated and output. By capturing infrared and ultraviolet images and establishing a corresponding relationship between the contamination levels, the present invention uses the maximum inter-class variance method to effectively enhance the segmentation effect and improve image segmentation accuracy. Degraded insulators are identified based on statistical analysis, significantly reducing workload and improving detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0066] Figure 1 A flowchart of a live non-contact insulator degradation detection method provided by an embodiment of the present invention;

[0067] Figure 2 A flowchart of preprocessing an infrared image provided by an embodiment of the present invention;

[0068] Figure 3 A flowchart of calculating the inter-class variance based on the target grayscale image to determine the target threshold provided by an embodiment of the present invention;

[0069] Figure 4 A flow chart for determining the average temperature of a steel foot region and each steel foot region according to a target threshold value provided in an embodiment of the present invention;

[0070] Figure 5This is a structural block diagram of a live non-contact insulator degradation detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0071] 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.

[0072] In this application, the terms "comprises," "comprising," or any other variations thereof are intended to encompass 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. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0073] As can be seen from the background, the number of live non-contact insulators is enormous, and the location and distribution of deterioration are random and irregular, making the detection of deteriorated insulators a daunting task. With the continuous expansion of the power grid and the gradual construction of ultra-high voltage (UHV) lines, the requirements for insulator performance have become increasingly stringent, and detection work has become even more challenging. The current tower-climbing inspection method is inefficient, and operators may misjudge due to fatigue, resulting in unreliable test results.

[0074] While a significant proportion of deteriorated insulators cannot be identified through visual imaging, the characteristics of deteriorated transmission line insulators, such as localized overheating, partial discharge, and physical damage to the dielectric surface, can be measured and assessed using ultraviolet (UV) imagers, infrared (IR) imagers, or ultrasonic transmitters and receivers. Furthermore, deteriorating insulators cause changes in the spatial electric field distribution, enabling non-contact electric field measurement methods to detect deteriorated insulators. However, ultrasonic and UV testing are not effective in determining insulator deterioration and are only effective in determining the degree of contamination. Electric field measurement methods, however, require a significant workload to measure the electric field distribution along each insulator string. Infrared detection of deteriorated insulators allows for rapid detection, but variations in the heating characteristics of deteriorated insulators under different contamination conditions, coupled with the inherent limitations of current infrared detection technology, can lead to unreliable test results.

[0075] Therefore, an embodiment of the present invention provides a live, non-contact insulator deterioration detection method and device. The method receives infrared images, ultraviolet images, and ambient humidity of an insulator string to be detected; preprocesses each infrared image to obtain a target grayscale image; determines a target threshold using the maximum inter-class variance method based on all grayscale pixels of the target grayscale image; determines the steel leg area of ​​each insulator in the infrared image and the average temperature of each steel leg area based on the target threshold, thereby determining the characteristic temperature of each insulator; constructs a grayscale-temperature relationship curve and then exports axial temperature data files of each insulator in the infrared image. A temperature difference threshold is determined based on all axial temperature data files to identify deteriorated insulators in the insulator string to be detected; further, the contamination level of the deteriorated insulator is estimated based on the corresponding ultraviolet image and ambient humidity of the deteriorated insulator, and a deteriorated insulator detection report is generated and output. The present invention uses drones to capture characteristic parameters such as infrared spectra and ultraviolet images, establishes a corresponding relationship between pollution levels, and performs data processing and statistical analysis to determine the deterioration of insulators. The use of the maximum inter-class variance method can effectively enhance the segmentation effect, improve the clarity and accuracy of image segmentation, and reduce the probability of misclassification. The present invention significantly improves the performance and accuracy of detection.

[0076] See also Figure 1 , shows a flow chart of a method for detecting deterioration of an energized non-contact insulator provided by an embodiment of the present invention. The method comprises:

[0077] As you can understand, an insulator consists of a steel cap, a steel leg, and a porcelain plate, and is axially symmetrical. By photographing its side view, the axial thermal field distribution of the insulator can be effectively captured and analyzed.

[0078] In practical applications, this method is applicable to a live non-contact insulator degradation detection system, which includes but is not limited to a flight module, a detection module, a ground analysis and processing module, and a wireless communication module.

[0079] The flight module includes, but is not limited to, a drone capable of carrying equipment. The detection module includes, but is not limited to, infrared thermal imagers, ultraviolet imagers, hygrometers, and other detection equipment, which are mounted on the flight module.

[0080] It should be noted that the multiple infrared images, multiple ultraviolet images and ambient humidity taken by the detection module are transmitted to the ground analysis and processing module through the wireless communication module; the ground analysis and processing module is used to execute a live non-contact insulator degradation detection method provided by an embodiment of the present invention, analyze and process the received multiple infrared images, multiple ultraviolet images and ambient humidity, and thus output a detection report on the degraded insulator.

[0081] Step S101: receiving multiple infrared images, multiple ultraviolet images and environmental humidity of an insulator string to be inspected sent by a detection device.

[0082] In practical applications, by mounting infrared imagers, ultraviolet imagers, hygrometers and other detection equipment on a mountable drone, the shooting angle is continuously adjusted, and corresponding images of insulator strings along the transmission line are captured, their positions are marked, and the humidity of the environment is detected. Multiple infrared images, multiple ultraviolet images and environmental humidity of the insulator strings to be tested are obtained.

[0083] In the specific implementation of step S101 , multiple infrared images, multiple ultraviolet images and environmental humidity of the insulator string to be inspected sent by the inspection device are received, and all infrared images are stored in the insulator infrared image library.

[0084] For example, the infrared images are stored in the insulator infrared image library, and each infrared image is named Infrare(t,S i ,X j ,L k ,P m ,I n ).jpg. Where t is the shooting time; S is the line segment number, i is the sequence number; X is the side where the insulator string is located, with the power transmission side (small side) being 1 and the load side (large side) being 2, and j is the sequence number; L is the circuit number, k is the sequence number; P is the phase number, m is the sequence number; I is the string number, n is the sequence number.

[0085] Step S102: For each infrared image, preprocess the infrared image to obtain a target grayscale image.

[0086] In the specific implementation of step S102 , each received infrared image is preprocessed, including but not limited to converting the infrared image into a grayscale image, performing median filtering noise reduction and angle correction, and obtaining a target grayscale image.

[0087] It is understandable that converting infrared images into grayscale images, performing median filtering for noise reduction and angle correction can effectively improve image quality, reduce noise interference, and ensure image consistency and accuracy, which are important foundations for subsequent analysis and processing.

[0088] See the embodiment of the present invention. Figure 2 ,The process of preprocessing the infrared image includes:

[0089] Step S201: for each infrared image, obtain the RGB value of each pixel of the infrared image, and convert the infrared image into a grayscale image based on the RGB value.

[0090] In the specific implementation of step S201, for each infrared image, the RGB value of each pixel of the infrared image is obtained, and the infrared image is converted into a grayscale image according to formula (1).

[0091] gray=0.2989×R+0.5870×G+0.1140×B (1)

[0092] Among them, gray is the grayscale image; R is the R value of the pixel in the infrared image; G is the G value of the pixel in the infrared image; B is the B value of the pixel in the infrared image.

[0093] It should be noted that in the RGB color space, RGB stands for red, green, and blue, respectively.

[0094] As you can understand, because grayscale images only contain brightness information, they reduce the amount of data, making subsequent image processing and analysis more efficient. Furthermore, in infrared imaging, temperature differences between objects are reflected in the image's brightness, and grayscale images can visually display these differences, facilitating observation and analysis.

[0095] Step S202: performing median filtering on the grayscale image to reduce noise, thereby obtaining a grayscale image after noise reduction.

[0096] In the specific implementation process of step S202, the grayscale image is loaded, the filter window size is determined, and then the grayscale image is traversed. For each pixel in the grayscale image, the selected filter window is used to collect the values ​​of the surrounding pixels, and the median of all pixel values ​​in the filter window is calculated. Finally, the calculated median replaces the value of the current pixel to obtain the denoised grayscale image.

[0097] It is understandable that, compared with mean filtering, median filtering for grayscale image denoising can better remove outliers while preserving the edge features of the image and is less susceptible to extreme values. Furthermore, median filtering can adapt to different noise distributions and has a certain degree of robustness.

[0098] Step S203: performing tilt correction processing on the denoised grayscale image based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

[0099] In the specific implementation of step S203 , the grayscale image after noise reduction is subjected to tilt correction processing using a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

[0100] The preset edge detection operator is the Canny edge detection operator; the Hough transform operator is the Hough transform operator.

[0101] Specifically, the Canny edge detection operator is used to extract the image edge features in the denoised grayscale image, and then the Hough transform operator is used to detect the straight line features in the denoised grayscale image. The longest straight line and its tilt angle are calculated and found. Finally, the affine transformation matrix is ​​determined according to the calculated tilt angle. The denoised grayscale image is rotated based on the affine transformation matrix to perform tilt correction processing on it to obtain the target grayscale image.

[0102] Here, the affine transformation matrix is ​​such as formula (2).

[0103]

[0104] Here, θ is the tilt angle.

[0105] Combine Figure 2 As shown in the figure, after converting infrared images into grayscale images, the complexity of the data can be reduced, making subsequent image processing (such as filtering and edge detection) more efficient. In addition, median filtering can effectively remove dirty spots such as salt noise while retaining edge information, improving image quality, and enhancing the accuracy of subsequent analysis. Images processed by median filtering usually show clearer features, which is helpful for object recognition and other image analysis tasks. At the same time, through angle correction, the distortion caused by different shooting angles can be eliminated, making the image more consistent with the actual scene, thus facilitating further analysis. Overall, the denoised and corrected images can significantly improve the performance and accuracy of subsequent detection.

[0106] Step S103: Based on all grayscale pixels of the target grayscale image, calculate the inter-class variance corresponding to each threshold within the image grayscale threshold range, and mark the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold.

[0107] In the specific implementation of step S103, all grayscale image pixels are obtained from the target grayscale image, and the image grayscale threshold range (0-255) is obtained. All grayscale image pixels and each threshold within the image grayscale threshold range are combined to calculate the inter-class variance, thereby determining the target threshold.

[0108] It can be understood that using the maximum inter-class variance method (Otsu method) to calculate the inter-class variance of the target grayscale image and thus determine the target threshold can effectively improve the quality of image segmentation. It has many advantages such as strong adaptability, reduction of mis-segmentation, and solid theoretical basis.

[0109] See the embodiment of the present invention. Figure 3 ,The process of calculating the inter-class variance based on the target grayscale image to determine the target threshold includes:

[0110] Step S301: Obtain the number of all grayscale pixels of the target grayscale image.

[0111] In the specific implementation of step S301 , the number of all grayscale pixels in the target grayscale image is counted.

[0112] Step S302: Calculate the grayscale mean of the target grayscale image according to the quantity.

[0113] In the specific implementation of step S302, the grayscale mean of the target grayscale image is calculated according to the number of all grayscale image pixels, which can be calculated specifically according to formula (3).

[0114]

[0115] Among them, α av is the grayscale mean; n is the total number of pixels in the grayscale image; α(i) is the grayscale value of the i-th pixel in the target grayscale image.

[0116] It is understandable that calculating the grayscale mean of the target grayscale image can help quickly evaluate the overall brightness distribution of the image, which helps to subsequently determine the appropriate target threshold.

[0117] Step S303: for each threshold value in the image grayscale threshold range, calculate a first number of all grayscale image pixel points that are less than or equal to the threshold value, and a second number of all grayscale image pixel points that are greater than the threshold value.

[0118] It should be noted that, in the image grayscale threshold range, the grayscale value ranges from 0 to 255.

[0119] In the specific implementation process of step S303, the image grayscale threshold range is obtained, and for each threshold in the range (i.e., 0 to 255), the number of grayscale pixel points that are less than or equal to the threshold among all grayscale pixel points is counted, marked as the first number, and the number of grayscale pixel points that are greater than the threshold among all grayscale pixel points is counted, marked as the second number.

[0120] Step S304: Calculate the first between-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second between-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity.

[0121] In the specific implementation of step S304, the first between-class variance corresponding to the first quantity and the second between-class variance corresponding to the second quantity are calculated, specifically according to formula (4). The first category weight corresponding to the first quantity and the second category weight corresponding to the second quantity are calculated, specifically according to formula (5).

[0122] Formula (4) is as follows:

[0123]

[0124] Among them, t is any threshold in the image grayscale threshold range, α av1 is the first between-class variance corresponding to the first quantity; α av2 is the second between-class variance corresponding to the second quantity; α(i) is the grayscale value of the i-th pixel in the target grayscale image; N1 is the first quantity less than or equal to t; N2 is the second quantity greater than t.

[0125] Formula (5) is as follows:

[0126]

[0127] Wherein, t is any threshold in the grayscale threshold range of the image, N1 is a first number less than or equal to t; N2 is a second number greater than t; N is the number of all grayscale image pixels; ω1(t) is the first category weight corresponding to the first number; ω2(t) is the second category weight corresponding to the second number.

[0128] Step S305: Calculate the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range.

[0129] In the specific implementation of step S305 , the inter-class variance corresponding to the threshold is calculated based on the first inter-class variance, the first category weight, the second inter-class variance and the second category weight to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range.

[0130] Specifically, the calculation is performed according to formula (6).

[0131]

[0132] in, is the inter-class variance corresponding to the threshold t; α av is the grayscale mean; ω1(t) is the first category weight corresponding to the first quantity; ω2(t) is the second category weight corresponding to the second quantity; α av1 is the first between-class variance corresponding to the first quantity; α av2 is the second between-class variance corresponding to the second quantity.

[0133] Step S306: extracting the largest inter-class variance from all inter-class variances, and marking the threshold corresponding to the largest inter-class variance as the target threshold.

[0134] In the specific implementation of step S306 , the inter-class variance with the largest value is extracted from all inter-class variances, and the threshold corresponding to the largest inter-class variance is marked as the target threshold.

[0135] As shown in formula (7).

[0136]

[0137] Where t′ is the target threshold.

[0138] Combine Figure 3 The results show that using the maximum inter-class variance method (Otsu method) can effectively enhance the segmentation effect, improve the clarity and accuracy of image segmentation, and reduce the probability of misclassification. Furthermore, this method does not rely on prior knowledge of image content and can automatically find the optimal target threshold, making it applicable to images with different characteristics.

[0139] Step S104: converting the target grayscale image into a binary image according to the target threshold value to determine the steel foot area of ​​each insulator in the infrared image and the average temperature of each steel foot area, and determining the average temperature as the characteristic temperature of each insulator.

[0140] In the specific implementation process of step S104, the target grayscale image is first converted into a binary image according to the target threshold value, and the steel foot area of ​​each insulator in the infrared image is determined based on the binary image. Then, the average temperature corresponding to each steel foot area is determined based on the target grayscale image. Finally, the average temperature corresponding to each steel foot area is determined as the characteristic temperature of each insulator corresponding to each steel foot area.

[0141] As you can see, binarization of the target grayscale image clearly segments the steel foot area, reducing background interference and improving the accuracy of subsequent temperature measurements. This clear regional demarcation makes calculating the average temperature of each steel foot area more efficient, saving time and manual intervention. Furthermore, combining the advantages of image processing and temperature measurement improves both detection efficiency and accuracy.

[0142] See the embodiment of the present invention. Figure 4 The process of determining the average temperature of the steel leg region and each steel leg region according to the target threshold and determining the average temperature as the characteristic temperature of each insulator includes:

[0143] Step S401: converting the target grayscale image into a binary image according to the target threshold.

[0144] In the specific implementation of step S401 , the target grayscale image is converted based on the target threshold value to obtain a binary image.

[0145] Step S402: searching for the first n connected regions in descending order of area in the binary image, and marking the connected regions as the steel leg regions of each insulator in the infrared image.

[0146] It should be noted that, since the temperature rise in the steel leg area of ​​the insulator is relatively high, it often appears as a connected area in the binary image.

[0147] In the specific implementation of step S402, all connected regions in the binary image are searched and the areas of these connected regions are calculated; all connected regions are sorted in descending order of area, and the first n (for example, 37) connected regions are extracted from them, and the extracted connected regions are marked as the steel foot areas of each insulator in the infrared image.

[0148] It can be understood that after the target grayscale image is binarized, the steel foot area can be clearly segmented, thereby improving the accuracy of subsequent temperature measurement.

[0149] Step S403: extracting the image grayscale value corresponding to each steel foot region according to the target grayscale image.

[0150] In the specific implementation of step S403 , for each steel foot region, the image grayscale value corresponding to the steel foot region is extracted according to the target grayscale image.

[0151] It can be understood that each steel foot area corresponds to a plurality of image grayscale values.

[0152] Step S404: For each steel foot region, calculate the average grayscale value of the steel foot region based on the area of ​​the steel foot region and the image grayscale value.

[0153] In the specific implementation of step S404, for each steel foot region, the sum of the image grayscale values ​​corresponding to the steel foot region is calculated, and the quotient of the area of ​​the steel foot region and the sum is calculated to obtain the average grayscale value of the steel foot region.

[0154] Step S405: determining the average temperature corresponding to each steel leg region based on the average grayscale value, and determining the average temperature as the characteristic temperature of each insulator.

[0155] As you can understand, grayscale values ​​are often correlated with an object's temperature. Infrared imaging technology captures images by detecting infrared radiation emitted by an object. Different objects radiate infrared radiation at varying intensities at different temperatures, so their grayscale values ​​can reflect temperature variations.

[0156] In the specific implementation process of step S405, the temperature corresponding to the average grayscale value is the average temperature corresponding to the steel foot area. Therefore, based on the average grayscale value of each steel foot area, the average temperature corresponding to each steel foot area is determined, and each steel foot area corresponds to each insulator. The average temperature corresponding to each steel foot area is determined as the characteristic temperature of each insulator.

[0157] Combine Figure 4 As shown in the figure, binarization of the target grayscale image clearly segments the steel foot area, reducing background interference and improving the accuracy of subsequent temperature measurements. This clear regional demarcation makes calculating the average temperature of each steel foot area more efficient. Furthermore, by comparing the average temperatures of different areas, steel feet with abnormal temperatures can be quickly identified, allowing timely maintenance measures to ensure equipment safety.

[0158] Step S105: constructing a gray-temperature relationship curve based on the maximum temperature in the target gray-scale image and the maximum gray-scale in the binary image.

[0159] In the specific implementation of step S105, a gray-temperature relationship curve is constructed by linear interpolation based on the maximum temperature and minimum temperature of all temperatures in the target grayscale image and the maximum grayscale value and minimum grayscale value in the binary image, specifically according to formula (8).

[0160]

[0161] Among them, T(α) represents the temperature corresponding to the gray value α; α min is the minimum gray value in the image; α max is the maximum gray value in the image; T min is the minimum temperature in the target grayscale image; T max is the maximum temperature in the target grayscale image.

[0162] It should be noted that formula (8) uses linear interpolation to calculate the grayscale value range (from α min to α max ) and the corresponding temperature range (from T min to T max ) to calculate the temperature value corresponding to each grayscale value, and then establish an approximate linear relationship between grayscale and temperature based on the temperature values ​​corresponding to all grayscale values.

[0163] As you can see, by establishing a relationship between grayscale value and temperature, quantitative temperature measurement can be achieved. This extends image processing beyond visual analysis to provide specific temperature data, improving detection accuracy and reducing errors. Furthermore, insulators made of different materials or coatings may have different grayscale-temperature characteristics. By establishing corresponding relationship curves, this approach can better adapt to diverse application scenarios. By establishing grayscale-temperature relationship curves, deteriorated insulators can be detected promptly, enhancing the scientific nature and effectiveness of detection and management.

[0164] Step S106: deriving an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the gray-temperature relationship curve.

[0165] In the specific implementation of step S106, an axial temperature data file for each insulator in the infrared image is derived based on the characteristic temperature and grayscale temperature relationship curve of each insulator, for example: Infrare(t,Si,Xj,Lk,Pm,In).dat. Here, t is the capture time; S is the line segment number, i is the sequence number; X is the side of the insulator string, with the power transmission side (small side) being 1 and the load side (large side) being 2, j is the sequence number; L is the circuit number, k is the sequence number; P is the phase number, m is the sequence number; I is the string number, n is the sequence number.

[0166] It's important to note that the insulator axial temperature data file derived from infrared images and grayscale temperature curves offers significant advantages. First, it provides accurate temperature assessments, helping to quickly identify deteriorating insulators. Furthermore, the resulting data file facilitates subsequent analysis, supporting long-term monitoring and evaluation of insulator performance, and providing a scientific basis for equipment management and maintenance decisions. This helps reduce the risk of major insulator failures and provides intuitive information to maintenance personnel through visual displays, further optimizing maintenance plans.

[0167] Step S107: Calculate the temperature difference between each insulator and its adjacent insulator in the insulator string to be inspected based on all axial temperature data files.

[0168] It is understood that the thermal characteristic of a deteriorated insulator is that the temperature difference between it and its adjacent insulators is significantly greater than the temperature difference between intact insulators. In addition, the number of deteriorated insulators in a line is also much smaller than the number of intact insulators.

[0169] Therefore, based on all axial temperature data files, the temperature difference between each insulator in the insulator string to be tested and its adjacent insulator is calculated, as shown in formula (9).

[0170]

[0171] Where p is the total number of insulators in the insulator string to be tested, i is the sequence number of each insulator in the insulator string to be tested, and i ranges from 1 to p from the high-voltage side to the ground side; T i is the characteristic temperature of the i-th insulator.

[0172] Step S108: determining a temperature difference threshold based on the temperature difference, and determining insulators with temperature differences greater than the temperature difference threshold as degraded insulators in the insulator string to be detected.

[0173] In the specific implementation of step S108, the mean and standard deviation of the temperature differences corresponding to all axial temperature data files are first calculated based on the temperature difference values. A temperature difference threshold is then determined based on the mean and standard deviation. Insulators with temperature differences greater than the temperature difference threshold are then determined to be degraded insulators in the insulator string to be inspected. The specific process for calculating the mean and standard deviation of the temperature differences corresponding to all axial temperature data files is shown in Formulas (10) and (11).

[0174]

[0175] Where ΔTav is the mean temperature difference of all axial temperature data files; ΔTstd is the standard deviation of all axial temperature data files; G is the number of insulator strings to be tested; T i is the temperature of the i-th insulator of the insulator string to be tested; P g is the total number of insulators in the gth string.

[0176] It should be noted that the method for determining the temperature difference threshold according to the temperature difference mean and standard deviation corresponding to all axial temperature data files is shown in formula (12).

[0177] ΔT thresh =ΔT av +3·ΔT std (12)

[0178] That is to say, the sum of the temperature difference mean and three times the standard deviation corresponding to all axial temperature data files is used as the temperature difference threshold.

[0179] It is understandable that the degraded insulator can be determined based on the temperature difference threshold. For example, the temperature difference threshold t is 1.2°C, and the data file Infrare (2024.05.27.18.24.30, S 110 The temperature difference of the 30th element in the .dat file is 2°C. This means that the 30th element in the first string of the second phase c on the large side of the 110th image, taken at 18:24:30 on 2024.05.27, may be a deteriorated insulator.

[0180] Step S109: estimating the contamination level of the deteriorated insulator according to the UV image corresponding to the deteriorated insulator and the ambient humidity, and generating and outputting a detection report of the deteriorated insulator.

[0181] Understandably, insulators with varying degrees of contamination experience varying temperature rises, particularly in humid conditions, where higher levels of contamination result in a more pronounced temperature rise. Therefore, it's crucial to simultaneously capture infrared images with UV images and measure ambient humidity. By analyzing the photon counts in the UV images and the ambient humidity data provided by the hygrometer, we can roughly assess the degree of contamination on deteriorating insulators. This allows us to categorize insulator strings by contamination level, minimizing the impact of temperature rise differences between insulators with varying degrees of contamination on the accuracy of test results.

[0182] In the specific implementation of step S109, the number of photons in the ultraviolet image corresponding to the degraded insulator is obtained, the contamination level of the degraded insulator is determined based on the number of photons and the ambient humidity, and a detection report of the degraded insulator is generated and output.

[0183] It should be noted that, as shown in Table 1, the levels of ambient humidity are divided into: dry, low humidity, medium humidity and high humidity.

[0184] Table 1

[0185] Ambient humidity <0.7 0.7～0.8 0.8～0.9 0.9～1.0 relative humidity dry Low humidity Medium wet High humidity

[0186] Table 2 shows the classification of the number of photons in ultraviolet images, where the number of photons is divided into five levels: I, II, III, IV, and V.

[0187] Table 2

[0188] Number of photons <100 100～1000 1000～10000 10000～100000 >100000 Photon magnitude Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ

[0189] Table 3 shows the classification of pollution levels, which are clean, mild, moderate, heavy, and severe.

[0190] Table 3

[0191]

[0192]

[0193] It should be noted that when generating a test report on deteriorated insulators, the required information includes but is not limited to the pollution level of the deteriorated insulators, its relationship with the temperature difference threshold, and the location information of the deterioration.

[0194] In specific applications, the contamination level of all insulators in the insulator string to be inspected can be classified based on all UV images and ambient humidity. Temperature difference thresholds are determined for all insulator strings with different contamination levels. Degraded insulators in each string are then detected based on these temperature difference thresholds. This helps eliminate temperature rise differences caused by contamination and improves detection accuracy.

[0195] In an embodiment of the present invention, infrared images are preprocessed, including grayscale conversion, noise reduction filtering, and angle correction. Given the high noise and low contrast characteristics of infrared images, median filtering is applied to remove high noise from the infrared image. The steel foot region is identified through image threshold segmentation. A grayscale-temperature relationship curve is constructed using the average temperature of the steel foot region as the characteristic temperature of the insulator. The temperature difference between adjacent insulators is used to determine whether an insulator is degraded, and a degraded insulator inspection report is generated based on the contamination level to notify maintenance personnel. The present invention captures infrared and ultraviolet images and establishes a corresponding relationship between contamination levels. The maximum inter-class variance method is used to effectively enhance the segmentation effect and improve the accuracy of image segmentation. Degraded insulators are identified based on statistical analysis, significantly improving detection accuracy.

[0196] The above-mentioned embodiment of the present invention Figure 1 A live non-contact insulator degradation detection method is shown in FIG. Figure 5 , shows a structural block diagram of a live non-contact insulator degradation detection device shown in an embodiment of the present invention, the detection device includes a receiving unit 501, a preprocessing unit 502, a first calculation unit 503, a first determination unit 504, a construction unit 505, a derivation unit 506, a second calculation unit 507, a second determination unit 508 and an output unit 509.

[0197] The receiving unit 501 is configured to receive a plurality of infrared images, a plurality of ultraviolet images and environmental humidity of the insulator string to be inspected sent by the inspection device.

[0198] The preprocessing unit 502 is used to preprocess each infrared image to obtain a target grayscale image.

[0199] The first calculation unit 503 is configured to calculate the inter-class variance corresponding to each threshold within the image grayscale threshold range based on all grayscale image pixels of the target grayscale image, and mark the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold.

[0200] The first determining unit 504 is configured to convert the target grayscale image into a binary image according to a target threshold value, so as to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determine the average temperature as the characteristic temperature of each insulator.

[0201] The constructing unit 505 is configured to construct a gray-temperature relationship curve based on the maximum temperature in the target gray-scale image and the maximum gray-scale in the binary image.

[0202] The exporting unit 506 is configured to export an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the gray-scale temperature relationship curve.

[0203] The second calculation unit 507 is configured to calculate the temperature difference between each insulator in the insulator string to be inspected and its adjacent insulators according to all axial temperature data files.

[0204] The second determining unit 508 is configured to determine a temperature difference threshold based on the temperature difference, and determine insulators with temperature differences greater than the temperature difference threshold as degraded insulators in the insulator string to be detected.

[0205] The second determining unit 508 is specifically configured to calculate the temperature difference mean and standard deviation corresponding to all axial temperature data files based on the temperature difference, and determine the temperature difference threshold according to the temperature difference mean and standard deviation.

[0206] The output unit 509 is configured to estimate the contamination level of the deteriorated insulator according to the UV image corresponding to the deteriorated insulator and the ambient humidity, and generate and output a detection report of the deteriorated insulator.

[0207] In an embodiment of the present invention, infrared images are preprocessed, including grayscale conversion, noise reduction filtering, and angle correction. Given the high noise and low contrast characteristics of infrared images, median filtering is applied to remove high noise from the infrared image. The steel foot region is identified through image threshold segmentation. A grayscale-temperature relationship curve is constructed using the average temperature of the steel foot region as the characteristic temperature of the insulator. The temperature difference between adjacent insulators is used to determine whether an insulator is degraded, and a degraded insulator inspection report is generated based on the contamination level to notify maintenance personnel. The present invention captures infrared and ultraviolet images and establishes a corresponding relationship between contamination levels. The maximum inter-class variance method is used to effectively enhance the segmentation effect and improve the accuracy of image segmentation. Degraded insulators are identified based on statistical analysis, significantly improving detection accuracy.

[0208] Combine Figure 5 The content shown is that the pre-processing unit 502 includes: an acquisition module, a noise reduction module and a correction module.

[0209] The acquisition module is used to obtain the RGB value of each pixel of each infrared image and convert the infrared image into a grayscale image based on the RGB value.

[0210] The denoising module is used to perform median filtering on the grayscale image to reduce noise and obtain a denoised grayscale image.

[0211] The correction module is used to perform tilt correction processing on the denoised grayscale image based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

[0212] Combine Figure 5 The content shown is that the first calculation unit 503 includes: a quantity acquisition module, a first calculation module, a second calculation module, a third calculation module, a fourth calculation module and a marking module.

[0213] The number acquisition module is used to obtain the number of all grayscale pixels of the target grayscale image.

[0214] The first calculation module is used to calculate the grayscale mean of the target grayscale image according to the quantity.

[0215] The second calculation module is used to calculate, for each threshold in the image grayscale threshold range, a first number of all grayscale image pixel points that are less than or equal to the threshold, and a second number of all grayscale image pixel points that are greater than the threshold.

[0216] The third calculation module is used to calculate the first between-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second between-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity.

[0217] The fourth calculation module is used to calculate the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range.

[0218] The marking module is used to extract the largest inter-class variance from all inter-class variances and mark the threshold corresponding to the largest inter-class variance as the target threshold.

[0219] Combine Figure 5 As shown in the content, the first determination unit 504 includes: a conversion module, a search module, an extraction module, a fifth calculation module and a determination module.

[0220] The conversion module is used to convert the target grayscale image into a binary image according to the target threshold.

[0221] The search module is used to search for the first n connected regions with the largest area in descending order in the binary image, and mark the connected regions as the steel foot regions of each insulator in the infrared image.

[0222] The extraction module is used to extract the image grayscale value corresponding to each steel foot area according to the target grayscale image.

[0223] The fifth calculation module is configured to calculate, for each steel foot region, an average grayscale value of the steel foot region based on the area of ​​the steel foot region and the image grayscale value, and determine the average temperature as the characteristic temperature of each insulator.

[0224] The determination module is used to determine the average temperature corresponding to each steel foot area based on the average grayscale value.

[0225] In summary, the present invention provides a method and device for detecting deterioration of live non-contact insulators. By taking infrared images and ultraviolet images and establishing corresponding relationships between pollution levels, the maximum inter-class variance method is used to effectively enhance the segmentation effect and improve the accuracy of image segmentation. Deteriorated insulators are identified based on statistical analysis, significantly improving detection accuracy.

[0226] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple. For relevant parts, refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0227] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0228] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting deterioration of an electrically charged non-contact insulator, characterized in that: The method comprises: receiving multiple infrared images, multiple ultraviolet images and environmental humidity of the insulator string to be inspected sent by the inspection equipment; For each of the infrared images, preprocessing the infrared image to obtain a target grayscale image; Based on all grayscale pixels of the target grayscale image, calculating the inter-class variance corresponding to each threshold within the image grayscale threshold range, and marking the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold; Converting the target grayscale image into a binary image according to the target threshold value to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determining the average temperature as the characteristic temperature of each insulator; Constructing a grayscale-temperature relationship curve based on the maximum temperature in the target grayscale image and the maximum grayscale in the binary image; Deriving an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the gray-scale temperature relationship curve; Calculating the temperature difference between each insulator and its adjacent insulator in the insulator string to be inspected based on all axial temperature data files; determining a temperature difference threshold based on the temperature difference, and determining the insulator having a temperature difference greater than the temperature difference threshold as a degraded insulator in the insulator string to be detected; The contamination level of the degraded insulator is estimated according to the ultraviolet image corresponding to the degraded insulator and the ambient humidity, and a detection report of the degraded insulator is generated and output.

2. The method according to claim 1, characterized in that The preprocessing of each infrared image to obtain a target grayscale image includes: For each infrared image, obtaining the RGB value of each pixel of the infrared image, and converting the infrared image into a grayscale image based on the RGB value; Performing median filtering on the grayscale image to reduce noise, thereby obtaining a grayscale image after noise reduction; The denoised grayscale image is subjected to tilt correction processing based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

3. The method according to claim 1, characterized in that The step of calculating the inter-class variance corresponding to each threshold value within the grayscale threshold range based on all grayscale image pixels of the target grayscale image, and marking the threshold value corresponding to the largest inter-class variance among all inter-class variances as the target threshold value, includes: Obtain the number of all grayscale pixels of the target grayscale image; Calculating the grayscale mean of the target grayscale image according to the quantity; For each threshold value in the image grayscale threshold range, calculate a first number of all grayscale image pixel points that are less than or equal to the threshold value, and a second number of all grayscale image pixel points that are greater than the threshold value; Calculating a first between-class variance corresponding to the first quantity, a first category weight corresponding to the first quantity, a second between-class variance corresponding to the second quantity, and a second category weight corresponding to the second quantity; Calculating the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range; The maximum inter-class variance is extracted from all inter-class variances, and the threshold corresponding to the maximum inter-class variance is marked as the target threshold.

4. The method according to claim 1, wherein The step of converting the target grayscale image into a binary image according to the target threshold value to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determining the average temperature as the characteristic temperature of each insulator includes: Converting the target grayscale image into a binary image according to the target threshold; Finding the first n connected regions in descending order of area in the binary image, and marking the connected regions as the steel foot regions of each insulator in the infrared image; Extracting the image grayscale value corresponding to each steel foot area according to the target grayscale image; For each of the steel foot regions, calculating an average grayscale value of the steel foot region based on the area of ​​the steel foot region and the grayscale value of the image; An average temperature corresponding to each of the steel leg regions is determined based on the average grayscale value, and the average temperature is determined as a characteristic temperature of each insulator.

5. The method according to claim 1, wherein The determining of the temperature difference threshold based on the temperature difference includes: The temperature difference mean and standard deviation corresponding to all axial temperature data files are calculated based on the temperature difference, and the temperature difference threshold is determined according to the temperature difference mean and the standard deviation.

6. A live non-contact insulator degradation detection device, characterized in that: The device comprises: A receiving unit, configured to receive multiple infrared images, multiple ultraviolet images, and ambient humidity of the insulator string to be inspected sent by the inspection equipment; a preprocessing unit, configured to preprocess each infrared image to obtain a target grayscale image; A first calculation unit is configured to calculate, based on all grayscale image pixels of the target grayscale image, the inter-class variance corresponding to each threshold within the image grayscale threshold range, and mark the threshold corresponding to the largest inter-class variance among all inter-class variances as the target threshold; a first determining unit, configured to convert the target grayscale image into a binary image according to the target threshold value, so as to determine the steel foot region of each insulator in the infrared image and the average temperature of each steel foot region, and determine the average temperature as the characteristic temperature of each insulator; A construction unit, configured to construct a grayscale-temperature relationship curve based on the maximum temperature in the target grayscale image and the maximum grayscale in the binary image; an exporting unit, configured to export an axial temperature data file of each insulator in the infrared image based on the characteristic temperature of each insulator and the grayscale-temperature relationship curve; A second calculation unit is used to calculate the temperature difference between each insulator in the insulator string to be tested and its adjacent insulators according to all axial temperature data files; a second determining unit, configured to determine a temperature difference threshold based on the temperature difference, and determine the insulator having a temperature difference greater than the temperature difference threshold as a degraded insulator in the insulator string to be detected; The output unit is configured to estimate the contamination level of the deteriorated insulator according to the ultraviolet image corresponding to the deteriorated insulator and the ambient humidity, and generate and output a detection report of the deteriorated insulator.

7. The device according to claim 6, characterized in that The pre-processing unit comprises: an acquisition module, configured to acquire, for each infrared image, an RGB value of each pixel of the infrared image, and convert the infrared image into a grayscale image based on the RGB value; A noise reduction module, configured to perform median filtering on the grayscale image to reduce noise, thereby obtaining a grayscale image after noise reduction; The correction module is used to perform tilt correction processing on the denoised grayscale image based on a preset edge detection operator and a Hough transform operator to obtain a target grayscale image.

8. The device according to claim 6, characterized in that The first computing unit includes: A quantity acquisition module is used to obtain the quantity of all grayscale pixels of the target grayscale image; A first calculation module, configured to calculate a grayscale mean of the target grayscale image according to the quantity; A second calculation module is configured to calculate, for each threshold value in the grayscale threshold range of the image, a first number of pixels in all grayscale images that are less than or equal to the threshold value, and a second number of pixels in all grayscale images that are greater than the threshold value; a third calculation module, configured to calculate a first between-class variance corresponding to the first quantity, a first category weight corresponding to the first quantity, a second between-class variance corresponding to the second quantity, and a second category weight corresponding to the second quantity; a fourth calculation module, configured to calculate the inter-class variance corresponding to the threshold based on the first inter-class variance corresponding to the first quantity, the first category weight corresponding to the first quantity, the second inter-class variance corresponding to the second quantity, and the second category weight corresponding to the second quantity, to obtain the inter-class variance corresponding to each threshold in the image grayscale threshold range; The marking module is used to extract the maximum inter-class variance from all inter-class variances, and mark the threshold corresponding to the maximum inter-class variance as the target threshold.

9. The device according to claim 6, characterized in that The first determining unit includes: A conversion module, configured to convert the target grayscale image into a binary image according to the target threshold; A search module is used to search for the first n connected regions in the binary image in descending order of area, and mark the connected regions as the steel foot regions of each insulator in the infrared image; An extraction module, configured to extract an image grayscale value corresponding to each steel foot region according to the target grayscale image; a fifth calculation module, configured to calculate, for each of the steel foot regions, an average grayscale value of the steel foot region based on the area of ​​the steel foot region and the grayscale value of the image; A determination module is configured to determine an average temperature corresponding to each of the steel leg regions based on the average grayscale value and determine the average temperature as a characteristic temperature of each insulator.

10. The device according to claim 6, characterized in that The second determining unit is specifically configured to calculate a temperature difference mean and a standard deviation corresponding to all axial temperature data files based on the temperature difference, and determine a temperature difference threshold according to the temperature difference mean and the standard deviation.

Images