A phase resolved partial discharge pattern processing method and related device

Abstract

The application provides a phase-resolved partial discharge pattern processing method and related device, and relates to the field of image processing, comprising: acquiring a binary pattern of a PRPD pattern; acquiring distribution statistical characteristics of each row of pixels in the vertical coordinate direction and distribution statistical characteristics of each column of pixels in the horizontal coordinate direction in the binary pattern, the distribution statistical characteristics of each row of pixels in the vertical coordinate direction being used to indicate the distribution of each row of pixels in the phase, and the distribution statistical characteristics of each column of pixels in the horizontal coordinate direction being used to indicate the amplitude range of each column of pixels; identifying noise pixels in the binary pattern according to preset distribution statistical characteristics of the noise pixels, the distribution statistical characteristics of each row of pixels in the vertical coordinate direction and the distribution statistical characteristics of each column of pixels in the horizontal coordinate direction; and identifying interference pixels in the binary pattern according to preset distribution statistical characteristics of the interference pixels, so as to avoid the problem of "dimension disaster" and improve the accuracy and robustness.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a phase-resolved partial discharge spectrum processing method and related apparatus. Background Technology

[0002] To ensure the safe operation of gas-insulated switchgear (GIS) equipment, partial discharge type identification is crucial. When partial discharge occurs in GIS equipment, multiple signals continuously generated by the GIS equipment are collected by ultra-high frequency sensors. By statistically analyzing all signals, a phase-resolved partial discharge (PRPD) spectrum can be obtained. A partial discharge type identification model extracts features from the PRPD spectrum and identifies the partial discharge type based on the features extracted from the PRPD spectrum.

[0003] However, UHF sensors not only collect partial discharge signals but also environmental noise and interference signals. The corresponding PRPD map contains three types of pixels: partial discharge pixels, noise pixels, and interference pixels. Local pixels are pixels that reflect the characteristics of partial discharge types. Therefore, the PRPD map input into the partial discharge type recognition model contains not only the partial discharge pixels required for identifying the partial discharge type but also redundant pixels such as noise pixels and interference pixels. These redundant pixels interfere with the partial discharge pixels. Therefore, the current partial discharge type recognition suffers from the problems of "curse of dimensionality" and low accuracy.

[0004] Furthermore, the PRPD maps used to train the partial discharge type identification model are basically obtained during laboratory measurements. These PRPD maps exhibit high repetition and redundant pixels without noise interference, resulting in poor robustness of the partial discharge type identification model. Summary of the Invention

[0005] In view of the above problems, this application provides a phase-resolved partial discharge spectrum processing method and related apparatus to avoid the "curse of dimensionality" and improve accuracy and robustness. The specific solution is as follows:

[0006] The first aspect of this application provides a method for processing phase-resolved partial discharge spectra, including:

[0007] Obtain the binarized spectrum of phase-resolved partial discharge (PRPD) pattern;

[0008] The statistical distribution features of each row of pixels in the vertical axis and the statistical distribution features of each column of pixels in the horizontal axis are obtained in the binarized map. The statistical distribution features of each row of pixels in the vertical axis are used to indicate the phase distribution of each row of pixels, and the statistical distribution features of each column of pixels in the horizontal axis are used to indicate the amplitude range of each column of pixels.

[0009] Based on the preset statistical characteristics of noise pixel distribution, the statistical characteristics of pixel distribution in the vertical direction of each row, and the statistical characteristics of pixel distribution in the horizontal direction of each column, noise pixels in the binarized map are identified.

[0010] Based on the preset statistical characteristics of the distribution of interfering pixels, the interfering pixels in the binarized map are identified.

[0011] In one possible implementation, the preset distribution statistics of the noise pixels include: the noise pixels are uniformly distributed on the horizontal axis and the amplitude range of the noise pixels is smaller than the amplitude range of the partial discharge pixels, wherein the partial discharge pixels are pixels in the spectrum that embody the characteristics of partial discharge.

[0012] The preset statistical characteristics of the distribution of interference pixels include: the sparsity of the interference pixels is greater than the sparsity of the partial discharge pixels, and the sparsity of the interference pixels and the sparsity of the partial discharge pixels are represented by density values.

[0013] In one possible implementation, obtaining the distribution statistics of each row of pixels in the vertical axis and the distribution statistics of each column of pixels in the horizontal axis of the binarized map includes:

[0014] The pixels in the binarized map are projected onto the vertical and horizontal coordinates respectively to obtain the first total projection of each row of pixels in the vertical direction and the second total projection of each column of pixels in the horizontal direction. The first total projection is the distribution statistical feature of each row of pixels in the vertical direction, and the second total projection is the distribution statistical feature of each column of pixels in the horizontal direction.

[0015] In one possible implementation, identifying noise pixels in the binarized map based on preset statistical characteristics of noise pixel distribution, statistical characteristics of pixel distribution in the vertical direction of each row, and statistical characteristics of pixel distribution in the horizontal direction of each column includes:

[0016] If the first total projection is equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection are suspected noise pixels. The fact that the first total projection is equal to the width of the binarized map indicates that the phase of all pixels in the row corresponding to the first total projection is uniformly distributed.

[0017] If the first total projection amount is not equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection amount are not noise pixels;

[0018] Obtain the position coordinates of each suspected noise pixel, the position coordinates including the position in the horizontal axis and the position in the vertical axis;

[0019] Find the minimum position among all positions along the vertical axis, where the minimum position is the position with the smallest value among all positions.

[0020] Based on the minimum position and the minimum value of the second projection total among all second projection totals, the noise pixel filtering conditions are determined, and the noise pixel filtering conditions are used to limit the amplitude range of the noise pixels;

[0021] If the position of the suspected noise pixel in the vertical axis direction satisfies the noise pixel screening condition, the suspected noise pixel is determined to be a noise pixel.

[0022] If the position of the suspected noise pixel in the vertical axis direction does not meet the noise pixel screening condition, the suspected noise pixel is determined not to be a noise pixel.

[0023] In one possible implementation, identifying the interfering pixels in the binarized map based on the preset distribution statistical characteristics of the interfering pixels includes:

[0024] Determine the neighborhood region of the pixel to be identified in the binarized map, wherein the pixel to be identified is the pixel in the binarized map excluding the noise pixel;

[0025] The density value of the pixel to be identified is calculated based on the gray values ​​of each pixel in the neighborhood region of the pixel to be identified.

[0026] If the density value of the pixel to be identified is less than or equal to a preset density threshold, the pixel to be identified is determined to be an interference pixel.

[0027] If the density value of the pixel to be identified is greater than the preset density threshold, the pixel to be identified is determined to be a partial discharge pixel. The partial discharge pixel is a pixel in the spectrum that reflects the characteristics of partial discharge. The preset density threshold is the minimum density value when the pixel to be identified is the partial discharge pixel.

[0028] In one possible implementation, obtaining the binarized spectrum of the phase-resolved partial discharge (PRPD) map includes:

[0029] The effective pixel region in the PRPD map is determined, and the size of the effective pixel region is normalized to obtain an effective pixel map with a target size;

[0030] Convert the effective pixel map into a grayscale image;

[0031] A grayscale threshold is determined, and the grayscale image is binarized using the grayscale threshold to obtain the binarized image.

[0032] A second aspect of this application provides a phase-resolved partial discharge spectrum processing apparatus, comprising:

[0033] The conversion unit is used to acquire the binarized spectrum of the phase-resolved partial discharge (PRPD) map;

[0034] The acquisition unit is used to acquire the distribution statistics of each row of pixels in the vertical axis direction and the distribution statistics of each column of pixels in the horizontal axis direction in the binarized map. The distribution statistics of each row of pixels in the vertical axis direction is used to indicate the phase distribution of each row of pixels, and the distribution statistics of each column of pixels in the horizontal axis direction is used to indicate the amplitude range of each column of pixels.

[0035] The identification unit is used to identify noise pixels in the binarized map based on preset statistical characteristics of noise pixel distribution, statistical characteristics of pixel distribution in the vertical direction of each row, and statistical characteristics of pixel distribution in the horizontal direction of each column, and to identify interference pixels in the binarized map based on preset statistical characteristics of interference pixels distribution.

[0036] A third aspect of this application provides a computer program product including computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the phase-resolved partial discharge pattern processing method of the first aspect or any implementation thereof.

[0037] A fourth aspect of this application provides an electronic device, including at least one processor and a memory connected to the processor, wherein:

[0038] The memory is used to store computer programs;

[0039] The processor is used to execute the computer program so that the electronic device can implement the phase-resolved partial discharge spectrum processing method of the first aspect or any implementation thereof.

[0040] The fifth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to perform the phase-resolved partial discharge pattern processing method described in the first aspect or any implementation thereof.

[0041] By employing the above technical solutions, the phase-resolved partial discharge (PRPD) image processing method and related apparatus provided in this application can identify noise pixels and interference pixels in the binarized PRPD image. Therefore, only PRPD pixels are input into the partial discharge type identification model, reducing the number and types of pixels input, thus avoiding the "curse of dimensionality" and improving accuracy. Furthermore, noise pixels are identified based on preset statistical distribution characteristics, the statistical distribution characteristics of pixels in each row of the image along the vertical axis, and the statistical distribution characteristics of pixels in each column of the image along the horizontal axis. Interference pixels are identified based on preset statistical distribution characteristics. Thus, without requiring any dataset training, the statistical distribution characteristics of noise pixels and interference pixels (which can be considered expert knowledge) are effectively utilized for identification, equivalent to unsupervised pixel recognition with high robustness. Furthermore, after processing by the phase-resolved partial discharge map processing method provided in this application embodiment, the difference between the PRPD map input into the partial discharge type identification model and the PRPD map used for training the partial discharge type identification model is small, reducing the production cost of the PRPD map used for training and improving the robustness of the partial discharge type identification model. Attached Figure Description

[0042] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0043] Figure 1 A flowchart of a phase-resolved partial discharge spectrum processing method provided in this application;

[0044] Figure 2 A grayscale image of the PRPD spectrum provided in this application;

[0045] Figure 3 for Figure 2 The binarized spectrum of the grayscale image shown;

[0046] Figure 4 for Figure 3 A schematic diagram of the first and second projection totals of the binarized map shown;

[0047] Figure 5 for Figure 2 The diagram shows a view of noise-removed pixels from a binarized image.

[0048] Figure 6 for Figure 5 The diagram shown illustrates the removal of interfering pixels from the image.

[0049] Figure 7 A schematic diagram of a phase-resolved partial discharge spectrum processing device provided in this application;

[0050] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0051] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0052] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0053] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, apparatus, product, or device that comprises a series of units is not necessarily limited to those units, but may include other units not explicitly listed or inherent to such processes, methods, apparatus, products, or devices.

[0054] Currently, partial discharge type identification models take PRPD (partial discharge patch) maps as input and output the probability of the identified partial discharge type, thus determining the corresponding partial discharge type based on the probability of the partial discharge type. However, the input PRPD maps include three types of pixels: partial discharge pixels, noise pixels, and interference pixels. Noise and interference pixels are redundant pixels in the partial discharge type identification process, and these redundant pixels can interfere with the partial discharge pixels, leading to the "curse of dimensionality" and low accuracy problems in partial discharge type identification. Furthermore, the PRPD maps used during the training of partial discharge type identification models are mainly obtained from laboratory measurements. These PRPD maps exhibit high repetition and lack redundant pixels such as noise interference, resulting in poor robustness of the partial discharge type identification models.

[0055] To address the aforementioned technical problems, this application provides a phase-resolved partial discharge (PRPD) image processing method and related apparatus. This method can identify noise and interference pixels in the binarized PRPD image, thus ensuring that only PRPD pixels are input into the partial discharge type identification model. This reduces the number and types of pixels input into the model, avoiding the "curse of dimensionality" and improving accuracy. Furthermore, noise pixels are identified based on preset statistical distribution characteristics, the statistical distribution characteristics of pixels in each row of the image along the vertical axis, and the statistical distribution characteristics of pixels in each column of the image along the horizontal axis. Interference pixels are identified based on preset statistical distribution characteristics. Therefore, without requiring any dataset training, this method effectively utilizes the statistical distribution characteristics of noise and interference pixels (which can be considered expert knowledge) for identification, essentially performing unsupervised pixel recognition with high robustness. Furthermore, after the PRPD map is processed by the phase-resolved partial discharge map processing method provided in this application embodiment, the difference between the map input into the partial discharge type identification model and the PRPD map used to train the partial discharge type identification model is small, which reduces the production cost of the PRPD map used for training and improves the robustness of the partial discharge type identification model.

[0056] The following description, in conjunction with the accompanying drawings, illustrates a phase-resolved partial discharge pattern processing method provided in an embodiment of this application. Please refer to the accompanying drawings. Figure 1 This illustrates an optional flow of a phase-resolved partial discharge spectrum processing method provided in an embodiment of this application, which may include the following steps:

[0057] S101. Obtain the binarized map of the PRPD map.

[0058] The PRPD atlas is a color image. It is first converted to a grayscale image, and then binarized to obtain a binarized atlas. Neither the binarized atlas nor the grayscale image changes the pixel type in the atlas. Therefore, the type of any pixel in the binarized atlas is the same as the type of its pixel in the PRPD atlas. For example, if a pixel in the binarized atlas is a partial discharge pixel, then that pixel will also be a partial discharge pixel in the PRPD atlas.

[0059] In this embodiment, the grayscale values ​​of background pixels in the PRPD map are the same, while the discharge frequencies corresponding to different partial discharge pixels are different. Therefore, the grayscale values ​​of partial discharge pixels are different from those of background pixels. Furthermore, noise pixels and interference pixels are generated based on redundant signals (such as environmental noise signals and interference signals) collected during the partial discharge process; therefore, the grayscale values ​​of noise pixels and interference pixels are also different from those of background pixels. Based on the principle that the grayscale values ​​of background pixels in the PRPD map are different from those of partial discharge pixels, noise pixels, and interference pixels, this embodiment provides a method of setting a grayscale threshold to convert the grayscale image of the PRPD map into a binary map. This divides the pixels in the binary map into two parts: background pixels and signal pixels. The signal pixels include three types of pixels: partial discharge pixels, noise pixels, and interference pixels, facilitating the identification of partial discharge pixels, noise pixels, and interference pixels from the binary map.

[0060] One feasible way to set a grayscale threshold is to obtain the grayscale values ​​of background pixels in a grayscale image, multiply the preset grayscale threshold coefficient by the grayscale values ​​of the background pixels to obtain the grayscale threshold. A background grayscale range is then generated based on the grayscale threshold. If the grayscale value of a pixel in the grayscale image is not within the background grayscale range, the pixel is determined to be a signal pixel, and its grayscale value is 1; if the grayscale value of a pixel in the grayscale image is within the background grayscale range, the pixel is determined to be a background pixel, and its grayscale value is 0. Because the grayscale values ​​of background pixels in a grayscale image have certain differences (i.e., the grayscale values ​​of all background pixels are not unique), this embodiment identifies background pixels by using the background grayscale range.

[0061] The grayscale value of the background pixel can be determined manually, and this embodiment does not limit this. For example, if the grayscale value range of the grayscale image is [0, 255], and the grayscale value G of the background pixel is... b The value is 125, the preset grayscale threshold α is 0.06, and the grayscale threshold G is... th =G b ×α=125×0.06=7.5, the background grayscale range can be [G b -G th G b +G th ].

[0062] In this embodiment, the grayscale image can be the effective pixel region in the PRPD map. Correspondingly, a feasible way to obtain the binarized map of the PRPD map is as follows:

[0063] The effective pixel region in the PRPD map is determined, and the size of the effective pixel region is normalized to obtain an effective pixel map with the target size. The effective pixel map is converted into a grayscale image. The grayscale threshold is determined, and the grayscale image is binarized using the grayscale threshold to obtain a binarized map.

[0064] The effective pixel region in a PRPD map can be the area including suspected signal pixels in the phase amplitude coordinate system of the PRPD map. After determining the effective pixel region, a scaling transformation is used to standardize the size of the effective pixel region to obtain an effective pixel map with a target size. The purpose is to standardize the effective pixel regions of different PRPD maps to the target size, such as an r×c size, where r is the width and c is the length. This is because PRPD maps may be obtained from GIS equipment from different manufacturers when partial discharge occurs. The pixel resolution and aspect ratio of PRPD maps drawn by different manufacturers are different. If the size of the effective pixel region is not standardized, even for the same type of partial discharge, the relative distribution of pixels in the PRPD map will change. Therefore, to reduce this variation, this embodiment standardizes the size of the effective pixel region so that the size of the effective pixel map of different PRPD maps is the target size.

[0065] The effective pixel spectrum is converted into a grayscale image using a preset grayscale conversion method, which can be any one of, but is not limited to, the average method, weighted method, single-channel method, maximum value method, and minimum value method. Then, a grayscale threshold is determined using the aforementioned method for setting the grayscale threshold. This grayscale threshold is then used to binarize the grayscale image to obtain a binarized spectrum. The size of the binarized spectrum is the target size.

[0066] In the binarized map, the i-th pixel can be represented as g(x) i ,y i ), where g(x) i ,y i The value of x is either 0 or 1. i Let y be the x-coordinate of the i-th pixel. i Let x be the ordinate of the i-th pixel. i The value of y is related to the length of the binarized map. i The value of x is related to the width of the binarized image. As mentioned above, if the length of the binarized image is c and the width is r, then x i ∈[1,c],y i ∈[1,r], so it can be seen that the final binarized map is a two-dimensional matrix with r rows (width of the map) and c columns (length of the map). The size of the matrix is ​​represented as r×c. The values ​​in the matrix are the gray values ​​of each pixel in the binarized map, which can be 0 or 1.

[0067] S102. Obtain the statistical distribution features of each row of pixels in the vertical axis and the statistical distribution features of each column of pixels in the horizontal axis in the binarized map. The statistical distribution features of each row of pixels in the vertical axis are used to indicate the phase distribution of each row of pixels, and the statistical distribution features of each column of pixels in the horizontal axis are used to indicate the amplitude range of each column of pixels.

[0068] In this embodiment, the statistical characteristics of the distribution of pixels in each row along the vertical axis and the statistical characteristics of the distribution of pixels in each column along the horizontal axis can be obtained by using the pixel values ​​g(x) in the binarized image. i ,y i ) to obtain. One feasible approach is:

[0069] The pixels in the binarized image are projected onto the horizontal and vertical axes respectively, yielding the first total projection of pixels in each row along the vertical axis and the second total projection of pixels in each column along the horizontal axis. The first total projection represents the statistical distribution of pixels in each row along the vertical axis, and the second total projection represents the statistical distribution of pixels in each column along the horizontal axis. The formulas for calculating the first and second total projections are as follows:

[0070]

[0071]

[0072] Where y_add(m) is the total first projection of the pixels in the m-th row along the vertical axis, and x_add(n) is the total second projection of the pixels in the n-th column along the horizontal axis. The total first projection y_add(m) determines the number of background pixels and signal pixels in the m-th row, thus representing the distribution of pixels in the m-th row. The total second projection of the pixels in the n-th column represents the amplitude range of the pixels in the n-th column.

[0073] S103. Based on the preset statistical characteristics of noise pixel distribution, the statistical characteristics of pixel distribution in the vertical direction of each row and the statistical characteristics of pixel distribution in the horizontal direction of each column, identify noise pixels in the binarized map.

[0074] Research revealed that the expert knowledge condition for noise pixels is that the noise distribution has no phase correlation. This manifests as the noise pixel distribution within the binarized image being uniformly distributed along the horizontal axis. In other words, if there are no background pixels in the m-th row of the binarized image, then the pixels in the m-th row are uniformly distributed along the horizontal axis. The horizontal axis corresponds to the phase of the image; the uniform distribution of pixels in the m-th row indicates that the pixels in the m-th row are uniformly distributed along the phase from 0 to 360°. If the first total projection y_add(m) is used as the statistical characteristic of the distribution of pixels in the m-th row along the vertical axis, then the expert knowledge condition could be: the first total projection y_add(m) is the same as the width of the binarized image.

[0075] However, besides noise pixels satisfying the aforementioned expert knowledge conditions, when the local discharge signal is strong or the local discharge frequency is high, the partial discharge pixels in the binarized spectrum can also satisfy the aforementioned expert knowledge conditions. Therefore, in order to distinguish between noise pixels and partial discharge pixels, other expert knowledge conditions need to be further provided. It is understood that noise pixels and partial discharge pixels have amplitude in addition to phase. Therefore, this embodiment can also distinguish noise pixels and partial discharge pixels based on amplitude.

[0076] Research has shown that the amplitude range of noise pixels is smaller than that of partial discharge pixels. Therefore, this embodiment identifies noise pixels by their amplitude range and whether they are uniformly distributed along the horizontal axis. Specifically, in this embodiment, the preset statistical distribution characteristics of noise pixels can be: noise pixels are uniformly distributed along the horizontal axis, and the amplitude range of noise pixels is smaller than that of partial discharge pixels. Correspondingly, based on these statistical distribution characteristics, the optional process for identifying noise pixels in a binarized image is as follows, including the following steps:

[0077] Step 1) If the first total projection is equal to the width of the binarized map, determine that all pixels under the row corresponding to the first total projection are suspected noise pixels. The first total projection being equal to the width of the binarized map indicates that the phase of all pixels under the row corresponding to the first total projection is uniformly distributed.

[0078] Step 2) If the first projection total is not equal to the width of the binarized map, determine that all pixels in the row corresponding to the first projection total are not noise pixels.

[0079]

[0080] Where N is the set of suspected noise pixels, if If all pixels in the m-th row are suspected noise pixels, then the suspected noise pixels are divided into a suspected noise pixel set. Then, the amplitude of the suspected noise pixels is used to determine whether the suspected noise pixels are noise pixels. The amplitude corresponds to the vertical coordinate, so the position of the suspected noise pixels on the vertical coordinate can be used to determine whether they are noise pixels.

[0081] Step 3) Obtain the position coordinates of each suspected noise pixel. The position coordinates include the position in the horizontal axis and the position in the vertical axis. In this embodiment, the position coordinates of the pixel with a phase of 0 and the smallest amplitude in the binarized map are defined as (0, 0), and the position coordinates of the pixel with a phase of 360° and the largest amplitude are defined as (c, r). Then, the positions of the suspected noise pixels in the suspected noise pixel set are marked within this range to mark the position coordinates of each suspected noise pixel. For example, the position coordinates of the i-th suspected noise pixel in the suspected noise pixel set can be represented as (x... i y i ).

[0082] Step 4) Obtain the minimum position from all positions along the vertical axis. The minimum position is the position with the smallest value among all positions.

[0083] Step 5) Determine the noise pixel filtering conditions based on the minimum position and the minimum value of the second projection total among all second projection totals. The noise pixel filtering conditions are used to limit the amplitude range of noise pixels.

[0084] Step 6) If the position of the suspected noise pixel in the vertical axis direction meets the noise pixel screening condition, the suspected noise pixel is determined to be a noise pixel.

[0085] Step 7) If the position of the suspected noise pixel in the vertical axis does not meet the noise pixel screening criteria, determine that the suspected noise pixel is not a noise pixel.

[0086] One possible way to filter noisy pixels is: ,in For the minimum position, This represents the minimum total amount of the second projection. Correspondingly, the method for filtering suspected noise pixels is as follows:

[0087]

[0088] The noise pixel set is defined by dividing all noise pixels into the noise pixel set using the method described above.

[0089] S104. Identify the interfering pixels in the binarized map based on the preset statistical characteristics of the distribution of interfering pixels.

[0090] Research has revealed that the partial discharge time is correlated throughout the entire voltage cycle. Therefore, partial discharge pixels exhibit clustering in the PRPD spectrum, while interfering pixels do not have phase correlation and are thus sparser in the PRPD spectrum. Consequently, this embodiment can identify interfering pixels based on their sparsity. The method for identifying interfering pixels is less complex, which can improve efficiency and reduce redundant pixels for partial discharge type identification.

[0091] Based on the above research findings, the preset statistical characteristics of the distribution of interference pixels can be: the sparsity of interference pixels is greater than that of partial discharge pixels, the sparsity of interference pixels and partial discharge pixels are represented by density values, and interference pixels in the binarized map are identified by density values.

[0092] One possible process is as follows: Determine the neighborhood region of the pixel to be identified in the binarized image. The pixel to be identified is the pixel in the binarized image excluding noise pixels. Calculate the density value of the pixel to be identified based on the grayscale values ​​of each pixel in the neighborhood region. If the density value of the pixel to be identified is less than or equal to a preset density threshold, the pixel to be identified is determined to be an interference pixel. If the density value of the pixel to be identified is greater than the preset density threshold, the pixel to be identified is determined to be a partial discharge pixel. The preset density threshold is the minimum density value when the pixel to be identified is a partial discharge pixel.

[0093] One definition of the neighborhood region of the pixel to be identified in the binarized map is: with radius r, the coordinate position (x, y) of the i-th pixel. i y i The area defined by the circle ) is R. i The region R i This is the neighborhood region, and the value of r can be, but is not limited to, 4. One formula for calculating the density value is as follows:

[0094]

[0095] ρ i Let ρ be the density value of the i-th pixel. i ρ th If ρ i> ρ th Then the i-th pixel is a partial discharge pixel. ρ th For a preset density threshold, such as ρ th =4. The pixel to be identified is g(x) in the binarized map. i ,y i The pixel with g(x) = 1 is identified because noise pixels in the binarized map were identified in step S103, and g(x) was identified in step S104. i ,y iIf the interfering pixels are those where )=1, then the remaining g(x) in the binarized map are... i ,y i Pixels with )=1 are partial discharge pixels. Thus, noise pixels and interference pixels in the PRPD map are identified through steps S103 and S104, reducing the number of pixels input to the partial discharge type identification model. Since most of the pixels input to the partial discharge type identification model are partial discharge pixels, the diversity of features in the PRPD map is reduced, thereby improving accuracy, precision and identification efficiency.

[0096] The phase-resolved partial discharge spectrum processing method described above is used to... Figure 2 The grayscale image of the PRPD spectrum shown is processed. Figure 2 The grayscale image shown has had its background pixels removed (the background pixel value is 0). Figure 2 The binarized spectrum of the grayscale image shown is as follows: Figure 3 As shown, for Figure 3 The pixels on the binarized map shown are projected, and the total first projection along the vertical axis and the total second projection along the horizontal axis are as follows: Figure 4 As shown. Then use Figure 4 The first and second projection totals in the image identify noisy pixels, and the binarized image after removing the noisy pixels is as follows: Figure 5 As shown. Further from Figure 5 The binarized map of the noise-removed pixels shows the identification of interfering pixels. The binarized maps of the noise-removed pixels and interfering pixels are shown below. Figure 6 As shown, the phase-resolved partial discharge map processing method described above can eliminate noise and interference pixels in the binarized map, thereby reducing redundant pixels in partial discharge type identification.

[0097] As can be seen from the above technical solution, the phase-resolved partial discharge (PRPD) map processing method provided in this application can identify noise pixels and interference pixels in the binarized PRPD map. Therefore, only PRPD pixels are input into the partial discharge type identification model, reducing the number and types of pixels input, thus avoiding the "curse of dimensionality" and improving accuracy. Furthermore, noise pixels are identified based on preset distribution statistics of noise pixels, the distribution statistics of pixels in each row of the map along the vertical axis, and the distribution statistics of pixels in each column of the map along the horizontal axis. Interference pixels are identified based on preset distribution statistics of interference pixels. Thus, without requiring any dataset training, the distribution statistics of noise pixels and interference pixels (which can be considered expert knowledge) are effectively utilized for identification, equivalent to unsupervised pixel identification with high robustness. Furthermore, after the PRPD map is processed by the phase-resolved partial discharge map processing method provided in this application embodiment, the difference between the map input into the partial discharge type identification model and the PRPD map used to train the partial discharge type identification model is small, which reduces the production cost of the PRPD map used for training and improves the robustness of the partial discharge type identification model.

[0098] The above describes a phase-resolved partial discharge pattern processing method provided by the embodiments of this application. The following describes the apparatus for performing the above-described phase-resolved partial discharge pattern processing method.

[0099] Please see Figure 7 , Figure 7 This is a schematic diagram of a phase-resolved partial discharge pattern processing device provided in an embodiment of this application. For detailed descriptions of each unit in this device, please refer to the above method embodiment; they will not be repeated here. Figure 7 As shown, the phase-resolved partial discharge spectrum processing device may include: a conversion unit 10, an acquisition unit 20, and an identification unit 30.

[0100] The conversion unit 10 is used to acquire a binarized image of the phase-resolved partial discharge (PRPD) map. One method for the conversion unit 10 to acquire the binarized image is as follows: determining the effective pixel region in the PRPD map, standardizing the size of the effective pixel region to obtain an effective pixel map with a target size; converting the effective pixel map into a grayscale image; determining a grayscale threshold, and using the grayscale threshold to perform binarization processing on the grayscale image to obtain the binarized image.

[0101] The acquisition unit 20 is used to acquire the distribution statistics of each row of pixels in the vertical axis direction and the distribution statistics of each column of pixels in the horizontal axis direction in the binarized map. The distribution statistics of each row of pixels in the vertical axis direction is used to indicate the phase distribution of each row of pixels, and the distribution statistics of each column of pixels in the horizontal axis direction is used to indicate the amplitude range of each column of pixels.

[0102] In one possible implementation, the acquisition unit 20 projects the pixels in the binarized map onto the vertical and horizontal coordinates respectively to obtain the first total projection of each row of pixels in the vertical direction and the second total projection of each column of pixels in the horizontal direction. The first total projection is the distribution statistical feature of each row of pixels in the vertical direction, and the second total projection is the distribution statistical feature of each column of pixels in the horizontal direction.

[0103] The identification unit 30 is used to identify noise pixels in the binarized map based on the preset distribution statistical characteristics of noise pixels, the distribution statistical characteristics of each row of pixels in the vertical direction and the distribution statistical characteristics of each column of pixels in the horizontal direction, and to identify interference pixels in the binarized map based on the preset distribution statistical characteristics of interference pixels.

[0104] The preset statistical characteristics of noise pixel distribution include: noise pixels are uniformly distributed on the horizontal axis and the amplitude range of noise pixels is smaller than that of partial discharge pixels. Partial discharge pixels are pixels in the spectrum that exhibit the characteristics of partial discharge. The preset statistical characteristics of interference pixel distribution include: interference pixels are more sparse than partial discharge pixels. The sparseness of interference pixels and partial discharge pixels are represented by density values.

[0105] Correspondingly, the process by which the recognition unit 30 identifies noisy pixels and interfering pixels is as follows:

[0106] If the first total projection is equal to the width of the binarized map, all pixels in the row corresponding to the first total projection are determined to be suspected noise pixels. The first total projection being equal to the width of the binarized map indicates that the phase of all pixels in the row corresponding to the first total projection is uniformly distributed. If the first total projection is not equal to the width of the binarized map, all pixels in the row corresponding to the first total projection are determined not to be noise pixels. The position coordinates of each suspected noise pixel are obtained, including the position in the horizontal axis and the position in the vertical axis. The minimum position is obtained from all positions in the vertical axis, which is the position with the smallest value among all positions. Based on the minimum position and the second total projection with the smallest value among all second total projections, the noise pixel screening conditions are determined. The noise pixel screening conditions are used to limit the amplitude range of noise pixels. If the position of the suspected noise pixel in the vertical axis meets the noise pixel screening conditions, the suspected noise pixel is determined to be a noise pixel. If the position of the suspected noise pixel in the vertical axis does not meet the noise pixel screening conditions, the suspected noise pixel is determined not to be a noise pixel.

[0107] The neighborhood region of the pixel to be identified in the binarized image is determined. The pixel to be identified is the pixel in the binarized image excluding noise pixels. The density value of the pixel to be identified is calculated based on the gray values ​​of each pixel in the neighborhood region. If the density value of the pixel to be identified is less than or equal to a preset density threshold, the pixel to be identified is determined to be an interference pixel. If the density value of the pixel to be identified is greater than the preset density threshold, the pixel to be identified is determined to be a partial discharge pixel. A partial discharge pixel is a pixel in the image that reflects the characteristics of partial discharge. The preset density threshold is the minimum density value when the pixel to be identified is a partial discharge pixel.

[0108] This application also provides an electronic device in its embodiments. (See reference...) Figure 8 The diagram illustrates a structural schematic suitable for implementing the electronic device in the embodiments of this application. The electronic device in the embodiments of this application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 8 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0109] like Figure 8As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 into a random access memory (RAM) 803. When the electronic device is powered on, the RAM 803 also stores various programs and data required for the operation of the electronic device. The processing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804. The ROM 802, RAM 803, and storage device 808 can serve as the memory of the electronic device, and the processing unit 801 can serve as the processor of the electronic device. The memory is used to store computer programs; the processor is used with the computer programs to enable the electronic device to implement any of the phase-resolved partial discharge pattern processing methods provided in the embodiments of this application.

[0110] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 808 including, for example, memory cards, hard drives, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0111] This application also provides a computer program product including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the phase-resolved partial discharge spectrum processing methods provided in this application.

[0112] This application also provides a computer-readable storage medium carrying one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the phase-resolved partial discharge pattern processing methods provided in this application.

[0113] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0114] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0115] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0116] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A method for processing phase-resolved partial discharge maps, characterized in that, include: Obtain the binarized spectrum of phase-resolved partial discharge (PRPD) pattern; The statistical distribution features of each row of pixels in the vertical axis and the statistical distribution features of each column of pixels in the horizontal axis are obtained in the binarized map. The statistical distribution features of each row of pixels in the vertical axis are used to indicate the phase distribution of each row of pixels, and the statistical distribution features of each column of pixels in the horizontal axis are used to indicate the amplitude range of each column of pixels. Based on the preset statistical characteristics of noise pixel distribution, the statistical characteristics of pixel distribution in the vertical direction of each row, and the statistical characteristics of pixel distribution in the horizontal direction of each column, noise pixels in the binarized map are identified. Based on the preset statistical characteristics of the distribution of interfering pixels, the interfering pixels in the binarized map are identified; The step of obtaining the distribution statistics of each row of pixels in the vertical axis and the distribution statistics of each column of pixels in the horizontal axis of the binarized map includes: projecting the pixels in the binarized map onto the vertical and horizontal axes respectively to obtain the first total projection of each row of pixels in the vertical axis and the second total projection of each column of pixels in the horizontal axis of the binarized map. The first total projection is the distribution statistics of each row of pixels in the vertical axis, and the second total projection is the distribution statistics of each column of pixels in the horizontal axis. The step of identifying noise pixels in the binarized map based on the preset distribution statistical characteristics of noise pixels, the distribution statistical characteristics of pixels in each row along the vertical axis, and the distribution statistical characteristics of pixels in each column along the horizontal axis includes: If the first total projection is equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection are suspected noise pixels. The fact that the first total projection is equal to the width of the binarized map indicates that the phase of all pixels in the row corresponding to the first total projection is uniformly distributed. If the first total projection amount is not equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection amount are not noise pixels; Obtain the position coordinates of each suspected noise pixel, the position coordinates including the position in the horizontal axis and the position in the vertical axis; Find the minimum position among all positions along the vertical axis, where the minimum position is the position with the smallest value among all positions. Based on the minimum position and the minimum value of the second projection total among all second projection totals, the noise pixel filtering conditions are determined, and the noise pixel filtering conditions are used to limit the amplitude range of the noise pixels; If the position of the suspected noise pixel in the vertical axis direction satisfies the noise pixel screening condition, the suspected noise pixel is determined to be a noise pixel. If the position of the suspected noise pixel in the vertical axis direction does not meet the noise pixel screening condition, the suspected noise pixel is determined not to be a noise pixel.

2. The method according to claim 1, characterized in that, The preset distribution statistical characteristics of the noise pixels include: the noise pixels are uniformly distributed on the horizontal axis and the amplitude range of the noise pixels is smaller than the amplitude range of the partial discharge pixels; the partial discharge pixels are pixels in the spectrum that reflect the characteristics of partial discharge. The preset statistical characteristics of the distribution of interference pixels include: the sparsity of the interference pixels is greater than the sparsity of the partial discharge pixels, and the sparsity of the interference pixels and the sparsity of the partial discharge pixels are represented by density values.

3. The method according to claim 1 or 2, characterized in that, The step of identifying interference pixels in the binarized map based on the preset distribution statistical characteristics of interference pixels includes: Determine the neighborhood region of the pixel to be identified in the binarized map, wherein the pixel to be identified is the pixel in the binarized map excluding the noise pixel; The density value of the pixel to be identified is calculated based on the gray values ​​of each pixel in the neighborhood region of the pixel to be identified. If the density value of the pixel to be identified is less than or equal to a preset density threshold, the pixel to be identified is determined to be an interference pixel. If the density value of the pixel to be identified is greater than the preset density threshold, the pixel to be identified is determined to be a partial discharge pixel. The partial discharge pixel is a pixel in the spectrum that reflects the characteristics of partial discharge. The preset density threshold is the minimum density value when the pixel to be identified is the partial discharge pixel.

4. The method according to claim 1 or 2, characterized in that, The process of obtaining the binarized spectrum of the phase-resolved partial discharge (PRPD) map includes: The effective pixel region in the PRPD map is determined, and the size of the effective pixel region is normalized to obtain an effective pixel map with a target size; Convert the effective pixel map into a grayscale image; A grayscale threshold is determined, and the grayscale image is binarized using the grayscale threshold to obtain the binarized image.

5. A phase-resolved partial discharge spectrum processing device, characterized in that, include: The conversion unit is used to acquire the binarized spectrum of the phase-resolved partial discharge (PRPD) map; The acquisition unit is used to acquire the distribution statistics of each row of pixels in the vertical axis direction and the distribution statistics of each column of pixels in the horizontal axis direction in the binarized map. The distribution statistics of each row of pixels in the vertical axis direction is used to indicate the phase distribution of each row of pixels, and the distribution statistics of each column of pixels in the horizontal axis direction is used to indicate the amplitude range of each column of pixels. The identification unit is used to identify noise pixels in the binarized map based on preset distribution statistical characteristics of noise pixels, distribution statistical characteristics of each row of pixels in the vertical direction and distribution statistical characteristics of each column of pixels in the horizontal direction, and to identify interference pixels in the binarized map based on preset distribution statistical characteristics of interference pixels. The acquisition unit acquires the distribution statistics of each row of pixels in the vertical axis and the distribution statistics of each column of pixels in the horizontal axis of the binarized map by: projecting the pixels in the binarized map onto the vertical and horizontal axes respectively to obtain the first total projection of each row of pixels in the vertical axis and the second total projection of each column of pixels in the horizontal axis of the binarized map. The first total projection is the distribution statistics of each row of pixels in the vertical axis and the second total projection is the distribution statistics of each column of pixels in the horizontal axis. The identification unit identifies noise pixels in the binarized map based on preset statistical characteristics of noise pixel distribution, statistical characteristics of pixel distribution in the vertical direction of each row, and statistical characteristics of pixel distribution in the horizontal direction of each column. If the first total projection is equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection are suspected noise pixels. The fact that the first total projection is equal to the width of the binarized map indicates that the phase of all pixels in the row corresponding to the first total projection is uniformly distributed. If the first total projection amount is not equal to the width of the binarized map, it is determined that all pixels in the row corresponding to the first total projection amount are not noise pixels; Obtain the position coordinates of each suspected noise pixel, the position coordinates including the position in the horizontal axis and the position in the vertical axis; Find the minimum position among all positions along the vertical axis, where the minimum position is the position with the smallest value among all positions. Based on the minimum position and the minimum value of the second projection total among all second projection totals, the noise pixel filtering conditions are determined, and the noise pixel filtering conditions are used to limit the amplitude range of the noise pixels; If the position of the suspected noise pixel in the vertical axis direction satisfies the noise pixel screening condition, the suspected noise pixel is determined to be a noise pixel. If the position of the suspected noise pixel in the vertical axis direction does not meet the noise pixel screening condition, the suspected noise pixel is determined not to be a noise pixel.

6. A computer program product, characterized in that, It includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the phase-resolved partial discharge pattern processing method as described in any one of claims 1 to 4.

7. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the electronic device to implement the phase-resolved partial discharge pattern processing method as described in any one of claims 1 to 4.

8. A computer storage medium, characterized in that, The storage medium carries one or more computer programs that, when executed by an electronic device, enable the electronic device to implement the phase-resolved partial discharge pattern processing method as described in any one of claims 1 to 4.

Images