10kV cable intermediate joint impurity type identification method

Abstract

The 10kV cable intermediate joint impurity type identification method provided by the invention comprises the following steps: radially transmitting a microwave signal with a frequency band of 22GHz-30GHz to a 10kV to-be-identified cable intermediate joint, and converting an obtained microwave reflection amplitude curve and a phase curve into a Smith circle image; constructing a normalized gray gradient co-occurrence matrix corresponding to the Smith circle image, and calculating a large gradient advantage feature, a gradient average feature and a mixed entropy feature according to the normalized gray gradient co-occurrence matrix; according to the large gradient advantage feature, the gradient average feature and the mixed entropy feature, calculating a type factor used for representing the type of the cable intermediate joint wrapping tape impurity; and determining the impurity type of the to-be-identified cable intermediate joint according to the type factor. Therefore, the types of the impurities in the intermediate joint can be quickly and accurately identified.

Description

Technical Field

[0001] The present application relates to the technical field of cable intermediate joint detection and evaluation, and in particular to a method for identifying impurity types in 10kV cable intermediate joints. Background Art

[0002] Cable intermediate joints play an important role in connecting cable segments and ensuring continuous power transmission in power systems. At the same time, they must have good electrical insulation performance and environmental adaptability to ensure long-term stable operation under various laying conditions. Their manufacturing and installation quality are directly related to the safety and reliability of system operation.

[0003] During actual construction, impurities often enter cable joints due to an unclean environment or improper operation. For example, contaminants such as dust and semi-conductive particles can easily adhere to the joint if not thoroughly removed from the construction site. Improper cable stripping can damage the insulation layer. Failure to wrap the insulating tape tightly can also create gaps. These impurities or defects can disrupt the uniformity of the electric field within the joint, causing electric field distortion and triggering partial discharge, which in turn accelerates the aging of the insulation material and reduces the joint's sealing performance. Over time, these hidden dangers can lead to insulation breakdown, causing cable failure, shortening service life, and in severe cases, even compromising the safe operation of the entire line. Therefore, being able to quickly and accurately identify the specific type of impurities within cable joints is crucial for improving the safety and stability of cable operation. Summary of the Invention

[0004] The purpose of this application is to solve at least one of the above-mentioned technical deficiencies, especially the technical deficiency in the prior art of how to quickly and accurately identify the specific type of impurities inside the cable intermediate joint.

[0005] In a first aspect, the present application provides a method for identifying impurity types in a 10kV cable intermediate joint, the method comprising:

[0006] The microwave reflection amplitude and phase curves of the 10kV cable to be identified are converted into Smith circle images by radially transmitting a microwave signal with a frequency range of 22GHz to 30GHz in the middle of the cable.

[0007] Construct the normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculate the large gradient advantage feature, gradient average feature and mixing entropy feature based on the normalized grayscale gradient co-occurrence matrix;

[0008] According to the large gradient advantage feature, gradient average feature and mixing entropy feature, the type factor used to represent the impurity type of the cable intermediate joint wrapping tape is calculated;

[0009] According to the type factor, the impurity type of the intermediate joint of the cable to be identified is determined.

[0010] In one embodiment, the step of constructing a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image includes:

[0011] Read the grayscale value of each pixel in the Smith circle image, and for each pixel, calculate the square root of the sum of the squares of the grayscale value differences between the pixel and the adjacent right pixel and the pixel below, obtain the gradient value corresponding to the pixel, and generate a gradient matrix;

[0012] After discretizing the gradient matrix, the initial grayscale gradient co-occurrence matrix is ​​constructed;

[0013] The initial grayscale gradient co-occurrence matrix is ​​normalized to obtain a normalized grayscale gradient co-occurrence matrix.

[0014] In one embodiment, the calculation expression of the large gradient advantage feature is:

[0015]

[0016] in, represents the large gradient advantage feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0017] In one embodiment, the calculation expression of the gradient average feature is:

[0018]

[0019] in, represents the gradient average feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0020] In one embodiment, the calculation expression of the mixed entropy feature is:

[0021]

[0022] in, represents the mixed entropy feature, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient, Represents a positive number used to avoid taking the logarithm of zero.

[0023] In one embodiment, the step of calculating a type factor for representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixing entropy feature includes:

[0024] The type factor is calculated as follows:

[0025]

[0026] in, represents the type factor, represents the large gradient advantage feature, represents the gradient average feature, Represents the mixing entropy feature.

[0027] In one embodiment, the step of determining the type of impurity in the cable intermediate joint to be identified based on the type factor includes:

[0028] If the type factor is greater than zero and not greater than 0.24, the impurity type of the cable intermediate joint to be identified is air;

[0029] If the type factor is greater than 0.24 and not greater than 0.63, the impurity type of the cable intermediate joint to be identified is sand and gravel;

[0030] If the type factor is greater than 0.63 and not greater than 1, the impurity type of the cable intermediate joint to be identified is metal residual particles.

[0031] In a second aspect, the present application provides a device for identifying impurity types in a 10kV cable intermediate joint, the device comprising:

[0032] The Smith circle image determination module is used to convert the obtained microwave reflection amplitude curve and phase curve of the radial microwave signal with a frequency range of 22GHz to 30GHz in the middle of the 10kV cable to be identified into a Smith circle image;

[0033] The feature calculation module is used to construct the normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculate the large gradient advantage feature, gradient average feature and mixed entropy feature based on the normalized grayscale gradient co-occurrence matrix;

[0034] A type factor calculation module is used to calculate a type factor representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixed entropy feature;

[0035] The impurity type determination module is used to determine the impurity type of the cable intermediate joint to be identified based on the type factor.

[0036] In a third aspect, the present application provides a storage medium: the storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the 10kV cable intermediate joint impurity type identification method as described in any one of the above embodiments.

[0037] In a fourth aspect, the present application provides a computer device, comprising: one or more processors, and a memory;

[0038] The memory stores computer-readable instructions, which, when executed by one or more processors, execute the steps of the method for identifying impurity types of 10kV cable intermediate joints in any one of the above embodiments.

[0039] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0040] The impurity type identification method for 10kV cable intermediate joints provided in the present application can realize the rapid and accurate identification of the impurity type inside the intermediate joint. Specifically, by radially emitting microwave signals in the frequency band of 22GHz~30GHz and constructing a Smith circle image, the changes in the electromagnetic characteristics inside the joint are effectively reflected; then, a normalized grayscale gradient co-occurrence matrix is ​​constructed based on the image, and large gradient advantage features, gradient average features and mixed entropy features are extracted, which helps to comprehensively characterize the microscopic interference characteristics caused by impurities on the internal structure of the joint; by calculating the type factor and determining the impurity type, the characteristic differences caused by different impurities can be distinguished, thereby providing support for subsequent fault warning and quality traceability, improving the intelligence and refinement level of intermediate joint defect detection, and enhancing the safety and reliability of cable line operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application 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 only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0042] Figure 1 A schematic flow chart of a method for identifying impurity types in a 10kV cable intermediate joint provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of the structure of a 10kV cable intermediate joint impurity type identification device provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0046] This application provides a method for identifying impurity types in 10kV cable intermediate joints. The following embodiments illustrate this method by applying it to a computer device. It is understood that the computer device can be any device with data processing capabilities, including but not limited to a single server, a server cluster, a personal laptop computer, a desktop computer, etc. Figure 1 As shown, the method may include the following steps:

[0047] S101: Radially transmit a microwave signal with a frequency range of 22 GHz to 30 GHz in the middle of a 10 kV cable to be identified, and convert the obtained microwave reflection amplitude and phase curves into Smith circle images.

[0048] Radially transmitting microwave signals refers to emitting microwave signals perpendicularly toward the cross-section of the cable's intermediate joint, ensuring that electromagnetic waves propagate radially within the cable's internal dielectric. A 10kV cable intermediate joint is a mid-section connector used to connect two sections of 10kV cross-linked polyethylene (XLPE) cable, providing electrical insulation and sealing protection. Microwave signals between 22GHz and 30GHz are high-frequency electromagnetic wave signals with a frequency range of 22GHz to 30GHz. They can effectively penetrate the cable's internal structure and respond sensitively to its dielectric properties. The microwave reflection amplitude curve and phase curve, respectively, represent the amplitude and phase changes of the microwave signal during reflection, reflecting the electromagnetic response characteristics of the cable joint's internal structure. The Smith circle diagram is a graphical representation that maps the complex reflection coefficient onto a two-dimensional circular diagram. It is used to analyze the impedance changes and matching of high-frequency signals in a dielectric medium.

[0049] Specifically, a microwave port is attached to the cable surface, and a vector network analyzer is used to transmit a continuous-wave microwave signal in the 22 GHz to 30 GHz frequency range to the 10 kV cable connector to be identified. To achieve radial transmission, the microwave signal is coupled to a rectangular waveguide port via a coaxial cable. A fixture ensures that the transmission port is perpendicularly aligned with the center of the cable connector surface, achieving a stable and repeatable radial incidence angle. The reflected echo is received simultaneously with the signal transmission, and the vector network analyzer collects the complex reflection coefficient data corresponding to each frequency point.

[0050] The computer then normalizes the collected data according to the frequency sequence and decomposes it into amplitude response curves and phase response curves. The amplitude curve describes the reflection intensity characteristics at different frequencies, while the phase curve depicts the phase shift of the electromagnetic wave along the signal propagation path. This process is automatically completed by the data preprocessing module to ensure the stability and consistency of the input data.

[0051] Furthermore, a computer maps the complex reflection coefficient data (including amplitude and phase) into a Smith circle image. To do this, each complex reflection coefficient is projected onto the Smith circle in the complex plane, resulting in a complete and continuous two-dimensional image. This image uses different colors or grayscale values ​​to indicate impedance trends, visually reflecting the impact of dielectric changes within the cable connector on microwave reflection characteristics.

[0052] It can be understood that by radially emitting microwaves at the cable's intermediate joint within the 22GHz to 30GHz frequency range, the electromagnetic response characteristics of the cable's internal materials can be fully stimulated. The resulting microwave reflection amplitude and phase curves truly reflect the distribution of the internal dielectric structure. Further converting these curves into Smith circle images can map the original one-dimensional signal curves into a two-dimensional image representation with spatial characteristics, allowing subsequent image processing algorithms to utilize the spatial information in the image structure to mine subtle differences in impurities or defects. Therefore, this step not only improves the ability to characterize cable joint impurities, but also enhances the diversity and discriminability of data input, providing the data foundation and image support for high-precision defect identification, thereby improving detection efficiency and recognition accuracy.

[0053] S102: constructing a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculating a large gradient advantage feature, a gradient average feature, and a mixed entropy feature based on the normalized grayscale gradient co-occurrence matrix.

[0054] The grayscale gradient co-occurrence matrix is ​​a two-dimensional matrix constructed based on the spatial relationship between pixel grayscale gradients in an image. It is used to quantify the structural distribution characteristics of image texture. The normalized grayscale gradient co-occurrence matrix is ​​a matrix obtained by normalizing the original grayscale gradient co-occurrence matrix based on the total frequency. It reflects the proportion of each gradient pair in the entire image. The large gradient dominance feature is a numerical indicator that measures the density of strong gradients in an image. The gradient average feature represents the average level of grayscale variation in the entire image. The mixed entropy feature is used to measure the uncertainty and complexity of information in image texture.

[0055] Specifically, after generating the Smith circle image, the computer converts the image into a grayscale image and calculates the grayscale gradient values ​​in the horizontal and vertical directions for each pixel in the image. To improve the stability of the image gradient features, the system uses the Sobel operator to perform edge gradient enhancement processing, obtaining horizontal and vertical gradient matrices, and then calculates the gradient amplitude of each pixel based on these matrices.

[0056] Next, the computer constructs a grayscale gradient co-occurrence matrix based on the gradient magnitude matrix. This process involves setting discrete gradient level intervals, mapping the gradient values ​​of all pixels to the preset gradient level intervals, and counting the frequency of different gradient value combinations that appear within a given direction and distance range, thereby forming a two-dimensional gradient co-occurrence relationship matrix. Here, adjacency relationships with angles of 0° and 90° can be used to fully reflect the horizontal and vertical texture changes in the image.

[0057] The computer then normalizes the grayscale gradient co-occurrence matrix by dividing each element by the total number of gradient combinations, so that each entry represents the relative probability distribution of its occurrence in the image. This normalization reduces interference caused by differences in gradient amplitude distribution between different images.

[0058] Furthermore, the computer equipment extracts large gradient dominance features, gradient average features, and mixed entropy features based on the normalized grayscale gradient co-occurrence matrix. The large gradient dominance feature is calculated by accumulating the probability values ​​of high gradient regions and performing a weighted calculation, reflecting the presence of local strong reflective areas in the image; the gradient average feature is obtained by taking the weighted average of the matrix and reflects the stability of the overall reflective field texture distribution; the mixed entropy feature is used to measure the texture complexity and randomness in the image by traversing all non-zero elements in the normalized matrix and performing a logarithmic calculation on their probability values ​​before taking the weighted sum.

[0059] It can be understood that by constructing a normalized grayscale gradient co-occurrence matrix based on the Smith circle image and further extracting large gradient dominance features, gradient average features, and mixed entropy features, it is possible to effectively quantify the texture change patterns caused by impurities in the image, converting the originally subjective image features into quantitative indicators that can be used for identification modeling. The large gradient dominance feature can identify areas of electric field disturbance caused by local strong impurities, the gradient average feature helps determine the uniformity of the overall insulation structure, and the mixed entropy feature enhances the ability to perceive differences in texture complexity caused by different types of impurities. Therefore, the fine-grained analysis capability and robustness of image feature extraction are improved, providing a reliable numerical basis for accurately distinguishing the types of impurities within cable joints, thereby significantly enhancing recognition accuracy.

[0060] S103: Calculating a type factor for representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixing entropy feature.

[0061] Among them, the type factor is a composite numerical index calculated by combining multiple image features, which is used to characterize the specific type or morphological characteristics of impurities in the cable intermediate joint wrapping tape in the feature space.

[0062] Specifically, the computer receives and loads the large gradient advantage feature, the gradient average feature, and the mixed entropy feature. At this stage, the three types of features are preferably stored uniformly in a set of feature vectors and normalized to eliminate differences in numerical dimensions between the different features, ensuring numerical consistency and physical interpretation consistency in subsequent combined calculations.

[0063] Next, a weighted fusion of the three image features can be performed. Fusion methods can be based on principal component analysis, weighted linear combination, support vector embedding, and other methods. In one specific embodiment, large gradient advantage features can be given a higher weight to highlight local perturbations caused by impurity edges. Meanwhile, the gradient average feature and the mixed entropy feature are assigned different degrees of secondary weights to construct a weighted formula.

[0064] Furthermore, the computer equipment can perform normalization mapping processing on the initially generated type factors to make their distribution more concentrated and improve the clarity of the classification boundaries. At the same time, multiple type factors can be mapped to a multidimensional feature space to construct a feature cluster model, and a similarity comparison can be performed with a sample library corresponding to known impurity types to form a clear feature-type label mapping relationship. In order to improve the calculation accuracy, fuzzy clustering or distance measurement algorithms, such as K-Means or Mahalanobis distance, can also be introduced to enhance the discriminative ability of type factors, so that they are not only used for numerical representation, but also have the potential for classification judgment.

[0065] It can be understood that by calculating the type factor through a weighted fusion of the large gradient advantage feature, the gradient average feature, and the mixed entropy feature, it is possible to effectively integrate the local strong disturbance caused by impurities in the image, the overall uniformity change, and the texture complexity information, forming a numerical expression with high discrimination of the impurity type. This type factor not only enhances the clustering separability of each impurity type in the feature space, but also provides a stable and quantifiable input basis for subsequent classification and recognition algorithms, thereby significantly improving the accuracy of impurity type recognition.

[0066] S104: Determine the impurity type of the intermediate joint of the cable to be identified based on the type factor.

[0067] Among them, impurity types refer to foreign matter or defects of different properties that may be mixed in the wrapping tape structure of the cable intermediate joint, including dust, semi-conductive particles, oil stains, water vapor residue or loose wrapping, which have different dielectric properties and image texture manifestations.

[0068] Specifically, the computer receives and analyzes the calculated type factors and can input them into a pre-built impurity identification model, which can be a supervised classifier such as a support vector machine (SVM), decision tree, random forest, or deep neural network model. During the training phase, this identification model has learned the mapping relationship between different type factors and impurity types based on a large number of labeled samples, and is capable of quickly classifying new input factors.

[0069] The recognition model then projects the input type factor into a multidimensional feature space and calculates its Euclidean or Mahalanobis distance from the known sample center, or, in the case of a neural network, calculates its classification probability after activation. A confidence threshold can be set. When the recognition confidence exceeds the preset threshold, the specific impurity type label is output; otherwise, it is marked as "uncertain" or "requires manual review," ensuring a balance between recognition accuracy and system stability.

[0070] Furthermore, the computer can compare and verify the identification results with typical impurity samples in the database and simultaneously display the identified impurity type and its corresponding image feature information on a graphical user interface, allowing engineers to quickly determine whether on-site rectification or repeated testing is necessary. In a preferred embodiment, real-time optimization of the recognition model based on an incremental learning mechanism is also supported. That is, after the recognition result is manually confirmed, the confirmed type factor and its corresponding label can be fed back into the model to fine-tune the model parameters, thereby achieving self-learning and continuous evolution of the impurity type recognition system.

[0071] It can be understood that by determining the impurity type of the cable intermediate joint according to the type factor, the complex image feature information can be converted into a structured and classifiable specific impurity label, thereby achieving clear identification of the impurity properties.

[0072] In the above embodiment, it is possible to quickly and accurately identify the type of impurities inside the intermediate joint. Specifically, by radially emitting microwave signals in the frequency band of 22GHz~30GHz and constructing a Smith circle image, the changes in the electromagnetic characteristics inside the joint are effectively reflected; then, a normalized grayscale gradient co-occurrence matrix is ​​constructed based on the image, and large gradient advantage features, gradient average features and mixed entropy features are extracted, which helps to comprehensively characterize the microscopic interference characteristics caused by impurities on the internal structure of the joint; by calculating the type factor and determining the impurity type, the characteristic differences caused by different impurities can be distinguished, thereby providing support for subsequent fault warning and quality traceability, improving the intelligence and refinement level of intermediate joint defect detection, and enhancing the safety and reliability of cable line operation.

[0073] In one embodiment, the step of constructing a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image includes:

[0074] Read the grayscale value of each pixel in the Smith circle image, and for each pixel, calculate the square root of the sum of the squares of the grayscale value differences between the pixel and the adjacent right pixel and the pixel below, obtain the gradient value corresponding to the pixel, and generate a gradient matrix;

[0075] After discretizing the gradient matrix, the initial grayscale gradient co-occurrence matrix is ​​constructed;

[0076] The initial grayscale gradient co-occurrence matrix is ​​normalized to obtain a normalized grayscale gradient co-occurrence matrix.

[0077] The grayscale value refers to the brightness intensity represented by each pixel in the image, typically an integer between 0 and 255, representing the brightness or darkness of the pixel. The gradient value refers to the change in grayscale value between a pixel and its adjacent pixels in the image, and is used to reflect local texture changes in the image. The gradient matrix is ​​a two-dimensional numerical matrix in which each element represents the grayscale gradient intensity of the corresponding pixel in the image. The initial grayscale gradient co-occurrence matrix is ​​a two-dimensional matrix constructed by counting the frequency of adjacent occurrences of different gradient values ​​in the image. It is used to capture the directionality and repetitive patterns of image texture.

[0078] Specifically, the computer device loads the Smith circle image data to be processed from memory or cache and calls the image reading module to convert the image into a two-dimensional grayscale matrix. Each element in this matrix corresponds to the grayscale value of a pixel in the image. In this embodiment, 8-bit grayscale encoding is preferably used for image representation to ensure sufficient image detail and facilitate subsequent computational processing.

[0079] Next, perform the gradient calculation operation on each pixel in the grayscale image. For each non-edge pixel in the image, read its current grayscale value and the grayscale values ​​of the pixel to its right and the pixel below it, and calculate the square root of the sum of the squares of the grayscale differences between the current pixel and its two adjacent pixels, that is:

[0080]

[0081] in, Indicates location The gray value at is the corresponding gradient value. This formula can effectively reflect the intensity of the grayscale change direction of the current pixel, forming a gradient matrix with the same size as the original image.

[0082] The computer then discretizes the gradient matrix. Gradient values ​​are divided into several levels (e.g., 015, 1631, etc.) based on preset interval boundaries. Each gradient value is replaced with the level number of the interval it falls into, generating a discrete gradient matrix. This step facilitates the subsequent construction of a statistically significant symbiotic relationship matrix and avoids statistical sparseness caused by overly detailed numerical differences.

[0083] On this basis, we scan all adjacent pixel pairs in the discrete gradient matrix, such as in the horizontal and vertical directions, and count the number of occurrences of the gradient level combination for each pair of adjacent pixels to construct an initial grayscale gradient co-occurrence matrix. Each row and column of this matrix corresponds to a gradient level, and the element value in the matrix is ​​the co-occurrence count of that level combination.

[0084] Finally, the computer normalizes the initial grayscale gradient co-occurrence matrix. This process divides each element by the total number of occurrences, so that the sum of all elements in the matrix is ​​1. This results in a normalized grayscale gradient co-occurrence matrix. Each element of the normalized matrix can be considered the relative probability distribution of a specific gradient combination in the entire image, facilitating comparability in subsequent feature calculations.

[0085] In one example, each pixel coordinate on the Smith chart The formula Calculated, where is the modulus of the microwave reflection coefficient, is the phase in radians, and are the horizontal and vertical coordinates of the image, respectively. After reading the Smith circle image, for each pixel, the square root of the sum of the squares of the grayscale value differences between it and its adjacent pixels to the right and below is calculated to obtain the corresponding gradient value, forming a gradient matrix. After discretizing this gradient matrix, a grayscale gradient co-occurrence matrix is ​​constructed, and this co-occurrence matrix is ​​normalized to obtain the relative distribution information of different gradient combinations in the image.

[0086] In this embodiment, by reading the grayscale value of each pixel in the Smith circle image and calculating the grayscale difference with adjacent pixels to form a gradient matrix, it is possible to accurately extract the local grayscale change information of the image, reflecting the dielectric disturbance traces caused by tiny impurities inside the cable intermediate joint; discretizing the gradient matrix and constructing an initial grayscale gradient co-occurrence matrix helps capture texture directionality and structural repeatability; the normalized grayscale gradient co-occurrence matrix obtained after further normalization not only unifies the scale differences between different images, but also provides a stable probabilistic basis for subsequent feature extraction and type recognition. Therefore, it can effectively convert image texture structure information into a quantifiable statistical model, improve the accuracy of subsequent classification judgment and the robustness of the model, and enhance the reliability of impurity identification under complex working conditions.

[0087] In one embodiment, the calculation expression of the large gradient advantage feature is:

[0088]

[0089] in, represents the large gradient advantage feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0090] Specifically, Represents the first in the normalized gray gradient co-occurrence matrix Gray level and The matrix element value corresponding to the gradient level, that is, the grayscale is , the gradient is The probability of symbiosis; A numerical value indicating the current gradient level; Indicates in The cumulative occurrence probability of all gray levels under the gradient level.

[0091] This formula multiplies the probability of co-occurrence at each gradient level by the corresponding gradient level value and then performs a weighted summation of all gradient levels to produce a numerical indicator that reflects the dominance of high-gradient regions in the image. A larger value indicates a more pronounced high-gradient texture feature in the image, potentially indicating prominent impurity edges or structural discontinuities.

[0092] Therefore, this calculation method can effectively extract the gradient enhancement area features caused by impurities in the Smith circle image of the cable intermediate joint, thereby enhancing the differences between different impurity types in the image texture space, improving the perception ability of the impurity classification and recognition model for abnormal disturbance areas, and ultimately helping to improve the accuracy of impurity recognition and the stability of the system.

[0093] In one embodiment, the calculation expression of the gradient average feature is:

[0094]

[0095] in, represents the gradient average feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0096] It can be understood that this formula reflects the balance of the grayscale gradient distribution across the entire image by calculating the weighted average of all grayscale levels. A high gradient average feature value indicates a dramatic change in the overall grayscale of the image, possibly with large areas of impurity disturbance or complex texture areas. Conversely, a low gradient average indicates a relatively smooth image with weak grayscale texture changes. Therefore, this calculation method can fully reflect the impact of impurities on the overall texture distribution of the Smith circle image of the cable intermediate joint, assisting in the identification of impurity types that lack distinct edges but exhibit statistically significant disturbance characteristics. This improves the sensitivity of impurity detection and the integrity of image feature expression, enhancing the system's recognition capabilities in low-contrast scenarios.

[0097] In one embodiment, the calculation expression of the mixed entropy feature is:

[0098]

[0099] in, represents the mixed entropy characteristic, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient, Represents a positive number used to avoid taking the logarithm of zero.

[0100] As can be understood, this formula calculates the complexity of the image texture distribution by calculating entropy based on the cumulative value of the co-occurrence probability of all gradient levels at each grayscale level. Larger values ​​indicate a more dispersed and uncertain gradient distribution corresponding to each grayscale level in the image, a higher information entropy, and a more complex texture structure. Smaller values ​​indicate a more concentrated texture distribution and a simpler structure. Therefore, using this mixed entropy feature as an auxiliary feature helps characterize the variations in texture complexity caused by different impurities in cable intermediate joint images. By quantitatively reflecting the degree of order in the information distribution within the image, the impurity recognition model can further enhance its ability to detect abnormal features with blurred edges and non-centralized distribution, thereby improving overall recognition robustness and adaptability to complex impurity interference scenarios.

[0101] In one embodiment, the step of calculating a type factor for representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixing entropy feature includes:

[0102] The type factor is calculated as follows:

[0103]

[0104] in, represents the type factor, represents the large gradient advantage feature, represents the gradient average feature, Represents the mixing entropy feature.

[0105] It can be understood that this calculation method fuses multiple texture statistical features in a linear weighted manner, while retaining the multi-dimensional information of the image, and extracts a single scalar indicator that can be used for classification and discrimination. The positive weight coefficient in the formula is used to amplify the response to the target feature, and the negative weight coefficient is used to suppress the interference of features with low discrimination, thereby achieving the coordinated optimization of feature dimensionality reduction and discrimination ability. Therefore, through the calculation of this type of factor, not only can the quantitative expression of the degree of influence of impurities in the image of the cable intermediate joint be achieved, but also the discrimination between different impurity types can be effectively improved, which is conducive to the subsequent accurate classification. Compared with the traditional method that relies on a single image indicator, this method has stronger robustness and adaptability on the basis of taking into account the overall gradient distribution, texture complexity and strong edge features of the image, and can significantly improve the system's recognition accuracy of different types of impurities, enhancing the practicality and reliability of the intelligent cable quality detection system.

[0106] In one embodiment, the step of determining the type of impurity in the cable intermediate joint to be identified based on the type factor includes:

[0107] If the type factor is greater than zero and not greater than 0.24, the impurity type of the cable intermediate joint to be identified is air;

[0108] If the type factor is greater than 0.24 and not greater than 0.63, the impurity type of the cable intermediate joint to be identified is sand and gravel;

[0109] If the type factor is greater than 0.63 and not greater than 1, the impurity type of the cable intermediate joint to be identified is metal residual particles.

[0110] Specifically, the computer device obtains the type factor value calculated above and, before performing recognition judgment, calls the preset type recognition rules to perform interval matching judgment on the factor value. When the type factor is greater than 0 and no greater than 0.24, the recognition result is marked as "air-type impurities," corresponding to defects such as loose wrapping and gas mixing; when the type factor is greater than 0.24 and no greater than 0.63, the recognition result is marked as "gravel-type impurities," usually caused by external sand and dust pollution; when the type factor is greater than 0.63 and no greater than 1, the recognition result is marked as "metal particle-type impurities," which may come from wire stripping tool residue, construction debris, etc. Each classification threshold can be obtained through statistical modeling or supervised learning of a large amount of labeled sample data. The computer device compares the type factor with the interval rules one by one, and automatically outputs the corresponding impurity type result after matching the qualified interval.

[0111] It can be understood that by establishing a one-to-one correspondence between the value range of the type factor and the impurity type, the classification and identification of impurities inside the cable intermediate joint can be completed quickly and clearly after the feature calculation is completed, thereby improving the recognition efficiency and the intelligence level of the system. The use of the numerical interval judgment method avoids subjective interpretation or repeated processing of the original image, and enhances the consistency and reproducibility of the recognition results. In addition, the factor judgment method driven by image features has good scalability and generalization capabilities, and can adapt to changes in impurity types in different batches of cables and different construction scenarios, ultimately improving the accuracy, stability and engineering practical value of the quality assessment of cable intermediate joints.

[0112] The following describes the 10kV cable intermediate joint impurity type identification device provided by the embodiment of the present application. The 10kV cable intermediate joint impurity type identification device described below and the 10kV cable intermediate joint impurity type identification method described above can be used for reference. Figure 2 As shown, the present application provides a 10kV cable intermediate joint impurity type identification device, the device comprising:

[0113] The Smith circle image determination module 201 is used to convert the obtained microwave reflection amplitude curve and phase curve of the radial microwave signal in the middle of the 10kV cable to be identified, which has a frequency range of 22GHz to 30GHz, into a Smith circle image;

[0114] The feature calculation module 202 is used to construct a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculate the large gradient advantage feature, the gradient average feature and the mixed entropy feature based on the normalized grayscale gradient co-occurrence matrix;

[0115] A type factor calculation module 203 is used to calculate a type factor representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixed entropy feature;

[0116] The impurity type determination module 204 is configured to determine the impurity type of the cable intermediate joint to be identified based on the type factor.

[0117] In one embodiment, the feature calculation module includes:

[0118] A gradient matrix generation unit is used to read the grayscale value of each pixel in the Smith circle image and, for each pixel, calculate the square root of the sum of the squares of the grayscale value differences between the pixel and the adjacent right pixel and the pixel below it to obtain the gradient value corresponding to the pixel and generate a gradient matrix;

[0119] An initial grayscale gradient co-occurrence matrix construction unit is used to construct an initial grayscale gradient co-occurrence matrix after discretizing the gradient matrix;

[0120] The normalized grayscale gradient co-occurrence matrix determining unit is used to normalize the initial grayscale gradient co-occurrence matrix to obtain a normalized grayscale gradient co-occurrence matrix.

[0121] In one embodiment, the calculation expression of the large gradient advantage feature is:

[0122]

[0123] in, represents the large gradient advantage feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0124] In one embodiment, the calculation expression of the gradient average feature is:

[0125]

[0126] in, represents the gradient average feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

[0127] In one embodiment, the calculation expression of the mixed entropy feature is:

[0128]

[0129] in, represents the mixed entropy feature, Represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient, Represents a positive number used to avoid taking the logarithm of zero.

[0130] In one embodiment, the type factor calculation module includes:

[0131] The type factor calculation unit is used to calculate the type factor according to the following expression:

[0132]

[0133] in, represents the type factor, represents the large gradient advantage feature, represents the gradient average feature, Represents the mixing entropy feature.

[0134] In one embodiment, the impurity type determination module includes:

[0135] a first impurity type determination unit, configured to determine that the impurity type of the cable intermediate joint to be identified is air if the type factor is greater than zero and not greater than 0.24;

[0136] a second impurity type determination unit, configured to determine that the impurity type of the cable intermediate joint to be identified is sand and gravel if the type factor is greater than 0.24 and not greater than 0.63;

[0137] The third impurity type determination unit is configured to determine that the impurity type of the cable intermediate joint to be identified is residual metal particles if the type factor is greater than 0.63 and not greater than 1.

[0138] In one embodiment, the present application also provides a storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the 10kV cable intermediate joint impurity type identification method as described in any of the above embodiments.

[0139] In one embodiment, the present application also provides a computer device, which stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the 10kV cable intermediate joint impurity type identification method as described in any of the above embodiments.

[0140] Schematically, as Figure 3 As shown, Figure 3 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 can be provided as a server. Figure 3 Computer device 300 includes a processing component 302, which further includes one or more processors and memory resources represented by memory 301 for storing instructions executable by processing component 302, such as application programs. The application programs stored in memory 301 may include one or more modules, each corresponding to a set of instructions. Furthermore, processing component 302 is configured to execute the instructions to perform the 10 kV cable intermediate joint impurity type identification method according to any of the above-described embodiments.

[0141] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or the like.

[0142] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0143] Finally, it should be noted that, in this article, relational terms such as first and second are merely used to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprise," "include," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. Without further restriction, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element. Herein, "one," "said," "the," and "its" may also include plural forms unless the context clearly indicates otherwise. A plurality refers to at least two, such as 2, 3, 5, or 8. "And / or" includes any and all combinations of the relevant listed items.

[0144] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referenced to each other.

[0145] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. 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 spirit or scope of the present application. Therefore, the present application 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 identifying impurity types in 10kV cable intermediate joints, characterized in that: The method comprises: The microwave reflection amplitude and phase curves of the 10kV cable to be identified are converted into Smith circle images by radially transmitting a microwave signal with a frequency range of 22GHz to 30GHz in the middle of the cable. Constructing a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculating a large gradient advantage feature, a gradient average feature, and a mixed entropy feature based on the normalized grayscale gradient co-occurrence matrix; Calculating a type factor for representing the impurity type of the cable intermediate joint wrapping tape according to the large gradient advantage feature, the gradient average feature, and the mixing entropy feature; The impurity type of the to-be-identified cable intermediate joint is determined according to the type factor.

2. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The step of constructing a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image includes: Read the grayscale value of each pixel in the Smith circle image, and for each pixel, calculate the square root of the sum of the squares of the grayscale value differences between the pixel and the adjacent right pixel and the pixel below it to obtain the gradient value corresponding to the pixel and generate a gradient matrix; After discretizing the gradient matrix, an initial grayscale gradient co-occurrence matrix is ​​constructed; The initial grayscale gradient co-occurrence matrix is ​​normalized to obtain the normalized grayscale gradient co-occurrence matrix.

3. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The calculation expression of the large gradient advantage feature is: in, represents the large gradient advantage feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

4. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The calculation expression of the gradient average feature is: in, represents the gradient average feature, represents the sum of all elements in the normalized grayscale gradient co-occurrence matrix, represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient.

5. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The calculation expression of the mixed entropy feature is: in, represents the mixing entropy characteristic, represents each element in the normalized grayscale gradient co-occurrence matrix, Indicates the Grayscale, Indicates the Level gradient, Represents a positive number used to avoid taking the logarithm of zero.

6. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The step of calculating a type factor for representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixing entropy feature comprises: The type factor is calculated according to the following expression: in, represents the type factor, represents the large gradient advantage feature, represents the gradient average feature, represents the mixing entropy characteristic.

7. The method for identifying impurity types in 10kV cable intermediate joints according to claim 1, characterized in that: The step of determining the impurity type of the cable intermediate joint to be identified based on the type factor includes: If the type factor is greater than zero and not greater than 0.24, the impurity type of the cable intermediate joint to be identified is air; If the type factor is greater than 0.24 and not greater than 0.63, the impurity type of the cable intermediate joint to be identified is gravel; If the type factor is greater than 0.63 and not greater than 1, the impurity type of the cable intermediate joint to be identified is residual metal particles.

8. A 10kV cable intermediate joint impurity type identification device, characterized in that: The device comprises: The Smith circle image determination module is used to convert the obtained microwave reflection amplitude curve and phase curve of the radial microwave signal with a frequency range of 22GHz to 30GHz in the middle of the 10kV cable to be identified into a Smith circle image; A feature calculation module is used to construct a normalized grayscale gradient co-occurrence matrix corresponding to the Smith circle image, and calculate a large gradient advantage feature, a gradient average feature and a mixed entropy feature based on the normalized grayscale gradient co-occurrence matrix; A type factor calculation module, configured to calculate a type factor representing the impurity type of the cable intermediate joint wrapping tape based on the large gradient advantage feature, the gradient average feature, and the mixing entropy feature; The impurity type determination module is used to determine the impurity type of the cable intermediate joint to be identified according to the type factor.

9. A storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed by one or more processors, enable the one or more processors to perform the steps of the method for identifying impurity types in a 10kV cable intermediate joint as described in any one of claims 1 to 7.

10. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, execute the steps of the method for identifying impurity types in a 10 kV cable intermediate joint as claimed in any one of claims 1 to 7.

Images