Cable intermediate joint defect detection method, system, equipment and medium

Abstract

The invention discloses a cable intermediate joint defect detection method, system and device, and a medium. The method comprises the following steps: acquiring an original image of a cable intermediate joint; performing feature extraction on the original image to obtain a first feature, and segmenting the original image based on the first feature to obtain a segmented image; calculating a pixel point value of the segmented image, and obtaining a second feature; and inputting the second feature into a recognition model for analysis and judgment, and recognizing and acquiring defect information of the cable intermediate joint. According to the method, the second feature is obtained by performing image segmentation, pixel value calculation and other processing on the original image of the cable intermediate joint, and the defect is identified by utilizing structural similarity comparison and comprehensive score based on the second feature, so that the defect detection precision and efficiency of the cable intermediate joint can be effectively improved, and internal defects can be found in time; and safe and stable operation of a power system is ensured.

Description

Technical Field

[0001] This invention relates to the field of power equipment testing technology, and in particular to a method, equipment and medium for detecting defects in cable intermediate joints. Background Technology

[0002] Power supply is a critical infrastructure ensuring the normal operation of production and daily life. Power cables, as an important carrier of power transmission, are widely used in various fields such as urban underground power grids, power plant transmission lines, internal power supply systems of industrial and mining enterprises, and underwater power transmission projects across rivers and seas. With the continuous growth of electricity demand, the proportion of cables in power lines is constantly increasing, and the voltage levels used are also getting higher and higher, from common low voltage to ultra-high voltage of 1500KV and above. Cable joints, as key components connecting various cable sections and ensuring the continuous and stable operation of power lines, directly affect the safe and reliable operation of the entire power system. Once a cable joint fails, it is highly likely to cause a large-scale power outage, resulting in huge economic losses to society. Therefore, accurately detecting the condition of cable joints and promptly identifying potential defects is crucial.

[0003] Currently, cable joint inspection technology faces numerous challenges. Traditional methods such as manual visual inspection, infrared thermography, and ultrasonic testing have significant limitations. Manual visual inspection relies entirely on the experience and eyesight of the inspectors, making it difficult to detect hidden defects within the cable. Infrared thermography primarily relies on temperature anomalies and is ineffective for early defects that have not yet caused noticeable temperature changes. Ultrasonic testing is easily affected by interference when dealing with complex cable joint structures, making it difficult to guarantee accuracy. Furthermore, the ultra-low frequency dielectric loss factor method is insufficient for detecting weak defects, the partial discharge method is difficult to locate in complex electrical environments, and the time-domain reflectometry method is ineffective for detecting internal defects such as insulation aging or moisture. These shortcomings of existing technologies create numerous safety hazards in power system operation and fail to meet the ever-increasing demands for power reliability. There is an urgent need to develop new inspection technologies to improve the accuracy and efficiency of cable joint inspection and ensure the safe and stable operation of the power system. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for detecting defects in cable intermediate joints, which solves the problems of low detection accuracy, low efficiency, and inability to detect internal defects in a timely manner in existing cable intermediate joint detection methods.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for detecting defects in cable joints, comprising: Obtain the original image of the cable mid-joint; Feature extraction is performed on the original image to obtain a first feature. Based on the first feature, the original image is segmented to obtain a segmented image. Calculate the pixel values ​​of the segmented image to obtain the second feature; The second feature is input into the recognition model for analysis and judgment to identify and obtain cable joint defect information.

[0007] As a preferred embodiment of the cable joint defect detection method of the present invention, the method includes: extracting features from the original image to obtain a first feature, including: The original image is converted to grayscale to obtain the first feature; The grayscale image is filtered using a weighted smoothing template to remove noise from the original image. The elements in the original image are reorganized proportionally to obtain the first image; Based on the first feature, the first image is segmented to obtain a segmented image.

[0008] The beneficial effects of this preferred technical solution are as follows: By converting a color image to a single-channel grayscale image through grayscale processing, the data dimensionality is reduced while preserving the main structural information of the image, thus lowering the computational complexity of subsequent processing. The weighted smoothing filter employs an adaptive weight strategy, effectively suppressing Gaussian noise and salt-and-pepper noise while better preserving image edge and detail information, avoiding the edge blurring problem caused by traditional mean filtering.

[0009] As a preferred embodiment of the cable joint defect detection method of the present invention, the method includes: calculating the pixel values ​​of the segmented image to obtain the second feature, including: The brightness and color deviation of the segmented image are preprocessed; The preprocessed segmented image is corrected using a gamma correction factor, and the gradient magnitude and gradient direction of the segmented image pixels are calculated to obtain the second feature.

[0010] The beneficial effects of this preferred technical solution are as follows: Brightness and color deviation preprocessing, through adaptive histogram equalization, enhances the local contrast of the image, making defect features more prominent. The gamma correction factor is automatically adjusted according to the image's grayscale distribution, performing non-linear enhancement on low-contrast areas, effectively solving the common problem of uneven brightness in X-ray images. Gradient magnitude and direction calculations can accurately extract the edge contour features of defects, providing rich geometric features for subsequent identification.

[0011] As a preferred embodiment of the cable joint defect detection method of the present invention, the identification model includes: The recognition model is trained using samples of defect-free cable joints and samples of defective cable joints to identify and obtain the features of defect-free cable joints and the features of defective cable joints. Calculate the cross-entropy loss function, focus loss function, and structural similarity loss function of the recognition model, and optimize the parameters of the recognition model.

[0012] As a preferred embodiment of the cable joint defect detection method of the present invention, the method involves: inputting the second feature into an identification model for analysis and judgment to identify and obtain cable joint defect information, including: The second feature is input into the recognition model, and the structural similarity image comparison method is used to compare the second feature with the features of defective cable joints and defective cable joints to obtain the contrast. Based on the brightness, color, structure, and contrast of the second feature, a comprehensive score is obtained to identify and obtain cable intermediate joint defect information; Based on the defect information, a defect report is generated and sent to the staff.

[0013] The beneficial effects of this preferred technical solution are as follows: the structural similarity image comparison method not only considers the differences in pixel values, but also evaluates the structural information, brightness, and contrast of the image, enabling it to more accurately capture the feature differences of minute defects. The comprehensive scoring mechanism adopts a multi-feature fusion strategy, effectively improving the recognition accuracy of various defect types such as cracks, bubbles, and insulation aging by weighted calculation of the contribution of different features.

[0014] In a preferred embodiment of the cable joint defect detection method of the present invention, acquiring the original image of the cable joint includes: Install and deploy X-ray transmitters and X-ray detectors at the cable mid-joint locations; Set and calibrate the appropriate emission power, start the X-ray emitter to allow the X-ray rays to penetrate the cable intermediate joint, and allow the X-ray detector to receive the rays and convert them into electrical signals for transmission to the image processing system; The image processing system processes the electrical signal to obtain the original image of the cable mid-joint.

[0015] The beneficial effects of this preferred technical solution are as follows: Secondly, the present invention provides a cable intermediate joint defect detection system, comprising: a data acquisition module for acquiring the original image of the cable intermediate joint; The feature extraction module is used to extract features from the original image to obtain a first feature; to segment the original image based on the first feature to obtain a segmented image; and to calculate the pixel values ​​of the segmented image to obtain a second feature. The analysis and identification module is used to input the second feature into the identification model for analysis and judgment, and to identify and obtain cable intermediate joint defect information.

[0016] Thirdly, the present invention provides an electronic device, comprising: A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of a method for detecting defects in cable joints.

[0017] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the cable joint defect detection method.

[0018] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention acquires the original image of the cable intermediate joint, and through a series of processes such as grayscale conversion, filtering, image segmentation, and pixel value calculation, obtains a second feature containing rich defect features. Then, using a recognition model trained with defect-free and defective samples and optimized by multiple loss functions, the defect is identified based on structural similarity comparison and comprehensive scoring. Finally, a report is generated and sent to the staff, which can effectively improve the accuracy and efficiency of cable intermediate joint defect detection, timely detect internal defects, and ensure the safe and stable operation of the power system. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the overall process of a cable joint defect detection method according to an embodiment of the present invention. Detailed Implementation

[0021] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0022] Example 1, referring to Figure 1As an embodiment of the present invention, a method for detecting defects in cable intermediate joints is provided, comprising: S100: acquiring an original image of the cable intermediate joint; S102: Extract features from the original image to obtain the first feature, and segment the original image based on the first feature to obtain the segmented image; S104: Calculate the pixel values ​​of the segmented image to obtain the second feature; S106: Input the second feature into the recognition model for analysis and judgment, and identify and obtain cable intermediate joint defect information.

[0023] It should be noted that in power systems, the condition of cable joints affects the safe and stable operation of the power system, making timely and accurate defect detection extremely crucial. This invention first acquires the original image, then extracts and segments its features, simplifying complex images, highlighting key parts, and enhancing the identification of potential defect features, making defect features clearer. Segmentation based on the first feature separates different regions, facilitating targeted analysis. The pixel values ​​of the segmented image are calculated to obtain a second feature, further mining image details and providing rich data for defect identification. The second feature is input into a recognition model trained on a large number of samples and with optimized parameters. Using a structural similarity image comparison method and a comprehensive scoring mechanism, defects can be accurately identified. This invention improves the accuracy and efficiency of cable joint defect detection, enabling timely detection of minor or hidden defects, allowing staff to take preventative measures to avoid power outages caused by cable joint defects, and ensuring the safe and stable operation of the power system.

[0024] Example 2, refer to Figure 1 As an embodiment of the present invention, a method for detecting defects in cable intermediate joints is provided based on the above embodiment.

[0025] In this embodiment of the invention, obtaining the original image of the cable intermediate joint in step S100 above further includes the following sub-steps A1-A2: A1: Install and deploy an X-ray transmitter and an X-ray detector at the cable mid-joint location; set and calibrate the appropriate transmission power, start the X-ray transmitter to allow the X-ray rays to penetrate the cable mid-joint, and allow the X-ray detector to receive the rays and convert them into electrical signals for transmission to the image processing system. A2: The image processing system processes electrical signals to produce the original image of the cable mid-joint.

[0026] Specifically, when installing the X-ray emitter and detector, it is crucial to ensure precise alignment between their relative positions to guarantee that the X-rays can accurately penetrate the critical inspection area of ​​the cable joint. The emission power calibration requires fine-tuning based on the material and thickness of the cable joint, as well as surrounding environmental interference, to avoid overexposure and loss of detail due to excessive power, or blurry image information due to insufficient power. The image processing system first performs preliminary amplification and filtering of the electrical signal to remove noise interference introduced during transmission, and then converts it into a digital image format suitable for subsequent analysis.

[0027] In one alternative implementation, raw images of the cable joint can be acquired using an industrial CT scanner. Specifically, the industrial CT scanner utilizes computed tomography (CT) technology to perform multi-angle X-ray scans around the cable joint. During the scan, the equipment moves precisely at specific rotation angles and step distances to acquire projection data from multiple angles. This projection data is then reconstructed using complex algorithms to generate two-dimensional images of different sections of the cable joint, which are then stitched together into a three-dimensional image using software.

[0028] In another alternative implementation, a portable digital X-ray imaging device can be used to acquire the raw image of the cable joint. Specifically, this portable digital X-ray imaging device integrates a miniaturized X-ray source and a digital detector. In use, the device is placed close to the cable joint, and X-rays are emitted after selecting appropriate exposure parameters. The detector quickly captures the rays passing through the cable joint and directly converts them into digital signals. These signals are then processed in real time by the built-in image processing chip, and the raw image of the cable joint is quickly displayed on the device's screen.

[0029] It should be noted that the precise installation and deployment of the X-ray emitter and detector at the cable joint location, along with the calibration of the emission power, ensured that the X-rays could effectively penetrate the joint, and that the detector could stably receive the X-rays and convert them into electrical signals. Setting an appropriate emission power allowed the X-rays to fully penetrate the internal structure of the cable joint while avoiding overexposure or underexposure, thus ensuring the integrity of the image details.

[0030] It should also be noted that the image processing system converts the electrical signals transmitted by the detector into raw images that are easy to analyze, removing noise and interference generated during signal transmission and enhancing image clarity and contrast. The resulting high-quality raw images provide a precise data foundation for subsequent steps such as feature extraction, image segmentation, and defect identification, thereby improving the accuracy and reliability of defect detection.

[0031] In this embodiment of the invention, step S102, which involves extracting features from the original image to obtain a first feature and segmenting the original image based on the first feature to obtain a segmented image, further includes sub-steps B1-B3: B1: Convert the original image to grayscale to obtain the first feature; B2: Use a weighted smoothing template to filter the grayscale image and remove noise from the original image.

[0032] B3: Reorganize the elements in the original image proportionally to obtain the first image; segment the first image based on the first feature to obtain the segmented image.

[0033] In an embodiment of the present invention, the first feature includes the grayscale, color, and geometric attributes of the original image; In one alternative implementation, the first feature may further include frequency features, texture features, and statistical features of the image.

[0034] In this embodiment of the invention, since the original image contains a large amount of complex information, direct processing can easily lead to information redundancy and analysis difficulties. The elements are reorganized proportionally, that is, the image pixels or image blocks are adjusted according to certain rules based on the image's grayscale, color, and geometric attributes, which are considered primary features.

[0035] For example, based on grayscale features, pixels with similar grayscale values ​​are grouped into similar regions; based on color features, parts with similar colors are grouped together; and based on geometric properties, elements with similar shapes and textures are merged. After this reorganization, the image structure becomes clearer, different feature regions are highlighted, and the originally scattered and messy information becomes orderly, laying a good foundation for subsequent segmentation of the first image based on the first feature and accurate acquisition of the segmented image.

[0036] In this embodiment of the invention, pixels near the target defect area and those at the image edge of a cable joint have higher value for accurately identifying defects in the image. Therefore, a weighted smoothing template is used to assign relatively large weights to these pixels; while pixels far from the target area and with less impact on defect identification are assigned smaller weights. When the weighted smoothing template slides across the grayscale image, it calculates the grayscale value of each pixel according to these weights and performs a weighted average of the grayscale values ​​of adjacent pixels, thereby effectively suppressing randomly distributed noise. Compared to ordinary mean filtering, the weighted smoothing template can better preserve the image's detailed information, avoiding blurring or loss of defect features due to over-filtering, and providing a clearer and more accurate image data foundation for subsequent defect identification based on image features.

[0037] In one optional implementation, wavelet transform can be used to obtain the first feature and perform denoising. Specifically, the original image is decomposed into wavelet coefficients for different frequency sub-bands. In the low-frequency sub-band, the main image contour information is retained as part of the first feature; in the high-frequency sub-band, an appropriate threshold is set to threshold the wavelet coefficients. Wavelet coefficients smaller than the threshold are considered to be caused by noise and are set to zero, while wavelet coefficients larger than the threshold are retained. Then, wavelet reconstruction is performed to obtain the denoised image.

[0038] In another alternative implementation, principal component analysis (PCA) can be used to obtain the first feature and perform denoising. Specifically, the pixel values ​​of the original image are arranged into a data matrix by rows or columns. PCA is then performed on this data matrix to calculate the eigenvalues ​​and eigenvectors of its covariance matrix. Based on the magnitude of the eigenvalues, the principal eigenvectors are selected. These eigenvectors correspond to principal components that contain the main information of the image and constitute the first feature.

[0039] It should be noted that this invention converts the original color image into a grayscale image through grayscale processing, removing interference from color information and simplifying complex color data into a single grayscale value, thus reducing the complexity of subsequent processing while retaining the key structural information of the image used for defect detection. Using weighted smoothing template filtering, noise in the image can be removed more effectively. The original image elements are reorganized proportionally, and based on primary features such as grayscale, color, and geometric attributes, the scattered and disordered elements in the image are integrated into a clearer first image, highlighting different feature regions. The resulting segmented image can more accurately represent the different parts of the cable joint, making subsequent defect identification and analysis more efficient and accurate, greatly improving the reliability and accuracy of cable joint defect detection, and ensuring the stable operation of the power system.

[0040] In this embodiment of the invention, the calculation of pixel values ​​in the segmented image and the acquisition of the second feature in step S104 are as follows: The brightness and color deviations of the segmented image are normalized; the normalized segmented image is corrected by a gamma correction factor, and the gradient magnitude and gradient direction of the pixels in the segmented image are calculated to obtain the second feature.

[0041] In this embodiment of the invention, the vertical and horizontal coordinate distribution of the input cable joint segmentation image is assumed to be as follows: and Input cable mid-joint segmentation image The gamma correction factor is The color-corrected image Represented as: In an embodiment of the present invention, the pixels of the segmented image of the cable intermediate joint are defined. The horizontal and vertical gradient values ​​are and The gradient magnitude of the cable joint image points Represented as: In an embodiment of the present invention, the gradient direction of the segmented image of the cable joint is... for: In an optional implementation, step S104, obtaining the second feature, can be achieved using a deep learning-based method, constructing a convolutional neural network model. The segmented image is used as input to the model, and multiple convolutional and pooling layers are set in the first few layers of the network. The convolutional layers extract image features, while the pooling layers downsample the feature maps, reducing the amount of data while retaining key features. Through the combination of these layers, the complex relationships between features such as brightness, color, and gradient in the image are automatically learned. During network training, a large number of cable joint segmentation image samples labeled with defect information are used, with mean squared error as the loss function. The backpropagation algorithm is used to continuously adjust the network's weight parameters, enabling the model to accurately extract the second feature for defect identification from the input image.

[0042] In another optional implementation, step S104, obtaining the second feature, can also employ a wavelet transform-based method. The segmented image is decomposed into wavelet coefficients of different frequency sub-bands. In the low-frequency sub-band, which mainly contains the approximate contour information of the image, the low-frequency wavelet coefficients are normalized for brightness and color deviations, and then adjusted using gamma correction to improve the overall brightness and contrast of the image. For the high-frequency sub-band, which contains edge and detail information of the image, the gradient magnitude and gradient direction of the high-frequency sub-band wavelet coefficients are calculated. The processed low-frequency and high-frequency sub-band wavelet coefficients are combined, and wavelet reconstruction is used to obtain an image containing features such as brightness, color, and gradient, thereby obtaining the second feature.

[0043] It should be noted that this invention normalizes the brightness and color deviations of the segmented images, unifying the brightness and color standards of different images and eliminating image inconsistencies caused by differences in imaging equipment or environmental factors, making subsequent analysis more stable and reliable. Gamma correction further optimizes the visual effect of the image, adjusting brightness and contrast according to the actual situation of the image, making defect features in the image more prominent, facilitating subsequent analysis and identification. Calculating the gradient magnitude and gradient direction of pixels in the segmented image can effectively extract edge and texture information. This information is crucial for identifying defects in cable joints; for example, minute cracks or structural changes will be clearly reflected in the gradient features. The final obtained second feature contains key information on the image's brightness, color, edges, and texture.

[0044] In this embodiment of the invention, step S106 above, in which the second feature is input into the recognition model for analysis and judgment to identify and obtain cable intermediate joint defect information, also includes sub-steps C1-C3: C1: Use samples of defect-free cable joints and samples of defective cable joints to train an identification model, and identify and obtain the features of defect-free cable joints and the features of defective cable joints. C2: Input the second feature into the recognition model, and use the structural similarity image comparison method to compare the second feature with the features of defective cable joints and defective cable joints to obtain the contrast. C3: Based on the brightness, color, structure, and contrast of the second feature, obtain a comprehensive score to identify and obtain cable intermediate joint defect information; based on the defect information, generate a defect report and send it to the staff.

[0045] In this embodiment of the invention, a convolutional neural network is used as the basic architecture to construct the recognition model, and transfer learning techniques are combined to accelerate model convergence. The training samples include defect-free samples taken from normally operating cable joints, covering different models, installation environments, and operating durations; samples containing typical defects from man-made or real-world cases, such as cracks, bubbles, and aging; and augmented samples expanded through data augmentation techniques such as rotation, scaling, and adding noise. During training, the model automatically learns key features of the joint area using an attention mechanism, such as the uniformity and thickness variation of the insulation layer, the structural integrity of the conductor connection, and the continuity and damage characteristics of the shielding layer.

[0046] In one alternative implementation, a support vector machine (SVM) algorithm can be used to construct the recognition model. The extracted second feature is used as the input vector, and the SVM algorithm searches for the optimal classification hyperplane in high-dimensional space to distinguish between defect-free and defective cable joint samples. A radial basis function is used as the kernel function, and cross-validation is employed to optimize the penalty parameters and kernel coefficients. During the training phase, the feature vectors are first standardized to eliminate the dimensionality differences between different feature dimensions; then, grid search combined with 5-fold cross-validation is used to determine the optimal parameter combination.

[0047] In another alternative implementation, a random forest algorithm can be used to construct the recognition model. Specifically, multi-dimensional features such as brightness, color, and gradient are used as input, and a classifier composed of multiple decision trees is constructed through ensemble learning. During the feature processing stage, principal component analysis is performed on high-dimensional features to reduce dimensionality, decrease feature redundancy, and retain key information. During decision tree construction, bootstrap sampling is used to generate different training subsets, and each decision tree is trained on a randomly selected feature subset. By adjusting hyperparameters such as the number of decision trees and their maximum depth, a balance between accuracy and computational efficiency is achieved.

[0048] In this embodiment of the invention, after inputting the second feature into the trained model, a structural similarity image comparison method is used for feature comparison. This algorithm performs multi-scale analysis, calculating structural similarity at different resolutions, taking into account both macroscopic and microscopic defects; it implements weighted feature fusion, assigning different weights to feature dimensions such as brightness, color, and gradient, highlighting features sensitive to defects; and it can also dynamically adjust the threshold, adaptively adjusting the similarity threshold according to the joint type and detection environment to reduce the false alarm rate.

[0049] In this embodiment of the invention, the comprehensive scoring model integrates multiple dimensions: brightness anomaly scoring calculates the brightness deviation from the standard image and quantifies overly bright or dark areas; color deviation scoring analyzes color changes caused by the aging of insulating materials; structural deformation scoring detects geometric anomalies based on gradient direction histograms; and the structural similarity image comparison method contrast scoring maps similarity values ​​to a defect confidence level of 0-100.

[0050] In this embodiment of the invention, the generated defect report includes the location, type, degree of defect in the joint, and related historical case library, etc. The types include cracks, bubbles, wear, etc. It provides targeted repair or replacement suggestions and a visualization report comparing the original image with the segmentation result with the defect mark superimposed.

[0051] In this embodiment of the invention, the cross-entropy loss function, focus loss function, and structural similarity loss function of the recognition model are calculated to optimize the recognition model parameters.

[0052] It should be noted that this invention utilizes diverse samples to train the recognition model, combining transfer learning and attention mechanisms to enable the model to automatically capture key features of the joint area, such as insulation uniformity and conductor connection integrity, thereby improving sensitivity to different types of defects. An improved structural similarity image comparison method is employed, using multi-scale analysis, weighted feature fusion, and dynamic threshold adjustment to achieve high-precision comparison of defect features, effectively reducing misjudgments caused by environmental noise and imaging differences. This enables real-time, intelligent, and remote defect detection, improving operational efficiency and reducing manual inspection costs.

[0053] Example 3 illustrates a schematic scheme for a cable joint defect detection method. It should be noted that the technical solution of this AI-based communication method belongs to the same concept as the above-described cable joint defect detection method. Details not described in detail in this embodiment of the cable joint defect detection system can be found in the description of the above-described cable joint defect detection method.

[0054] This embodiment also provides a cable joint defect detection system, including: The data acquisition module is used to acquire the original image of the cable intermediate joint; The feature extraction module is used to extract features from the original image, obtain the first feature, segment the original image based on the first feature, obtain the segmented image, and calculate the pixel values ​​of the segmented image to obtain the second feature. The analysis and identification module is used to input the second feature into the identification model for analysis and judgment, and to identify and obtain information on defects in cable intermediate joints.

[0055] This embodiment also provides an electronic device suitable for detecting defects in cable joints, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the cable joint defect detection method proposed in the above embodiment.

[0056] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the cable joint defect detection method proposed in the above embodiments.

[0057] The storage medium proposed in this embodiment and the method for detecting defects in cable intermediate joints proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0058] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0059] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method of detecting defects in a cable intermediate joint, characterized in that, The method comprises the following steps: obtaining an original image of a cable intermediate joint; performing feature extraction on the original image to obtain first features, segmenting the original image based on the first features, and obtaining a segmented image; calculating pixel point values of the segmented image to obtain second features; inputting the second features into a recognition model for analysis and judgment to obtain defect information of the cable intermediate joint.

2. The method of cable intermediate joint defect detection of claim 1, wherein, The feature extraction on the original image to obtain first features comprises: performing grayscale processing on the original image to obtain first features; using a weighted smoothing template to filter the grayscale image to remove noise points in the original image.

3. The method of cable intermediate joint defect detection of claim 2, wherein, The segmented image is obtained by: reorganizing elements in the original image in proportion to obtain a first image; segmenting the first image based on the first features to obtain a segmented image.

4. The method of cable intermediate joint defect detection of claim 3, wherein, The calculation of the pixel point values of the segmented image to obtain second features comprises: preprocessing the brightness and color deviation of the segmented image; correcting the preprocessed segmented image through a gamma correction factor, calculating the gradient amplitude and gradient direction of the pixel points of the segmented image, and obtaining second features.

5. The method of cable intermediate joint defect detection of claim 4, wherein, The recognition model comprises: training the recognition model using defect-free cable intermediate joint samples and defective cable intermediate joint samples to obtain defect-free cable intermediate joint features and defective cable intermediate joint features; calculating the cross-entropy loss function, focal loss function and structural similarity loss function of the recognition model to optimize the parameters of the recognition model.

6. The method of cable intermediate joint defect detection of claim 5, wherein, The inputting of the second features into the recognition model for analysis and judgment to obtain defect information of the cable intermediate joint comprises: inputting the second features into the recognition model, using a structural similarity image comparison method to compare the second features with the defect-free cable intermediate joint features and the defective cable intermediate joint features to obtain contrast; based on the brightness, color, structure and contrast of the second features, obtaining a comprehensive score to obtain defect information of the cable intermediate joint; generating a defect report based on the defect information and sending the defect report to a worker.

7. The method of cable intermediate joint defect detection of claim 1, wherein, The original image of the cable intermediate joint is obtained by: installing and deploying an X-ray emitter and an X-ray detector at the position of the cable intermediate joint; setting and calibrating appropriate transmission power, starting the X-ray emitter to make X-ray penetrate the cable intermediate joint, and making the X-ray detector receive the rays and convert them into electrical signals transmitted to an image processing system; the image processing system processes the electrical signals to obtain the original image of the cable intermediate joint.

8. A cable mid-joint defect detection system applying the cable mid-joint defect detection method according to any one of claims 1 to 7, characterized by, The method comprises the following steps: a data acquisition module for obtaining an original image of a cable intermediate joint; a feature extraction module for performing feature extraction on the original image to obtain first features, segmenting the original image based on the first features, and obtaining a segmented image; calculating pixel point values of the segmented image to obtain second features; an analysis and recognition module for inputting the second features into a recognition model for analysis and judgment to obtain defect information of the cable intermediate joint.

9. An electronic device, comprising: The method comprises the following steps: a memory and a processor; The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, and the computer executable instructions, when executed by the processor, implement the steps of the method for detecting defects of a cable intermediate joint according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program product has computer executable instructions stored therein, and the computer executable instructions, when executed by the processor, implement the steps of the method for detecting defects of a cable intermediate joint according to any one of claims 1 to 7.

Images