State detection method and device for transformer bushing in power grid and electronic equipment

Abstract

The invention discloses a state detection method and device for a transformer bushing in a power grid and electronic equipment. The method comprises the steps that an optical image of a transformer bushing to be subjected to state detection in a power grid is acquired, and the optical image at least comprises a visible light image and an infrared image of the transformer bushing; feature extraction is carried out on the optical image, a target image feature descriptor of the optical image is obtained, and the target image descriptor is used for representing edge contour image features of the optical image; based on the target image feature descriptor, fusing the visible light image and the infrared image in the optical image to obtain a fused image; and analyzing the fused image to obtain a state detection result of the transformer bushing. According to the invention, the technical problem of low efficiency of state detection of the transformer bushing in the power grid in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of transformer bushing detection, and in particular to a state detection method, device and electronic equipment for a transformer bushing in a power grid. Background Art

[0002] By detecting the insulation condition of the transformer bushing, including whether there is leakage, damage or aging, that is, performing status detection on the transformer bushing in the power grid, the normal and safe operation of the transformer bushing can be ensured.

[0003] At present, the method of performing status detection on the transformer bushing in the power grid in related technologies usually requires the use of professional equipment such as insulation resistance testers for detection, which is more dependent on professional equipment or professional technicians, or adopts a regular inspection method. The timeliness of the status detection of the transformer bushing is low, resulting in low efficiency in the status detection of the transformer bushing in the power grid.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method, a device and an electronic device for detecting the state of a transformer bushing in a power grid, so as to at least solve the technical problem of low efficiency in detecting the state of a transformer bushing in a power grid in the related art.

[0006] According to one aspect of an embodiment of the present invention, a method for detecting the state of a transformer bushing in a power grid is provided, comprising: acquiring an optical image of a transformer bushing to be detected in the power grid, wherein the optical image at least comprises a visible light image and an infrared image of the transformer bushing; performing feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize edge contour image features of the optical image; based on the target image feature descriptor, fusing a visible light image and an infrared image in the optical image to obtain a fused image; and analyzing the fused image to obtain a result of detecting the state of the transformer bushing.

[0007] Optionally, feature extraction is performed on the optical image to obtain a target image feature descriptor of the optical image, including: feature extraction is performed on the optical image to obtain a feature corner point set of the optical image, wherein the feature corner point set is used to characterize edge intersection positions in the optical image; a main direction set of each feature corner point in the feature corner point set is determined, wherein the main direction set is used to characterize gradient directional features of each feature corner point; and a target image feature descriptor is constructed based on the feature corner point set and the main direction set.

[0008] Optionally, feature extraction is performed on the optical image to obtain a feature corner point set of the optical image, including: feature extraction of multiple pixel points on a preset image edge contour in the optical image to obtain an initial feature corner point set of the optical image; determining corner point responses of the multiple pixel points, wherein the corner point responses are used to characterize the degree of change in the gradient direction of the multiple pixel points; based on the corner point responses, determining an adaptive threshold value of the initial feature corner point set, wherein the adaptive threshold value is used to characterize a threshold value of corner point detection obtained based on adjustment of multiple pixel points; and screening the initial feature corner point set based on the adaptive threshold value to obtain a feature corner point set.

[0009] Optionally, determining the main direction set of each feature corner point in the feature corner point set includes: constructing a feature triangle based on each feature corner point in the feature corner point set and a feature vector of each feature corner point; and determining the main direction set based on a triangle midline of the feature triangle.

[0010] Optionally, based on a target image feature descriptor, a visible light image and an infrared image in the optical image are fused to obtain a fused image, including: determining geometric transformation parameters of a preset geometric transformation model based on the target image feature descriptor, wherein the preset geometric transformation model is used to represent a preset mathematical model for geometric transformation of the optical image; constructing a target geometric transformation model based on the preset geometric transformation model and the geometric transformation parameters; and based on the target geometric transformation model, fusing the visible light image and the infrared image to obtain a fused image.

[0011] Optionally, based on the target geometric transformation model, the visible light image and the infrared image are fused to obtain a fused image, including: based on the target geometric transformation model, the visible light image and the infrared image are interpolated in the target direction to obtain a fused image, wherein the target direction includes at least a horizontal direction and a vertical direction.

[0012] According to another aspect of an embodiment of the present invention, a state detection device for a transformer bushing in a power grid is provided, comprising: an acquisition module, used to acquire an optical image of a transformer bushing to be subjected to state detection in the power grid, wherein the optical image at least comprises a visible light image and an infrared image of the transformer bushing; an extraction module, used to perform feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize the edge contour image features of the optical image; a fusion module, used to fuse the visible light image and the infrared image in the optical image based on the target image feature descriptor to obtain a fused image; and an analysis module, used to analyze the fused image to obtain a state detection result of the transformer bushing.

[0013] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0014] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0015] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0016] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0017] According to another aspect of the embodiments of the present invention, a computer program is further provided. When the computer program is executed by a processor, the methods in the embodiments of the present invention are implemented.

[0018] In an embodiment of the present invention, a state detection method for a transformer bushing in a power grid is provided, comprising: acquiring an optical image of a transformer bushing to be detected in the power grid, wherein the optical image at least comprises a visible light image and an infrared image of the transformer bushing; performing feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize edge contour image features of the optical image; based on the target image feature descriptor, fusing a visible light image and an infrared image in the optical image to obtain a fused image; and analyzing the fused image to obtain a state detection result of the transformer bushing. It is easy to notice that in the process of transformer bushing status detection, the optical image is feature extracted to obtain the target image feature descriptor used to characterize the edge contour image features of the optical image. By analyzing the edge contour, the shape, size and structure of the transformer bushing can be obtained. The optical image and the infrared image can be fused based on the target image feature descriptor. The visible light image can show the appearance and surface features of the object, and is not sensitive to information such as the temperature distribution inside the transformer bushing; the infrared image can reflect the thermal distribution of the object and can detect thermal anomalies in the transformer bushing. Fusion of the visible light image and the infrared image can comprehensively utilize the advantages of the visible light image and the infrared image to improve the detection accuracy and reliability of the transformer bushing status, so that the fused image can be further analyzed and processed, and the status of the transformer bushing can be more comprehensively understood, the accuracy and efficiency of fault detection can be improved, and it is helpful to take repair or replacement measures in time to ensure the safety and stable operation of the power grid, thereby solving the technical problem of low efficiency in status detection of transformer bushings in the power grid in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 is a flow chart of a method for detecting the state of a transformer bushing in a power grid according to an embodiment of the present invention;

[0021] Figure 2 is a flow chart of an optional method for detecting the state of a transformer bushing in a power grid according to an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of a state detection device for a transformer bushing in a power grid according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] According to an embodiment of the present invention, an embodiment of a method for detecting the status of a transformer bushing in a power grid is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0026] Figure 1 is a flow chart of a method for detecting the state of a transformer bushing in a power grid according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0027] Step S102: acquiring an optical image of a transformer bushing to be inspected in a power grid.

[0028] The optical image includes at least a visible light image and an infrared image of the transformer bushing.

[0029] The above-mentioned transformer bushing may refer to a bushing made of insulating material in a transformer, which is used to protect the windings and other components inside the transformer. The transformer bushing can prevent the external environment from affecting the insulating material and wires inside the transformer, and can also play a role in protection and isolation. The transformer bushing has good high temperature resistance, corrosion resistance and pressure resistance, and can adapt to harsh environmental conditions.

[0030] The above-mentioned optical image may refer to image information of the transformer bushing surface obtained by optical technology, which may include visible light images, infrared images, thermal imaging images, etc. The thermal imaging image may display the temperature distribution inside the transformer bushing. By detecting the differences in different temperature areas, potential fault points or abnormal conditions may be discovered. The specific type of the optical image may also be determined according to actual needs and is not limited here.

[0031] The above-mentioned visible light image may refer to an image obtained by a visible light camera device, which can be used to detect the physical structure and appearance characteristics of the transformer bushing surface, such as surface damage, corrosion, oil leakage, etc. The visible light image can intuitively display the overall condition of the transformer bushing and help with surface inspection and fault diagnosis. For example, the visible light image can show whether there are cracks, deformations, or accumulation of foreign matter on the surface of the transformer bushing, so as to timely discover and resolve potential safety hazards.

[0032] The above-mentioned infrared image may refer to an image obtained by infrared thermal imaging technology, which can be used to detect the temperature distribution inside the transformer bushing. Since the transformer bushing will be affected by the heating of the current during operation, the heat distribution of the transformer bushing can be detected through the infrared image, and local overheating or abnormal temperature phenomena can be discovered in time to prevent overheating failure of the transformer bushing. For example, the infrared image can show whether there is a local high-temperature area inside the transformer bushing, which helps to adjust the operating parameters in time to ensure the safe and stable operation of the transformer.

[0033] In an optional embodiment, an optical image of the transformer bushing can be obtained, which can help identify damage, aging or other problems on the surface of the transformer bushing, so that maintenance or replacement measures can be taken in time to ensure the safety and stability of the power grid operation. Specifically, an optical camera can be used to capture a visible light image of the transformer bushing; an infrared camera or an infrared imaging device can be used to capture an infrared image of the transformer bushing. The optical camera, the infrared camera or the infrared imaging device can be installed in an appropriate position so as to be able to completely capture the surface image of the transformer bushing, ensuring that the angle and distance of the camera can meet the requirements for obtaining a clear image; the obtained visible light image and infrared image can be imported into computer software for processing, and an image processing algorithm can be used to enhance, segment and analyze the image so as to more clearly observe the status of the transformer bushing. Through the above steps, the visible light image and infrared image of the transformer bushing can be obtained, and the status detection and analysis can be effectively performed, providing important reference information for the maintenance and management of power grid equipment.

[0034] Step S104: extract features from the optical image to obtain a target image feature descriptor of the optical image.

[0035] Among them, the target image descriptor is used to characterize the edge contour image features of the optical image.

[0036] The above-mentioned target image feature descriptor may refer to a set of numerical feature values ​​obtained by analyzing and processing an image, which is used to describe specific features or structures of the image. The target image feature descriptor may include an edge histogram, an edge gradient histogram, an edge shape descriptor, etc. The target image feature descriptor is used to characterize the numerical description of the edge contour features of the optical image. By extracting and analyzing the target image feature descriptor, automatic detection and recognition of the transformer bushing status can be achieved.

[0037] In an optional embodiment, during the transformer bushing state detection process, feature extraction can be performed on the optical image, and key features that can characterize the transformer bushing state can be extracted from the optical image for subsequent image analysis and processing. Specifically, edge detection can be performed on the optical image to accurately extract edge information in the image; the detected edges can be connected to form a closed contour, and key points and line segment information in the closed contour can be extracted to subsequently describe the contour; based on the extracted contour information, an image feature descriptor describing the contour shape and structure can be calculated; the extracted image feature descriptors can be integrated together to form a target image feature descriptor for characterizing the edge wheel of the optical image. The target image feature descriptor is used to characterize the edge contour of the optical image, which can be the contour feature of the transformer bushing. The edge contour is an important feature of the image that represents the shape of the object. By analyzing the edge contour, the shape, size, structure and other information of the object can be obtained. The target image feature descriptor can be obtained by edge detection, gradient calculation and shape analysis of the image. By analyzing and comparing the target image feature descriptors, the transformer bushings in different states can be effectively distinguished, and the detection and recognition of the transformer bushing state can be realized. Through the above steps, the feature extraction of the optical image can be realized, and the target image descriptor that can characterize the state of the transformer bushing can be obtained, thereby realizing the detection and analysis of the state of the transformer bushing.

[0038] Step S106: Based on the target image feature descriptor, the visible light image and the infrared image in the optical image are fused to obtain a fused image.

[0039] In an optional embodiment, during the state detection process of the transformer bushing in the power grid, the optical image and the infrared image can be fused based on the target image feature descriptor. The fusion method may include weighted averaging, feature splicing, feature weighted multiplication, etc., which can be determined according to actual needs and is not limited here. The visible light image can display the appearance and surface features of the object, and is not sensitive to information such as the temperature distribution inside the transformer bushing. The infrared image can reflect the thermal distribution of the object and can detect thermal anomalies in the transformer bushing, such as potential fault locations or aging conditions. Fusion of the visible light image and the infrared image can comprehensively utilize the advantages of the visible light image and the infrared image to improve the detection accuracy and reliability of the transformer bushing state. Visible light images and infrared images can be registered to ensure that they are aligned in the same spatial coordinate system for subsequent fusion processing. Image fusion algorithms can be used to fuse visible light images and infrared images to generate fused images. Image fusion methods can include pixel-level fusion, feature-level fusion, etc., to facilitate further analysis and processing of the fused images and extract characteristic information of the transformer bushing, such as thermal anomaly areas, temperature distribution, etc., so as to evaluate and diagnose the status of the transformer bushing. By fusing visible light images and infrared images, the status of the transformer bushing can be more comprehensively understood, the accuracy and efficiency of fault detection can be improved, and timely repair or replacement measures can be taken to ensure the safe and stable operation of the power grid.

[0040] Step S108: Analyze the fused image to obtain a state detection result of the transformer bushing.

[0041] In an optional embodiment, during the status detection process of the transformer bushing in the power grid, the fused image can be preprocessed, including image enhancement, denoising and other operations to improve the clarity and accuracy of the image. Machine learning or deep learning algorithms can be used for image classification and recognition to determine whether the transformer bushing is abnormal or damaged. The training model can be used to identify transformer bushings in different states and give corresponding prediction results. According to the analysis results, corresponding measures can be taken in time to maintain and repair the transformer bushing to ensure the safe operation of the power grid. The analysis of the fused image can comprehensively use image processing, machine learning, deep learning and other technologies. The method of analyzing the fused image can be determined according to actual needs and is not limited here to achieve accurate detection and evaluation of the status of the transformer bushing.

[0042] In an embodiment of the present invention, a state detection method for a transformer bushing in a power grid is provided, comprising: acquiring an optical image of a transformer bushing to be detected in the power grid, wherein the optical image at least comprises a visible light image and an infrared image of the transformer bushing; performing feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize edge contour image features of the optical image; based on the target image feature descriptor, fusing a visible light image and an infrared image in the optical image to obtain a fused image; and analyzing the fused image to obtain a state detection result of the transformer bushing. It is easy to notice that in the process of transformer bushing status detection, the optical image is feature extracted to obtain the target image feature descriptor used to characterize the edge contour image features of the optical image. By analyzing the edge contour, the shape, size and structure of the transformer bushing can be obtained. The optical image and the infrared image can be fused based on the target image feature descriptor. The visible light image can show the appearance and surface features of the object, and is not sensitive to information such as the temperature distribution inside the transformer bushing; the infrared image can reflect the thermal distribution of the object and can detect thermal anomalies in the transformer bushing. Fusion of the visible light image and the infrared image can comprehensively utilize the advantages of the visible light image and the infrared image to improve the detection accuracy and reliability of the transformer bushing status, so that the fused image can be further analyzed and processed, and the status of the transformer bushing can be more comprehensively understood, the accuracy and efficiency of fault detection can be improved, and it is helpful to take repair or replacement measures in time to ensure the safety and stable operation of the power grid, thereby solving the technical problem of low efficiency in status detection of transformer bushings in the power grid in related technologies.

[0043] Optionally, feature extraction is performed on the optical image to obtain a target image feature descriptor of the optical image, including: feature extraction is performed on the optical image to obtain a feature corner point set of the optical image, wherein the feature corner point set is used to characterize edge intersection positions in the optical image; a main direction set of each feature corner point in the feature corner point set is determined, wherein the main direction set is used to characterize gradient directional features of each feature corner point; and a target image feature descriptor is constructed based on the feature corner point set and the main direction set.

[0044] In an optional embodiment, during the state detection process of the transformer bushing in the power grid, feature extraction can be performed on the optical image, and a corner point detection algorithm can be used to detect corner points of the optical image to obtain a feature corner point set, which can be used to characterize the edges and intersection positions in the image; for each corner point in the feature corner point set, the gradient direction of the pixels around each corner point can be calculated, and these gradient directions can be counted to obtain a main direction set, which is used to characterize the gradient direction characteristics of each feature corner point; a target image feature descriptor can be constructed based on the feature corner point set and the main direction set, and the generation process of the target image feature descriptor can include the steps of sampling the area around the corner point, normalizing the direction, calculating the feature vector, etc. Finally, the obtained target image feature descriptor can be used to realize the detection of the transformer bushing state. The above steps can effectively characterize the local structure and texture information of the image by extracting the feature corner point set and the main direction set of the image, providing strong support for subsequent state detection tasks.

[0045] Optionally, feature extraction is performed on the optical image to obtain a feature corner point set of the optical image, including: feature extraction of multiple pixel points on a preset image edge contour in the optical image to obtain an initial feature corner point set of the optical image; determining corner point responses of the multiple pixel points, wherein the corner point responses are used to characterize the degree of change in the gradient direction of the multiple pixel points; based on the corner point responses, determining an adaptive threshold value of the initial feature corner point set, wherein the adaptive threshold value is used to characterize a threshold value of corner point detection obtained based on adjustment of multiple pixel points; and screening the initial feature corner point set based on the adaptive threshold value to obtain a feature corner point set.

[0046] In an optional embodiment, during the state detection process of the transformer bushing in the power grid, feature extraction can be performed on multiple pixel points on the edge contour of a preset image in the optical image to obtain an initial feature corner point set of the optical image. The edge information of the image can be obtained by an edge detection algorithm, and then multiple pixel points are selected on the edge contour as candidate feature points for feature extraction, such as using a corner point detection algorithm. The corner point responses of the multiple pixel points can be determined, wherein the corner point responses are used to characterize the degree of change in the gradient direction of the multiple pixel points. The corner point response values ​​of each pixel point are calculated, and the corner point response values ​​can reflect the change in the gradient around the pixel point, which can be achieved by calculating the characteristic value or the corner point response function. Based on the corner point response, the adaptive threshold value of the initial feature corner point set can be determined. Among them, the adaptive threshold value is used to characterize the threshold value of corner point detection obtained based on the adjustment of multiple pixel points. According to the distribution of corner point response values, an adaptive threshold value can be determined to screen out pixel points with higher corner point response as the final feature corner points; the initial feature corner point set can be screened based on the adaptive threshold value to obtain a feature corner point set, and the initial feature corner point set can be screened according to the adaptive threshold value to remove those pixel points whose corner point response values ​​are lower than the threshold to obtain the final feature corner point set for subsequent state detection and analysis. Through the above steps, representative feature points can be effectively extracted from the image of the transformer bushing in the power grid, and the detection and analysis of its state can be realized, which is helpful to improve the monitoring efficiency and accuracy of power grid equipment.

[0047] Optionally, determining the main direction set of each feature corner point in the feature corner point set includes: constructing a feature triangle based on each feature corner point in the feature corner point set and a feature vector of each feature corner point; and determining the main direction set based on a triangle midline of the feature triangle.

[0048] In an optional embodiment, during the state detection process of the transformer bushing in the power grid, image processing and computer vision technology can be used to construct a feature triangle based on each feature corner point in the feature corner point set and the feature vector of each feature corner point; the main direction set is determined based on the triangle midline of the feature triangle. Specifically, the feature vector of each feature corner point in the feature corner point set can be calculated, which can be achieved through some descriptor algorithms. The descriptor can describe the image information around the feature point and help distinguish different feature points; after obtaining each feature corner point and the feature vector of each feature corner point, a triangle can be constructed through each feature corner point and the feature vector of each feature corner point; for the constructed feature triangle, the main direction set can be determined by calculating the midline of the triangle. The midline can help find the main direction of the triangle, so as to better understand the state of the transformer bushing. Through the above steps, the main feature information of the transformer bushing image can be obtained, which helps to detect the state of the transformer bushing more accurately.

[0049] Optionally, based on a target image feature descriptor, a visible light image and an infrared image in the optical image are fused to obtain a fused image, including: determining geometric transformation parameters of a preset geometric transformation model based on the target image feature descriptor, wherein the preset geometric transformation model is used to represent a preset mathematical model for geometric transformation of the optical image; constructing a target geometric transformation model based on the preset geometric transformation model and the geometric transformation parameters; and based on the target geometric transformation model, fusing the visible light image and the infrared image to obtain a fused image.

[0050] The aforementioned preset geometric change model may refer to a mathematical model capable of performing geometric transformation of an image, and may be determined according to actual needs, and is not limited here.

[0051] In an optional embodiment, the geometric transformation parameters of the preset geometric transformation model can be determined based on the target image feature descriptor: feature extraction can be performed on the visible light image and the infrared image, and an algorithm can be used to extract the feature descriptor of the image. By matching the feature descriptors, the relationship between the visible light image and the infrared image is determined, and the geometric transformation parameters of the preset geometric transformation model, such as the rotation angle, the translation distance, etc., can be calculated; the target geometric transformation model can be constructed based on the preset geometric transformation model and the geometric transformation parameters; the target geometric transformation model can be constructed according to the determined geometric transformation parameters, and the visible light image and the infrared image can be geometrically transformed so that the visible light image and the infrared image are aligned in space and maintain a consistent scale and direction; based on the target geometric transformation model, the visible light image and the infrared image can be fused to obtain a fused image, and the fused image can be obtained by fusing the visible light image and the infrared image after the geometric transformation, which contains the integration of two different spectral information, and the image fusion can be performed using weighted average, multi-scale fusion and other methods to improve the image quality and information content. Through the above steps, the detection of the status of the transformer bushing in the power grid can be realized, the accuracy and reliability of the detection can be improved, and the safety and stability of the power grid equipment can be helped to be improved.

[0052] Optionally, based on the target geometric transformation model, the visible light image and the infrared image are fused to obtain a fused image, including: based on the target geometric transformation model, the visible light image and the infrared image are interpolated in the target direction to obtain a fused image, wherein the target direction includes at least a horizontal direction and a vertical direction.

[0053] In an optional embodiment, the visible light image and the infrared image can be fused based on the target geometric transformation model to improve the accuracy and reliability of detection. The target detection algorithm can be used to locate the transformer bushing in the visible light image and the infrared image to obtain the position and bounding box of the target. According to the position and bounding box of the target, geometric transformation can be performed in the visible light image and the infrared image respectively so that the position of the target in the two images is consistent; the transformed visible light image and the infrared image can be interpolated to obtain a fused image, and the interpolation can include interpolation in the horizontal and vertical directions, and methods such as bilinear interpolation or bicubic interpolation can be used; the interpolated visible light image and the infrared image can be weighted fused to obtain a final fused image, and the fusion weight can be adjusted according to different application scenarios to improve the detection effect. Through the above steps, the fusion processing of the visible light image and the infrared image can be realized, so as to obtain the information of the transformer bushing more comprehensively and improve the accuracy and reliability of state detection.

[0054] The technical solution proposed in this application is explained below in combination with an optional application scenario. This application proposes a bushing fault diagnosis method based on infrared visible light. Under the trend of accelerating development of new power systems, the contradiction between insufficient grassroots equipment management personnel and the continuous growth of equipment scale is more prominent. In order to solve the contradiction between insufficient operation and maintenance personnel and the increasing amount of equipment maintenance, substations rely on drones, robots, high-definition videos, etc. to replace part of the substation operation and maintenance work, which greatly improves the overall operation and maintenance efficiency of substations; high-voltage capacitive insulating bushings are used as a line-out device. Their main function is to lead the high-voltage line in the transformer to the outside of the oil tank. They are important accessories of the transformer and one of the more important power transmission and transformation equipment. High-voltage bushings operate in high-voltage and harsh working conditions for a long time. During operation, they are easily affected by various factors such as external environment, power grid fluctuations, insulation aging, internal moisture, poor contact or mechanical fatigue, and are prone to internal or external defects. In severe cases, they will cause mechanical damage to the bushing, insulation failure, fire and explosion, and other serious consequences. At present, the requirements for the stability of the main grid power supply are increasing, and high-voltage bushings, as a weak area of ​​power grid operation, urgently need to strengthen equipment status monitoring and reliability assessment capabilities. This application proposes to improve the high-voltage bushing status monitoring method by studying the unified collection technology and reliability assessment method of bushing multi-source parameters, combining equipment asset information and power grid information, and realizing online dynamic assessment of bushing health status; through the unified management of data obtained by live detection, power outage testing and online monitoring, the evaluation algorithm is used to form a collaborative management model of data, indicators and samples, and the maintenance strategy optimization suggestions are proposed to strengthen the bushing status management and control capabilities.

[0055] The present application can set up cameras and handheld infrared detectors around the bushings of substations to obtain visible light images and infrared images of high-voltage bushings, and realize the unified acquisition of multi-source images; the input visible light images and infrared images can be grayed out, and the gray value range can be linearly stretched to enhance some details of the image, and the image resolution can be adjusted to the same vertical resolution; feature points can be detected, and a dual adaptive gradient descent method (DualAdaptiveGradient Descent (abbreviated as Dual-AGDD) feature corner point supervision screening and orientation mechanism can input the visible light image and infrared image of the high-voltage bushing into the anisotropic Gaussian kernel function corner point detection network; Adaptive Gradient Descent (abbreviated as AGDD) adaptive threshold corner point supervision screening can be performed, and the detector (Canny) can be used to extract the edge contours of the heterogeneous images respectively, and then the anisotropic Gaussian kernel function is used to traverse and smooth the edge contours, and the gradient descent response of each pixel point on the image edge contour is calculated in turn and standardized; an adaptive threshold threshold can be set, and the pixel points are controlled by the activation function (ReLU) to extract the candidate corner point set, and the candidate corner points are subjected to non-maximum suppression to determine the correct corner point set, that is, the feature corner point set.

[0056] Furthermore, the main direction of the feature point can be specified, and the target image feature descriptor can be constructed. The gradient descent contour pixel feature point orientation can be performed. The maximum and bilateral minimum values ​​of the gradient descent measurement value of the edge contour pixel can be used to construct the Gaussian feature triangle, and the feature direction of the corner point set can be selected to construct the pixel edge contour gradient descent (Pixel Edge Contour Gradient Descent Feature Descriptor, referred to as PIIFD) feature descriptor and perform feature matching, which can be the above-mentioned target image feature descriptor; geometric transformation and image fusion can be performed, and a suitable geometric transformation model can be selected, and then the transformation parameters of the geometric model are estimated by the obtained target image feature descriptor, that is, the target geometric transformation model, that is, the above-mentioned preset geometric transformation model, the specific preset geometric transformation model can be determined according to actual needs, and is not limited here; then according to the transformation parameters of the geometric model H*, the bicubic interpolation algorithm can be used to effectively fuse the infrared image into the visible light image; finally, the thermal image temperature, texture and shape features can be extracted, and the point, surface and line features of the heating origin can be diagnosed by using a neural network to realize accurate alarm of heating faults. The present application creatively applies the fault diagnosis method based on infrared visible light to high-voltage bushings. The present application applies the Dual-AGDD feature extraction algorithm to the state detection of transformer bushings to overcome the defects of the original technology and improve the accuracy of diagnosis.

[0057] The method provided by the present application may include the following steps: Dual-AGDD feature point screening and orientation mechanism; the power equipment image set, that is, the optical image can be expressed as Q, qi(x)∈Q, i=1, 2, 3..., x=[nx, ny]T, where x is the pixel position and i is the image identification number; the filter window can be generated using an anisotropic Gaussian kernel (ANGKs) function, and the anisotropic Gaussian kernel function can be expressed as:

[0058]

[0059] Where θ is the Gaussian kernel direction, θ = (l-1)2π / L, l = 1, 2, ... L represents different sampling directions, and L is the total number of directional derivatives; R (θ) represents the rotation matrix with direction θ; H is the anisotropy factor matrix; δ∈(0,∞] is the Gaussian scale. The larger the scale, the stronger the kernel function's ability to suppress noise; ρ∈[1,∞] is the anisotropy factor. The larger the anisotropy factor, the stronger the directional selectivity of the kernel function and the stronger the ability to distinguish edge pixels and corner points. However, as the scale δ and the anisotropy factor ρ increase, the corner point positioning performance decreases, resulting in overlapping calculated positions of adjacent corner points. In order to balance noise stability and corner point positioning accuracy, ρ 2 With δ 2 The selection range is [4, 8] and can also be determined according to actual needs and is not limited here.

[0060] The image data exists in the form of a two-dimensional discrete signal, so the Gaussian kernel function and its directional derivative are discretized to obtain:

[0061]

[0062] Corresponding to qi(x)∈Q, x=[nx,ny]T, at this time n=[n x , n y ]T∈Z represents each pixel point in the image. The anisotropic Gaussian directional derivative can be obtained by taking the derivative of the above formula with respect to π / 2+θ:

[0063]

[0064] The anisotropic Gaussian directional derivative response of the high-voltage bushing image qi(x) can be expressed as:

[0065]

[0066] Where: n1 represents the coordinates of the anisotropic Gaussian convolution kernel.

[0067] Gradient descent adaptive threshold corner point supervision screening can be performed. The corner point type can be represented by superposition of multiple simple corner point models, and the general corner point model can be expressed as:

[0068]

[0069] Where: C(r, α) represents a simple L-shaped corner point model; r is the polar radius corresponding to the corner point; α represents the polar angle; T represents the grayscale information of different regions; m is the number of simple corner point models contained in the corner point.

[0070] Then the gradient descent response of the universal corner point is expressed as:

[0071]

[0072] When α1-α2=π, the corner point model is the edge model, and its gradient descent response can be expressed as:

[0073]

[0074] The response has two amplitude extremes at θ=α1 and θ=π+α1:

[0075]

[0076] Pixels whose grayscale changes dramatically in multiple directions in the image are defined as corner points. It can be seen that the gradient descent response amplitude is proportional to the grayscale difference around the pixel. Therefore, the gradient descent response amplitude spectrum of edge pixels only has a large response amplitude in the direction of the grayscale change gradient. In order to extract a complete, effective and accurately corresponding set of corner points from infrared images and visible light images, the present application can traverse each pixel point on the edge contour of the heterogeneous image, obtain the gradient descent response of each pixel, standardize it and sum it, set an adaptive threshold value, and select an adaptive limit for the response of each pixel, thereby overcoming the limitation of the empirical threshold and reducing the corner point missed and misjudgment rate.

[0077] Among them, F is the corner point adaptive threshold, which is obtained by gradient descent response normalization and integral summation. The calculation formula can be:

[0078]

[0079] ξ is the standardized measurement value of the actual gradient descent response in each direction, and the calculation formula can be expressed as:

[0080]

[0081] Where I is the gate value of the corner gate signal, and the pixel is flow-controlled through the activation function (ReLU). When the measurement value ξ is less than the threshold value F, the neuron is not activated, the dot product of the gate value I and the measurement value ξ is 0, the flow gate is closed, and the pixel is identified as an edge pixel; when the measurement value ξ is greater than the threshold value F, the neuron is activated, the corner point is verified, and a candidate corner point set is generated. Non-maximum suppression is performed on the candidate corner point set to screen out the final corner point set of the heterogeneous image.

[0082] The gradient descent contour pixel feature point orientation can be performed. After determining the corner point set, the feature main direction can be selected for each corner point, thereby realizing the construction of the feature descriptor; the gradient descent response of the corner point set and each edge pixel of the infrared image and visible light image has been obtained. Taking each feature corner point as the vertex, the minimum value ξ of the measurement value ξ on both sides of the feature corner point is min1 and min2 The characteristic vector of is taken as the edge to construct a Gaussian characteristic triangle. The direction pointed by the midline of the Gaussian characteristic triangle can be defined as the main direction of the characteristic corner point. The characteristic corner point and the characteristic main direction vector can be used to construct a PIIFD descriptor. That is, based on the characteristic corner point set and the main direction set, a target image feature descriptor is constructed, and the PIIFD descriptors of the visible light image and the infrared image are matched.

[0083] A geometric transformation can be performed, and the geometric transformation process is to select a suitable geometric transformation model, that is, the preset geometric transformation model mentioned above; and then estimate the transformation parameters of the geometric model through the obtained matching feature vector, that is, the target image feature descriptor. Assume that N matching point pairs (xi, yi) and (x'i, y'i) are obtained by matching the visible light image and the infrared image. These matching point pairs are associated by an unknown homography matrix H, as shown below:

[0084]

[0085] Among them, w is an arbitrary non-zero constant; H is a plane mapping transformation described by a 3×3 matrix, that is, a geometric transformation model, that is, a target geometric transformation model, as shown below:

[0086]

[0087] Classical model parameter estimation methods, such as the least squares method, can estimate and optimize model parameters according to a given objective equation to maximize the fit of all given data sets. None of these methods include the process of detecting and eliminating abnormal data. They are all based on the smoothing assumption: ignoring the size of the given data set, assuming that there are enough accurate data values ​​to eliminate the impact of abnormal data. However, in many practical situations, the smoothing assumption cannot be established, and the data may contain serious erroneous data that cannot be compensated, such as the on-site operation images of substation equipment obtained in bad weather. Due to the presence of a large amount of noise, such model parameter estimation methods cannot correctly estimate or optimize the parameters of the geometric transformation model. In order to distinguish the inlier data (data that can correctly describe the model) and the outlier data (i.e., noise data, generated by incorrect detection or matching), and effectively estimate the geometric transformation matrix H through the inlier set.

[0088] This application uses the random sampling consistency RANSAC (Random Sample Consensus) algorithm to obtain the optimal geometric transformation model so that the infrared image can obtain the optimal transformation relative to the visible light image. The basic assumption of the RANSAC algorithm is that the sample set contains both inlier data and outlier data, and given a set of inlier data sets, the geometric transformation model parameters can be estimated and optimized. The algorithm flow is as follows: first, obtain a matching point set from the visible light image and the infrared image; randomly select 4 pairs of matching points from the matching point set as the inlier consistency set, obtain the null space through the consistency set, and calculate the geometric transformation matrix H between the visible light image and the infrared image; use the obtained geometric transformation matrix H to map the feature points in the infrared image to the visible light image; calculate the distance between the mapped position and the actual position of each feature point, and when this distance is less than the allowable error (threshold 1), the feature point is considered is an inlier, otherwise it is regarded as an outlier; the new inlier and the original 4 pairs of matching points together form a new inlier set, and the geometric transformation matrix H′ is re-estimated by the inlier set, and all feature points are remapped using H′, and the iteration is continued in sequence; when a certain number (threshold 2) of inliers are selected, that is, when the mapping distance of these feature points is less than the allowable error (threshold 1), the iteration ends and a maximum inlier set is generated; otherwise, return to execute "randomly select 4 pairs of matching points from the matching point set as the consistent inlier set, obtain the null space through the consistent set, and calculate the geometric transformation matrix H between the visible light image and the infrared image"; the maximum inlier set selected by the RANSAC algorithm generates the optimal geometric transformation matrix H*.

[0089] Image fusion can be performed. Image fusion is to realize the coordinate transformation and grayscale interpolation of the infrared image relative to the visible light image according to the transformation parameters of the geometric model H*. The bicubic interpolation algorithm can be used. This algorithm realizes the interpolation of data points in a two-dimensional regular grid, which is equivalent to the fusion of two one-dimensional cubic interpolation algorithms, that is, it realizes cubic interpolation in the horizontal and vertical directions at the same time. The interpolation surface obtained by this algorithm is smoother than the bilinear interpolation and nearest neighbor interpolation algorithms. After obtaining the optimal parameters of the geometric transformation model through the RANSAC algorithm, the bicubic interpolation algorithm is used to effectively fuse the infrared image into the visible light image. The temperature, texture and shape features of the thermal image are extracted, and the point, surface and line features of the heating origin are diagnosed by the neural network to realize the accurate alarm of heating faults.

[0090] Figure 2 is a flow chart of an optional method for detecting the state of a transformer bushing in a power grid according to an embodiment of the present invention. Figure 2 As shown, it may include: step S202, acquiring an optical image of a transformer bushing to be tested for status in a power grid; step S204, extracting features from the optical image to obtain a target image feature descriptor of the optical image; step S206, determining geometric transformation parameters of a preset geometric transformation model based on the target image feature descriptor; constructing a target geometric transformation model based on the preset geometric transformation model and the geometric transformation parameters; based on the target geometric transformation model, fusing the visible light image and the infrared image to obtain a fused image; step S208, analyzing the fused image to obtain a status detection result of the transformer bushing.

[0091] According to another aspect of an embodiment of the present invention, a state detection device for a transformer bushing in a power grid is also provided. The device can execute the state detection method for a transformer bushing in a power grid of the above embodiment. The specific implementation method and preferred application scenario are the same as those of the above embodiment and will not be described in detail here.

[0092] Figure 3 is a schematic diagram of a state detection device for a transformer bushing in a power grid according to an embodiment of the present application, such as Figure 3 As shown, the device includes the following: an acquisition module 302 , an extraction module 304 , a fusion module 306 and an analysis module 308 .

[0093] Among them, the acquisition module is used to acquire the optical image of the transformer bushing to be tested for status in the power grid, wherein the optical image at least includes the visible light image and the infrared image of the transformer bushing; the extraction module is used to extract features from the optical image to obtain the target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize the edge contour image features of the optical image; the fusion module is used to fuse the visible light image and the infrared image in the optical image based on the target image feature descriptor to obtain the fused image; the analysis module is used to analyze the fused image to obtain the status detection result of the transformer bushing.

[0094] Among them, the extraction module is also used to extract features from the optical image to obtain a feature corner point set of the optical image, wherein the feature corner point set is used to characterize the edge intersection position in the optical image; determine the main direction set of each feature corner point in the feature corner point set, wherein the main direction set is used to characterize the gradient direction characteristics of each feature corner point; based on the feature corner point set and the main direction set, construct a target image feature descriptor.

[0095] Among them, the extraction module is also used to extract features from multiple pixel points on the preset image edge contour in the optical image to obtain an initial feature corner point set of the optical image; determine the corner point responses of the multiple pixel points, wherein the corner point responses are used to characterize the degree of change in the gradient direction of the multiple pixel points; based on the corner point responses, determine the adaptive threshold value of the initial feature corner point set, wherein the adaptive threshold value is used to characterize the threshold value of corner point detection obtained based on the adjustment of multiple pixel points; based on the adaptive threshold value, the initial feature corner point set is screened to obtain a feature corner point set.

[0096] The extraction module is also used to construct a feature triangle based on each feature corner point in the feature corner point set and the feature vector of each feature corner point; and determine a main direction set based on the triangle midline of the feature triangle.

[0097] Among them, the fusion module is also used to determine the geometric transformation parameters of the preset geometric transformation model based on the target image feature descriptor, wherein the preset geometric transformation model is used to represent a preset mathematical model for geometric transformation of the optical image; based on the preset geometric transformation model and the geometric transformation parameters, a target geometric transformation model is constructed; based on the target geometric transformation model, the visible light image and the infrared image are fused to obtain a fused image.

[0098] The fusion module is also used to perform interpolation processing on the visible light image and the infrared image in the target direction based on the target geometric transformation model to obtain a fused image, wherein the target direction includes at least a horizontal direction and a vertical direction.

[0099] An embodiment of the present application further provides an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention when running.

[0100] The above-mentioned memory may refer to a device inside a computer for storing data and programs, and may include memory, hard disk, etc., wherein the memory may be used to temporarily store running programs and data, the hard disk may be used to store programs and data for a long time, and the memory may be used to enable the computer to read and write data, and execute programs; the above-mentioned processor may be responsible for executing instructions in computer programs and performing data processing, and may be responsible for controlling and executing various operations, including arithmetic operations, logical operations, data transmission, etc.

[0101] An embodiment of the present application further provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0102] The above-mentioned computer storage medium may refer to a medium in a computer memory used to store certain discontinuous physical quantities. Computer storage media mainly include semiconductors, magnetic cores, magnetic drums, magnetic tapes, laser disks, etc.; the stored program included in the computer-readable storage medium may be a set of instructions that can be recognized and executed by a computer, running on an electronic computer, and is an information tool that meets certain needs of people.

[0103] An embodiment of the present application further provides a computer program product, including a computer program, which implements the methods in various embodiments of the present invention when executed by a processor.

[0104] The above-mentioned computer program product may refer to a software program that has been written, tested and released and can be run on a computer or other device. The computer program product may include an application, an operating system, tool software, etc., which is used to implement specific functions or solve specific problems.

[0105] An embodiment of the present application further provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0106] The above-mentioned non-volatile computer-readable storage medium may refer to a medium for storing data. The non-volatile computer-readable storage medium can keep the data from being lost when the power is off, and can be used to store long-term data, such as operating systems, applications, and user files. The non-volatile storage medium may include hard disk drives, solid-state drives, optical disks, and flash memory storage devices, etc.

[0107] The embodiments of the present application further provide a computer program, which implements the methods in the above-mentioned embodiments of the present invention when executed by a processor.

[0108] The above-mentioned computer program may refer to a collection of instructions used to tell a computer to perform a specific task or operation. A computer program may be written by a programmer using a specific programming language and may include algorithms, data structures, logic, and control flows. A computer program may be used for a variety of purposes, including application software, operating systems, and the like.

[0109] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0111] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0112] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0113] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0114] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting the state of a transformer bushing in a power grid, characterized in that: include: Acquire an optical image of a transformer bushing to be inspected in a power grid, wherein the optical image includes at least a visible light image and an infrared image of the transformer bushing; Performing feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize edge contour image features of the optical image; Based on the target image feature descriptor, fusing the visible light image and the infrared image in the optical image to obtain a fused image; The fused image is analyzed to obtain a state detection result of the transformer bushing.

2. The method for detecting the state of a transformer bushing in a power grid according to claim 1, characterized in that: Extracting features from the optical image to obtain a target image feature descriptor of the optical image includes: Performing feature extraction on the optical image to obtain a feature corner point set of the optical image, wherein the feature corner point set is used to characterize edge intersection positions in the optical image; Determine a main direction set of each feature corner point in the feature corner point set, wherein the main direction set is used to characterize the gradient direction feature of each feature corner point; The target image feature descriptor is constructed based on the feature corner point set and the main direction set.

3. The method for detecting the state of a transformer bushing in a power grid according to claim 2, characterized in that: Performing feature extraction on the optical image to obtain a feature corner point set of the optical image includes: Extracting features from a plurality of pixel points on a preset image edge contour in the optical image to obtain an initial feature corner point set of the optical image; Determine corner point responses of the plurality of pixel points, wherein the corner point responses are used to characterize the degree of change of the gradient direction of the plurality of pixel points; Based on the corner point response, determining an adaptive threshold value of the initial feature corner point set, wherein the adaptive threshold value is used to characterize a threshold value of corner point detection obtained by adjusting the plurality of pixel points; The initial feature corner point set is screened based on the adaptive threshold value to obtain the feature corner point set.

4. The method for detecting the state of a transformer bushing in a power grid according to claim 2, characterized in that: Determining the main direction set of each feature corner point in the feature corner point set includes: Constructing a feature triangle based on each feature corner point in the feature corner point set and a feature vector of each feature corner point; The set of main directions is determined based on a triangle midline of the characteristic triangle.

5. The method for detecting the state of a transformer bushing in a power grid according to claim 1, characterized in that: Based on the target image feature descriptor, the visible light image and the infrared image in the optical image are fused to obtain a fused image, including: Determining geometric transformation parameters of a preset geometric transformation model based on the target image feature descriptor, wherein the preset geometric transformation model is used to represent a preset mathematical model for geometrically transforming the optical image; Based on the preset geometric transformation model and the geometric transformation parameters, construct a target geometric transformation model; Based on the target geometric transformation model, the visible light image and the infrared image are fused to obtain the fused image.

6. The method for detecting the state of a transformer bushing in a power grid according to claim 5, characterized in that: Based on the target geometric transformation model, the visible light image and the infrared image are fused to obtain the fused image, including: Based on the target geometric transformation model, the visible light image and the infrared image are interpolated in a target direction to obtain the fused image, wherein the target direction includes at least a horizontal direction and a vertical direction.

7. A state detection device for a transformer bushing in a power grid, characterized in that: include: An acquisition module, used for acquiring an optical image of a transformer bushing to be inspected in a power grid, wherein the optical image at least includes a visible light image and an infrared image of the transformer bushing; An extraction module, used for performing feature extraction on the optical image to obtain a target image feature descriptor of the optical image, wherein the target image descriptor is used to characterize edge contour image features of the optical image; A fusion module, used for fusing the visible light image and the infrared image in the optical image based on the target image feature descriptor to obtain a fused image; The analysis module is used to analyze the fused image to obtain the status detection result of the transformer bushing.

8. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein the program executes the state detection method of the transformer bushing in the power grid according to any one of claims 1 to 6 when running.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the storage medium is located is controlled to execute the state detection method for the transformer bushing in the power grid according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises a computer program, which, when executed by a processor, implements the method for detecting the state of a transformer bushing in a power grid according to any one of claims 1 to 6.

Images