Electric power device sleeve discharge detection method, device, equipment and storage medium

Abstract

The invention relates to a discharge detection method, device and equipment for a power device sleeve and a storage medium. The method comprises the following steps: acquiring an ultraviolet image and a visible light image for the power device sleeve, and fusing the ultraviolet image and the visible light image to obtain an ultraviolet fused image; performing edge detection on the ultraviolet fusion image to obtain an edge image of the power device sleeve; performing global detection on the edge image, determining a first discharge part of the power device sleeve in the edge image, and further obtaining a first detection result of the power device sleeve in a discharge state; carrying out image segmentation on the edge image to obtain a plurality of segmented images, respectively determining a second discharge part of the electric power device sleeve in each segmented image, and further obtaining a second detection result of the electric power device sleeve in the discharge state; according to the first detection result and the second detection result, the discharge detection result of the electric power device sleeve is determined, and the accuracy of discharge defect identification of the discharge part is improved.

Description

Technical Field

[0001] The present application relates to the technical field of power system discharge detection, and in particular to a method, device, computer equipment, computer-readable storage medium, and computer program product for detecting discharge of a bushing of a power device. Background Art

[0002] With the continuous development of power systems, the safety and stability of their operation have become a key concern for the power industry. Partial discharge, a common electrical phenomenon in power systems, poses a serious threat to the insulation performance and operational reliability of equipment. Partial discharge not only degrades equipment insulation but can also cause equipment failures and even system breakdowns. Therefore, accurate and timely detection of partial discharge defects in power systems is crucial for preventing equipment failures and ensuring safe power system operation.

[0003] In recent years, with the continuous development of ultraviolet imaging technology, its application in partial discharge detection in power systems has gradually attracted attention. Ultraviolet imaging technology can capture the ultraviolet radiation generated by partial discharge. By analyzing the UV image, it can accurately detect partial discharge defects. However, existing discharge detection algorithms have low detection accuracy for the discharge location, resulting in large errors in the identification of discharge type and location, which affects the subsequent discharge analysis results of power equipment. Summary of the Invention

[0004] Based on this, it is necessary to provide a discharge detection method, device, computer equipment, computer-readable storage medium and computer program product for the bushing of an electric power device, which can improve the discharge detection accuracy, in order to address the above technical problems.

[0005] In a first aspect, the present application provides a method for detecting discharge of a bushing of an electric device, the method comprising:

[0006] collecting an ultraviolet image and a visible light image of a bushing of an electric power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image;

[0007] performing edge detection on the casing of the electric device in the ultraviolet fusion image to obtain an edge image of the casing of the electric device;

[0008] performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, and obtaining a first detection result of the bushing of the power device in a discharge state based on the first discharge location;

[0009] performing image segmentation on the edge image to obtain a plurality of segmented images, determining a second discharge location of the bushing of the power device in each segmented image, and obtaining a second detection result of the bushing of the power device in the discharge state based on each second discharge location;

[0010] A discharge detection result of the bushing of the power device is determined according to the first detection result and the second detection result.

[0011] In one embodiment, collecting an ultraviolet image and a visible light image of the bushing of the power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image, includes:

[0012] Acquire an optical signal from a bushing of an electric power device, and sequentially perform ultraviolet filtering and photoelectric conversion on the optical signal along a preset optical axis direction to obtain at least one ultraviolet image, and acquire a visible light image of the bushing of the electric power device;

[0013] When there is one ultraviolet image, fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image;

[0014] When there are at least two ultraviolet images, for each ultraviolet image, the ultraviolet image is fused with the visible light image to obtain at least two candidate fused images; based on the historical fused image of the bushing of the power device, a qualified candidate fused image is determined from the at least two candidate fused images, and the qualified candidate fused image is determined as the ultraviolet fused image.

[0015] In one embodiment, fusing the ultraviolet image with the visible light image includes:

[0016] Performing grayscale processing on the ultraviolet image to obtain a first grayscale image, generating a first scale space of the first grayscale image, extracting first feature points of the ultraviolet image based on the first scale space, and generating a first feature descriptor corresponding to the first feature points;

[0017] performing grayscale processing on the visible light image to obtain a second grayscale image, generating a second scale space for the second grayscale image, extracting second feature points of the visible light image based on the second scale space, and generating second feature descriptors corresponding to the second feature points;

[0018] Matching the first feature point with the second feature point based on the first feature descriptor and the second feature descriptor to obtain a feature matching matrix;

[0019] A transformation matrix between the ultraviolet image and the visible light image is determined based on the feature matching matrix, and the ultraviolet image and the visible light image are fused using the transformation matrix.

[0020] In one embodiment, performing edge detection on the power device casing in the UV fusion image to obtain an edge image of the power device casing includes:

[0021] Preprocessing the ultraviolet fusion image, and performing gradient calculation on the preprocessed ultraviolet fusion image to obtain gradient information of each pixel in the ultraviolet fusion image;

[0022] Performing edge pixel detection on each pixel point based on the gradient information to obtain edge pixel points;

[0023] An edge image of the bushing of the power device is generated based on the edge pixel points.

[0024] In one embodiment, performing edge pixel detection on each pixel based on the gradient information to determine the edge pixel includes:

[0025] Determine the gradient direction and gradient amplitude of each pixel point according to the gradient information;

[0026] Traversing the gradient direction and the gradient amplitude of each pixel point;

[0027] When the gradient direction of the pixel point meets a preset direction condition and the gradient amplitude of the pixel point meets a preset amplitude condition, determining the pixel point as a candidate pixel point;

[0028] Traversing the gradient magnitude of each candidate pixel point, and comparing the gradient magnitude with a first threshold and a second threshold;

[0029] When the gradient magnitude is greater than the first threshold, the candidate pixel is determined as a first edge pixel; when the gradient magnitude is greater than or equal to the second threshold and less than or equal to the first threshold, a neighborhood detection is performed on the candidate pixel, and when the number of the first edge pixels in the neighborhood of the candidate pixel meets the condition, the candidate pixel is determined as a second edge pixel;

[0030] The first edge pixel and the second edge pixel are integrated to obtain an edge pixel point.

[0031] In one embodiment, performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, and obtaining a first detection result of the bushing of the power device in a discharge state based on the first discharge location includes:

[0032] performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, extracting first feature information of the first discharge location, and obtaining a first discharge type and a first discharge position of the first discharge location based on the first feature information;

[0033] A first detection result of the electric power device bushing in a discharge state is obtained based on the first discharge type and the first discharge position.

[0034] In one embodiment, segmenting the edge image to obtain a plurality of segmented images, determining a second discharge location of the bushing of the power device in each segmented image, and obtaining a second detection result of the bushing of the power device in the discharge state based on each second discharge location includes:

[0035] Determining the number of image segmentations, and performing image segmentation on the edge image according to the number of image segmentations to obtain a plurality of segmented images;

[0036] for each segmented image, determining a second discharge location of the bushing of the power device in each segmented image, extracting second feature information of the second discharge location, and obtaining a second discharge type and a second discharge position of the second discharge location based on the second feature information;

[0037] A second detection result of the electric device bushing in a discharge state is obtained based on the second discharge type and the second discharge position.

[0038] In a second aspect, the present application further provides a discharge detection device for a bushing of an electric device, the device comprising:

[0039] An image acquisition module, configured to acquire an ultraviolet image and a visible light image of the bushing of the power device, and fuse the ultraviolet image with the visible light image to obtain an ultraviolet fused image;

[0040] an edge detection module, configured to perform edge detection on the casing of the power device in the ultraviolet fusion image to obtain an edge image of the casing of the power device;

[0041] a global detection module, configured to perform global detection on the edge image, determine a first discharge location of the bushing of the power device in the edge image, and obtain a first detection result of the bushing of the power device in a discharge state based on the first discharge location;

[0042] a partition detection module, configured to segment the edge image to obtain a plurality of segmented images, determine a second discharge location of the bushing of the power device in each segmented image, and obtain a second detection result of the bushing of the power device in the discharge state based on each second discharge location;

[0043] The result fusion module is used to determine the discharge detection result of the bushing of the power device according to the first detection result and the second detection result.

[0044] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0045] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above-described method when executed by a processor.

[0046] In a fifth aspect, the present application also provides a computer program product, comprising a computer program, which implements the steps of the above method when executed by a processor.

[0047] The above-mentioned discharge detection method, device, computer equipment, computer-readable storage medium and computer program product of the power device bushing collect an ultraviolet image of the power device bushing and fuse the ultraviolet image with a visible light image of the power device bushing to obtain an ultraviolet fused image; perform edge detection on the power device bushing in the ultraviolet fused image to obtain an edge image of the power device bushing. Based on the edge image, the edge position of the power device bushing can be accurately identified to provide a data basis for subsequent detection of the discharge position; then, by performing global detection on the edge image, the first discharge position of the power device bushing in the edge image is determined, and the position of the power device bushing at the edge of the edge image is obtained. The invention provides a first detection result under the discharge state; performing image segmentation on the edge image to obtain multiple segmented images, respectively determining the second discharge position of the power device bushing in each segmented image, and obtaining the second detection result of the power device bushing under the discharge state based on each second discharge position; and screening the first detection result and the second detection result according to preset conditions, and determining the screened detection result as the discharge detection result of the power device bushing. Combining global detection with segmented detection is conducive to improving the accuracy of discharge defect identification at the discharge position, thereby realizing rapid positioning and classification of discharge defects on the power device bushing, and providing data support for fault warning and maintenance of the power device bushing. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 A diagram illustrating an application environment of a method for detecting a discharge of a bushing of an electric power device according to an embodiment;

[0050] Figure 2 1 is a flow chart of a method for detecting discharge of a bushing of an electric power device according to an embodiment;

[0051] Figure 3 1 is a schematic diagram of a process for fusing an ultraviolet image with a visible light image in one embodiment;

[0052] Figure 4 1 is a schematic diagram of a process of edge detection in one embodiment;

[0053] Figure 5 A schematic diagram of a process for determining edge pixels in one embodiment;

[0054] Figure 6 is a structural block diagram of a discharge detection device for a bushing of an electric power device according to one embodiment;

[0055] Figure 7 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0057] The discharge detection method for the bushing of the power device provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the terminal 102 communicates with the server 104 via a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The server 104 collects ultraviolet and visible light images of the power device bushing and fuses the ultraviolet and visible light images to obtain an ultraviolet fused image. Next, the server 104 performs edge detection on the power device bushing in the ultraviolet fused image to obtain an edge image of the power device bushing. Subsequently, the server 104 performs global detection on the edge image to determine a first discharge location of the power device bushing in the edge image, and obtains a first detection result of the power device bushing in a discharged state based on the first discharge location. The server 104 performs image segmentation to obtain multiple segmented images, and determines a second discharge location of the power device bushing in each segmented image. Based on each second discharge location, a second detection result of the power device bushing in a discharged state is obtained. Finally, the server 104 determines a discharge detection result of the power device bushing based on the first and second detection results.

[0058] Terminal 102 may include, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, smart car devices, and projectors. Portable wearable devices may include smart watches, smart bracelets, and head-mounted devices. Head-mounted devices may include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, and the like. Server 104 may be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing cloud computing services.

[0059] In one embodiment, Figure 2 As shown, a method for detecting discharge of a bushing of an electric device is provided, and the method is applied to Figure 1 The server 104 in FIG is used as an example to illustrate, and it is understood that the method can also be applied to Figure 1 The terminal 102 in the embodiment may also be applied to a system including the terminal 102 and the server 104, and implemented through interaction between the terminal 102 and the server 104. The method of this embodiment includes:

[0060] Step 201 : collecting an ultraviolet image and a visible light image of a bushing of an electric power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image.

[0061] Among them, the power device bushing refers to a structure used to protect the connection between cables, wires and equipment in the power system to prevent current leakage and short circuit. The power device can be a high-voltage device, and the bushing can be a wall bushing, insulator bushing, etc.

[0062] Among them, the ultraviolet image refers to the ultraviolet radiation image generated by discharge around the casing of the power device captured by the ultraviolet imaging equipment or system. The ultraviolet image can show the specific discharge position and discharge intensity corresponding to the discharge part on the casing of the electronic device.

[0063] Among them, the visible light image refers to the visible light image of the power device bushing captured by an ordinary camera. The visible light image can provide image information of the power device bushing and its surrounding environment.

[0064] The UV fusion image is a fused image that contains information from both the UV and visible light images, facilitating a more comprehensive analysis of the discharge conditions in power plant bushings. Fusion methods include weighted fusion and maximum fusion.

[0065] Fusion of the UV image and the visible light image refers to combining the UV image and the visible light image in a certain manner to generate a single image containing information from both images, namely, a UV fused image. In this embodiment, methods for fusing the UV image and the visible light image include, but are not limited to, splicing, weighting, concatenation, mosaicking, and embedding, so as to simultaneously display the discharge phenomenon and visual characteristics of the bushing of the power device in the UV fused image.

[0066] In this embodiment, when image fusion is performed, the power device bushing targeted by the visible light image and the power device bushing targeted by the ultraviolet image are the same or the same type of power device bushing, so as to facilitate subsequent accurate analysis of the discharge condition of the power device bushing.

[0067] For example, when a discharge analysis is required for a bushing of a certain power device, the server collects an ultraviolet image and a visible light image of the bushing of the power device, and fuses the ultraviolet image with the visible light image to obtain an ultraviolet fused image.

[0068] Among them, when collecting ultraviolet images and visible light images, an imaging system with a dual-optical path design can be used to synchronously shoot the ultraviolet images and visible light images to obtain ultraviolet images and visible light images at the same shooting angle and the same shooting time, so as to reduce the difficulty of subsequent image fusion. Of course, in some other embodiments, the ultraviolet image can also be shot by an imaging system involving a single optical path to reduce the light attenuation caused by the dual optical path and ensure the imaging quality of the ultraviolet image. At this time, the visible light image can also be a pre-stored image of the same or the same type of power device bushing at different angles, so that when fusion is performed, a matching image can be selected from multiple visible light images according to the position and angle of the power device bushing in the ultraviolet image for fusion.

[0069] Step 202 : performing edge detection on the bushing of the power device in the UV fusion image to obtain an edge image of the bushing of the power device.

[0070] Among them, edge detection refers to the process of identifying the boundary between the power device casing and background information in the ultraviolet fusion image to detect the outline of the power device casing.

[0071] The edge image refers to an image obtained after edge detection processing, and the edge image contains contour information of the bushing of the power device for subsequent discharge location identification.

[0072] Exemplarily, the server uses an edge detection algorithm to identify the boundary between the power device bushing and the background information in the ultraviolet fusion image, determine the contour information of the power device bushing, and then obtain an edge image of the power device bushing.

[0073] The edge detection algorithm may be implemented by at least one of the following algorithms, including but not limited to the Canny algorithm, the Sobel algorithm, and the Laplace algorithm.

[0074] Step 203 : Perform global detection on the edge image, determine a first discharge location of the bushing of the power device in the edge image, and obtain a first detection result of the bushing of the power device in a discharge state based on the first discharge location.

[0075] Global detection involves analyzing the entire edge image to identify the overall characteristics and locations of discharges on the bushings of power equipment. Specifically, machine learning and neural networks can be used to train a model to identify discharge locations.

[0076] The first discharge location refers to a region where a discharge phenomenon occurs, which is determined in the edge image by global detection.

[0077] The first detection result refers to a global detection result of the bushing of the power device in a discharge state obtained based on the first discharge position.

[0078] Exemplarily, the server performs global detection of discharge locations on the edge image to determine a first discharge location of the power device bushing in the discharge state in the edge image. Subsequently, the server obtains a first detection result of the power device bushing in the discharge state based on the first discharge location.

[0079] Step 204 : segment the edge image to obtain multiple segmented images, determine a second discharge location of the bushing of the power device in each segmented image, and obtain a second detection result of the bushing of the power device in the discharge state based on each second discharge location.

[0080] Image segmentation refers to dividing the edge image into multiple smaller, relatively independent regions or images according to a predetermined number and size, resulting in multiple segmented images. Each segmented image can then be processed separately to achieve more accurate identification of the discharge location and analysis of the discharge state of the power device bushing. Segmented images refer to multiple smaller image regions obtained after image segmentation, balancing both discharge location identification accuracy and efficiency. Each segmented image contains a portion of the power device bushing. However, it should be understood that since the power device bushing may only be present in certain regions during discharge, this does not necessarily mean that a corresponding second discharge location can be detected from the segmented image just because a portion of the power device bushing is present. In some cases, the second discharge location may not be present in the segmented image. It is understood that in other embodiments, when precise analysis of the discharge location is required, the edge image can be divided into more fine-grained segmented images. Since the energy of the power device bushing radiates outward during discharge, some segmented images may contain the second discharge location but not the power device bushing.

[0081] The second discharge location refers to a local position of the discharge phenomenon determined in each divided image.

[0082] The second detection result refers to a detailed detection result of the bushing of the power device under discharge conditions, derived based on each second discharge location. The second detection result can be a global detection result corresponding to the first detection result, or a partitioned detection result corresponding to each second discharge location. When the second detection result is a global detection result, the second detection result is obtained by fusing each second discharge location, using methods including but not limited to weighting, concatenation, splicing, and screening. When the second detection result is a partitioned detection result, the second detection result corresponds to a second discharge location or a segmented image, and is obtained based on the detection results of each second discharge location or each segmented image.

[0083] Exemplarily, the server segments the edge image according to a set number or size to obtain multiple segmented images. Subsequently, the server determines the second discharge location of the power device bushing in each segmented image, and obtains a second detection result of the power device bushing in the discharge state based on each second discharge location.

[0084] Step 205 : Determine a discharge detection result of the bushing of the power device according to the first detection result and the second detection result.

[0085] The discharge detection result refers to the final discharge detection result of the power device bushing, obtained by fusing the first and second detection results. This result accurately reflects the discharge phenomenon on the power device bushing and provides a basis for subsequent maintenance and processing. The fusion method for fusing the first and second detection results includes, but is not limited to, one or more of weighting, concatenation, splicing, and filtering.

[0086] For example, when the second detection result is a global detection result, the weights of the first and second detection results can be set based on information such as the type, operating environment, and operating time of the power device bushing, and one of the first and second detection results can be selected as the final discharge detection result. For another example, when the second detection result is a partitioned detection result, the second detection result can be compared one by one with the detection results of the corresponding positions in the first detection result, and the discharge locations in the first detection result can be retained or replaced using a weighted approach, so that the two detection results can be combined to form the final discharge detection result.

[0087] Exemplarily, the server fuses the first detection result and the second detection result to determine a discharge detection result of the bushing of the power device.

[0088] In the above-mentioned method for detecting discharge of a power device bushing, an ultraviolet image of the power device bushing is collected and fused with a visible light image of the power device bushing to obtain an ultraviolet fused image. Edge detection is performed on the power device bushing in the ultraviolet fused image to obtain an edge image of the power device bushing. Based on the edge image, the edge position of the power device bushing can be accurately identified, providing a data basis for subsequent detection of discharge locations. Subsequently, a global detection is performed on the edge image to determine a first discharge location of the power device bushing in the edge image, thereby obtaining a first detection result of the power device bushing in a discharged state. The edge image is segmented to obtain multiple segmented images, and the second discharge location of the power device bushing in each segmented image is determined respectively. Based on each second discharge location, a second detection result of the power device bushing in the discharge state is obtained; the first detection result and the second detection result are filtered according to preset conditions, and the filtered detection result is determined as the discharge detection result of the power device bushing. Combining global detection with segmented detection is conducive to improving the accuracy of discharge defect identification at the discharge location, thereby realizing rapid positioning and classification of discharge defects on the power device bushing, and providing data support for fault warning and maintenance of the power device bushing.

[0089] In one embodiment, in step 201, collecting an ultraviolet image and a visible light image of the bushing of the power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image includes:

[0090] An optical signal from the bushing of the power device is obtained, and ultraviolet filtering and photoelectric conversion are sequentially performed on the optical signal along a preset optical axis direction to obtain at least one ultraviolet image and obtain a visible light image of the bushing of the power device.

[0091] The optical signal refers to a visible light signal reflected by the surface of the bushing of the power device when light is irradiated onto the bushing of the power device, or an ultraviolet light signal generated by the bushing of the power device during discharge.

[0092] Among them, the preset optical axis direction refers to the optical axis direction of the ultraviolet imaging device or system, which ensures that the captured light signal comes from a specific direction or area, helps to reduce the interference of stray light, light attenuation, etc., so as to improve the imaging quality of the image.

[0093] Among them, UV filtering refers to the use of specific filters or filters to filter the optical signal, allowing only ultraviolet light to pass through while blocking light of other wavelengths to separate the ultraviolet light component from the optical signal.

[0094] Among them, photoelectric conversion refers to the process of converting optical signals into electrical signals, which is usually achieved using photoelectric sensors such as CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor), and ICCD (Intensified CCD).

[0095] Exemplarily, the server obtains a light signal generated by reflection or discharge from the bushing of the power device, and performs ultraviolet filtering on the light signal according to a preset optical axis direction. After obtaining the ultraviolet signal, the server performs photoelectric conversion on the ultraviolet signal to obtain at least one ultraviolet image, and simultaneously obtains a visible light image of the bushing of the power device. The visible light image can be a visible light image obtained by synchronously capturing an image of the same power device bushing as the ultraviolet image, or it can be a visible light image of the same power device bushing or the same type of power device bushing pre-stored in a data storage system.

[0096] When there is only one ultraviolet image, the ultraviolet image is fused with the visible light image to obtain an ultraviolet fused image. When there are at least two ultraviolet images, the ultraviolet image is fused with the visible light image for each ultraviolet image to obtain at least two candidate fused images. Based on the historical fused images of the power device bushing, a candidate fused image that meets the conditions is determined from the at least two candidate fused images, and the qualified candidate fused image is determined as the ultraviolet fused image.

[0097] Among them, the candidate fusion image refers to the intermediate image obtained by fusing each ultraviolet image with the corresponding visible light image when there are multiple ultraviolet images. The candidate fusion image may be different due to factors such as shooting time, angle, and lighting conditions. The image that meets the conditions among the candidate fusion images can be used as the final ultraviolet fusion image.

[0098] Among them, the historical fusion image refers to a fusion image of the same or the same type of power device bushing that has been previously generated and stored in a data storage system or an external storage device. The historical fusion image can be used as a reference to evaluate whether the quality of the current candidate fusion image meets the requirements, so as to determine one or more candidate fusion images that meet the requirements from multiple candidate fusion images as the ultraviolet fusion image.

[0099] The qualified candidate fusion image is an image that meets the requirements and is selected from multiple candidate fusion images based on set standards or conditions. The set standards or conditions may be image clarity, contrast, visibility of discharge phenomena, etc.

[0100] Exemplarily, the server determines the number of UV fusion images. If it is determined that there is only one UV image, this indicates that there is only one candidate image. In this case, the UV image only needs to be fused with the visible light image to obtain the UV fusion image. If it is determined that there are at least two UV images, the server fuses each UV image with a corresponding visible light image. The visible light image can be the same image or a visible light image corresponding to the UV image, thereby obtaining at least two candidate fusion images. Subsequently, the server determines a candidate fusion image that meets the requirements from the at least two candidate fusion images based on historical fusion images of the power device bushing, and determines the qualified candidate fusion image as the UV fusion image.

[0101] In this embodiment, by acquiring an optical signal from the bushing of an electric power device and sequentially performing ultraviolet filtering and photoelectric conversion on the optical signal along a preset optical axis direction, at least one ultraviolet image is obtained, and a visible light image of the bushing of the electric power device is obtained; then, the ultraviolet image and the visible light image are fused and processed in a corresponding manner according to the number of ultraviolet images to improve the quality of the ultraviolet fused image.

[0102] In one embodiment, Figure 3 As shown, in step 201, the ultraviolet image and the visible light image are fused, including:

[0103] Step 301 : grayscale processing is performed on the ultraviolet image to obtain a first grayscale image, a first scale space of the first grayscale image is generated, and based on the first scale space, first feature points of the ultraviolet image are extracted and a first feature descriptor corresponding to the first feature points is generated.

[0104] The grayscale processing in this step refers to the process of converting the ultraviolet image into a grayscale image. The first grayscale image refers to the image obtained after the ultraviolet image is grayscale processed.

[0105] The first scale space is used to capture features of the first grayscale image at different scales. For the first grayscale image, the first scale space represents grayscale information of the first grayscale image at different scales.

[0106] The first feature point refers to a point with significant features extracted in the first scale space of the first grayscale image. The first feature point can be a corner point, edge point, or spot in the first grayscale image, and is used to characterize the characteristics of the power device bushing in the ultraviolet image. The first feature descriptor refers to a vector used to describe the local features of the image around the first feature point. The first feature descriptor contains information such as the grayscale gradient and direction around the first feature point, which is used for subsequent feature matching.

[0107] In this embodiment, at least one of the methods including but not limited to SuperPoint (a feature point detection and description algorithm based on deep learning), SIFT (Scale-invariant feature transform), and ORB (Oriented FAST and Rotated BRIEF) can be used for implementation.

[0108] Exemplarily, taking the SIFT algorithm as an example (the following steps are also explained using the SIFT algorithm), the server performs grayscale processing on the acquired ultraviolet image to obtain a first grayscale image; by constructing a Gaussian pyramid, the first grayscale image is Gaussian blurred at different scales, and sampled layer by layer from the top to form a series of images with different resolutions and scales to obtain a first scale space. Subsequently, based on the first scale space, the server extracts the first feature point in the first grayscale image by calculating the Gaussian difference, and compares the changes in pixel brightness around the feature point in the neighborhood around the first feature point, assigns a main direction to each first feature point, and determines the position and scale of the feature point. Subsequently, a window grid of corresponding size is constructed, such as a 4×4 window grid, and an 8-bit gradient direction histogram is calculated for each window grid to form a 128-dimensional feature vector to describe the local image features around the first feature point, thereby obtaining the corresponding first feature descriptor.

[0109] Step 302 : grayscale processing is performed on the visible light image to obtain a second grayscale image, a second scale space of the second grayscale image is generated, and based on the second scale space, second feature points of the visible light image are extracted and second feature descriptors corresponding to the second feature points are generated.

[0110] The grayscale processing in this step refers to the process of converting the visible light image into a grayscale image. The second grayscale image refers to the image obtained after grayscale processing of the visible light image.

[0111] The second scale space is used to capture features of the second grayscale image at different scales. For the second grayscale image, the second scale space represents grayscale information of the second grayscale image at different scales.

[0112] The second feature point refers to a point with significant features extracted in the second scale space of the second grayscale image. The second feature point can be a corner point, edge point, or spot in the second grayscale image, and is used to characterize the characteristics of the power device bushing in the visible light image. The second feature descriptor refers to a vector used to describe the local image features around the second feature point. The second feature descriptor contains information such as the grayscale gradient and direction around the second feature point and is used for subsequent feature matching.

[0113] Exemplarily, the server performs grayscale processing on the acquired visible light image to obtain a second grayscale image; by constructing a Gaussian pyramid, the second grayscale image is Gaussian blurred at different scales, and is sampled layer by layer from the top to form a series of images with different resolutions and scales to obtain a second scale space. Subsequently, based on the second scale space, the server extracts the second feature point in the second grayscale image by calculating the Gaussian difference, and compares the changes in the brightness of pixels around the feature point in the neighborhood around the second feature point, assigns a main direction to each second feature point, and determines the position and scale of the feature point. Subsequently, a window grid of corresponding size is constructed, such as a 4×4 window grid, and an 8-bit gradient direction histogram is calculated for each window grid, thereby forming a 128-dimensional feature vector to describe the local image features around the second feature point, and obtain the corresponding second feature descriptor.

[0114] Step 303: Match the first feature point with the second feature point based on the first feature descriptor and the second feature descriptor to obtain a feature matching matrix.

[0115] The feature matching matrix is ​​a matrix formed by comparing the first feature descriptor and the second feature descriptor to find the correspondence between the first feature descriptor and its corresponding first feature point and the second feature descriptor and its corresponding second feature point. During matching, the correlation and matching degree between the two feature points can be determined by calculating the Euclidean distance or the nearest neighbor distance ratio between the two feature points.

[0116] Exemplarily, the server matches the first feature descriptor and the second feature descriptor to screen out matching pairs that meet the matching condition to obtain a feature matching matrix.

[0117] Step 304 : determining a transformation matrix between the ultraviolet image and the visible light image based on the feature matching matrix, and fusing the ultraviolet image and the visible light image using the transformation matrix.

[0118] Among them, the transformation matrix refers to a matrix used to store the transformation relationship between the ultraviolet image and the visible light image, and the transformation relationship refers to a relationship used to describe how the ultraviolet image and the visible light image are converted or aligned with each other; the mutual conversion or alignment can be through affine transformation, homography transformation, etc.

[0119] Exemplarily, the server determines the transformation matrix between the ultraviolet image and the visible light image based on the feature matching matrix, and uses the transformation matrix to transform the first feature point in the ultraviolet image into the coordinate system of the visible light image to achieve fusion of the ultraviolet image and the visible light image.

[0120] In this embodiment, the first feature point in the ultraviolet image and the second feature point in the visible light image are extracted respectively, and the first feature point and the second feature point are used to fuse the ultraviolet image and the visible light image. Since the ultraviolet image is focused on the ultraviolet band, other interference information can be reduced, so that when fused with the visible light image, it can complement the visible light image, thereby forming an information-rich ultraviolet fused image.

[0121] In one embodiment, Figure 4 As shown, in step 202, edge detection is performed on the bushing of the power device in the UV fusion image to obtain an edge image of the bushing of the power device, including:

[0122] Step 401 : preprocessing the ultraviolet fusion image, and performing gradient calculation on the preprocessed ultraviolet fusion image to obtain gradient information of each pixel in the ultraviolet fusion image.

[0123] Preprocessing refers to a series of operations performed on the UV-fused image before gradient calculation to improve image quality and reduce the impact of noise on subsequent edge detection. Preprocessing includes, but is not limited to, at least one of importing the UV-fused image, grayscale processing, and filtering the UV-fused image.

[0124] The gradient represents the rate and direction of change in pixel values ​​in the preprocessed UV-fused image. Gradient calculation is the process of calculating the grayscale value change of each pixel within its neighborhood in the preprocessed UV-fused image. This process is used to reveal locations in the preprocessed UV-fused image where grayscale values ​​change rapidly. These locations are often potential edges and can be used to subsequently identify edge pixels.

[0125] The gradient information refers to the gradient value and direction information of each pixel in the preprocessed UV fusion image obtained after gradient calculation, which is used for subsequent edge pixel detection.

[0126] Exemplarily, the server performs grayscale processing and Gaussian filtering on the ultraviolet fusion image in sequence to obtain a preprocessed ultraviolet fusion image, and performs gradient calculation on the preprocessed ultraviolet fusion image to obtain gradient information of each pixel in the ultraviolet fusion image.

[0127] Step 402: perform edge pixel detection on each pixel based on gradient information to obtain edge pixel points.

[0128] Edge pixel detection refers to the process of determining whether each pixel in the preprocessed UV-fused image is an edge pixel. Edge pixels are pixels where the gradient value in the gradient information exceeds a certain threshold and the gradient direction varies significantly within a local range.

[0129] Exemplarily, the server judges each pixel point according to the gradient value of each pixel point in the gradient information. When a pixel point is an edge pixel, the pixel point is determined as an edge pixel point. Otherwise, the pixel point is a non-edge pixel and is not considered until all pixel points are judged and all edge pixel points are obtained.

[0130] Step 403: Generate an edge image of the bushing of the power device based on the edge pixels.

[0131] The edge image is generated based on the determined edge pixels and contains only the edge information of the power device bushing. In the edge image, the edge pixels are retained while other non-edge pixels are discarded to clearly display the edge features of the power device bushing in the image.

[0132] Exemplarily, the server connects or fuses the edge pixel points based on the determined edge pixel points to generate an edge image for the bushing of the power device.

[0133] In this embodiment, by performing edge detection on the bushing of the power device in the ultraviolet fusion image, the edge pixel points of the bushing of the power device in the ultraviolet fusion image can be accurately detected and extracted, thereby generating a high-quality edge image of the bushing of the power device, which is beneficial to providing a data basis for subsequent detection of the discharge location.

[0134] In one embodiment, Figure 5 As shown, in step 402, edge pixel detection is performed on each pixel based on gradient information to determine edge pixel points, including:

[0135] Step 501: Determine the gradient direction and gradient amplitude of each pixel point according to the gradient information.

[0136] The gradient direction refers to the direction in which the pixel value in the preprocessed UV fusion image changes fastest. In specific implementation, the gradient direction is usually expressed as an angle to determine whether a pixel point is located on a potential edge.

[0137] The gradient amplitude refers to the intensity of the pixel value change in the preprocessed UV fusion image, which is usually expressed as a scalar value and is used to measure the possibility of a pixel point becoming an edge point.

[0138] Exemplarily, the server determines the gradient direction and gradient magnitude of each pixel point based on the gradient information.

[0139] Step 502: traverse the gradient direction and gradient amplitude of each pixel point.

[0140] Step 503 : When the gradient direction of the pixel point meets the preset direction condition and the gradient amplitude of the pixel point meets the preset amplitude condition, the pixel point is determined as a candidate pixel point.

[0141] Among them, the preset direction condition refers to the gradient direction condition used to screen potential edge pixels. The preset direction condition can be determined based on characteristics such as the continuity or consistency of the gradient direction. The preset amplitude condition refers to the gradient amplitude condition used to screen potential edge pixels. The preset amplitude condition can be determined based on the gradient amplitude value and is used to exclude pixels with too small gradient amplitudes or unlikely to be edge pixels. In this embodiment, edge pixels are generally required to meet both the preset direction condition and the preset amplitude condition to ensure the accuracy of the edge pixels; of course, in other embodiments, the screening conditions or condition values ​​can also be changed as needed to determine the expected edge pixels.

[0142] Among them, candidate pixel points refer to pixel points that meet the preset direction conditions and preset amplitude conditions. These pixel points are considered to be potential edge points, but further processing is required to determine whether they are indeed edge pixels. Therefore, they are used as candidates for further judgment.

[0143] Exemplarily, when traversing the gradient direction and gradient amplitude of pixel points, the server determines the gradient direction of each pixel point according to a preset direction condition and determines the gradient amplitude of each pixel point according to a preset amplitude condition. When it is determined that the gradient direction of a pixel point meets the preset direction condition and the gradient amplitude meets the preset amplitude condition, the pixel point is determined as a candidate pixel point.

[0144] Step 504 , traverse the gradient magnitude of each candidate pixel point, and compare the gradient magnitude with a first threshold and a second threshold.

[0145] Among them, the first threshold and the second threshold refer to the threshold values ​​for the gradient amplitude that are customized according to the need to filter edge pixels, which are used to further filter out pixels that are more likely to become edge pixels from candidate pixels; the first threshold is greater than the second threshold.

[0146] Exemplarily, the server traverses the gradient amplitude of each candidate pixel point, and during the traversal, compares the gradient amplitude of each candidate pixel with the first threshold and the second threshold value. If the gradient amplitude of a candidate pixel point is greater than the first threshold value, the candidate pixel point is considered to be a strong edge point, that is, the candidate pixel point is an edge pixel point. If the gradient amplitude of a pixel point is less than the second threshold value, the candidate pixel point is considered to be a non-edge point. If the gradient amplitude of a pixel point is between the first threshold value and the second threshold value, the pixel point is considered to be a weak edge point, and further determination is required.

[0147] Step 505: When the gradient amplitude is greater than the first threshold, the candidate pixel is determined as a first edge pixel; when the gradient amplitude is greater than or equal to the second threshold and less than or equal to the first threshold, a neighborhood detection is performed on the candidate pixel. When the number of first edge pixels in the neighborhood of the candidate pixel meets the condition, the candidate pixel is determined as a second edge pixel.

[0148] Among them, the first edge pixel refers to a candidate pixel point whose gradient amplitude is greater than the first threshold, that is, a candidate pixel point considered to be a strong edge point. The first edge pixel has a higher gradient amplitude and is more likely to be a true edge point. The second edge pixel refers to a candidate pixel point whose gradient amplitude is between the first and second thresholds and meets the conditions after neighborhood detection, that is, a candidate pixel point considered to be a weak edge point. The second edge pixel has a lower gradient amplitude, but because the first edge pixel exists in the neighborhood relationship of the second edge pixel, it is used as a pixel point for constructing the edge image to ensure a smooth transition between the two first edge pixels and ensure the accuracy of edge image generation.

[0149] Neighborhood detection involves determining whether there are first edge pixels (i.e., strong edge points) within the neighborhood of a candidate pixel whose gradient magnitude is between a first threshold and a second threshold. If a sufficient number of first edge pixels exist within the neighborhood of a candidate pixel, the candidate pixel is considered a second edge pixel (i.e., weak edge point). This sufficient number can be one or more.

[0150] Exemplarily, when the server detects that the gradient amplitude of a candidate pixel point is greater than a first threshold, the candidate pixel point is determined as a first edge pixel and temporarily stored or marked; when the server detects that the gradient amplitude of a candidate pixel point is greater than or equal to a second threshold and less than or equal to the first threshold, the server performs a neighborhood detection on the candidate pixel point. If the number of first edge pixels in the neighborhood of the candidate pixel point meets the conditions, the candidate pixel point is determined as a second edge pixel. Otherwise, the candidate pixel point is considered to be a non-edge point and is not considered.

[0151] Step 506: Integrate the first edge pixel and the second edge pixel to obtain an edge pixel point.

[0152] Integration refers to a process of merging the first edge pixel and the second edge pixel into a complete edge pixel set, where the edge pixel set includes all pixels determined to be edge points, and is used to generate a final edge image.

[0153] Exemplarily, the server integrates the determined first edge pixel and the second edge pixel to obtain edge pixel points for subsequent generation of an edge image.

[0154] In this embodiment, by integrating the gradient direction and gradient amplitude information of the pixel points, candidate pixel points can be accurately screened out, and based on the gradient amplitude of the candidate pixel points and the number of first edge pixels in the neighborhood, the first edge pixels and the second edge pixels can be more finely distinguished and determined, and finally the precise edge pixel points are integrated, which is conducive to improving the accuracy and detail of edge detection, so as to achieve more accurate edge recognition and feature extraction in the subsequent processing process.

[0155] In one embodiment, in step 203, performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, and obtaining a first detection result of the bushing of the power device in a discharge state based on the first discharge location includes:

[0156] A global detection is performed on the edge image, a first discharge location of the bushing of the power device is determined in the edge image, and first feature information of the first discharge location is extracted. A first discharge type and a first discharge position of the first discharge location are obtained based on the first feature information; and a first detection result of the bushing of the power device in a discharge state is obtained based on the first discharge type and the first discharge position.

[0157] The first characteristic information refers to information extracted from the first discharge site and used to describe the characteristics of the first discharge site. The first characteristic information includes but is not limited to the color, shape, texture, brightness, etc. of the first discharge site, which is used to identify the corresponding discharge type and discharge location.

[0158] The first discharge type refers to the type of discharge phenomenon determined based on the first characteristic information, including but not limited to surface discharge, internal discharge, and creeping discharge. Each discharge type has its own specific characteristics and hazards. By identifying the discharge type, response and improvement measures can be quickly determined to ensure the safe and stable operation of the power device. The first discharge location refers to the specific location of the first discharge site on the bushing of the power device, which facilitates the rapid location of the discharge during subsequent maintenance, inspection, and troubleshooting.

[0159] In an optional embodiment, when extracting the first feature information of the first discharge location, one or more of a deep learning algorithm (such as a convolutional neural network), a corner detection algorithm, a histogram feature extraction algorithm, and a frequency domain feature extraction algorithm may be used.

[0160] Exemplarily, the server performs global detection on the edge image, determines the first discharge location of the power device bushing in the edge image, and extracts first feature information of the first discharge location. Subsequently, the server obtains the first discharge type and the first discharge position of the first discharge location based on the first feature information; and obtains the first detection result of the power device bushing in the discharge state based on the first discharge type and the first discharge position.

[0161] In this embodiment, by performing global detection on the edge image, the first discharge location of the power device bushing can be accurately identified and located, and key feature information can be extracted to determine the discharge type and location, thereby obtaining a global detection result of the power device bushing in the discharge state, which is conducive to reflecting the discharge status of the power device bushing from a holistic perspective.

[0162] In one embodiment, in step 204, the edge image is segmented to obtain a plurality of segmented images, a second discharge location of the bushing of the power device is determined in each segmented image, and a second detection result of the bushing of the power device in the discharge state is obtained based on each second discharge location, including:

[0163] The number of image segmentations is determined, and the edge image is segmented according to the number of image segmentations to obtain a plurality of segmented images; for each segmented image, a second discharge location of the bushing of the power device is determined in each segmented image, and second characteristic information of the second discharge location is extracted, and a second discharge type and a second discharge position of the second discharge location are obtained based on the second characteristic information; and a second detection result of the bushing of the power device in a discharged state is obtained based on the second discharge type and the second discharge position.

[0164] The number of image segmentations refers to the number of segmented images expected to be obtained when the edge image is segmented. The number of image segmentations is determined based on factors such as actual needs, image size and complexity, etc.

[0165] The second characteristic information refers to data extracted from the second discharge site and used to describe the characteristics of that site. Similar to the first characteristic information, this characteristic information may include color, shape, texture, brightness, etc., which are crucial for identifying the discharge type and location.

[0166] The second characteristic information, similar to the first characteristic information, refers to information extracted from the second discharge location and used to describe the characteristics of the second discharge location. The second characteristic information also includes, but is not limited to, the color, shape, texture, brightness, etc. of the second discharge location, and is used to identify the corresponding discharge type and location.

[0167] The second discharge type refers to the type of discharge phenomenon determined based on the second characteristic information, including but not limited to surface discharge, internal discharge, and creeping discharge. Each discharge type has its own specific characteristics and hazards. By identifying the discharge type, response and improvement measures can be quickly determined to ensure the safe and stable operation of the power device. The second discharge location refers to the specific location of the second discharge site on the power device bushing, facilitating the rapid location of the discharge during subsequent maintenance, repair, and troubleshooting.

[0168] Exemplarily, the server performs image segmentation on the edge image based on the determined number of image segmentations to obtain a corresponding number of segmented images; then, for each segmented image, the server determines the second discharge location of the bushing of the power device in each segmented image, extracts second feature information of the second discharge location, and obtains a second discharge type and a second discharge position of the second discharge location based on the second feature information; and obtains a second detection result of the bushing of the power device in a discharge state based on the second discharge type and the second discharge position.

[0169] In this embodiment, by determining the number of image segmentations and performing segmentation processing on the edge image, the discharge condition of the power device bushing in each segmented image can be analyzed in more detail, the second discharge location can be accurately identified and feature information can be extracted, and the discharge type and location can be determined. Ultimately, the partitioned detection results of the power device bushing in the discharge state are obtained, which is conducive to improving the accuracy and reliability of discharge detection. After being integrated with the global detection results, it helps to conduct a more comprehensive assessment of the status of the power device bushing.

[0170] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0171] Based on the same inventive concept, embodiments of the present application also provide a power device bushing discharge detection device for implementing the aforementioned power device bushing discharge detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more power device bushing discharge detection device embodiments provided below can be found in the aforementioned limitations of the power device bushing discharge detection method, and will not be further elaborated here.

[0172] In an exemplary embodiment, Figure 6 As shown, a discharge detection device for bushings of power devices is provided, comprising: an image acquisition module 601, an edge detection module 602, a global detection module 603, a partition detection module 604, and a result fusion module 605, wherein:

[0173] An image acquisition module 601 is configured to acquire an ultraviolet image and a visible light image of a bushing of an electric power device, and fuse the ultraviolet image with the visible light image to obtain an ultraviolet fused image;

[0174] An edge detection module 602 is configured to perform edge detection on the bushing of the power device in the UV fusion image to obtain an edge image of the bushing of the power device;

[0175] a global detection module 603 configured to perform global detection on the edge image, determine a first discharge location of the bushing of the power device in the edge image, and obtain a first detection result of the bushing of the power device in a discharge state based on the first discharge location;

[0176] a partition detection module 604 configured to segment the edge image to obtain a plurality of segmented images, determine a second discharge location of the bushing of the power device in each segmented image, and obtain a second detection result of the bushing of the power device in a discharged state based on each second discharge location;

[0177] The result fusion module 605 is configured to determine a discharge detection result of the bushing of the power device according to the first detection result and the second detection result.

[0178] In an optional embodiment, the image acquisition module 601 is further used to obtain an optical signal from the bushing of the power device, and perform ultraviolet filtering and photoelectric conversion on the optical signal in sequence according to a preset optical axis direction to obtain at least one ultraviolet image, and obtain a visible light image of the bushing of the power device; when there is one ultraviolet image, the ultraviolet image and the visible light image are fused to obtain an ultraviolet fused image; when there are at least two ultraviolet images, the ultraviolet image and the visible light image are fused for each ultraviolet image to obtain at least two candidate fused images; based on the historical fused image of the bushing of the power device, a qualified candidate fused image is determined from the at least two candidate fused images, and the qualified candidate fused image is determined as the ultraviolet fused image.

[0179] In an optional embodiment, the image acquisition module 601 is also used to perform grayscale processing on the ultraviolet image to obtain a first grayscale image, generate a first scale space of the first grayscale image, extract a first feature point of the ultraviolet image based on the first scale space, and generate a first feature descriptor corresponding to the first feature point; perform grayscale processing on the visible light image to obtain a second grayscale image, generate a second scale space of the second grayscale image, extract a second feature point of the visible light image based on the second scale space, and generate a second feature descriptor corresponding to the second feature point; match the first feature point with the second feature point based on the first feature descriptor and the second feature descriptor to obtain a feature matching matrix; determine a transformation matrix between the ultraviolet image and the visible light image based on the feature matching matrix, and use the transformation matrix to fuse the ultraviolet image and the visible light image.

[0180] In an optional embodiment, the edge detection module 602 is also used to preprocess the ultraviolet fusion image, and perform gradient calculation on the preprocessed ultraviolet fusion image to obtain gradient information of each pixel point in the ultraviolet fusion image; perform edge pixel detection on each pixel point based on the gradient information to obtain edge pixel points; and generate an edge image for the power device bushing based on the edge pixel points.

[0181] In an optional embodiment, the edge detection module 602 is further used to determine the gradient direction and gradient amplitude of each pixel point based on the gradient information; traverse the gradient direction and gradient amplitude of each pixel point; when the gradient direction of the pixel point meets the preset direction condition and the gradient amplitude of the pixel point meets the preset amplitude condition, determine the pixel point as a candidate pixel point; traverse the gradient amplitude of each candidate pixel point, and compare the gradient amplitude with a first threshold and a second threshold; when the gradient amplitude is greater than the first threshold, determine the candidate pixel point as a first edge pixel; when the gradient amplitude is greater than or equal to the second threshold and less than or equal to the first threshold, perform neighborhood detection on the candidate pixel point, and when the number of first edge pixels in the neighborhood of the candidate pixel point meets the condition, determine the candidate pixel point as a second edge pixel; integrate the first edge pixel and the second edge pixel to obtain an edge pixel point.

[0182] In an optional embodiment, the global detection module 603 is further configured to perform global detection on the edge image, determine a first discharge location of the bushing of the power device in the edge image, extract first feature information of the first discharge location, obtain a first discharge type and a first discharge position of the first discharge location based on the first feature information; and obtain a first detection result of the bushing of the power device in a discharge state based on the first discharge type and the first discharge position.

[0183] In an optional embodiment, the partition detection module 604 is further configured to determine a number of image segmentations, and perform image segmentation on the edge image according to the number of image segmentations to obtain a plurality of segmented images; for each segmented image, determine a second discharge location of the bushing of the power device in each segmented image, extract second feature information of the second discharge location, and obtain a second discharge type and a second discharge position of the second discharge location based on the second feature information; and obtain a second detection result of the bushing of the power device in a discharged state based on the second discharge type and the second discharge position.

[0184] Each module in the aforementioned power device bushing discharge detection device may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0185] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, memory and input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data such as ultraviolet images, visible light images, ultraviolet fusion images, historical fusion images, and discharge detection results. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a discharge detection method for a bushing of an electric power device is implemented.

[0186] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0187] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the method of the above embodiment when executing the computer program.

[0188] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method of the above embodiment are implemented.

[0189] In one embodiment, a computer program product is provided, comprising a computer program, which implements the steps of the method of the above embodiment when executed by a processor.

[0190] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0191] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0192] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0193] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for detecting discharge of a bushing of an electric device, characterized in that: The method comprises: Collecting an ultraviolet image and a visible light image of a bushing of an electric power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image; Performing edge detection on the casing of the electric device in the ultraviolet fusion image to obtain an edge image of the casing of the electric device; Performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, and obtaining a first detection result of the bushing of the power device in a discharge state based on the first discharge location; Performing image segmentation on the edge image to obtain a plurality of segmented images, determining a second discharge location of the bushing of the power device in each of the segmented images, and obtaining a second detection result of the bushing of the power device in the discharge state based on each of the second discharge locations; A discharge detection result of the bushing of the power device is determined according to the first detection result and the second detection result.

2. The method according to claim 1, characterized in that The collecting of the ultraviolet image and the visible light image of the bushing of the power device and fusing the ultraviolet image with the visible light image to obtain the ultraviolet fused image includes: Acquire an optical signal from a bushing of an electric device, and sequentially perform ultraviolet filtering and photoelectric conversion on the optical signal in a preset optical axis direction to obtain at least one ultraviolet image, and acquire a visible light image of the bushing of the electric device; When there is one ultraviolet image, fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image; When there are at least two ultraviolet images, for each of the ultraviolet images, the ultraviolet image is fused with the visible light image to obtain at least two candidate fused images; based on the historical fused image of the bushing of the power device, a qualified candidate fused image is determined from the at least two candidate fused images, and the qualified candidate fused image is determined as the ultraviolet fused image.

3. The method according to claim 2, characterized in that The fusing the ultraviolet image with the visible light image comprises: Performing grayscale processing on the ultraviolet image to obtain a first grayscale image, generating a first scale space of the first grayscale image, extracting first feature points of the ultraviolet image based on the first scale space, and generating a first feature descriptor corresponding to the first feature points; Performing grayscale processing on the visible light image to obtain a second grayscale image, generating a second scale space of the second grayscale image, and extracting second feature points of the visible light image based on the second scale space and generating second feature descriptors corresponding to the second feature points; Matching the first feature point with the second feature point based on the first feature descriptor and the second feature descriptor to obtain a feature matching matrix; A transformation matrix between the ultraviolet image and the visible light image is determined based on the feature matching matrix, and the ultraviolet image and the visible light image are fused using the transformation matrix.

4. The method according to claim 1, characterized in that The performing edge detection on the bushing of the electric device in the ultraviolet fusion image to obtain an edge image of the bushing of the electric device includes: Preprocessing the ultraviolet fusion image, and performing gradient calculation on the preprocessed ultraviolet fusion image to obtain gradient information of each pixel in the ultraviolet fusion image; Perform edge pixel detection on each pixel point based on the gradient information to obtain edge pixel points; An edge image for the bushing of the electric device is generated based on the edge pixel points.

5. The method according to claim 4, characterized in that The performing edge pixel detection on each pixel point based on the gradient information to determine the edge pixel point includes: Determine the gradient direction and gradient amplitude of each pixel point according to the gradient information; Traversing the gradient direction and the gradient amplitude of each pixel point; When the gradient direction of the pixel point meets the preset direction condition and the gradient amplitude of the pixel point meets the preset amplitude condition, determining the pixel point as a candidate pixel point; Traversing the gradient magnitude of each of the candidate pixel points, and comparing the gradient magnitude with a first threshold and a second threshold; When the gradient amplitude is greater than the first threshold, the candidate pixel is determined as a first edge pixel; when the gradient amplitude is greater than or equal to the second threshold and less than or equal to the first threshold, a neighborhood detection is performed on the candidate pixel, and when the number of the first edge pixels in the neighborhood of the candidate pixel meets the condition, the candidate pixel is determined as a second edge pixel; The first edge pixel and the second edge pixel are integrated to obtain an edge pixel point.

6. The method according to any one of claims 1 to 5, characterized in that: The performing global detection on the edge image, determining a first discharge position of the bushing of the power device in the edge image, and obtaining a first detection result of the bushing of the power device in a discharge state based on the first discharge position, includes: Performing global detection on the edge image, determining a first discharge location of the bushing of the power device in the edge image, extracting first feature information of the first discharge location, and obtaining a first discharge type and a first discharge position of the first discharge location based on the first feature information; A first detection result of the bushing of the electric device in a discharge state is obtained based on the first discharge type and the first discharge position.

7. The method according to any one of claims 1 to 5, characterized in that: The step of performing image segmentation on the edge image to obtain a plurality of segmented images, determining a second discharge location of the bushing of the power device in each of the segmented images, and obtaining a second detection result of the bushing of the power device in the discharge state based on each of the second discharge locations includes: Determining the number of image segmentations, and performing image segmentation on the edge image according to the number of image segmentations to obtain a plurality of segmented images; For each segmented image, determining a second discharge location of the bushing of the power device in each segmented image, extracting second feature information of the second discharge location, and obtaining a second discharge type and a second discharge position of the second discharge location based on the second feature information; A second detection result of the bushing of the electric device in a discharge state is obtained based on the second discharge type and the second discharge position.

8. A discharge detection device for a bushing of an electric power device, characterized in that: The device comprises: An image acquisition module, used for acquiring an ultraviolet image and a visible light image of the bushing of the power device, and fusing the ultraviolet image with the visible light image to obtain an ultraviolet fused image; An edge detection module, used for performing edge detection on the casing of the electric device in the ultraviolet fusion image to obtain an edge image of the casing of the electric device; a global detection module, configured to perform global detection on the edge image, determine a first discharge position of the bushing of the power device in the edge image, and obtain a first detection result of the bushing of the power device in a discharge state based on the first discharge position; a partition detection module, configured to perform image segmentation on the edge image to obtain a plurality of segmented images, determine a second discharge location of the bushing of the power device in each of the segmented images, and obtain a second detection result of the bushing of the power device in the discharge state based on each of the second discharge locations; The result fusion module is used to determine the discharge detection result of the bushing of the power device according to the first detection result and the second detection result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images