Power equipment fault detection method and device, equipment, storage medium and program product
By extracting the timing and spatial characteristics of power equipment and generating repair images using repair models, the problem of low accuracy of power equipment fault detection is solved, and higher quality fault detection is achieved.
Patent Information
- Application Number
- CN202410996048.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-15
AI Technical Summary
In the prior art, the accuracy of power equipment fault detection is low, mainly due to the influence of image quality by external environmental factors such as rain, fog, low light, and backlight, resulting in a decrease in image quality and affecting the accuracy of detection.
By collecting the initial image of the device to be detected, extracting its timing characteristics and spatial characteristics, and generating repair images in combination with the repair model, thereby performing fault detection. The method includes feature fusion, information extraction and the use of repair models to improve image quality and accurately analyze device status.
By generating high-quality repair images, potential faults can be identified more accurately, and the accuracy of power equipment fault detection is improved, which is more effective than traditional contrast enhancement methods.
Smart Images

Figure CN118967597A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power equipment detection, and in particular to a method, device, equipment, storage medium and program product for power equipment fault detection. Background Art
[0002] With the development of automation and intelligence of power systems, the use of image monitoring technology for fault detection of power equipment has become an important means to ensure the safe operation of power systems. However, due to the complex and changeable power production environment, the quality of collected images is often affected by various factors, such as rain, fog, low light, backlight and other external environments, which leads to a decrease in image quality and affects the accuracy of fault detection of power equipment.
[0003] In the prior art, image quality is enhanced by adjusting the contrast and brightness of an image, and the fault of the electric power equipment is detected using the enhanced image.
[0004] However, the above technology has the problem of low accuracy in detecting power equipment faults. Summary of the invention
[0005] Based on this, it is necessary to provide a method, device, equipment, storage medium and program product for power equipment fault detection that can improve the accuracy of power equipment fault detection in response to the above technical problems.
[0006] In a first aspect, the present application provides a method for detecting a fault in an electric power device, the method comprising:
[0007] Collecting an initial image of the device to be inspected;
[0008] Extract features from the initial image to determine the temporal features and spatial features corresponding to the initial image;
[0009] According to the time series characteristics, spatial characteristics and the repair model, the repair image corresponding to the initial image is determined, and fault detection is performed on the device to be inspected according to the repair image.
[0010] In one embodiment, determining the restored image corresponding to the initial image according to the temporal features, the spatial features and the restoration model includes:
[0011] Perform feature fusion on temporal features and spatial features to obtain spatiotemporal features;
[0012] The restoration image is determined according to the spatiotemporal characteristics and the restoration model.
[0013] In one embodiment, determining the repaired image according to the spatiotemporal features and the repair model includes:
[0014] Obtaining prior information corresponding to the spatiotemporal features, the prior information is related to the degradation process of the initial image;
[0015] The prior information and spatiotemporal features are input into the restoration model to obtain the restoration image output by the restoration model.
[0016] In one embodiment, the obtaining of prior information corresponding to the spatiotemporal features includes:
[0017] Extract information from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features;
[0018] Prior information is determined based on the degradation parameter information.
[0019] In one embodiment, the above-mentioned performing fault detection on the device to be detected according to the repaired image includes:
[0020] Determine a region to be detected from the repaired image, mark fault points in the region to be detected, and determine at least one fault point;
[0021] A fault detection result is determined according to a fault type and a fault location of at least one fault point.
[0022] In one embodiment, the method further comprises:
[0023] Generate alarm information based on the fault detection result and output the alarm information.
[0024] In one embodiment, the method further comprises:
[0025] Acquire a sample image set of a plurality of power devices;
[0026] Degrading the sample images in the sample image set to obtain a low-quality sample image set;
[0027] The model is trained using a set of sample images and a set of low-quality sample images to obtain a repair model.
[0028] In a second aspect, the present application also provides a power equipment fault detection device, the device comprising:
[0029] An acquisition module, used for acquiring an initial image of the device to be detected;
[0030] An extraction module is used to extract features from the initial image and determine the temporal features and spatial features corresponding to the initial image;
[0031] The repair detection module is used to determine the repair image corresponding to the initial image according to the time series characteristics, spatial characteristics and the repair model, and perform fault detection on the device to be detected according to the repair image.
[0032] In a third aspect, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method described in the first aspect are implemented.
[0033] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in the first aspect above are implemented.
[0034] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the method described in the first aspect above.
[0035] The above-mentioned power equipment fault detection method, device, equipment, storage medium and program product collect the initial image of the equipment to be detected, extract features from the initial image, determine the time series features and spatial features corresponding to the initial image, determine the repair image corresponding to the initial image according to the time series features, spatial features and repair model, and perform fault detection on the equipment to be detected according to the repair image. In this method, the time series features corresponding to the initial image reflect the dynamic features of the equipment to be detected, while the spatial features reflect the structure and appearance of the equipment. Comprehensive equipment status information is obtained according to the time series features and spatial features, which is helpful to identify potential fault signs or abnormalities. Moreover, the initial image is repaired to obtain a repaired image, which can more clearly show the features and details of the equipment to be detected, and can more accurately analyze the operating status and potential faults of the equipment to be detected. Compared with the method of contrast enhancement of the image before fault detection in the traditional technology, a high-quality repaired image can be obtained. Fault detection based on the high-quality repaired image can effectively improve the accuracy of power equipment fault detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0037] FIG1 is a schematic diagram of a flow chart of a method for detecting a fault in an electric power device according to an embodiment;
[0038] FIG2 is a schematic diagram of a flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0039] FIG3 is a schematic diagram of a flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0040] FIG4 is a schematic flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0041] FIG5 is a schematic flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0042] FIG6 is a schematic flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0043] FIG7 is a schematic diagram of a flow chart of a method for detecting a fault in an electric power device in another embodiment;
[0044] FIG8 is a block diagram of a device for detecting a fault in an electric power device according to an embodiment;
[0045] FIG. 9 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0046] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0047] With the development of automation and intelligence of power systems, the use of image monitoring technology for fault detection of power equipment has become an important means to ensure the safe operation of power systems. However, due to the complex and changeable power production environment, the quality of the collected images is often affected by various factors. The collected images are often affected by external environments such as rain, fog, low light, and backlight, resulting in a decrease in image quality, which in turn affects the accuracy of image analysis. For example, rainy and foggy weather will cause blurred images, low light environments will cause increased image noise and loss of details, and backlight environments will cause reduced image contrast, resulting in a decrease in image quality, affecting the accuracy of fault detection of power equipment.
[0048] In the prior art, the image of the power equipment is restored and the restored image is used to detect the fault of the power equipment. The traditional image restoration method can be divided into three categories: denoising, deblurring and enhancement. Among them, the denoising methods mainly include mean filtering, median filtering, Gaussian filtering, etc. These methods can remove image noise to a certain extent, but often lead to loss of image details in the denoising process; deblurring methods include inverse filtering, Wiener filtering, blind deconvolution, etc. These methods require the blur kernel of the image to be known or estimated, but in practical applications, the blur kernel is often difficult to obtain accurately; enhancement methods include histogram equalization, adaptive histogram equalization, etc. These methods enhance the image quality by adjusting the contrast and brightness of the image, but the effect is limited when processing low-light and backlit images. In summary, the above technologies have the problem of low accuracy in power equipment fault detection.
[0049] Therefore, the embodiments of the present application provide a method, device, equipment, storage medium and program product for detecting faults of electric power equipment, which can solve the above-mentioned technical problems.
[0050] The power equipment fault detection method provided in the embodiment of the present application can be used for computer equipment, which can be a server or a terminal. The server can be a single server or a server cluster composed of multiple servers. The terminal can be but is not limited to various personal computers, laptops, smart phones and tablet devices; the power equipment fault detection method provided in the embodiment of the present application can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server.
[0051] In one embodiment, as shown in FIG1 , a method for detecting a fault in an electric power device is provided. The method is described by taking the method for a computer device as an example. The method comprises the following steps:
[0052] S202, collecting an initial image of the device to be detected.
[0053] The device to be detected may be an electric power device in an electric power production scene, such as a generator set, a high-voltage transmission line, etc. The initial image of the device to be detected is an image of the device to be detected captured by a camera.
[0054] In an embodiment of the present application, the computer device may collect an initial image of the device to be detected through a camera installed around the power equipment, or the computer device may capture the initial image of the device to be detected from a monitoring video corresponding to the device to be detected.
[0055] S204, extracting features from the initial image to determine temporal features and spatial features corresponding to the initial image.
[0056] The spatial features corresponding to the initial image are used to characterize the static information in the image, that is, the content and structure of the image. The spatial features may include edge and texture features, shape features, color features, etc. The temporal features corresponding to the initial object are used to characterize the dynamic features of the image that change in time.
[0057] In the embodiment of the present application, the computer device can preprocess the initial image, including normalization, denoising, cropping, etc. of the image, and extract spatial features from the preprocessed initial image, for example, using an edge detection operator to detect the edge in the image to extract the edge features of the image, calculating the color distribution of the image, and extracting the color features of the image. A corner detector can be used to determine the corner points in the image and extract the shape features in the image. A texture extraction network can also be used to extract the texture features of the image to obtain texture features.
[0058] In an embodiment of the present application, when capturing an initial image, the computer device can obtain image frames that are continuous in time sequence with the initial image, perform motion detection on the initial image and the image frames continuous with the initial image, determine the dynamic features corresponding to the initial image by comparing the differences between consecutive frames, and extract the time sequence features corresponding to the initial image.
[0059] S206, determining a repaired image corresponding to the initial image according to the temporal features, the spatial features and the repair model, and performing fault detection on the device to be detected according to the repaired image.
[0060] Among them, the repair model is a pre-trained image restoration model, which is used to repair low-quality images into high-quality images. The repair model can be a repair model built based on a mean regression stochastic differential equation model.
[0061] In one possible implementation, a computer device may utilize a variational autoencoder to transform spatial features into a latent space, obtain a latent vector corresponding to the spatial features, input the latent vector into a restoration model, and during the process of the restoration model generating a restoration image, dynamically adjust the restoration image in combination with the temporal features to generate a final restoration image.
[0062] In another possible implementation, the computer device may fuse the initial features and the temporal features to obtain the spatiotemporal features corresponding to the initial image, input the spatiotemporal features into the restoration model, and generate the restoration image by the restoration model.
[0063] After the repair image is generated as described above, the computer device may perform fault detection on the device to be detected according to the repair image to determine whether the device to be detected has a fault.
[0064] In a possible implementation, a computer device can obtain a fault mark image corresponding to the device to be detected, wherein the fault mark image includes historical fault points of the device to be detected and regional coordinates of each historical fault point in the image. The computer device can align the device area to be detected in the repair image with the device area to be detected in the fault mark image, and extract a first image block within the corresponding regional coordinates from the repair image according to the regional coordinates of each historical fault point in the image, and compare the extracted first image block with a second image block corresponding to the historical fault point in the fault mark image. If the feature similarity between the extracted first image block and the second image block is greater than a similarity threshold, it is determined that a fault exists in the area corresponding to the first image block, thereby determining that a fault exists in the device to be detected.
[0065] In another possible implementation, the computer device may input the repaired image into a fault detection model, and the fault detection model determines whether there is a faulty area in the repaired image and determines whether there is a fault in the device to be detected.
[0066] In the above-mentioned power equipment fault detection method, by collecting the initial image of the device to be detected, the feature extraction of the initial image is performed, the time series feature and spatial feature corresponding to the initial image are determined, and the repair image corresponding to the initial image is determined according to the time series feature, the spatial feature and the repair model, and the fault detection of the device to be detected is performed according to the repair image. In this method, the time series feature corresponding to the initial image reflects the dynamic features of the device to be detected, and the spatial feature reflects the structure and appearance of the device. The comprehensive device status information is obtained according to the time series feature and the spatial feature, which is helpful to identify potential fault signs or abnormalities. Moreover, the initial image is repaired to obtain the repaired image, and the repair can more clearly show the features and details of the device to be detected, and can more accurately analyze the operating status and potential faults of the device to be detected. Compared with the method of contrast enhancement of the image and then fault detection in the traditional technology, a high-quality repaired image can be obtained, and fault detection based on the high-quality repaired image can effectively improve the accuracy of power equipment fault detection.
[0067] In the above embodiment, it is mentioned that the restoration image corresponding to the initial image is determined according to the temporal features, spatial features and restoration model. Next, an implementation method of determining the restoration image corresponding to the initial image according to the temporal features, spatial features and restoration model is described.
[0068] Based on the embodiment shown in FIG. 1 , referring to FIG. 2 , the above S106 may include the following steps:
[0069] S202, performing feature fusion on the temporal features and the spatial features to obtain spatiotemporal features.
[0070] Among them, the spatiotemporal features include time series features and space features. In the embodiment of the present application, the computer device can connect the time series features and the space features in the form of feature vectors to achieve feature serial fusion and obtain the spatiotemporal features. Alternatively, the computer device can perform feature selection on the time series features and the space features respectively, perform feature dimensionality reduction, and merge the time series features after dimensionality reduction and the space features after dimensionality reduction through an autoencoder to obtain the spatiotemporal features.
[0071] S204: Determine a restoration image according to the spatiotemporal characteristics and the restoration model.
[0072] In the embodiment of the present application, the restoration model includes a forward degradation model and a reverse restoration model, wherein the forward degradation model is used to convert a high-quality image into a low-quality image, and the reverse restoration model is used to convert a low-quality model into a high-quality image. The forward degradation model can be expressed by formula (1), and the reverse restoration model can be expressed by formula (2).
[0073] Formula (1)
[0074] in, is the diffusion term, which controls the flow direction of the stochastic differential equation; is the diffusion term, which controls the speed of noise addition; represents Brownian motion, i.e. Gaussian noise.
[0075] Formula (2)
[0076] in, is an unknown scoring function. In order to determine the scoring function, the computer device can input the degraded low-quality image into the residual network, estimate the scoring function through the residual network, and output the corresponding scoring function.
[0077] In a possible implementation, the computer device may input the spatiotemporal features into a reverse restoration model, and the reverse restoration model may generate a restoration image according to the input spatiotemporal features and an estimated scoring function.
[0078] In another possible implementation, the computer device may perform parameter estimation based on the spatiotemporal characteristics, estimate each model parameter in the reverse repair model, and use the estimated model parameters to input the spatiotemporal characteristics into the reverse repair model to generate a repaired image.
[0079] In this way, through the above method, by reasonably integrating temporal features and spatial features, spatiotemporal features are obtained, and a repaired image is generated according to the spatiotemporal features, which can reduce artifacts, distortion or other unreal parts in the repair process and make the repaired image clearer and more realistic.
[0080] In the above embodiment, it is mentioned that the restoration image is determined according to the spatiotemporal characteristics and the restoration model. Next, an implementation method of determining the restoration image according to the spatiotemporal characteristics and the restoration model by a computer device is described.
[0081] Based on the embodiment shown in FIG. 2 , referring to FIG. 3 , the above S204 may include the following steps:
[0082] S302, obtaining prior information corresponding to the spatiotemporal features.
[0083] Among them, the prior information is related to the degradation process of the initial image, and the prior information may be an influencing parameter that affects the image quality of the initial image of the device to be detected. For example, the prior information may include the noise level corresponding to the initial image, and / or the environmental information when the initial image is taken, such as rain level, visibility, low light moments, backlight moments, etc.
[0084] In an embodiment of the present application, the computer device can determine the device parameters of the shooting device when the initial image is taken, determine the noise level corresponding to the initial image based on the device parameters, and can also obtain the time information when the initial image is taken, determine the environmental information corresponding to the initial image based on the time information, so as to obtain prior information corresponding to the spatiotemporal characteristics.
[0085] In a possible implementation, a computer device may obtain a noise level corresponding to an initial image and / or environmental information corresponding to the initial image from a process of capturing the initial image, and determine the noise level and / or environmental information as prior information corresponding to the spatiotemporal features.
[0086] In another possible implementation, the computer device may perform a priori estimation of the spatiotemporal features to obtain a priori information corresponding to the spatiotemporal features.
[0087] S304, inputting the prior information and the spatiotemporal features into the restoration model to obtain a restoration image output by the restoration model.
[0088] In an embodiment of the present application, prior information can be used to determine each model parameter in the repair model. The computer device can input the prior information into the repair model to determine each model parameter in the repair model. Based on each model parameter in the repair model, the computer device can input spatiotemporal characteristics into the repair model, and the repair model generates a repaired image.
[0089] In this way, through the above method, the prior information corresponding to the spatiotemporal features is obtained. The prior information can help the restoration model better understand the content and structure of the image. By understanding the possible process of image degradation, the model can infer the content and features of the missing area based on the prior information, thereby generating a more accurate restoration image.
[0090] The above embodiment mentions obtaining the prior information corresponding to the spatiotemporal features. Next, an implementation method of obtaining the prior information corresponding to the spatiotemporal features by a computer device is described.
[0091] Based on the embodiment shown in FIG. 3 , referring to FIG. 4 , the above S302 may include the following steps:
[0092] S402: extract information from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features.
[0093] The degradation parameter information includes the noise level and / or degradation factor corresponding to the spatiotemporal characteristics, and the degradation factor is a factor that affects the image quality due to the external environment.
[0094] In the embodiment of the present application, the computer device can pre-train the information extraction model, the input of the information extraction model is the spatiotemporal features of the image, and the output is the degradation parameter information corresponding to the spatiotemporal features. The information extraction model can be used to characterize the mapping relationship between the spatiotemporal features and the degradation parameter information corresponding to the spatiotemporal features. The computer device can input the spatiotemporal features into the information extraction model to obtain the degradation parameter information corresponding to the spatiotemporal features.
[0095] S404: Determine prior information according to the degradation parameter information.
[0096] In an embodiment of the present application, the computer device may perform statistical analysis on the degradation parameter information to obtain prior information. Exemplarily, the noise level in the degradation parameter information is used as the noise level in the prior information, and the degradation factor in the degradation parameter information is statistically analyzed to determine the environmental information in the prior information.
[0097] In this way, through the above-mentioned method, by extracting spatiotemporal features and analyzing degradation parameter information, the prior information can be determined more accurately, so that the image restoration can be carried out in a targeted manner according to the prior information, thereby improving the quality of the image restoration.
[0098] In the above embodiment, it is mentioned that fault detection is performed on the device to be detected according to the repaired image. Next, an implementation method of performing fault detection on the device to be detected according to the repaired image by a computer device is described.
[0099] Based on the embodiment shown in FIG1 , referring to FIG5 , the above S106 may further include the following steps:
[0100] S502, determining a region to be detected from the repaired image, marking fault points in the region to be detected, and determining at least one fault point.
[0101] Wherein, the area to be detected is the area of the device to be detected in the image, and the device to be detected is at the center of the area to be detected. In an embodiment of the present application, the computer device can segment the area to be detected from the repaired image using an image segmentation network. After determining the area to be detected, the computer device can obtain a fault mark image, which includes the historical fault points of the device to be detected and their coordinate information in the image. The historical fault points may be a sign of a known device problem. Using image registration technology, the area of the device to be detected in the fault mark image is aligned with the area to be detected in the repaired image. According to the coordinate information of the historical fault points in the fault mark image, the corresponding first image block is extracted from the area to be detected in the repaired image, and the image area of each historical fault point in the fault mark image is extracted as a second image block, and the size and position of the first image block correspond to the second image block. Features are extracted from the first image block and the second image block, such as local features, global features, or feature vectors extracted by a deep learning model. Whether the first image block is similar to the second image block is determined by comparing the similarity of these features. Commonly used similarity metrics include Euclidean distance, cosine similarity, etc. If the feature similarity is greater than a preset similarity threshold, it is considered that there is a fault feature similar to the second image block in the first image block. The computer device can mark the first image block similar to the second image block as a fault point to determine at least one fault point in the area to be detected.
[0102] S504: Determine a fault detection result according to the fault type and fault location of at least one fault point.
[0103] In the embodiment of the present application, a fault point corresponds to a first image block in the area to be detected, and the first image block corresponds to the second image block, and each second image block corresponds to a historical fault point, so a fault point can correspond to a historical fault point, and the computer device can determine the historical fault type of the historical fault point corresponding to the fault point as the fault type of the fault point in the area to be detected, thereby obtaining the fault type of each fault point. In addition, the computer device can determine the position of the first image block in the area to be detected as the position of the fault point in the area to be detected, and use the coordinate relationship between the area to be detected and the repaired image to convert the position of the fault point in the area to be detected to the position of the fault point in the repaired image, and determine the position of the fault point in the repaired image as the fault position, and determine the fault type and fault position as the fault detection result.
[0104] As an optional embodiment, the computer device may generate warning information according to the fault detection result and output the warning information, wherein the warning information includes the initial image, the repaired image, the fault type and the fault location.
[0105] In this way, by marking the fault point in the detection area in the above manner, the accuracy of fault detection can be ensured. By obtaining the fault type and fault location of the fault point, the fault can be quickly located, improving the timeliness of fault detection. By giving an alarm for the fault detection result, a quick fault repair response can be made to avoid further expansion of the fault and serious damage to the equipment.
[0106] The above embodiment mentions a repair model. Next, an implementation method of training the repair model by a computer device is described.
[0107] Referring to FIG. 6 , the method for detecting a fault in an electric power device in this embodiment may include the following steps:
[0108] S602: Acquire a set of sample images of multiple power devices.
[0109] In an embodiment of the present application, a computer device can collect image data of various power equipment through a camera to obtain a sample image set of multiple power equipment, and the sample image set of multiple power equipment includes images of multiple power equipment in normal or faulty states.
[0110] S604: Degrade the sample images in the sample image set to obtain a low-quality sample image set.
[0111] In an embodiment of the present application, the computer device may preprocess the images in the sample image set, including image normalization, denoising, cropping, etc., to ensure that the image data quality meets the requirements of model training and reasoning. The preprocessed image data set is used to train the neural network model, and the network parameters are continuously optimized so that the neural network model can accurately estimate the score loss function. At the same time, for the images in the preprocessed sample image set, the forward degradation model is used to degrade the images in the sample image set, and Gaussian noise is added to obtain a low-quality sample image set. The obtained low-quality sample image set and the images in the sample image set are used as training sample pairs to obtain training data for the reverse repair model.
[0112] S606: Perform model training using the sample image set and the low-quality sample image set to obtain a repair model.
[0113] The estimated score loss function and training data are used to train the reverse restoration model, and the restoration model is trained to learn the mapping process from low-quality images to high-quality images. During the training process, the model is trained based on the maximum likelihood loss function, and the formula of the maximum likelihood loss function is shown in formula (3).
[0114] Formula (3)
[0115] in, is the image at time step i, is the noise predicted by the neural network, is the reverse state obtained through optimization, is the weight coefficient, and T is the total number of time steps.
[0116] During the iterative training process, common image quality evaluation indicators (such as PSNR, SSIM, etc.) are used to quantify the image restoration effect. According to the restoration results, the restoration model and model parameters are further optimized to improve the accuracy and stability of image restoration. The trained restoration model is deployed to the power equipment monitoring platform to achieve real-time image restoration and fault detection. In the actual power production scenario, the images collected by the real-time monitoring camera are repaired in time to ensure image quality and improve the monitoring and fault detection capabilities of power equipment.
[0117] In this way, through the above method, the image degradation process can be simulated, a set of low-quality sample images can be constructed, and the model training can be carried out using the sample image set and the low-quality sample image set. The restoration model can learn the image distribution of low-quality images in real scenes, thereby improving the image restoration effect of the restoration model.
[0118] In one embodiment, a method for detecting a fault in an electric power device is provided, which is used for a computer device. Referring to FIG. 7 , the method comprises the following steps:
[0119] S701, collecting an initial image of the device to be detected;
[0120] S702, extracting features from the initial image to determine temporal features and spatial features corresponding to the initial image;
[0121] S703, performing feature fusion on the temporal features and the spatial features to obtain spatiotemporal features;
[0122] S704, extracting information from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features;
[0123] S705, determining prior information according to the degradation parameter information;
[0124] S706, inputting the prior information and the spatiotemporal features into the restoration model to obtain a restoration image output by the restoration model;
[0125] S707, determining a region to be detected from the repaired image, marking fault points in the region to be detected, and determining at least one fault point;
[0126] S708, determining a fault detection result according to a fault type and a fault location of at least one fault point;
[0127] S709, generating alarm information according to the fault detection result, and outputting the alarm information.
[0128] The following is an exemplary description of the implementation of the power equipment fault detection method according to the above embodiment.
[0129] As shown in FIG7 , an initial image of the device to be detected is acquired, and feature extraction is performed on the initial image to determine the temporal features and spatial features corresponding to the initial image. The temporal features and spatial features are feature fused to obtain spatiotemporal features. Information extraction is performed on the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features. Prior information is determined based on the degradation parameter information. The prior information and the spatiotemporal features are input into the repair model to obtain a repaired image output by the repair model. The area to be detected is determined from the repaired image, and the fault point of the area to be detected is marked to determine at least one fault point. According to the fault type and fault location of at least one fault point, the fault detection result is determined. Alarm information is generated based on the fault detection result, and the alarm information is output.
[0130] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed 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 part of the steps or stages in other steps.
[0131] Based on the same inventive concept, the embodiment of the present application also provides an electric equipment fault detection device for implementing the electric equipment fault detection method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the electric equipment fault detection device provided below can refer to the limitations of the electric equipment fault detection method above, and will not be repeated here.
[0132] In one embodiment, as shown in FIG8 , a power equipment fault detection device is provided, including: a collection module 801, an extraction module 802, and a repair detection module 803, wherein:
[0133] The acquisition module 801 is used to acquire an initial image of the device to be detected;
[0134] An extraction module 802 is used to extract features from the initial image and determine temporal features and spatial features corresponding to the initial image;
[0135] The repair detection module 803 is used to determine the repair image corresponding to the initial image according to the time series characteristics, the spatial characteristics and the repair model, and perform fault detection on the device to be detected according to the repair image.
[0136] In another embodiment, another power equipment fault detection device is provided. Based on the above embodiment, the above repair detection module 803 includes a fusion unit and a repair unit, wherein:
[0137] A fusion unit is used to fuse the temporal features and spatial features to obtain spatiotemporal features;
[0138] The restoration unit is used to determine the restoration image according to the spatiotemporal characteristics and the restoration model.
[0139] In another embodiment, another power equipment fault detection device is provided. Based on the above embodiment, the above repair unit includes an acquisition subunit and a repair subunit, wherein:
[0140] An acquisition subunit, used to acquire prior information corresponding to the spatiotemporal features, the prior information being related to the degradation process of the initial image;
[0141] The restoration subunit is used to input the prior information and the spatiotemporal features into the restoration model to obtain the restoration image output by the restoration model.
[0142] Optionally, the acquisition subunit is specifically used to: extract information from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features; and determine prior information based on the degradation parameter information.
[0143] In another embodiment, another power equipment fault detection device is provided. Based on the above embodiment, the above repair detection module 803 further includes a detection unit and a determination unit, wherein:
[0144] A detection unit, used to determine a region to be detected from the repaired image, mark a fault point in the region to be detected, and determine at least one fault point;
[0145] The determination unit is used to determine the fault detection result according to the fault type and fault location of at least one fault point.
[0146] In another embodiment, another power equipment fault detection device is provided. Based on the above embodiment, the device further includes an alarm module, wherein:
[0147] The alarm module is used to generate alarm information according to the fault detection result and output the alarm information.
[0148] In another embodiment, another power equipment fault detection device is provided. Based on the above embodiment, the device further includes a sample acquisition module, a degradation module and a training module, wherein:
[0149] A sample acquisition module, used to acquire a sample image set of multiple power devices;
[0150] A degradation module, used for degrading the sample images in the sample image set to obtain a low-quality sample image set;
[0151] The training module is used to perform model training using a sample image set and a low-quality sample image set to obtain a repair model.
[0152] Each module in the above-mentioned power equipment fault detection device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.
[0153] In an exemplary embodiment, a computer device is provided, which can be a terminal or a server. Taking a terminal as an example, its internal structure diagram can be shown in FIG9. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, 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 and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for detecting faults of power equipment is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a key, a trackball or a touch pad set on the housing of the computer device, or an external keyboard, touch pad or mouse, etc.
[0154] Those skilled in the art will understand that the structure shown in FIG. 9 is merely a block diagram of a partial 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 arrangement of components.
[0155] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0156] Collect the initial image of the device to be detected; extract features from the initial image to determine the temporal features and spatial features corresponding to the initial image; determine the repair image corresponding to the initial image based on the temporal features, spatial features and repair model, and perform fault detection on the device to be detected based on the repair image.
[0157] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0158] The temporal features and spatial features are fused to obtain the spatiotemporal features; the repaired image is determined based on the spatiotemporal features and the repair model.
[0159] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0160] Prior information corresponding to the spatiotemporal features is obtained, where the prior information is related to the degradation process of the initial image; the prior information and the spatiotemporal features are input into a restoration model to obtain a restoration image output by the restoration model.
[0161] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0162] Information is extracted from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features; and prior information is determined based on the degradation parameter information.
[0163] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0164] Determine a region to be detected from the repaired image, mark fault points in the region to be detected, and determine at least one fault point; and determine a fault detection result according to the fault type and fault position of the at least one fault point.
[0165] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0166] Generate alarm information based on the fault detection result and output the alarm information.
[0167] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0168] A sample image set of multiple power equipment is obtained; sample images in the sample image set are degraded to obtain a low-quality sample image set; and a model is trained using the sample image set and the low-quality sample image set to obtain a repair model.
[0169] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0170] Collect the initial image of the device to be detected; extract features from the initial image to determine the temporal features and spatial features corresponding to the initial image; determine the repair image corresponding to the initial image based on the temporal features, spatial features and repair model, and perform fault detection on the device to be detected based on the repair image.
[0171] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0172] The temporal features and spatial features are fused to obtain the spatiotemporal features; the repaired image is determined based on the spatiotemporal features and the repair model.
[0173] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0174] Prior information corresponding to the spatiotemporal features is obtained, where the prior information is related to the degradation process of the initial image; the prior information and the spatiotemporal features are input into a restoration model to obtain a restoration image output by the restoration model.
[0175] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0176] Information is extracted from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features; and prior information is determined based on the degradation parameter information.
[0177] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0178] A region to be detected is determined from the repaired image, and a fault point is marked on the region to be detected to determine at least one fault point; and a fault detection result is determined according to the fault type and fault position of the at least one fault point.
[0179] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0180] Generate alarm information based on the fault detection result and output the alarm information.
[0181] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0182] A sample image set of multiple power equipment is obtained; sample images in the sample image set are degraded to obtain a low-quality sample image set; and a model is trained using the sample image set and the low-quality sample image set to obtain a repair model.
[0183] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0184] Collect the initial image of the device to be detected; extract features from the initial image to determine the temporal features and spatial features corresponding to the initial image; determine the repair image corresponding to the initial image based on the temporal features, spatial features and repair model, and perform fault detection on the device to be detected based on the repair image.
[0185] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0186] The temporal features and spatial features are fused to obtain the spatiotemporal features; the repaired image is determined based on the spatiotemporal features and the repair model.
[0187] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0188] Prior information corresponding to the spatiotemporal features is obtained, where the prior information is related to the degradation process of the initial image; the prior information and the spatiotemporal features are input into a restoration model to obtain a restoration image output by the restoration model.
[0189] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0190] Information is extracted from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features; and prior information is determined based on the degradation parameter information.
[0191] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0192] Determine a region to be detected from the repaired image, mark fault points in the region to be detected, and determine at least one fault point; and determine a fault detection result according to the fault type and fault position of the at least one fault point.
[0193] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0194] Generate alarm information based on the fault detection result and output the alarm information.
[0195] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0196] A sample image set of multiple power equipment is obtained; sample images in the sample image set are degraded to obtain a low-quality sample image set; and a model is trained using the sample image set and the low-quality sample image set to obtain a repair model.
[0197] It should be noted that the data involved in this application (including but not limited to data used for analysis, storage, display, etc.) are all data fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0198] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and 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. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present 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. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0199] The technical features of the above embodiments may be combined arbitrarily. 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.
[0200] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for detecting a fault in an electric power device, characterized in that: The method comprises: Collecting an initial image of the device to be inspected; Extracting features from the initial image to determine temporal features and spatial features corresponding to the initial image; A repair image corresponding to the initial image is determined according to the time series feature, the spatial feature and the repair model, and fault detection is performed on the device to be detected according to the repair image.
2. The method according to claim 1, characterized in that The determining, according to the temporal features, the spatial features and the restoration model, a restoration image corresponding to the initial image comprises: Performing feature fusion on the temporal features and the spatial features to obtain spatiotemporal features; The restored image is determined according to the spatiotemporal characteristics and the restoration model.
3. The method according to claim 2, characterized in that Determining the restored image according to the spatiotemporal features and the restoration model includes: Acquire prior information corresponding to the spatiotemporal feature, wherein the prior information is related to a degradation process of the initial image; The prior information and the spatiotemporal features are input into the restoration model to obtain the restoration image output by the restoration model.
4. The method according to claim 3, characterized in that The obtaining of the prior information corresponding to the spatiotemporal features includes: Extracting information from the spatiotemporal features to obtain degradation parameter information corresponding to the spatiotemporal features; The prior information is determined according to the degradation parameter information.
5. The method according to claim 1, characterized in that The performing fault detection on the device to be detected according to the repair image includes: Determine a region to be detected from the repaired image, and mark the fault points of the region to be detected to determine at least one fault point; A fault detection result is determined according to the fault type and the fault location of the at least one fault point.
6. The method according to claim 5, characterized in that The method further comprises: Generate warning information according to the fault detection result, and output the warning information.
7. The method according to claim 1, characterized in that The method further comprises: Acquire a sample image set of a plurality of power devices; Degrading the sample images in the sample image set to obtain a low-quality sample image set; The sample image set and the low-quality sample image set are used to perform model training to obtain the restoration model.
8. A power equipment fault detection device, characterized in that: The device comprises: An acquisition module, used for acquiring an initial image of the device to be detected; An extraction module, used to extract features from the initial image and determine temporal features and spatial features corresponding to the initial image; A repair detection module is used to determine a repair image corresponding to the initial image according to the temporal characteristics, the spatial characteristics and the repair model, and perform fault detection on the device to be detected according to the repair image.
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.
11. A computer program product, comprising a computer program, 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.
Citation Information
Patent Citations
Intelligent power plant electrical equipment fault diagnosis method based on width learning and temporal and spatial characteristics of infrared images
CN109870617A
Infrared image fault detection method and device for power equipment
CN111986172A
Weather type image detection and foggy day image processing method and device
CN115620119A
Power equipment monitoring method based on image restoration
CN116311040A
