CN116129345A - Method and device for detecting oil level of conservator of transformer and computer equipment - Google Patents
Method and device for detecting oil level of conservator of transformer and computer equipment Download PDFInfo
- Publication number
- CN116129345A CN116129345A CN202211516889.XA CN202211516889A CN116129345A CN 116129345 A CN116129345 A CN 116129345A CN 202211516889 A CN202211516889 A CN 202211516889A CN 116129345 A CN116129345 A CN 116129345A
- Authority
- CN
- China
- Prior art keywords
- infrared image
- image
- oil
- enhanced
- preset
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The application relates to a method and a device for detecting the oil level of a conservator of a transformer and computer equipment. The method comprises the following steps: acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image; if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm. By adopting the scheme, the oil level detection accuracy can be improved.
Description
Technical Field
The present application relates to the field of transformers, and in particular, to a method, an apparatus, a computer device, a storage medium, and a computer program product for detecting an oil level of a conservator of a transformer.
Background
The power transformer is used as an important component of a power system and plays a role in junction of power transmission and electric energy level conversion, while most of transformers in China are oil-immersed transformers, the oil level is important characteristic data for judging the running state of the transformers, and the transformers are possibly not normally operated due to the fact that the oil level is too high or too low.
At present, an oil level gauge is adopted in an oil immersed transformer, the oil level is detected through a connectivity principle, and the phenomenon of false oil level is easily caused by breather blockage, pointer blocking and the like.
The existing method for detecting the oil level of the transformer by using the infrared image is greatly influenced by weather, particularly the infrared image is difficult to clearly identify in rainy days, and the algorithm effect for extracting the oil conservator area is poor, so that the accuracy of the oil level detection result is low.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, a computer-readable storage medium, and a computer program product for detecting the oil level of a wick of a transformer capable of improving the accuracy of oil level detection.
In a first aspect, the present application provides a method for detecting a conservator oil level of a transformer. The method comprises the following steps:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image;
if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image;
extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm;
and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
In one embodiment, if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model, and obtaining the enhanced infrared image includes:
acquiring a total loss function of a preset CycleGAN model;
training a preset CycleGAN model according to the total loss function; if the denoised infrared image is an infrared image corresponding to a rainy day, the denoised infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In one embodiment, the obtaining the total loss function of the preset CycleGAN model includes:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model;
and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
In one embodiment, the extracting, by the YOLO target detection algorithm, the image of the region where the oil pillow is located in the enhanced infrared image includes:
extracting a characteristic image in the enhanced infrared image through a characteristic extraction part of a YOLO model, wherein the characteristic extraction part comprises a plurality of module layers, and each module layer consists of a convolutional neural network layer, a Batch-nomination layer and a residual error connection layer;
performing enhanced extraction on the feature image by adopting a feature pyramid to obtain an enhanced feature image;
and extracting an area image of the oil pillow in the enhanced feature image through an effective feature layer of the YOLO model, wherein the effective feature layer consists of at least two module layers on the output side of the YOLO model.
In one embodiment, the extracting, by the YOLO target detection algorithm, the image of the region where the oil pillow is located in the enhanced infrared image includes:
acquiring a total loss function of the YOLO model;
training the YOLO model by a gradient descent method according to the total loss function;
and extracting an area image of the oil pillow in the enhanced infrared image through a trained YOLO model.
In one embodiment, the detecting the oil level of the oil conservator by using a temperature measurement algorithm includes:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along a preset direction in the region images of the oil conservers;
extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line;
and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
In a second aspect, the application also provides a device for detecting the oil level of the oil conservator of the transformer. The device comprises:
the noise-reducing infrared image acquisition module is used for acquiring an infrared image of the transformer, filtering random noise in the infrared image through a bilateral filtering algorithm, and obtaining a noise-reducing infrared image;
the enhanced infrared image acquisition module is used for enhancing the noisy infrared image through a preset CycleGAN model if the noisy infrared image is an infrared image corresponding to a rainy day, so as to obtain an enhanced infrared image;
the image extraction module is used for extracting an area image of the oil conservator in the enhanced infrared image through a YOLO target detection algorithm;
and the oil level detection module is used for detecting the oil level of the oil conservator through a temperature measurement algorithm on the image of the area where the oil conservator is located.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image;
if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image;
extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm;
and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image;
if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image;
extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm;
and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image;
if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image;
extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm;
and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
According to the oil level detection method, the oil level detection device, the computer equipment, the storage medium and the computer program product of the transformer, the infrared image of the transformer is obtained, random noise in the infrared image is filtered through a bilateral filtering algorithm, and a noisy infrared image is obtained; if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located by a temperature measurement algorithm. In the scheme, through presetting a CycleGAN model, the over-noise infrared image corresponding to the rainy day is enhanced, the influence of the rainy day on the infrared image is reduced, then, through a YOLO target detection algorithm, the image of the region where the oil pillow is located in the enhanced infrared image is extracted, and the interference of the edge useless information on oil level detection is reduced, so that the oil level detection accuracy is improved.
Drawings
FIG. 1 is an application environment diagram of a method for detecting the oil level of a conservator of a transformer in one embodiment;
FIG. 2 is a flow chart of a method of detecting a wick oil level of a transformer in one embodiment;
FIG. 3 is a schematic diagram of a network structure of the YOLO model;
fig. 4 is a schematic diagram of a specific flow of a method for detecting the oil level of a conservator of a transformer;
fig. 5 is a flow chart of a method for detecting the oil level of a conservator of a transformer in yet another embodiment;
FIG. 6 is a block diagram of a wick oil level detection apparatus of a transformer in one embodiment;
fig. 7 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The oil level detection method for the oil conservator of the transformer can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The terminal 102 acquires an infrared image of the transformer, sends the infrared image of the transformer to the server 104, the server 104 filters random noise in the infrared image sent by the terminal 102 through a bilateral filtering algorithm to obtain a denoised infrared image, if the denoised infrared image is an infrared image corresponding to rainy days, the denoised infrared image is enhanced through a preset CycleGAN model to obtain an enhanced infrared image, an area image of the oil conservator in the enhanced infrared image is extracted through a YOLO target detection algorithm, and oil level detection is carried out on the area image of the oil conservator through a temperature measurement algorithm. Further, the server 104 may also feed back the results of the detection of the oil level of the wick to the terminal 102. The terminal 102 may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.
In one embodiment, as shown in fig. 2, there is provided a method for detecting the oil level of a conservator of a transformer, which is described by taking the application of the method to the server 104 in fig. 1 as an example, and includes the following steps:
s200, acquiring an infrared image of the transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noisy infrared image.
The bilateral filtering algorithm adopts a weighted average method, the weighted average of the brightness values of the peripheral pixels is used for representing the intensity of a certain pixel, and the weighted average is based on Gaussian distribution, so that the effects of edge maintenance and noise reduction smoothness can be achieved; noise is unnecessary or redundant interference information present in the infrared image data of the transformer.
Specifically, in the infrared image of the transformer, a view size is selected, for example, 3×3, the center of the view is a pixel point to be updated, other pixels in the view include pixel values to be given with weights, coordinate axes are respectively established along the horizontal direction and the vertical downward direction, the horizontal X-axis is called, the vertical downward Y-axis is called, each pixel has a coordinate value and a pixel value, and the filtering formula is shown in the following formula.
I(p)＝f(i,j)
I(q)＝f(m,n)
Where p is the coordinates of the pixel value to be updated, q is the coordinates of the pixel value to be weighted, I (p) is the magnitude of the pixel value to be updated, I (q) is the magnitude of the pixel value to be weighted, wp is the normalization parameter, and G is a gaussian function.
And S400, if the denoised infrared image is an infrared image corresponding to a rainy day, enhancing the denoised infrared image through a preset CycleGAN model to obtain an enhanced infrared image.
The CycleGAN (Cycle Generative Adversarial Networks) model can realize mutual conversion of images, convert a rainy day infrared image into a sunny day infrared image, and convert the sunny day infrared image into a rainy day infrared image.
Specifically, whether the weather in the process of the over-noise infrared image is rainy or not is judged, if so, the trained CycleGAN model is used for enhancing the over-noise infrared image, otherwise, the over-noise infrared image is not enhanced.
S600, extracting and enhancing an area image of the oil pillow in the infrared image through a YOLO target detection algorithm.
The YOLO target detection algorithm adopts a single CNN (Convolutional Neural Networks, convolutional neural network) model to realize end-to-end target detection, and the core idea is to directly return the position of a boundingbox (bounding box) and the category to which the boundingbox belongs at an output layer by using the whole graph as the input of the network.
Specifically, the feature extraction part of the YOLO model is used for extracting the features of the enhanced infrared image, then an FPN feature pyramid is constructed for enhancing the feature extraction of the enhanced infrared image, finally the effective feature layers of the convolutional neural network layer, the Batch-localization layer and the residual connecting layer are subjected to feature enhancement, and then the region image of the oil pillow in the enhanced infrared image is extracted, wherein the network structure of the YOLO model is shown in figure 3.
S800, detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
Specifically, in an area image where the oil conservator is located in the enhanced infrared image, temperature measuring lines are arranged at equal intervals along the preset longitudinal direction, adjacent pixel points with the largest change on the temperature measuring lines are extracted from the temperature measuring lines to serve as results of the temperature measuring lines, and an average value of all the results is taken as an oil level detection result of the oil conservator.
The flow of the method for detecting the oil level of the conservator of the whole transformer is shown in fig. 4.
According to the oil level detection method for the oil conservator of the transformer, the infrared image of the transformer is obtained, random noise in the infrared image is filtered through a bilateral filtering algorithm, and a noisy infrared image is obtained; if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; the oil level of the oil conservator is detected by the temperature measurement algorithm on the image of the area where the oil conservator is located, so that the influence of rainy days on the infrared image and the interference of the edge useless information on the oil level detection can be reduced, and the oil level detection accuracy is improved.
In one embodiment, if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image by presetting a CycleGAN model, and obtaining the enhanced infrared image includes:
acquiring a total loss function of a preset CycleGAN model; training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
The loss function is used for measuring the deviation degree of the enhanced infrared image output by the CycleGAN model and the real rainy infrared image.
Specifically, the total loss function includes three parts, namely a generator loss, a discriminator loss and a cycle consistency loss, wherein the generator loss hopes that an image generated by the generator makes the discriminator make mistakes as much as possible, and the image is discriminated as a real picture; the loss of the discriminator hopes that the discriminator can distinguish the generated picture and the real picture; the loss of the loop consistency hopes that the original information is kept as much as possible after the original picture is converted twice, so that the problem of mode collapse is prevented.
The method comprises the steps of enhancing an over-noise infrared image, firstly, respectively collecting infrared images of a transformer in a sunny day and a transformer in a rainy day, wherein the infrared image of the transformer in the rainy day is called input-A, the infrared image of the transformer in the sunny day is called input-B, the infrared image in the rainy day is converted into a trained CycleGAN model, the converted infrared image in the sunny day is called generate-B, the infrared image in the rainy day, which is converted into the code-A by the trained CycleGAN model, is called cycle-A, and the generated infrared image in the rainy day is converted into the cycle-B by the trained CycleGAN model.
Training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In this embodiment, the preset CycleGAN model is trained by acquiring the total loss function of the preset CycleGAN model, if the denoising infrared image is an infrared image corresponding to a rainy day, the denoising infrared image is enhanced by the trained CycleGAN model to obtain an enhanced infrared image, and the denoising infrared image in the rainy day can be enhanced by the CycleGAN model, so that the influence of the rainy day on the infrared image is reduced.
In one embodiment, obtaining the total loss function of the preset CycleGAN model includes:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model; and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
Specifically, the total loss function is as follows:
L(G,F,D X ,D Y )＝L GAN (G,D X ,X,Y)+L GAN (G,D Y ,X,Y)+λL cyc (G,F)
wherein G, F is a generator, G is a generator for converting a rainy day picture into a sunny day picture, F is a generator for converting a sunny day picture into a rainy day picture, D is a discriminator, D X 、D Y The device is a discriminator for discriminating the infrared image in rainy days and the infrared image in sunny days respectively. L (L) GAN (F,D X X, Y) represents the loss of a rainy picture to a sunny picture, L GAN (G,D Y X, Y) represents the loss of a clear picture to a rainy picture, L cyc (G, F) represents an original rainy day picture, the corresponding sunny day picture is generated through the generator G, and then the loss of the corresponding rainy day picture is generated through the generator F.
In this embodiment, by acquiring the generator loss, the discriminator loss and the cycle consistency loss of the preset CycleGAN model, a total loss function of the preset CycleGAN model is acquired according to the generator loss, the discriminator loss and the cycle consistency loss, and the trained CycleGAN model can be measured through the total loss function.
In one embodiment, as shown in fig. 5, S600 includes:
s620, extracting the characteristic image in the enhanced infrared image through the characteristic extraction part of the YOLO model.
And S640, performing enhanced extraction on the feature image by adopting the feature pyramid to obtain an enhanced feature image.
S660, extracting an area image of the oil pillow in the reinforced feature image through an effective feature layer of the YOLO model.
Specifically, firstly training a YOLO model, extracting and enhancing a characteristic image in an infrared image through a characteristic extraction part of the trained YOLO model, wherein the characteristic extraction part of the YOLO model consists of a convolutional neural network layer, a Batch-localization layer and a residual error connection layer, then constructing a FPN (Feature Pyramid Networks, characteristic pyramid network) characteristic pyramid to conduct enhanced characteristic extraction, and finally predicting an effective characteristic layer by utilizing a YOLO Head.
In this embodiment, the feature extraction portion of the YOLO model is used to extract the feature image in the enhanced infrared image, the feature pyramid is used to perform enhanced extraction on the feature image, so as to obtain an enhanced feature image, and the effective feature layer of the YOLO model is used to extract the region image of the oil conservator in the enhanced feature image, so that the region image of the oil level detection of the oil conservator can be obtained.
In one embodiment, extracting the image of the region where the oil pillow is located in the enhanced infrared image by the YOLO target detection algorithm includes:
acquiring a total loss function of the YOLO model; training a YOLO model by a gradient descent method according to the total loss function; and extracting and enhancing an area image of the oil pillow in the infrared image through a trained YOLO model.
Specifically, the total loss function of the model includes two parts, namely, a loss of confidence and a loss of coordinates of the frame, wherein the confidence represents the probability of an object in the frame, and the coordinates of the frame are represented by four numbers of a center point and an altitude.
loss＝l box +l cls
Wherein S is a grid in which the picture is divided into S x S, each grid predicts B anchor frames, and x, y, w and h are the center coordinates, height and width of the predicted anchor frames; l (L) box Representing predicted frame loss, l cls Representing the classification loss.
And marking the reinforced infrared image of the transformer, training the YOLO model by using a gradient descent method, and extracting the required region image of the oil conservator from the reinforced infrared image by using the trained YOLO model.
In this embodiment, the total loss function of the YOLO model is obtained; according to the total loss function, training a YOLO model by a gradient descent method, extracting an area image of the oil conservator in the enhanced infrared image by the trained YOLO model, and detecting the oil level of the oil conservator in the area image of the oil conservator.
In one embodiment, the detecting the oil level of the oil conservator on the image of the area where the oil conservator is located through a temperature measurement algorithm comprises:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along the preset direction in the region images of the oil conservers; extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line; and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
Specifically, a plurality of temperature measuring lines are arranged at equal intervals along a preset longitudinal direction in an extracted image of an area where the oil conservator is located, adjacent pixel values on each temperature measuring line are different from each other in pairs, adjacent pixel points with the largest change on the temperature measuring lines are obtained to serve as oil level detection results of the temperature measuring lines, and the oil level detection results obtained by each temperature measuring line are averaged to serve as final oil level detection results.
In this embodiment, through determining preset temperature measuring lines corresponding to the region image where the oil conservator is located, the preset temperature measuring lines are arranged at equal intervals in the region image where the oil conservator is located along the preset direction, adjacent pixel points with the largest change on each preset temperature measuring line are extracted along the preset direction, temperature measuring results corresponding to each preset temperature measuring line are obtained, an average value of the temperature measuring results is calculated, an oil level detection result of the oil conservator is obtained, and the oil level of the oil conservator of the transformer can be obtained.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a device for detecting the oil level of the oil conservator, which is used for realizing the method for detecting the oil level of the oil conservator of the transformer. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiments of the device for detecting the oil level of the conservator provided in the following may be referred to the limitation of the method for detecting the oil level of the conservator of the transformer hereinabove, and will not be repeated here.
In one embodiment, as shown in fig. 6, there is provided a wick oil level detection apparatus of a transformer, including: the device comprises a noise-passing infrared image acquisition module 100, an enhanced infrared image acquisition module 200, a region image extraction module 300 where a conservator is located and a conservator oil level detection module 400, wherein:
the noise-reducing infrared image acquisition module 200 is used for acquiring an infrared image of the transformer, filtering random noise in the infrared image through a bilateral filtering algorithm, and obtaining a noise-reducing infrared image;
the enhanced infrared image obtaining module 400 is configured to enhance the over-noise infrared image by presetting a CycleGAN model if the over-noise infrared image is an infrared image corresponding to a rainy day, so as to obtain an enhanced infrared image;
the image extraction module 600 of the region where the oil conservator is located is used for extracting and enhancing the region image of the oil conservator in the infrared image through a YOLO target detection algorithm;
the oil level detection module 800 is configured to detect the oil level of the oil conservator by using a temperature measurement algorithm.
In one embodiment, the enhanced infrared image acquisition module 400 is further configured to acquire a total loss function of the preset CycleGAN model; training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In one embodiment, the enhanced infrared image acquisition module 400 is further configured to acquire a generator loss, a arbiter loss, and a cycle consistency loss of the preset CycleGAN model; and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
In one embodiment, the image extraction module 600 of the region where the conservator is located is further configured to extract a feature image in the enhanced infrared image through a feature extraction part of the YOLO model, where the feature extraction part includes a plurality of module layers, and the module layers are composed of a convolutional neural network layer, a Batch-localization layer and a residual connection layer; adopting a feature pyramid to carry out enhanced extraction on the feature image to obtain an enhanced feature image; and extracting an area image of the oil pillow in the enhanced feature image through an effective feature layer of the YOLO model, wherein the effective feature layer consists of at least two module layers on the output side of the YOLO model.
In one embodiment, the image extraction module 600 of the region where the conservator is located is further configured to obtain a total loss function of the YOLO model; training a YOLO model by a gradient descent method according to the total loss function; and extracting and enhancing an area image of the oil pillow in the infrared image through a trained YOLO model.
In one embodiment, the oil level detection module 800 is further configured to determine a preset temperature line corresponding to an image of an area where the oil conservator is located, where the preset temperature line is disposed at equal intervals along a preset direction in the image of the area where the oil conservator is located; extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line; and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
The above-mentioned various modules in the oil conservator oil level detection device of the transformer can be realized in whole or in part by software, hardware and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 7. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured 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, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing infrared image data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program when executed by a processor is to implement a method of detecting a conservator oil level of a transformer.
It will be appreciated by those skilled in the art that the structure shown in fig. 7 is merely a block diagram of some of the structures associated with the present application and is not limiting of the computer device to which the present application may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noisy infrared image; if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located by a temperature measurement algorithm.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring a total loss function of a preset CycleGAN model; training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In one embodiment, the processor when executing the computer program further performs the steps of:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model; and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
In one embodiment, the processor when executing the computer program further performs the steps of:
extracting and enhancing a characteristic image in the infrared image through a characteristic extraction part of the YOLO model, wherein the characteristic extraction part comprises a plurality of module layers, and each module layer consists of a convolutional neural network layer, a Batch-nomination layer and a residual error connection layer; adopting a feature pyramid to carry out enhanced extraction on the feature image to obtain an enhanced feature image; and extracting an area image of the oil pillow in the enhanced feature image through an effective feature layer of the YOLO model, wherein the effective feature layer consists of at least two module layers on the output side of the YOLO model.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring a total loss function of the YOLO model; training a YOLO model by a gradient descent method according to the total loss function; and extracting and enhancing an area image of the oil pillow in the infrared image through a trained YOLO model.
In one embodiment, the processor when executing the computer program further performs the steps of:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along the preset direction in the region images of the oil conservers; extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line; and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noisy infrared image; if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located by a temperature measurement algorithm.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a total loss function of a preset CycleGAN model; training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In one embodiment, the computer program when executed by the processor further performs the steps of:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model; and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
In one embodiment, the computer program when executed by the processor further performs the steps of:
extracting and enhancing a characteristic image in the infrared image through a characteristic extraction part of the YOLO model, wherein the characteristic extraction part comprises a plurality of module layers, and each module layer consists of a convolutional neural network layer, a Batch-nomination layer and a residual error connection layer; adopting a feature pyramid to carry out enhanced extraction on the feature image to obtain an enhanced feature image; and extracting an area image of the oil pillow in the enhanced feature image through an effective feature layer of the YOLO model, wherein the effective feature layer consists of at least two module layers on the output side of the YOLO model.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a total loss function of the YOLO model; training a YOLO model by a gradient descent method according to the total loss function; and extracting and enhancing an area image of the oil pillow in the infrared image through a trained YOLO model.
In one embodiment, the computer program when executed by the processor further performs the steps of:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along the preset direction in the region images of the oil conservers; extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line; and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noisy infrared image; if the over-noise infrared image is an infrared image corresponding to a rainy day, enhancing the over-noise infrared image through a preset CycleGAN model to obtain an enhanced infrared image; extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm; and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located by a temperature measurement algorithm.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a total loss function of a preset CycleGAN model; training a preset CycleGAN model according to the total loss function; if the super-noise infrared image is an infrared image corresponding to a rainy day, the super-noise infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
In one embodiment, the computer program when executed by the processor further performs the steps of:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model; and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
In one embodiment, the computer program when executed by the processor further performs the steps of:
the characteristic extraction part of the YOLO model is used for extracting the characteristic of the enhanced infrared image, and the characteristic extraction part consists of a convolutional neural network layer, a Batch-normalization layer and a residual error connection layer; enhancing the infrared image feature extraction through feature pyramid enhancement; and on an effective feature layer of the feature extraction part, enhancing an image of an area where the oil pillow is located in the infrared image after feature enhancement, wherein the effective feature layer is an output prediction module layer formed by a convolutional neural network layer, a Batch-minimization layer and a residual error connection layer.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a total loss function of the YOLO model; training a YOLO model by a gradient descent method according to the total loss function; and extracting and enhancing an area image of the oil pillow in the infrared image through a trained YOLO model.
In one embodiment, the computer program when executed by the processor further performs the steps of:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along the preset direction in the region images of the oil conservers; extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line; and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
It should be noted that, user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.
Claims (10)
1. A method for detecting the oil level of a conservator of a transformer, the method comprising:
acquiring an infrared image of a transformer, and filtering random noise in the infrared image through a bilateral filtering algorithm to obtain a noise-filtered infrared image;
if the noise-exceeding infrared image is an infrared image corresponding to a rainy day, enhancing the noise-exceeding infrared image through a preset CycleGAN model to obtain an enhanced infrared image;
extracting an area image of the oil pillow in the enhanced infrared image by a YOLO target detection algorithm;
and detecting the oil level of the oil conservator according to the image of the area where the oil conservator is located through a temperature measurement algorithm.
2. The method of claim 1, wherein if the denoised infrared image is an infrared image corresponding to a rainy day, enhancing the denoised infrared image by a preset CycleGAN model, the obtaining an enhanced infrared image comprises:
acquiring a total loss function of a preset CycleGAN model;
training a preset CycleGAN model according to the total loss function; if the denoised infrared image is an infrared image corresponding to a rainy day, the denoised infrared image is enhanced through a trained CycleGAN model, and an enhanced infrared image is obtained.
3. The method of claim 2, wherein the obtaining the total loss function of the preset CycleGAN model comprises:
obtaining generator loss, discriminator loss and cycle consistency loss of a preset CycleGAN model;
and obtaining a total loss function of a preset CycleGAN model according to the generator loss, the discriminator loss and the cycle consistency loss.
4. The method of claim 1, wherein extracting the region image of the headrest in the enhanced infrared image by the YOLO target detection algorithm comprises:
extracting a characteristic image in the enhanced infrared image through a characteristic extraction part of a YOLO model, wherein the characteristic extraction part comprises a plurality of module layers, and each module layer consists of a convolutional neural network layer, a Batch-nomination layer and a residual error connection layer;
performing enhanced extraction on the feature image by adopting a feature pyramid to obtain an enhanced feature image;
and extracting an area image of the oil pillow in the enhanced feature image through an effective feature layer of the YOLO model, wherein the effective feature layer consists of at least two module layers on the output side of the YOLO model.
5. The method of claim 1, wherein extracting the region image of the headrest in the enhanced infrared image by the YOLO target detection algorithm comprises:
acquiring a total loss function of the YOLO model;
training the YOLO model by a gradient descent method according to the total loss function;
and extracting an area image of the oil pillow in the enhanced infrared image through a trained YOLO model.
6. The method of claim 1, wherein said detecting the oil level of the oil conservator from the image of the area of the oil conservator by a thermometry algorithm comprises:
determining preset temperature measuring lines corresponding to the region images of the oil conservers, wherein the preset temperature measuring lines are arranged at equal intervals along a preset direction in the region images of the oil conservers;
extracting adjacent pixel points with the largest change on each preset temperature measuring line along the preset direction to obtain a temperature measuring result corresponding to each preset temperature measuring line;
and calculating the average value of the temperature measurement results to obtain the oil level detection result of the oil conservator.
7. A device for detecting the oil level of a conservator of a transformer, said device comprising:
the noise-reducing infrared image acquisition module is used for acquiring an infrared image of the transformer, filtering random noise in the infrared image through a bilateral filtering algorithm, and obtaining a noise-reducing infrared image;
the enhanced infrared image acquisition module is used for enhancing the noisy infrared image through a preset CycleGAN model if the noisy infrared image is an infrared image corresponding to a rainy day, so as to obtain an enhanced infrared image;
the image extraction module is used for extracting an area image of the oil conservator in the enhanced infrared image through a YOLO target detection algorithm;
and the oil level detection module is used for detecting the oil level of the oil conservator through a temperature measurement algorithm on the image of the area where the oil conservator is located.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211516889.XA CN116129345A (en) | 2022-11-30 | 2022-11-30 | Method and device for detecting oil level of conservator of transformer and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211516889.XA CN116129345A (en) | 2022-11-30 | 2022-11-30 | Method and device for detecting oil level of conservator of transformer and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116129345A true CN116129345A (en) | 2023-05-16 |
Family
ID=86305391
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211516889.XA Pending CN116129345A (en) | 2022-11-30 | 2022-11-30 | Method and device for detecting oil level of conservator of transformer and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116129345A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876373A (en) * | 2024-03-12 | 2024-04-12 | 成都航空职业技术学院 | Transformer fault detection method |
-
2022
- 2022-11-30 CN CN202211516889.XA patent/CN116129345A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876373A (en) * | 2024-03-12 | 2024-04-12 | 成都航空职业技术学院 | Transformer fault detection method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11581130B2 (en) | Internal thermal fault diagnosis method of oil-immersed transformer based on deep convolutional neural network and image segmentation | |
| CN109685152B (en) | Image target detection method based on DC-SPP-YOLO | |
| US10803554B2 (en) | Image processing method and device | |
| CN110929607B (en) | Remote sensing identification method and system for urban building construction progress | |
| CN110991311B (en) | Target detection method based on dense connection deep network | |
| CN111028327B (en) | Processing method, device and equipment for three-dimensional point cloud | |
| TW202207077A (en) | Text area positioning method and device | |
| CN107506792B (en) | Semi-supervised salient object detection method | |
| JP6107531B2 (en) | Feature extraction program and information processing apparatus | |
| JPWO2011001817A1 (en) | Representative feature extraction system and method | |
| CN113537180B (en) | Tree obstacle identification method and device, computer equipment and storage medium | |
| CN111583148A (en) | Rock core image reconstruction method based on generation countermeasure network | |
| CN112766123A (en) | Crowd counting method and system based on criss-cross attention network | |
| CN116129345A (en) | Method and device for detecting oil level of conservator of transformer and computer equipment | |
| Zheng et al. | Stripe segmentation of oceanic internal waves in SAR images based on SegNet | |
| CN115953330B (en) | Texture optimization method, device, equipment and storage medium for virtual scene image | |
| CN112270289A (en) | Intelligent monitoring method based on graph convolution attention network | |
| CN117058554A (en) | Power equipment target detection method, model training method and device | |
| CN116206302A (en) | Three-dimensional object detection method, three-dimensional object detection device, computer equipment and storage medium | |
| CN115937395A (en) | Electrical equipment model rendering method and device, computer equipment and storage medium | |
| CN115063473A (en) | Object height detection method and device, computer equipment and storage medium | |
| CN114463764A (en) | Table line detection method and device, computer equipment and storage medium | |
| CN106469437B (en) | Image processing method and image processing apparatus | |
| CN113139540A (en) | Backboard detection method and equipment | |
| CN114998990B (en) | Method and device for identifying safety behaviors of personnel on construction site |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |