CN114757941A - Transformer substation equipment defect identification method and device, electronic equipment and storage medium - Google Patents

Transformer substation equipment defect identification method and device, electronic equipment and storage medium Download PDF

Info

Publication number
CN114757941A
CN114757941A CN202210556740.8A CN202210556740A CN114757941A CN 114757941 A CN114757941 A CN 114757941A CN 202210556740 A CN202210556740 A CN 202210556740A CN 114757941 A CN114757941 A CN 114757941A
Authority
CN
China
Prior art keywords
equipment
defect
image
convolution
result
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
Application number
CN202210556740.8A
Other languages
Chinese (zh)
Inventor
王�琦
王振利
王万国
徐康
孙志周
崔其会
郭修霄
张志�
刘晗
李振宇
杨月琛
王晓丹
刘海波
张海龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Intelligent Technology Co Ltd
Original Assignee
State Grid Intelligent Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by State Grid Intelligent Technology Co Ltd filed Critical State Grid Intelligent Technology Co Ltd
Priority to CN202210556740.8A priority Critical patent/CN114757941A/en
Publication of CN114757941A publication Critical patent/CN114757941A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/60—Rotation of whole images or parts thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/40—Image enhancement or restoration using histogram techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/73—Deblurring; Sharpening
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20112—Image segmentation details
    • G06T2207/20132—Image cropping

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

The embodiment of the invention discloses a method and a device for identifying defects of transformer substation equipment, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring a target equipment image of to-be-detected substation equipment; inputting the target equipment image into an equipment defect identification model which is trained in advance to obtain an identification result of the equipment defect of the transformer substation; the equipment defect identification model comprises a convolution module and a full-connection module which are connected in parallel, the convolution module comprises at least one convolution layer and at least one full-connection layer which is connected with the convolution layer in series, and the full-connection module comprises a full-connection module consisting of at least one full-connection layer. According to the technical scheme of the embodiment of the invention, the equipment defect identification model with the convolution module and the full-connection module connected in parallel is established in advance, so that the positions of the defects in the target equipment image can be identified and the defect types can be positioned at the same time, the manual participation degree is reduced, and the efficiency and the accuracy of the transformer substation equipment defect identification are improved.

Description

Transformer substation equipment defect identification method and device, electronic equipment and storage medium
Technical Field
The embodiment of the invention relates to the field of computer image technology, in particular to a method and a device for identifying defects of transformer substation equipment, electronic equipment and a storage medium.
Background
The transformer substation is one of important components in an electric power system, a plurality of electric power devices are arranged in a place, and after the electric power devices are used for a long time, the surfaces of the electric power devices are easily damaged, so that the corrosion phenomenon of internal metal is easily generated, and therefore, the defect detection of the transformer substation devices is required to be carried out at variable time to prevent serious electric power accidents.
The existing transformer substation equipment defect detection is that the corrosion degree inside the equipment is judged manually according to the damage condition, so that the detection mode is prone to misjudgment, the detection result is not accurate, meanwhile, the manual operation consumes unnecessary time and labor cost, and the actual application requirement is not met.
Disclosure of Invention
The embodiment of the invention provides a method and a device for identifying the defects of transformer substation equipment, electronic equipment and a storage medium, so as to realize intelligent analysis and identification of the defects of the transformer substation equipment.
In a first aspect, an embodiment of the present invention provides a method for identifying a substation device defect, where the method includes:
acquiring a target device image of to-be-detected substation equipment;
inputting the target equipment image into an equipment defect recognition model trained in advance to obtain a recognition result of the substation equipment defect;
the equipment defect identification model comprises a convolution module and a full-connection module which are connected in parallel, the convolution module comprises at least one convolution layer and at least one full-connection layer which is connected with the convolution layer in series, and the full-connection module comprises a full-connection module consisting of at least one full-connection layer.
In a second aspect, an embodiment of the present invention further provides a device for identifying a substation device defect, where the device includes:
the device image acquisition module is used for acquiring a target device image of the to-be-detected substation device;
the equipment image input module is used for inputting the target equipment image into an equipment defect identification model which is trained in advance to obtain an identification result of the equipment defect of the transformer substation;
the equipment defect identification model comprises a convolution module and a full-connection module which are connected in parallel, the convolution module comprises at least one convolution layer and at least one full-connection layer which is connected with the convolution layer in series, and the full-connection module comprises a full-connection module consisting of at least one full-connection layer.
In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, the one or more programs cause the one or more processors to implement the method for identifying a substation device defect provided by any embodiment of the present invention.
In a fourth aspect, the embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for identifying a defect in a substation device provided in any embodiment of the present invention.
The invention provides a method for identifying equipment defects of a transformer substation, which is an equipment defect identification method based on a neural network model, adopts a convolution module and a full-connection module which are connected in parallel, effectively improves the robustness of the equipment defect identification model compared with a model structure connected in series, and can effectively improve the accuracy of an output result by adopting a mode of fusing a full-connection output result and a convolution output result, thereby realizing the technical effect of accurate and efficient transformer substation defect identification, solving the problems of large detection workload, low detection efficiency and inaccuracy caused by manually detecting the defects of the transformer substation in the prior art, and further providing a method for constructing a training data set compared with the prior model training method in the training process of the equipment defect identification model, the acquired sample equipment images are expanded in an image expansion mode, so that the comprehensiveness of a training data set is improved, and meanwhile, the generalization of a neural network model is improved.
Drawings
In order to more clearly illustrate the technical solution of the exemplary embodiment of the present invention, a brief introduction will be made to the drawings required for describing the embodiment. It should be clear that the described figures are only views of some of the embodiments of the invention to be described, not all, and that for a person skilled in the art, other figures can be derived from these figures without inventive effort.
Fig. 1 is a schematic flow chart of a method for identifying a substation device defect according to a first embodiment of the present invention;
fig. 2 is a schematic structural diagram of an equipment defect identification model in an identification method for an equipment defect of a substation provided in a first embodiment of the present invention;
fig. 3 is a schematic flow chart of a method for identifying a substation device defect according to a second embodiment of the present invention;
fig. 4 is a schematic structural diagram of a device for identifying a substation device defect according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings. Before discussing exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart may describe the operations (or steps) as a sequential process, many of the operations can be performed in parallel, concurrently, or simultaneously. In addition, the order of the operations may be re-arranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figure. The processes may correspond to methods, functions, procedures, subroutines, and the like.
Example one
Fig. 1 is a schematic flow chart of a method for identifying a substation device defect according to an embodiment of the present invention, where the present embodiment is applicable to a case where a defect of a substation device is detected, the method may be executed by a device for identifying a substation device defect, the device may be implemented by software and/or hardware, and may be configured in a terminal and/or a server to implement the method for identifying a substation device defect according to the embodiment of the present invention.
As shown in fig. 1, the method of the embodiment may specifically include:
and S110, acquiring a target device image of the substation device to be detected.
In this embodiment, the target device image may be understood as an image obtained when the substation device to be detected is patrolled. For example, the target device image may be obtained by an image capturing device, or may be obtained by an image database, and the like, which is not limited in this embodiment. For example, the substation equipment to be detected can be shot by a fixed point shooting device or an inspection robot, and the shot image is used as a target equipment image.
In specific implementation, the image acquisition device can be used for acquiring an image of the substation equipment to be detected, and the acquired image is used as a target equipment image of the substation equipment to be detected.
And S120, inputting the target equipment image into the pre-trained equipment defect identification model to obtain the identification result of the substation equipment defect.
In this embodiment, the device defect identification model may be a neural network model for identifying the device defect of the substation device to be detected through the target device image. Illustratively, the device defect identification model may be formed by at least one of a convolutional neural network, a cyclic neural network, or a deep neural network, which is not limited in this embodiment.
Fig. 2 is a schematic structural diagram of an apparatus defect identification model in an embodiment of the present invention. In the present embodiment, a specific structure of the device defect identification model may be as shown in fig. 2.
Optionally, the device defect identification model may include a convolution module and a full-connection module connected in parallel, the convolution module may include at least one convolution layer and at least one full-connection layer connected in series with the convolution layer, and the full-connection module may include a full-connection module composed of at least one full-connection layer.
The convolution module can be understood as a feedforward neural network module which is used for realizing convolution calculation and has a depth structure. A fully-connected layer may be understood as a layered structure consisting of multiple layers of neurons, and each neuron has a weighted connection with all inputs. The convolution layer can play a role in extracting the characteristic information of the target equipment image in the equipment defect identification model, and the full-connection layer can play a role in a classifier in the equipment defect identification model, so that the convolution module can classify and analyze the characteristic information of the target equipment image on the basis of extracting the characteristic information of the target equipment image.
The proportion of the whole equipment image of the transformer substation in the target equipment image is large, the defect part needing to be identified is usually located in the transformer substation equipment, and the proportion of the defect part in the target equipment image is small, so that the equipment defect identification model adopts the convolution module and the full-connection module which are connected in parallel, the bias of the equipment defect identification model on characteristic information extraction and classification analysis can be enhanced, and compared with the mode that the convolution module is connected with the full-connection module in series, the sensitivity of the equipment defect identification model on the defect identification of the transformer substation equipment in the target equipment image can be effectively improved.
The invention provides an equipment defect identification method based on a neural network model, which adopts a mode of fusing a convolution module and a full-connection module which are connected in parallel and output results of all the modules, effectively improves the robustness of an equipment defect identification model compared with a serial-connection model structure, and can effectively improve the accuracy of an output result by adopting a mode of fusing the full-connection output result and the convolution output result, thereby realizing the technical effect of accurate and efficient transformer substation defect identification.
Optionally, the device defect identification model may further include a feature map extraction module, and the feature map extraction module may include at least one convolution layer.
The feature map extraction module may be understood as a neural network module for extracting feature information from the target device image. Illustratively, the feature map extraction module may be formed by at least one of a convolutional neural network structure, a deep neural network structure, or other form of neural network structure, such as a VGG-Net or RES-Net structure.
It should be noted that the convolution module may be used to extract feature information in the target device image, and the full-connection module may be used to extract spatial relationship information of the target device image, that is, the region where the defect portion of the target device image is located may be determined, and the type of the defect in the target device image may also be determined.
It should be further noted that before the target device image is input into the pre-trained device defect recognition model, a series of image preprocessing operations may be performed on the target device image to improve the diversification of the image data set, reduce the risk of overfitting, and improve the generalization of the device defect recognition model. Optionally, the preprocessing operation on the target device image may include, but is not limited to, geometric transformation (e.g., rotation, scaling, cropping, translation, affine change, etc.), color space change (e.g., contrast change, brightness change, saturation change, histogram enhancement, gray scale adjustment, etc.), or pixel relationship adjustment (e.g., blur, sharpening, noise, etc.), which is not limited in this embodiment.
In this embodiment, the device defect recognition model may be trained based on the sample device image and the expected recognition result corresponding to the sample device image.
On the basis of the technical scheme, the method further comprises the following steps: constructing a training data set for training an equipment defect recognition model, wherein the training data set comprises a sample equipment image and an expected recognition result corresponding to the sample equipment image; inputting the sample equipment image into a model to be trained to obtain a model output result corresponding to the sample equipment image; and performing model adjustment on the model to be trained according to the model output result corresponding to the sample equipment image, the expected identification result and a pre-constructed model loss function to obtain an equipment defect identification model.
The sample training image may be a standard image of the device used for model training. The expected identification result may be pre-labeled tag data for identifying a defect in the device. The expected recognition result may be used as a basis for evaluating subsequent predictions and classification results.
Optionally, constructing a training data set for training the device defect recognition model includes: acquiring historical equipment images of power transformation equipment in a transformer substation, and determining sample equipment images according to the historical equipment images; marking the sample equipment image to obtain an expected identification result corresponding to the sample equipment image; and constructing a training data set for training the equipment defect recognition model according to the sample equipment image and the expected recognition result corresponding to the sample equipment image.
The historical device image may be an image acquired by the power transformation device at a historical time point.
Specifically, when the training data set of the model is constructed, the historical device images of the power transformation devices in the transformer substation are obtained, the historical device images are screened to determine the sample device images, further, device defect labeling is performed on the sample device images to obtain expected identification results corresponding to the sample device images, and finally, the training data set of the device defect device model is constructed according to the sample device images and the expected identification results corresponding to the sample device images.
It should be noted that the historical device image may be acquired from the image acquisition device in real time, may also be acquired from the image database, and may also be acquired in other manners, and the acquisition manner of the sample device image is not limited in any way in the embodiment of the present invention; the process of labeling the expected identification result may be implemented by labeling the image of the sample device through a manual labeling program, or by using other methods, which is not limited in this embodiment.
Optionally, determining the sample device image from the historical device images comprises: and carrying out image expansion processing on the historical device image according to a preset image expansion mode to obtain an expanded device image, and taking the expanded device image and the historical device image as a sample device image.
The preset image expansion mode may be a preset image processing mode for performing data expansion on the image. Illustratively, the preset image expansion manner includes at least one of geometric transformation, color space variation and pixel relation adjustment.
Optionally, the geometric transformation comprises at least one of a rotation transformation, a scaling transformation, a cropping transformation, a translation transformation, and an affine transformation.
The affine transformation may be a process in which, in geometry, one vector space is subjected to linear transformation and then translated to another vector space.
Optionally, the color space variation comprises at least one of a contrast variation, a brightness variation, a saturation variation, a histogram enhancement, and a grey scale adjustment.
The histogram enhancement may be a processing method for achieving image enhancement by adjusting a histogram. Illustratively, histogram enhancement may include histogram equalization, histogram matching, and the like.
Optionally, the pixel relation adjustment includes at least one of a blurring process, a sharpening process, and a noise process.
Generally, when a sample device image for model training is determined, in order to expand the training data set, improve the comprehensiveness of the training data set, and improve the generalization of the device defect recognition model, an image expansion process may be performed on a historical device image according to a preset image expansion method to obtain an expanded device image, and further, the expanded device image and the historical device image may be used as the sample device image, and further, the sample device image may be labeled to obtain an expected recognition result corresponding to the sample device image, and a training data set for model training may be constructed according to the sample device image and the expected recognition result corresponding to the sample device image.
The invention provides a construction method of a training data set, which is used for expanding acquired sample equipment images in an image expansion mode, so that the comprehensiveness of the training data set is improved, and meanwhile, the generalization of a neural network model is improved.
In specific implementation, a sample device image is input into a device defect identification model to be trained, and a feature image extraction module can extract feature information of a training image data set to obtain a preliminary feature image data set; inputting the preliminary characteristic image data sets into a convolution module and a full-connection module which are connected in parallel respectively to obtain a convolution output result of the convolution module and a full-connection output result of the full-connection module respectively; and performing result fusion on the convolution output result and the full-connection output result to obtain a model output result corresponding to the sample equipment image, and then performing model parameter adjustment on the equipment defect identification model to be trained according to the model output result corresponding to the sample equipment image, the expected output result and the model loss function until a training end condition is met to obtain the trained equipment defect identification model.
Alternatively, the target loss function of the device defect identification model may be determined according to the loss values of the fully connected modules and the loss values of the convolution modules of the device defect identification model.
Alternatively, the target loss function of the equipment defect identification model can be determined by performing weighted calculation according to the loss value of the convolution module and the loss value of the fully-connected module.
Specifically, the specific way of determining the target loss function of the equipment defect identification model according to the loss value of the fully-connected module and the loss value of the convolution module of the equipment defect identification model is as follows:
Figure BDA0003652560430000101
wherein L represents an objective loss function of the equipment defect identification model, LfcRepresents the loss value, L, of a fully connected moduleconvRepresenting the loss value of the convolution module, ω representing a weighting factor, ωfcIs | ln (1-L)fc)|,ωconvIs composed of
Figure BDA0003652560430000102
Figure BDA0003652560430000103
Indicating the correction parameters.
The correction parameter may be a value of a target loss value for correcting the equipment defect identification model, and may be a constant set empirically, and the value is not limited in detail herein.
It should be noted that the training end condition of the equipment defect identification model may be that the target loss function calculated based on the above formula tends to converge in the training process, or that the identification accuracy of the training result output by the equipment defect identification model reaches a preset accuracy, and the like, which is not limited in this embodiment.
In specific implementation, a target device image is input into a device defect identification model which is trained in advance, feature information in the target device image is extracted through a feature map extraction module to obtain a target device feature image, the target device feature image is respectively input into a convolution module and a full-connection module which are connected in parallel, defect feature information in the target device feature image is extracted through the convolution module to determine position information of a defect area in the target device feature image and/or defect types of the defect area in the target device feature image, the defect types of the defect area in the target device feature image are classified through the full-connection module, and finally, a transformer substation device defect identification result is output.
According to the technical scheme of the embodiment of the invention, the target equipment image of the to-be-detected substation equipment is obtained and is input into the pre-trained equipment defect identification model to obtain the identification result of the substation equipment defect, wherein the equipment defect identification model adopts the convolution module and the full-connection module which are connected in parallel, compared with a serial connection model structure, the robustness of the equipment defect identification model can be effectively improved, the identification accuracy is improved, the problems of large detection workload, low detection efficiency and inaccuracy caused by manual detection of the substation equipment defect in the prior art are solved, and the technical effect of accurately identifying the substation equipment defect while the labor cost is reduced is achieved.
Example two
Fig. 2 is a schematic flow chart of a method for identifying a substation device defect according to a second embodiment of the present invention, and on the basis of the foregoing technical solution, the second embodiment further refines the technical solution. On the basis of any optional technical solution in the embodiment of the present invention, optionally, the inputting the target device image into a device defect recognition model which is trained in advance to obtain a recognition result of a substation device defect includes: inputting the target equipment image into an equipment defect identification model which is trained in advance, and respectively obtaining a convolution output result output by the convolution module and a full-connection output result output by the full-connection module; and determining the identification result of the defects of the transformer substation equipment according to the full connection output result and the convolution output result.
The technical terms that are the same as or corresponding to the above embodiments are not repeated herein.
As shown in fig. 2, the method of this embodiment specifically includes the following steps:
s210, receiving a target equipment image of the substation equipment to be detected, which is acquired by the target inspection equipment.
The target inspection equipment can be understood as field inspection equipment which is used for carrying out field inspection on substation equipment and acquiring equipment images. For example, the target inspection device may include, but is not limited to, an image acquisition device disposed at a fixed point or a substation inspection robot with a camera device, and the like, and this embodiment is not limited thereto.
In specific implementation, the substation equipment to be detected can be patrolled through the target patrol equipment, and the equipment state image of the substation equipment is acquired in the patrol process, so that the acquired equipment state image can be used as the target equipment image of the substation equipment to be detected, and the defect of the substation equipment is identified.
It should be noted that, the target device image acquired by the target inspection device in the inspection process may be acquired and simultaneously subjected to identification and analysis of the substation device defect, or the acquired target device image may be stored in the image database in advance, and after the inspection process is finished or at the substation device defect identification stage, the acquired target device image is subjected to identification and analysis of the substation device defect, and the like.
S220, inputting the target equipment image into the pre-trained equipment defect identification model, and respectively obtaining a convolution output result output by the convolution module and a full-connection output result output by the full-connection module.
In this embodiment, optionally, the convolution output result may include a convolution classification result of a substation equipment defect, and may further include a convolution identification result of a defect area of the substation equipment. The convolution classification result can be understood as a classification result obtained by classifying the transformer substation equipment defect types in the target equipment image by using a convolution module. The convolution identification result can be understood as an identification result of performing feature identification on a defective area of the substation equipment defect in the target equipment image by applying a convolution module.
Optionally, the full-connection output result may include a substation equipment defect full-connection classification result. The full-connection classification result can be understood as a classification result obtained by classifying the transformer substation equipment defect types in the target equipment image by using a full-connection module.
In this embodiment, optionally, inputting the target device image into a device defect recognition model trained in advance, and obtaining a convolution output result output by the convolution module and a full-connection output result output by the full-connection module respectively, includes: inputting the target equipment image into a feature map extraction module of an equipment defect recognition model which is trained in advance to obtain a preliminary feature map; and inputting the preliminary characteristic graph into a full-connection module to obtain a full-connection output result, and inputting the preliminary characteristic graph into a convolution module to obtain a convolution output result.
In some embodiments, optionally, inputting the target device image into a device defect recognition model trained in advance to obtain a convolution output result output by the convolution module and a full-connection output result output by the full-connection module, respectively, and may further include: respectively arranging a first characteristic diagram extraction module and a second characteristic diagram extraction module before the full-connection module and the convolution module, inputting the target equipment image into the first characteristic diagram extraction module to obtain a first preliminary characteristic diagram, and inputting the first preliminary characteristic diagram into the full-connection module to obtain a full-connection output result; and inputting the target equipment image into a second feature map extraction module to obtain a second preliminary feature map, and inputting the second preliminary feature map into a convolution module to obtain a convolution output result.
It should be noted that. References to "first" and "second" in the above embodiments do not denote any order, quantity, or importance, but rather their corresponding target entities are different. Illustratively, the "first feature map extraction module" and the "second feature map extraction module" may represent the same feature map extraction module or may represent different feature map extraction modules, and similarly, the "first preliminary feature map" and the "second preliminary feature map" may represent the same preliminary feature map or may represent different preliminary feature maps.
And S230, determining the identification result of the defects of the substation equipment according to the full connection output result and the convolution output result.
In this embodiment, optionally, the identification result of the substation equipment defect may include a target classification result of the substation equipment defect. The target classification result can be understood as a result obtained by classifying the defect type in the target equipment image by the equipment defect identification model.
It should be noted that, because the full-connection output result includes a full-connection classification result of a substation equipment defect, and the convolution output result includes a convolution classification result of a substation equipment defect, in some embodiments, the target classification result of a substation equipment defect may be a full-connection classification result, a convolution classification result, a classification result obtained by fusing a full-connection classification result and a convolution classification result, and the like.
In this embodiment, optionally, determining an identification result of the substation device defect according to the full-connection output result and the convolution output result includes: and performing result fusion on the full-connection classification result and the convolution classification result to obtain a target classification result of the defects of the substation equipment.
It should be further noted that, in the process of performing result fusion on the fully-connected classification result and the convolution classification result, since the fully-connected module can well implement classification analysis on the defect types of the substation equipment, when performing result fusion, it can be considered that the convolution classification result is corrected according to the fully-connected classification result, so that a more accurate target classification result can be obtained.
Optionally, performing result fusion on the full-connection classification result and the convolution classification result to obtain a target classification result of the defect of the substation equipment, where the target classification result includes: and performing result fusion on the full-connection classification result and the convolution classification result based on the following formula to obtain a target classification result of the defects of the substation equipment:
Figure BDA0003652560430000141
wherein p isiA defect classification result of the defect area of the substation equipment which is output by the equipment defect identification model is represented,
Figure BDA0003652560430000142
a defect classification result of a convolution module representing a device defect recognition model,
Figure BDA0003652560430000143
and representing the defect classification result of the fully-connected module of the equipment defect identification model.
The invention provides an equipment defect identification method based on a neural network model, which adopts a convolution module and a full-connection module which are connected in parallel, effectively improves the robustness of the equipment defect identification model compared with a model structure connected in series, and can effectively improve the accuracy of an output result by adopting a mode of fusing a full-connection output result and a convolution output result, thereby realizing the technical effect of accurate and efficient transformer substation defect identification and solving the problems of large detection workload, low detection efficiency and inaccuracy caused by manually detecting the defects of transformer substation equipment in the prior art.
In this embodiment, optionally, the identification result of the substation equipment defect may further include a target identification result of a defective area of the substation equipment. The target identification result can be understood as an identification result of the equipment defect identification model on the transformer substation equipment defect area in the target equipment image. The target recognition result can represent the approximate position of the defect area of the substation equipment in the target equipment image.
Optionally, determining an identification result of the substation equipment defect according to the full-connection output result and the convolution output result, including: and taking the convolution identification result of the defect area of the substation equipment as the target identification result of the defect area of the substation equipment.
In specific implementation, the acquired target equipment image is input into an equipment defect identification model which is trained in advance, a full-connection output result output by a full-connection module and a convolution output result output by a convolution module can be respectively obtained, the full-connection output result and the convolution output result are subjected to weighted result fusion, and the identification result of the equipment defect of the transformer substation can be obtained.
It should be noted that, in this embodiment, the device defect identification model may be docked with the background system of the target inspection device and the camera device in the substation in a manner of packaging the whole container, packaging the packaged container into a dynamic link library, compiling the dynamic link library into an executable file, burning the executable file into a firmware, integrating a front-end chip, and the like, so that synchronous analysis of the acquired video may be implemented, and the defect of the substation device may be identified in real time.
The technical scheme of the embodiment of the invention receives the target equipment image of the substation equipment to be detected, which is acquired by the target inspection equipment, and the target equipment image is input into the pre-trained equipment defect identification model to respectively obtain a full connection output result output by the full connection module and a convolution output result output by the convolution module, and further, the full connection output result and the convolution output result are subjected to result fusion to finally obtain the identification result of the equipment defect of the transformer substation, thereby solving the problem that the equipment defect is not found in time due to the fact that the transformer substation equipment defect is analyzed by adopting a manual studying and judging mode in the prior art, thereby affecting the safe operation of the power grid and the like, and by adopting a mode of fusing the full-connection output result and the convolution output result, the accuracy of the output result can be effectively improved, and therefore the technical effect of accurate and efficient transformer substation defect identification can be achieved.
EXAMPLE III
Fig. 3 is a schematic structural diagram of a device for identifying a substation device defect according to a third embodiment of the present invention, where the device for identifying a substation device defect provided in this embodiment may be implemented by software and/or hardware, and may be configured in a terminal and/or a server to implement the method for identifying a substation device defect according to the third embodiment of the present invention. The device may specifically comprise: a device image acquisition module 310 and a device image input module 320.
The device image acquiring module 310 is configured to acquire a target device image of a to-be-detected substation device;
the device image input module 320 is used for inputting the target device image into a device defect identification model which is trained in advance to obtain an identification result of the substation device defect; the equipment defect identification model comprises convolution modules and full-connection modules which are connected in parallel, each convolution module comprises at least one convolution layer and at least one full-connection layer connected with the convolution layer in series, and each full-connection module comprises a full-connection module formed by at least one full-connection layer.
According to the technical scheme of the embodiment of the invention, the target equipment image of the to-be-detected substation equipment is obtained and is input into the pre-trained equipment defect identification model to obtain the identification result of the substation equipment defect, wherein the equipment defect identification model adopts the convolution module and the full-connection module which are connected in parallel, compared with a serial connection model structure, the robustness of the equipment defect identification model can be effectively improved, the identification accuracy is improved, the problems of large detection workload, low detection efficiency and inaccuracy caused by manual detection of the substation equipment defect in the prior art are solved, and the technical effect of accurately identifying the substation equipment defect while the labor cost is reduced is achieved.
Optionally, the device image input module 320 further includes a device image input unit and a recognition result determination unit.
The device image input unit is used for inputting a target device image into a pre-trained device defect identification model to respectively obtain a convolution output result output by the convolution module and a full-connection output result output by the full-connection module; and the identification result determining unit is used for determining the identification result of the defects of the substation equipment according to the full-connection output result and the convolution output result.
Optionally, the device defect identification model further includes a feature map extraction module, where the feature map extraction module includes at least one convolution layer;
correspondingly, the equipment image input unit is also used for inputting the target equipment image into a feature map extraction module of the pre-trained equipment defect identification model to obtain a preliminary feature map; and inputting the preliminary characteristic graph into a full-connection module to obtain a full-connection output result, and inputting the preliminary characteristic graph into a convolution module to obtain a convolution output result.
Optionally, the full-connection output result comprises a full-connection classification result of the defect of the substation equipment, and the convolution output result comprises a convolution classification result of the defect of the substation equipment; the identification result of the substation equipment defect comprises a target classification result of the substation equipment defect;
correspondingly, the identification result determining unit is further used for performing result fusion on the full-connection classification result and the convolution classification result to obtain a target classification result of the defects of the substation equipment.
Optionally, the identification result determining unit is further configured to perform result fusion on the full-connection classification result and the convolution classification result based on the following formula to obtain a target classification result of the defect of the substation device:
Figure BDA0003652560430000181
wherein p isiA defect classification result of the defect region of the substation equipment which is output by the equipment defect identification model,
Figure BDA0003652560430000182
a defect classification result of a convolution module representing a device defect recognition model,
Figure BDA0003652560430000183
and representing the defect classification result of the fully-connected module of the equipment defect identification model.
Optionally, the convolution output result further includes a convolution identification result of a defect area of the substation device; the identification result of the substation equipment defect comprises a target identification result of a defect area of the substation equipment;
similarly, the identification result determination unit is further configured to use the convolution identification result of the defective area of the substation device as the target identification result of the defective area of the substation device.
Optionally, the target loss function of the equipment defect identification model is determined according to the loss value of the fully-connected module of the equipment defect identification model and the loss value of the convolution module.
Optionally, the specific manner of determining the target loss function of the equipment defect identification model according to the loss value of the fully-connected module of the equipment defect identification model and the loss value of the convolution module is as follows:
Figure BDA0003652560430000184
wherein L represents an objective loss function of the equipment defect identification model, LfcRepresents the loss value, L, of a fully connected moduleconvRepresents the loss value of the convolution module, ω represents a weighting factor, ωfcIs | ln (1-L)fc)|,ωconvIs composed of
Figure BDA0003652560430000185
Figure BDA0003652560430000186
Indicating the correction parameters.
Optionally, the device image obtaining module 310 is further configured to receive a target device image of the substation device to be detected, which is acquired by the target inspection device.
Optionally, the apparatus further comprises: the device comprises a data set building module, an image input module and a model adjusting module.
The device comprises a data set construction module, a data set identification module and a data processing module, wherein the data set construction module is used for constructing a training data set used for training a device defect identification model, and the training data set comprises a sample device image and an expected identification result corresponding to the sample device image; the image input module is used for inputting the sample equipment image into the model to be trained to obtain a model output result corresponding to the sample equipment image; and the model adjusting module is used for carrying out model adjustment on the model to be trained according to the model output result corresponding to the sample equipment image, the expected identification result and the pre-constructed model loss function so as to obtain the equipment defect identification model.
Optionally, the data set constructing module includes an apparatus image determining unit, an image labeling unit and a data set constructing unit.
The device image determining unit is used for acquiring historical device images of power transformation devices in a transformer substation and determining sample device images according to the historical device images; the image labeling unit is used for labeling the sample equipment image to obtain an expected identification result corresponding to the sample equipment image; and the data set construction unit is used for constructing a training data set for training the equipment defect recognition model according to the sample equipment image and the expected recognition result corresponding to the sample equipment image.
Optionally, the device image determining unit is further configured to perform image expansion processing on the historical device image according to a preset image expansion mode to obtain an expanded device image, and use the expanded device image and the historical device image as sample training images; the preset image expansion mode comprises at least one of geometric transformation, color space change and pixel relation adjustment.
Optionally, the geometric transformation comprises at least one of a rotation transformation, a scaling transformation, a cropping transformation, a translation transformation, and an affine transformation.
Optionally, the color space variation comprises at least one of a contrast variation, a brightness variation, a saturation variation, a histogram enhancement, and a gamma adjustment.
Optionally, the pixel relation adjustment includes at least one of a blurring process, a sharpening process, and a noise process.
The device for identifying the equipment defect of the transformer substation can execute the method for identifying the equipment defect of the transformer substation provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method for identifying the equipment defect of the transformer substation.
Example four
Fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention. FIG. 5 illustrates a block diagram of an exemplary electronic device 40 suitable for use in implementing embodiments of the present invention. The electronic device 40 shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiment of the present invention.
As shown in fig. 5, electronic device 40 is embodied in the form of a general purpose computing device. The components of electronic device 40 may include, but are not limited to: one or more processors or processing units 401, a system memory 402, and a bus 403 that couples various system components including the system memory 402 and the processing unit 401.
Bus 403 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, micro-channel architecture (MAC) bus, enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.
Electronic device 40 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by electronic device 40 and includes both volatile and nonvolatile media, removable and non-removable media.
The system memory 402 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)404 and/or cache memory 405. The electronic device 40 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 406 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 5, commonly referred to as a "hard drive"). Although not shown in FIG. 5, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 403 by one or more data media interfaces. Memory 402 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.
A program/utility 408 having a set (at least one) of program modules 407 may be stored, for example, in memory 402, such program modules 407 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 407 generally perform the functions and/or methods of the described embodiments of the invention.
The electronic device 40 may also communicate with one or more external devices 409 (e.g., keyboard, pointing device, display 410, etc.), with one or more devices that enable a user to interact with the electronic device 40, and/or with any devices (e.g., network card, modem, etc.) that enable the electronic device 40 to communicate with one or more other computing devices. Such communication may be through input/output (I/O) interfaces 411. Also, the electronic device 40 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 412. As shown, the network adapter 412 communicates with the other modules of the electronic device 40 over the bus 403. It should be appreciated that although not shown in FIG. 5, other hardware and/or software modules may be used in conjunction with electronic device 40, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
The processing unit 401 executes various functional applications and data processing by running the program stored in the system memory 402, for example, implementing the method for identifying a defect in a substation device according to the embodiment of the present invention.
EXAMPLE seven
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform a method for identifying a substation device defect, the method including:
acquiring a target device image of to-be-detected substation equipment;
inputting the target equipment image into an equipment defect identification model which is trained in advance to obtain an identification result of the equipment defect of the transformer substation;
the equipment defect identification model comprises convolution modules and full-connection modules, wherein the convolution modules are connected in parallel, each convolution module comprises at least one convolution layer and at least one full-connection layer connected with the convolution layers in series, and each full-connection module comprises a full-connection module formed by at least one full-connection layer.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for embodiments of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will appreciate that the present invention is not limited to the particular embodiments described herein, and that various obvious changes, rearrangements and substitutions will now be apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims (18)

1. A transformer substation equipment defect identification method is characterized by comprising the following steps:
acquiring a target device image of to-be-detected substation equipment;
inputting the target equipment image into an equipment defect identification model which is trained in advance to obtain an identification result of the equipment defect of the transformer substation;
the equipment defect identification model comprises convolution modules and full-connection modules, wherein the convolution modules are connected in parallel, each convolution module comprises at least one convolution layer and at least one full-connection layer connected with the convolution layers in series, and each full-connection module comprises a full-connection module formed by at least one full-connection layer.
2. The method of claim 1, wherein the inputting the target device image into a pre-trained device defect recognition model to obtain a recognition result of a substation device defect comprises:
inputting the target equipment image into an equipment defect identification model which is trained in advance, and respectively obtaining a convolution output result output by the convolution module and a full-connection output result output by the full-connection module;
and determining the identification result of the defects of the transformer substation equipment according to the full connection output result and the convolution output result.
3. The method of claim 2, wherein the device defect identification model further comprises a feature map extraction module, the feature map extraction module comprising at least one convolutional layer;
the inputting the target device image into a device defect recognition model trained in advance to obtain a convolution output result output by the convolution module and a full-connection output result output by the full-connection module respectively includes:
inputting the target equipment image into a feature map extraction module of an equipment defect recognition model which is trained in advance to obtain a preliminary feature map;
and inputting the preliminary characteristic diagram into the full-connection module to obtain a full-connection output result, and inputting the preliminary characteristic diagram into the convolution module to obtain a convolution output result.
4. The method of claim 2, wherein the fully-connected output results comprise fully-connected classification results of the defects of the substation equipment, and the convolved output results comprise convolved classification results of the defects of the substation equipment; the identification result of the substation equipment defect comprises a target classification result of the substation equipment defect;
the determining of the identification result of the substation equipment defect according to the full-connection output result and the convolution output result comprises the following steps:
and performing result fusion on the full-connection classification result and the convolution classification result to obtain a target classification result of the defects of the substation equipment.
5. The method of claim 4, wherein the fusing the results of the fully-connected classification result and the convolution classification result to obtain a target classification result of the defect of the substation equipment comprises:
and performing result fusion on the full-connection classification result and the convolution classification result based on the following formula to obtain a target classification result of the defects of the substation equipment:
Figure FDA0003652560420000021
wherein p isiA defect classification result of the defect area of the substation equipment which is output by the equipment defect identification model is represented,
Figure FDA0003652560420000022
a defect classification result of a convolution module representing a device defect identification model,
Figure FDA0003652560420000023
and representing the defect classification result of the fully-connected module of the equipment defect identification model.
6. The method of claim 4, wherein the convolution output further comprises a convolution identification of a defect region of the substation equipment; the identification result of the substation equipment defect comprises a target identification result of a defect area of the substation equipment;
the determining of the identification result of the substation equipment defect according to the full-connection output result and the convolution output result comprises the following steps:
and taking the convolution identification result of the defect area of the substation equipment as a target identification result of the defect area of the substation equipment.
7. The method of claim 1, wherein the target loss function of the plant defect identification model is determined according to a loss value of a fully connected module of the plant defect identification model and a loss value of a convolution module.
8. The method of claim 7, wherein the target loss function of the equipment defect identification model is determined according to the loss values of the fully-connected modules and the loss values of the convolution modules of the equipment defect identification model in the following manner:
Figure FDA0003652560420000031
wherein L represents an objective loss function of the equipment defect identification model, LfcRepresents the loss value, L, of a fully connected moduleconvRepresenting the loss value of the convolution module, ω representing a weighting factor, ωfcIs | ln (1-L)fc)|,ωconvIs composed of
Figure FDA0003652560420000032
Figure FDA0003652560420000033
Indicating the correction parameters.
9. The method according to claim 1, wherein the acquiring of the target device image of the substation device to be detected comprises:
and receiving a target equipment image of the to-be-detected substation equipment, which is acquired by the target inspection equipment.
10. The method of claim 1, further comprising:
constructing a training data set for training an equipment defect recognition model, wherein the training data set comprises a sample equipment image and an expected recognition result corresponding to the sample equipment image;
inputting the sample equipment image into a model to be trained to obtain a model output result corresponding to the sample equipment image;
and performing model adjustment on the model to be trained according to the model output result corresponding to the sample equipment image, the expected recognition result and a pre-constructed model loss function to obtain an equipment defect recognition model.
11. The method of claim 10, wherein constructing a training data set for training a device defect recognition model comprises:
acquiring historical equipment images of power transformation equipment in a transformer substation, and determining sample equipment images according to the historical equipment images;
labeling the sample equipment image to obtain an expected identification result corresponding to the sample equipment image;
and training a training data set of an equipment defect recognition model according to the sample equipment image and the expected recognition result corresponding to the sample equipment image.
12. The method of claim 10, wherein determining a sample training image from the historical device images comprises:
performing image expansion processing on the historical equipment image according to a preset image expansion mode to obtain an expanded equipment image, and taking the expanded equipment image and the historical equipment image as sample training images;
the preset image expansion mode comprises at least one of geometric transformation, color space change and pixel relation adjustment.
13. The method of claim 12, wherein the geometric transformation comprises at least one of a rotation transformation, a scaling transformation, a cropping transformation, a translation transformation, and an affine transformation.
14. The method of claim 12, wherein the color space variation comprises at least one of contrast variation, brightness variation, saturation variation, histogram enhancement, and gray scale adjustment.
15. The method of claim 12, wherein the pixel relationship adjustment comprises at least one of a blurring process, a sharpening process, and a noise process.
16. A transformer substation equipment defect recognition device, characterized in that includes:
the device image acquisition module is used for acquiring a target device image of the to-be-detected substation device;
the equipment image input module is used for inputting the target equipment image into an equipment defect recognition model which is trained in advance to obtain a recognition result of the equipment defect of the transformer substation;
the equipment defect identification model comprises convolution modules and full-connection modules, wherein the convolution modules are connected in parallel, each convolution module comprises at least one convolution layer and at least one full-connection layer connected with the convolution layers in series, and each full-connection module comprises a full-connection module formed by at least one full-connection layer.
17. An electronic device, characterized in that the electronic device comprises:
one or more processors;
a storage device to store one or more programs,
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of identifying substation equipment defects of any of claims 1-15.
18. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out a method of identifying a substation device defect according to any one of claims 1 to 15.
CN202210556740.8A 2022-05-19 2022-05-19 Transformer substation equipment defect identification method and device, electronic equipment and storage medium Pending CN114757941A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210556740.8A CN114757941A (en) 2022-05-19 2022-05-19 Transformer substation equipment defect identification method and device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210556740.8A CN114757941A (en) 2022-05-19 2022-05-19 Transformer substation equipment defect identification method and device, electronic equipment and storage medium

Publications (1)

Publication Number Publication Date
CN114757941A true CN114757941A (en) 2022-07-15

Family

ID=82335441

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210556740.8A Pending CN114757941A (en) 2022-05-19 2022-05-19 Transformer substation equipment defect identification method and device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN114757941A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117351001A (en) * 2023-11-16 2024-01-05 肇庆市大正铝业有限公司 Surface defect identification method for regenerated aluminum alloy template
CN117351001B (en) * 2023-11-16 2024-05-28 肇庆市大正铝业有限公司 Surface defect identification method for regenerated aluminum alloy template

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117351001A (en) * 2023-11-16 2024-01-05 肇庆市大正铝业有限公司 Surface defect identification method for regenerated aluminum alloy template
CN117351001B (en) * 2023-11-16 2024-05-28 肇庆市大正铝业有限公司 Surface defect identification method for regenerated aluminum alloy template

Similar Documents

Publication Publication Date Title
CN110059694B (en) Intelligent identification method for character data in complex scene of power industry
CN109086811B (en) Multi-label image classification method and device and electronic equipment
CN111080693A (en) Robot autonomous classification grabbing method based on YOLOv3
CN111652225B (en) Non-invasive camera shooting and reading method and system based on deep learning
CN112070134A (en) Power equipment image classification method and device, power equipment and storage medium
CN112766110A (en) Training method of object defect recognition model, object defect recognition method and device
CN113052295B (en) Training method of neural network, object detection method, device and equipment
CN117593304B (en) Semi-supervised industrial product surface defect detection method based on cross local global features
CN110334775B (en) Unmanned aerial vehicle line fault identification method and device based on width learning
CN111310837A (en) Vehicle refitting recognition method, device, system, medium and equipment
CN116823793A (en) Device defect detection method, device, electronic device and readable storage medium
CN115577768A (en) Semi-supervised model training method and device
CN115410024A (en) Power image defect detection method based on dynamic activation thermodynamic diagram
CN113989604A (en) Tire DOT information identification method based on end-to-end deep learning
CN117152094A (en) Method, device and system for analyzing surface defects of steel plate based on computer vision
CN110728316A (en) Classroom behavior detection method, system, device and storage medium
CN116152191A (en) Display screen crack defect detection method, device and equipment based on deep learning
CN114757941A (en) Transformer substation equipment defect identification method and device, electronic equipment and storage medium
CN115330705A (en) Skin paint surface defect detection method based on adaptive weighting template NCC
CN114782822A (en) Method and device for detecting state of power equipment, electronic equipment and storage medium
CN113283429A (en) Liquid level meter reading method based on deep convolutional neural network
CN114596244A (en) Infrared image identification method and system based on visual processing and multi-feature fusion
CN111259843B (en) Multimedia navigator testing method based on visual stability feature classification registration
US20220253995A1 (en) Method and system for checking data gathering conditions associated with image-data during ai enabled visual-inspection process
CN114943895A (en) Transformer substation ground oil stain identification method and device, electronic equipment and storage medium

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