CN118072152A - Power equipment operation fault detection method and system based on image fusion - Google Patents
Power equipment operation fault detection method and system based on image fusion Download PDFInfo
- Publication number
- CN118072152A CN118072152A CN202410464753.1A CN202410464753A CN118072152A CN 118072152 A CN118072152 A CN 118072152A CN 202410464753 A CN202410464753 A CN 202410464753A CN 118072152 A CN118072152 A CN 118072152A
- Authority
- CN
- China
- Prior art keywords
- image
- fault
- outline
- map
- texture
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 69
- 230000004927 fusion Effects 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 claims description 67
- 230000005856 abnormality Effects 0.000 claims description 38
- 238000010586 diagram Methods 0.000 claims description 30
- 238000013527 convolutional neural network Methods 0.000 claims description 23
- 238000001931 thermography Methods 0.000 claims description 23
- 230000002159 abnormal effect Effects 0.000 claims description 19
- 238000001914 filtration Methods 0.000 claims description 17
- 230000002146 bilateral effect Effects 0.000 claims description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000012935 Averaging Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 5
- 238000011084 recovery Methods 0.000 claims description 4
- 238000010438 heat treatment Methods 0.000 description 6
- 238000005452 bending Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 230000008030 elimination Effects 0.000 description 3
- 238000003379 elimination reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention is suitable for the technical field of power detection, and provides a power equipment operation fault detection method and system based on image fusion.
Description
Technical Field
The invention relates to the technical field of power detection and extraction, in particular to a power equipment operation fault detection method and system based on image fusion.
Background
During operation of the power plant, a wide variety of faults inevitably occur, with thermal faults being one of the most common types of faults. Once the thermal fault point in the equipment cannot be timely detected, the local temperature is abnormally increased, tripping and power failure of the equipment are easily caused, and a great challenge is brought to safe and stable operation of the power equipment. At present, the infrared thermal imaging technology is a common technical means for diagnosing the thermal faults of the power equipment, has the advantage of non-contact live detection, and can effectively detect the temperature field information of the power equipment. However, with the continuous development of the power system in China, the pressure of the infrared inspection task of the power equipment is gradually increased.
At present, the thermal fault diagnosis mode of the power equipment is mainly to judge and analyze through collected infrared images, so that a thermal fault area can be accurately locked, but when a certain part of the power equipment generates heat abnormally due to the quality problem (burrs, defects or bending and the like of a conductive joint), the infrared images cannot intuitively present the image information of the burrs, defects or bending and the like of the conductive joint to workers, so that the problem of the thermal fault of the power equipment cannot be effectively solved.
Disclosure of Invention
Aiming at the defects existing in the prior art, the invention aims to provide an electric equipment operation fault detection method and system based on image fusion, so as to solve the problems existing in the background art.
The invention is realized in such a way that the method for detecting the operation faults of the power equipment based on image fusion comprises the following steps:
Acquiring image data of a power equipment component, wherein the image data comprises an infrared thermal imaging image and a visible light image, and preprocessing the image data;
respectively extracting a thermal fault region distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
classifying and identifying a thermal fault area distribution map and an outline map by using a convolutional neural network algorithm;
The method for classifying and identifying the thermal fault regional distribution map and the outline map by utilizing the convolutional neural network algorithm comprises the following steps:
inputting the thermal fault region distribution diagram and the outline diagram into a convolutional neural network model;
The bilateral filtering principle is utilized to carry out filtering optimization detection on the thermal fault region distribution map, and a final detection result of the thermal fault region is obtained;
Identifying a texture abnormality detection result or a shape abnormality detection result in the outline map by utilizing a DCT-based hash method;
and carrying out image fusion analysis on the thermal fault region distribution diagram and the outline diagram, and outputting fault information.
As a further scheme of the invention: the step of preprocessing the image data specifically includes:
Gray sampling the image data on the M x N lattice, and quantizing to obtain a processed digital image;
eliminating random noise in the image data by using a median method or a local averaging method or a k-nearest neighbor averaging method;
the image data is corrected according to the least square method.
As a further scheme of the invention: the step of respectively extracting the thermal fault region distribution map and the outline map in the preprocessed infrared thermal imaging image and the preprocessed visible light image specifically comprises the following steps:
respectively extracting a thermal fault area distribution map and an outline map of the same position of the infrared thermal imaging image and the visible light image by using a low-rank matrix recovery algorithm;
brightness values of the thermal fault region profile and the outline profile are adjusted to be consistent using a brightness-contrast transfer technique.
As a further scheme of the invention: the convolutional neural network model stores a component visible light image standard library, and the step of identifying a texture abnormality detection result or a shape abnormality detection result in the outline profile by utilizing a DCT-based hash method specifically comprises the following steps:
the method comprises the steps of marking a part name on an outline map, and inputting the outline map marked with the part name into a visible light image standard library for matching, wherein the visible light image standard library comprises a plurality of texture abnormal image templates and shape abnormal image templates;
extracting a local texture image and a local shape image on the outline map;
Respectively calculating the hash values of the local texture image and the texture abnormal image template by using a DCT-based hash method to obtain h_1 and h_2;
Respectively calculating the hash values of the local shape image and the shape anomaly image template by using a DCT-based hash method to obtain h_3 and h_4;
calculating a hamming distance dis_h1 between h_1 and h_2, and calculating a hamming distance dis_h2 between h_3 and h_4;
Calculating according to the Hamming distance dis_h1 and the Hamming distance dis_h2 to obtain a first similarity between the local texture image and the texture abnormal image template and a second similarity between the local shape image and the shape abnormal image template, wherein the similarity is a matching value;
And outputting a texture abnormality detection result or a shape abnormality detection result.
As a further scheme of the invention: the step of carrying out image fusion analysis on the thermal fault region distribution diagram and the outline diagram and outputting fault information specifically comprises the following steps:
texture, shape and contour matching are carried out on the thermal fault area distribution diagram and the contour diagram;
On the premise of contour matching, whether textures or shapes are matched or not is determined, and when the textures are matched, a fault region distribution map and a corresponding texture abnormality detection result or shape abnormality detection result are output.
As a further scheme of the invention: the calculation formula using the brightness-contrast transfer technique:
：
Wherein M is an input image, F is an output image, mean is the mean of M, mean_ref is the mean of the reference image, ms is the mean variance of M, msref is the mean variance of the reference image, the mean is used for reflecting the average brightness of the image, and the variance is used for representing the contrast of the image.
An image fusion-based power equipment operational fault detection system, the system comprising:
the image acquisition module is used for acquiring image data of the power equipment component, wherein the image data comprises an infrared thermal imaging image and a visible light image;
the data preprocessing module is used for preprocessing the image data;
the image extraction module is used for respectively extracting a thermal fault area distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
The classification and identification module is used for classifying and identifying a thermal fault region distribution map and an outline map by utilizing a convolutional neural network algorithm;
the fault output module is used for carrying out image fusion analysis on the thermal fault regional distribution diagram and the outline diagram and outputting fault information.
As a further scheme of the invention: the classification and identification module comprises:
the input unit is used for inputting the thermal fault region distribution diagram and the outline drawing into the convolutional neural network model;
The bilateral filter is used for carrying out filtering optimization detection on the thermal fault region distribution map by utilizing the bilateral filtering principle to obtain a final detection result of the thermal fault region;
And the identification unit is used for identifying the texture abnormality detection result or the shape abnormality detection result in the outline profile by utilizing a DCT-based hash method.
Compared with the prior art, the method and the device have the advantages that the final detection result of the thermal fault area and the texture abnormality detection result or the shape abnormality detection result in the outline profile are respectively obtained by utilizing the convolutional neural network model, and then whether the textures or the shapes are matched is determined on the premise of profile matching, so that heating abnormality data of the power equipment part, part characterization information (burrs, texture deformation or shape bending on the surface of the part) causing the heating abnormality and the like can be simultaneously identified, a user can more intuitively and carefully know the part problem, and the method and the device have the characteristics of being specific and intuitive in fault analysis.
Drawings
Fig. 1 is a flowchart of a method for detecting an operation failure of a power device based on image fusion.
Fig. 2 is a flowchart of a thermal fault area distribution diagram and an outline diagram in an infrared thermal imaging image and a visible light image after pretreatment extracted respectively in an operation fault detection method of power equipment based on image fusion.
Fig. 3 is a flowchart of performing image fusion analysis on a thermal fault area distribution diagram and an outline diagram in an electric power equipment operation fault detection method based on image fusion, and outputting fault information.
Fig. 4 is a schematic structural diagram of an image fusion-based power equipment operation fault detection system.
Fig. 5 is a schematic structural diagram of a classification and identification module in a power equipment operation fault detection system based on image fusion.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more clear, the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Specific implementations of the invention are described in detail below in connection with specific embodiments.
As shown in fig. 1, an embodiment of the present invention provides a method for detecting an operation failure of an electrical device based on image fusion, where the method includes the following steps:
s100, acquiring image data of a power equipment component, wherein the image data comprises an infrared thermal imaging image and a visible light image, and preprocessing the image data;
S200, respectively extracting a thermal fault area distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
s300, classifying and identifying a thermal fault region distribution map and an outline map by using a convolutional neural network algorithm;
The method for classifying and identifying the thermal fault regional distribution map and the outline map by utilizing the convolutional neural network algorithm comprises the following steps:
s310, inputting a thermal fault region distribution diagram and an outline drawing into a convolutional neural network model;
s320, filtering optimization detection is carried out on the thermal fault region distribution map by utilizing a bilateral filtering principle, and a final detection result of the thermal fault region is obtained;
s330, identifying a texture abnormality detection result or a shape abnormality detection result in the outline profile by utilizing a DCT-based hash method;
S400, performing image fusion analysis on the thermal fault region distribution diagram and the outline diagram, and outputting fault information.
In the embodiment of the invention, the convolutional neural network comprises a feature extractor consisting of a convolutional layer and a sampling layer, a plurality of feature planes are usually contained in one convolutional layer of the CNN, each feature plane is initialized by a plurality of rectangular arrangement forms, a convolutional kernel is learned to obtain reasonable weight in the training process of the network, and a convolutional neural network model comprises two forms of mean sampling and maximum sampling, so that the model complexity is greatly simplified by convolution and sampling, and the parameters of the model are reduced.
Bilateral filtering (Bilateral filter) is a nonlinear filtering method, is essentially based on Gaussian filtering, and aims to solve edge blurring caused by Gaussian filtering. The spatial proximity of the image and the pixel value similarity are combined to perform data processing, and the spatial domain information and the gray level similarity are considered at the same time, so that the purposes of edge protection and denoising are achieved. Has the characteristics of simplicity, non-iteration and local.
The bilateral filter has the advantages that edge preservation can be carried out, the edge is obviously blurred by removing noise through Gaussian filtering, the protection effect on high-frequency details is not obvious, the bilateral filter is a Gaussian kernel which is based on pixel color distribution as the name implies, so that the influence is larger when two pixels are very close to each other near the edge only when the colors are very close to each other, otherwise, the color difference is larger although the distance is very close, the smoothing weight is very small, the pixel value near the edge is kept, and the edge protection effect is achieved.
As a preferred embodiment of the present invention, the step of preprocessing the image data specifically includes:
S110, gray sampling is carried out on the image data on the M-N dot matrix, and quantization is carried out to obtain a processed digital image;
s120, eliminating random noise in the image data by using a median method or a local averaging method or a k-nearest neighbor averaging method;
S130, correcting the image data according to a least square method.
In the embodiment of the invention, the image data is quantized (classified into one of 2b gray levels) while the gray level of the image data is sampled on an M x N lattice, so that a digital image which can be processed by a computer can be obtained, an algebraic method of image restoration is based on a least square method optimal criterion, an estimated value is sought, and a merit criterion function value is minimized.
As shown in fig. 2, as a preferred embodiment of the present invention, the step of extracting the thermal fault region distribution map and the outline map in the preprocessed infrared thermal imaging image and the visible light image respectively specifically includes:
s210, respectively extracting a thermal fault area distribution map and an outline map of the same position of an infrared thermal imaging image and a visible light image by using a low-rank matrix recovery algorithm;
S220, adjusting the brightness values of the thermal fault region distribution map and the outline map to be consistent by utilizing a brightness-contrast transmission technology.
In the embodiment of the invention, a low-rank matrix recovery algorithm: min l + lambda s1, s.t d-l-s < sigma,
Wherein d is an acquired infrared image of the power equipment, s is a thermal fault area in the infrared image of the power equipment, and l is a background area in the infrared image of the power equipment; the term @ represents a matrix kernel norm, the term @ 1 represents an L1 norm, λ is a regularization parameter for adjusting a weight size of the sparse term,M and n are the length and width of the input image, respectively; σ is a constant representing the intensity of gaussian noise and is an empirical set point.
The calculation formula using the brightness-contrast transfer technique:
：
Wherein M is an input image, F is an output image, mean is the mean of M, mean_ref is the mean of the reference image, ms is the mean variance of M, msref is the mean variance of the reference image, the mean is used for reflecting the average brightness of the image, and the variance is used for representing the contrast of the image.
As a preferred embodiment of the present invention, the convolutional neural network model stores a component visible light image standard library, and the step of identifying the texture anomaly detection result or the shape anomaly detection result in the contour map by using a DCT-based hash method specifically includes:
S3301, marking a part name on an outline map, and inputting the outline map marked with the part name into a visible light image standard library for matching, wherein the visible light image standard library comprises a plurality of abnormal texture image templates and abnormal shape image templates;
s3302, extracting a local texture image and a local shape image on the outline drawing.
S3303, respectively calculating the hash values of the local texture image and the texture abnormal image template by using a DCT-based hash method to obtain h_1 and h_2;
S3304, respectively calculating the hash values of the local shape image and the shape abnormal image template by using a DCT-based hash method to obtain h_3 and h_4;
s3305, calculating a Hamming distance dis_h1 between h_1 and h_2, and calculating a Hamming distance dis_h2 between h_3 and h_4;
S3306, calculating to obtain a first similarity between the local texture image and the texture abnormal image template and a second similarity between the local shape image and the shape abnormal image template according to the Hamming distance dis_h1 and the Hamming distance dis_h2, wherein the similarity is a matching value;
S3307, a texture abnormality detection result or a shape abnormality detection result is output.
In the embodiment of the invention, a DCT-based hash method is used for identifying pictures as an AI picture identification method in the prior art, the DCT-based hash method uses discrete cosine transform to extract low-frequency components of the pictures, firstly converts the pictures into gray level images with standard sizes, then carries out DCT transform on the gray level images, then extracts 64-bit hash values from coefficient matrixes as fingerprints, then calculates hamming distances dis_h between h_1 and h_2 and hamming distances dis_h2 between h_3 and h_4, and calculates to obtain a first similarity between a local texture image and a texture abnormal image template and a second similarity between a local shape image and a shape abnormal image template according to the hamming distances dis_h1 and the hamming distances dis_h2, wherein the method for calculating the similarity between the two pictures is the prior art and is not described in detail herein.
As shown in fig. 3, as a preferred embodiment of the present invention, the step of performing image fusion analysis on the thermal fault region distribution map and the outline map and outputting fault information specifically includes:
s410, performing texture, shape and contour matching on the thermal fault region distribution map and the outline map;
S420, on the premise of contour matching, determining whether textures or shapes are matched, and when the textures are matched, outputting a fault region distribution map and a corresponding texture abnormality detection result or shape abnormality detection result.
In the embodiment of the invention, when the contact surface of the circuit joint of the power equipment is improperly processed, such as burrs, uneven contact surface, bending or torsion angle of the conductor and the like, the contact surface can be analyzed from texture, shape and contour matching of the outline profile, and the factors can cause heating phenomena, and the heating phenomena can be reflected from a thermal fault area distribution diagram.
As shown in fig. 4, an image fusion-based power equipment operation fault detection system is characterized in that the system comprises:
an image acquisition module 100 for acquiring image data of the electrical equipment component, the image data including an infrared thermographic image and a visible light image;
a data preprocessing module 200 for preprocessing the above image data;
The image extraction module 300 is used for respectively extracting a thermal fault area distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
The classification and identification module 400 is used for classifying and identifying a thermal fault area distribution map and an outline map by using a convolutional neural network algorithm;
the fault output module 500 is configured to perform image fusion analysis on the thermal fault region distribution map and the outline map, and output fault information.
In the embodiment of the invention, the image acquisition module 100 comprises an infrared thermal imaging camera and a visible light camera, the infrared thermal imaging camera is used for acquiring an infrared thermal imaging image of a power component, the visible light camera is used for acquiring a visible light image of the power component, the data preprocessing module 200 comprises a digital processing unit 201, a noise elimination unit 202 and a data correction unit 203, the digital processing unit 201, the noise elimination unit 202 and the data correction unit 203 are in communication connection, the digital processing unit 201 is used for gray sampling image data on an M x N lattice and quantizing the image data to acquire a processed digital image, the noise elimination unit 202 eliminates random noise in the image data by using a median method or a local averaging method or a k nearest neighbor averaging method, and the data correction unit 203 corrects the image data according to a least square method;
The image extraction module 300 includes an image extraction unit 301 and a brightness value adjustment unit 302, where the image extraction unit 301 and the brightness value adjustment unit 302 are connected in communication, the image extraction unit 301 extracts a thermal fault area distribution map and an outline map of the same position of the infrared thermal imaging image and the visible light image respectively by using a low-rank matrix restoration algorithm, and the brightness value adjustment unit 302 adjusts brightness values of the thermal fault area distribution map and the outline map to be consistent by using a brightness-contrast transmission technology.
As shown in fig. 5, as a preferred embodiment of the present invention, the classification recognition module 400 includes:
An input unit 401 for inputting the thermal fault region distribution map and the outline map into a convolutional neural network model;
the bilateral filter 402 is configured to perform filtering optimization detection on the thermal fault region distribution map by using a bilateral filtering principle, so as to obtain a final detection result of the thermal fault region;
An identifying unit 403 for identifying a texture abnormality detection result or a shape abnormality detection result in the outline map using a DCT-based hash method.
In the embodiment of the present invention, the input unit 401, the bilateral filter 402, and the recognition unit 403 are in communication connection, the principle that the bilateral filter 402 obtains the monitoring result of the thermal fault area is widely applied in the prior art, so that the description is omitted, the recognition unit 403 includes a first hash value calculation subunit 4031, a second hash value calculation subunit 4032, a hamming distance calculation subunit 4033, a matching subunit 4034, and an output subunit 4035, where the first hash value calculation subunit 4031, the second hash value calculation subunit 4032, the matching subunit 4033, the hamming distance calculation subunit 4044, and the output subunit 4045 are in communication connection, the first hash value calculation subunit 4031 calculates the hash values of the local texture image and the texture abnormal image template respectively by using a DCT-based hash method, obtaining h_1 and h_2, respectively calculating the hash values of the local shape image and the shape anomaly image template by a hash method based on DCT (discrete cosine transform), obtaining h_3 and h_4, calculating the Hamming distance dis_h1 between h_1 and h_2 by a Hamming distance calculation subunit 4033, calculating the Hamming distance dis_h2 between h_3 and h_4, and obtaining the similarity I between the local texture image and the texture anomaly image template and the similarity II between the local shape image and the shape anomaly image template by a matching subunit 4034 according to the Hamming distance dis_h1 and the Hamming distance dis_h2 respectively, wherein the similarity is a matching value, and outputting a texture anomaly detection result or a shape anomaly detection result by an output subunit 4035;
the fault output module 500 comprises an image matching unit 501 and an information output unit 502, the image matching unit 501 and the information output unit 502 are in communication connection, the image matching unit 501 performs texture, shape and contour matching on the thermal fault region distribution map and the outline map, on the premise of contour matching, whether the texture or the shape is matched is determined, and when the texture is matched, the information output unit 502 simultaneously outputs the fault region distribution map and a corresponding texture abnormality detection result or shape abnormality detection result.
In summary, the final detection result of the thermal fault area and the texture abnormality detection result or the shape abnormality detection result in the outline profile are obtained by using the convolutional neural network model, and then whether the texture or the shape is matched is determined on the premise of profile matching, so that heating abnormality data of the power equipment component, component characterization information (burrs, texture deformation or shape bending on the surface of the component) causing the heating abnormality and the like can be identified at the same time, a user can more intuitively and carefully know the component problem, and the method has the characteristics of specific and intuitive fault analysis.
It should be understood that, although the steps in the flowcharts of the embodiments of the present invention are shown in order as indicated by the arrows, these steps are not necessarily performed in order as 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 various embodiments may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or other steps.
Those skilled in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by a computer program for instructing relevant hardware, where the program may be stored in a non-volatile computer readable storage medium, and where the program, when executed, may include processes in the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
Claims (8)
1. An electric power equipment operation fault detection method based on image fusion is characterized by comprising the following steps:
Acquiring image data of a power equipment component, wherein the image data comprises an infrared thermal imaging image and a visible light image, and preprocessing the image data;
respectively extracting a thermal fault region distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
classifying and identifying a thermal fault area distribution map and an outline map by using a convolutional neural network algorithm;
The method for classifying and identifying the thermal fault regional distribution map and the outline map by utilizing the convolutional neural network algorithm comprises the following steps:
inputting the thermal fault region distribution diagram and the outline diagram into a convolutional neural network model;
The bilateral filtering principle is utilized to carry out filtering optimization detection on the thermal fault region distribution map, and a final detection result of the thermal fault region is obtained;
Identifying a texture abnormality detection result or a shape abnormality detection result in the outline map by utilizing a DCT-based hash method;
and carrying out image fusion analysis on the thermal fault region distribution diagram and the outline diagram, and outputting fault information.
2. The method for detecting the operation fault of the electric equipment based on the image fusion according to claim 1, wherein the step of preprocessing the image data specifically comprises the following steps:
Gray sampling the image data on the M x N lattice, and quantizing to obtain a processed digital image;
eliminating random noise in the image data by using a median method or a local averaging method or a k-nearest neighbor averaging method;
the image data is corrected according to the least square method.
3. The method for detecting the operation fault of the electric equipment based on the image fusion according to claim 2, wherein the step of respectively extracting the thermal fault region distribution map and the outline map in the preprocessed infrared thermal imaging image and the preprocessed visible light image specifically comprises the following steps:
respectively extracting a thermal fault area distribution map and an outline map of the same position of the infrared thermal imaging image and the visible light image by using a low-rank matrix recovery algorithm;
brightness values of the thermal fault region profile and the outline profile are adjusted to be consistent using a brightness-contrast transfer technique.
4. The method for detecting the operation fault of the electrical equipment based on the image fusion according to claim 3, wherein the convolutional neural network model stores a component visible light image standard library, and the step of identifying the texture abnormality detection result or the shape abnormality detection result in the outline map by using a hash method based on DCT specifically comprises the following steps:
the method comprises the steps of marking a part name on an outline map, and inputting the outline map marked with the part name into a visible light image standard library for matching, wherein the visible light image standard library comprises a plurality of texture abnormal image templates and shape abnormal image templates;
extracting a local texture image and a local shape image on the outline map;
Respectively calculating the hash values of the local texture image and the texture abnormal image template by using a DCT-based hash method to obtain h_1 and h_2;
Respectively calculating the hash values of the local shape image and the shape anomaly image template by using a DCT-based hash method to obtain h_3 and h_4;
calculating a hamming distance dis_h1 between h_1 and h_2, and calculating a hamming distance dis_h2 between h_3 and h_4;
Calculating according to the Hamming distance dis_h1 and the Hamming distance dis_h2 to obtain a first similarity between the local texture image and the texture abnormal image template and a second similarity between the local shape image and the shape abnormal image template, wherein the similarity is a matching value;
And outputting a texture abnormality detection result or a shape abnormality detection result.
5. The method for detecting the operation fault of the power equipment based on the image fusion according to claim 4, wherein the step of performing the image fusion analysis on the thermal fault region distribution map and the outline map and outputting the fault information specifically comprises the following steps:
texture, shape and contour matching are carried out on the thermal fault area distribution diagram and the contour diagram;
On the premise of contour matching, whether textures or shapes are matched or not is determined, and when the textures are matched, a fault region distribution map and a corresponding texture abnormality detection result or shape abnormality detection result are output.
6. A method for detecting an operation failure of an electrical device based on image fusion according to claim 3, wherein the calculation formula using a brightness-contrast transfer technique is as follows:
：
Wherein M is an input image, F is an output image, mean is the mean of M, mean_ref is the mean of the reference image, ms is the mean variance of M, msref is the mean variance of the reference image, the mean is used for reflecting the average brightness of the image, and the variance is used for representing the contrast of the image.
7. An image fusion-based power equipment operation fault detection system, the system comprising:
the image acquisition module is used for acquiring image data of the power equipment component, wherein the image data comprises an infrared thermal imaging image and a visible light image;
the data preprocessing module is used for preprocessing the image data;
the image extraction module is used for respectively extracting a thermal fault area distribution map in the preprocessed infrared thermal imaging image and an outline map in the visible light image;
The classification and identification module is used for classifying and identifying a thermal fault region distribution map and an outline map by utilizing a convolutional neural network algorithm;
the fault output module is used for carrying out image fusion analysis on the thermal fault regional distribution diagram and the outline diagram and outputting fault information.
8. The image fusion-based power equipment operation fault detection system of claim 7, wherein the classification recognition module comprises:
the input unit is used for inputting the thermal fault region distribution diagram and the outline drawing into the convolutional neural network model;
The bilateral filter is used for carrying out filtering optimization detection on the thermal fault region distribution map by utilizing the bilateral filtering principle to obtain a final detection result of the thermal fault region;
And the identification unit is used for identifying the texture abnormality detection result or the shape abnormality detection result in the outline profile by utilizing a DCT-based hash method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410464753.1A CN118072152A (en) | 2024-04-18 | 2024-04-18 | Power equipment operation fault detection method and system based on image fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410464753.1A CN118072152A (en) | 2024-04-18 | 2024-04-18 | Power equipment operation fault detection method and system based on image fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118072152A true CN118072152A (en) | 2024-05-24 |
Family
ID=91104323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410464753.1A Pending CN118072152A (en) | 2024-04-18 | 2024-04-18 | Power equipment operation fault detection method and system based on image fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118072152A (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150269745A1 (en) * | 2014-03-21 | 2015-09-24 | Nice-Systems Ltd. | Robust change-detection system and method |
| CN105760883A (en) * | 2016-02-15 | 2016-07-13 | 西安科技大学 | Belt conveyer key component automatic identification method based on infrared thermography |
| CN108254077A (en) * | 2018-01-02 | 2018-07-06 | 国网上海市电力公司 | The GIS thermal fault detection methods merged based on part with global characteristics information |
| CN110266268A (en) * | 2019-06-26 | 2019-09-20 | 武汉理工大学 | A kind of photovoltaic module fault detection method based on image co-registration identification |
| CN112733950A (en) * | 2021-01-18 | 2021-04-30 | 湖北工业大学 | Power equipment fault diagnosis method based on combination of image fusion and target detection |
| CN113034465A (en) * | 2021-03-23 | 2021-06-25 | 国网湖南省电力有限公司 | Power equipment thermal fault monitoring method, device and medium based on infrared image |
| US20220319011A1 (en) * | 2020-06-08 | 2022-10-06 | Shanghai Jiaotong University | Heterogeneous Image Registration Method and System |
| CN115861268A (en) * | 2022-12-20 | 2023-03-28 | 无锡四海电子科技有限公司 | Detection method and system of motor controller, computer equipment and storage medium |
| CN116977907A (en) * | 2023-06-25 | 2023-10-31 | 国网山东省电力公司金乡县供电公司 | Image recognition-based power inspection fault diagnosis method and system |
| CN117227553A (en) * | 2023-09-01 | 2023-12-15 | 山东云电网络科技有限公司 | Integrated super charging application system |
-
2024
- 2024-04-18 CN CN202410464753.1A patent/CN118072152A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150269745A1 (en) * | 2014-03-21 | 2015-09-24 | Nice-Systems Ltd. | Robust change-detection system and method |
| CN105760883A (en) * | 2016-02-15 | 2016-07-13 | 西安科技大学 | Belt conveyer key component automatic identification method based on infrared thermography |
| CN108254077A (en) * | 2018-01-02 | 2018-07-06 | 国网上海市电力公司 | The GIS thermal fault detection methods merged based on part with global characteristics information |
| CN110266268A (en) * | 2019-06-26 | 2019-09-20 | 武汉理工大学 | A kind of photovoltaic module fault detection method based on image co-registration identification |
| US20220319011A1 (en) * | 2020-06-08 | 2022-10-06 | Shanghai Jiaotong University | Heterogeneous Image Registration Method and System |
| CN112733950A (en) * | 2021-01-18 | 2021-04-30 | 湖北工业大学 | Power equipment fault diagnosis method based on combination of image fusion and target detection |
| CN113034465A (en) * | 2021-03-23 | 2021-06-25 | 国网湖南省电力有限公司 | Power equipment thermal fault monitoring method, device and medium based on infrared image |
| CN115861268A (en) * | 2022-12-20 | 2023-03-28 | 无锡四海电子科技有限公司 | Detection method and system of motor controller, computer equipment and storage medium |
| CN116977907A (en) * | 2023-06-25 | 2023-10-31 | 国网山东省电力公司金乡县供电公司 | Image recognition-based power inspection fault diagnosis method and system |
| CN117227553A (en) * | 2023-09-01 | 2023-12-15 | 山东云电网络科技有限公司 | Integrated super charging application system |
Non-Patent Citations (2)
| Title |
|---|
| YINGNAN GENG; WEIXUAN DIAO; YAN ZHAO: "Infrared-RGB Image Registration for Power Thermal Fault Detection Based on Gradient Hash Matching", 2022 IEEE 22ND INTERNATIONAL CONFERENCE ON COMMUNICATION TECHNOLOGY (ICCT), 27 March 2023 (2023-03-27) * |
| 吕志宁;: "输电线路常见故障分析与检测方法综述", 自动化与仪器仪表, no. 01, 25 January 2020 (2020-01-25) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021000524A1 (en) | Hole protection cap detection method and apparatus, computer device and storage medium | |
| CN110458839B (en) | Effective wire and cable monitoring system | |
| CN110033431B (en) | Non-contact detection device and detection method for detecting corrosion area on surface of steel bridge | |
| KR20180109665A (en) | A method and apparatus of image processing for object detection | |
| CN111199523B (en) | Power equipment identification method, device, computer equipment and storage medium | |
| CN114902279A (en) | Automated defect detection based on machine vision | |
| CN113160200B (en) | Industrial image defect detection method and system based on multi-task twin network | |
| WO2021012647A1 (en) | Face verification method and apparatus, server and readable storage medium | |
| CN115436382B (en) | Defect detection system and method for preventing strip breakage of lithium battery pole piece | |
| CN111563896A (en) | Image processing method for catenary anomaly detection | |
| CN116797977A (en) | Method and device for identifying dynamic target of inspection robot and measuring temperature and storage medium | |
| CN113516619A (en) | Product surface flaw identification method based on image processing technology | |
| CN115512381A (en) | Text recognition method, text recognition device, text recognition equipment, storage medium and working machine | |
| Li et al. | Printed label defect detection using twice gradient matching based on improved cosine similarity measure | |
| CN118072152A (en) | Power equipment operation fault detection method and system based on image fusion | |
| CN112541478A (en) | Insulator string stain detection method and system based on binocular camera | |
| CN116758425A (en) | Automatic acceptance checking method and device for large-base photovoltaic power station | |
| KR101311309B1 (en) | Detecting method of image operation and detecting device of image operation | |
| CN114821035A (en) | Distance parameter identification method for infrared temperature measurement equipment of power equipment | |
| Okarma et al. | A method supporting fault-tolerant optical text recognition from video sequences recorded with handheld cameras | |
| CN115272952A (en) | Safety monitoring method, device and system for new energy capital construction and storage medium | |
| CN115345842A (en) | Automobile key component similarity comparison method based on deep learning | |
| Dai et al. | Stripe noise removal for the thermal infrared spectrometer of the SDGSAT-1 | |
| CN113344987A (en) | Infrared and visible light image registration method and system for power equipment under complex background | |
| Sathyan et al. | Infrared Thermal Image Enhancement in Cold Spot Detection of Condenser Air Ingress. |
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 |