CN117173703B - Isolating switch state identification method - Google Patents
Isolating switch state identification method Download PDFInfo
- Publication number
- CN117173703B CN117173703B CN202311443231.5A CN202311443231A CN117173703B CN 117173703 B CN117173703 B CN 117173703B CN 202311443231 A CN202311443231 A CN 202311443231A CN 117173703 B CN117173703 B CN 117173703B
- Authority
- CN
- China
- Prior art keywords
- knife
- pixel
- pixel point
- area
- region
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000007547 defect Effects 0.000 claims abstract description 50
- 230000008859 change Effects 0.000 claims abstract description 30
- 238000005286 illumination Methods 0.000 claims abstract description 30
- 230000008447 perception Effects 0.000 claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims abstract description 14
- 239000011159 matrix material Substances 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 8
- 230000002950 deficient Effects 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 5
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 description 6
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000003064 k means clustering Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000020169 heat generation Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to a method for identifying the state of an isolating switch, which comprises the following steps: acquiring a gray level image of the isolating switch; setting the region of the knife and the connection part as an ROI region; dividing the ROI area on a space domain and a frequency domain respectively; combining energy characteristics on a frequency domain to obtain an overall energy value of pixel points in the ROI; constructing a knife switch illumination perception change of the pixel points; according to the gradient amplitude variation in the airspace, constructing a knife texture characteristic difference of the pixel points; acquiring a growth criterion of region growth; therefore, the complete segmentation of the knife defect area is effectively realized, the identification of the abnormal state of the isolating switch is completed, and the detection quality is improved.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a method for identifying the state of an isolating switch.
Background
In recent years, the development of the power industry in China is very rapid, and along with the construction and perfection of the smart grid, important operation equipment in a power system is also more and more emphasized. The isolating switch is used as a device capable of rapidly and safely cutting off a circuit, and can separate the circuit under the condition of not influencing current, so that the safety of maintenance personnel in the process of maintaining the circuit is ensured. Over time, the design of the isolating switch is gradually improved, and the isolating switch can be divided into a single-column type isolating switch, a double-column type isolating switch and a three-column type isolating switch according to the structure of the isolating column, so that the isolating switch can be applied to more scenes, the current can be controlled more accurately, and the operation risk is reduced.
For the isolating switch, the connection and closing of the knife are very important parts, and are necessary conditions for realizing the connection and disconnection of a circuit, but the isolating switch can be exposed outdoors for a long time to cause the problems of corrosion or mechanical damage of the connection part between the knife and the support column, and the like, so that the electrical performance of the isolating switch can be reduced, and high resistance can be generated on a current path, and unnecessary energy loss and heat generation are caused. Over time, excessive heating may cause risks of failure, damage to equipment, and the like. Even causing an arc to occur at the contact surface of the contacts or blades, causing sparks and explosions, and safety accidents when the personnel involved operate. Therefore, the abnormal state identification of the disconnecting switch knife is very important, however, parts are projected to the knife at the connection part under different illumination conditions, so that various characteristics of pixels of an image area are changed, the accuracy of distinguishing the defect and the normal area is reduced, the separated defect image of the area growth is incoherent and even incomplete, and the growth criterion of the area growth needs to be improved.
Disclosure of Invention
In order to solve the technical problems, the invention provides a method for identifying the state of an isolating switch, which aims to solve the existing problems.
The invention discloses a method for identifying the state of an isolating switch, which adopts the following technical scheme:
one embodiment of the invention provides a method for identifying the state of an isolating switch, which comprises the following steps:
acquiring a gray level image of the isolating switch;
acquiring a region of a knife and a connecting part according to a gray image, and marking the region as an ROI region; dividing the ROI area into a normal guillotine area, a shadow guillotine area and a connection area in an airspace; dividing the ROI area into a low-frequency sub-band and three high-frequency sub-bands in the frequency domain; obtaining coefficient matrixes corresponding to the high-frequency sub-bands; acquiring the total energy value of the pixel points in the ROI according to coefficient matrixes corresponding to the three high-frequency sub-bands; using the gray scale run matrix to construct a gray scale run matrix by taking the total energy values of the pixel points in three high-frequency sub-bands as input; acquiring the length of a run length code where a pixel point is located according to a gray scale run matrix; acquiring the switch blade illumination perception change of the pixel point according to the length of the run length code of the pixel point; acquiring the consistency of the horizontal gradient of the guillotine of the pixel point according to the amplitude variation of the horizontal gradient in the neighborhood window of the pixel point; acquiring longitudinal gradient consistency of a knife of a pixel point; acquiring the difference of the texture characteristics of the guillotine of the pixel point according to the consistency of the transverse gradient and the consistency of the longitudinal gradient of the guillotine of the pixel point;
acquiring a growth criterion of regional growth according to the switch blade illumination perception change of the pixel points and the switch blade texture characteristic difference; completing the segmentation of the knife switch defect area according to the growth criterion; and completing identification of the abnormal state of the isolating switch according to the knife defect area.
Preferably, the acquiring the regions of the knife and the connecting portion according to the gray image specifically includes:
segmenting the image by using a semantic segmentation network, marking a knife region and a joint as 1, and marking other regions as 0; the semantic segmentation network then extracts the regions of the blades and the connected portions.
Preferably, the ROI area is spatially divided into a normal guillotine area, a shadow guillotine area and a connection area, and the specific method is as follows:
dividing the ROI into three areas by adopting a clustering algorithm;
and (3) marking the region with the largest gray average value in the three regions as a normal knife switch region, marking the second largest gray average value as a connecting region, and marking the region with the smallest gray average value as a shadow knife switch region.
Preferably, the ROI area is divided into one low frequency subband and three high frequency subbands in the frequency domain, specifically:
a layer of two-dimensional wavelet transform is performed on the ROI area using a wavelet basis function to divide the ROI area into a low frequency subband, a horizontal high frequency subband, a vertical high frequency subband, and a diagonal high frequency subband.
Preferably, the obtaining the total energy value of the pixel point in the ROI area according to the coefficient matrix corresponding to the three high-frequency subbands specifically includes:
taking the values of the half of the horizontal coordinates and the vertical coordinates of the pixel points which are rounded upwards as the row and column values of the corresponding coefficient matrix elements; and taking the square average value of the element values of the corresponding coefficient matrix of the pixel points in the three high-frequency sub-bands as the total energy value of the pixel points.
Preferably, the obtaining the switch blade illumination sensing change of the pixel point according to the length of the run length code where the pixel point is located specifically includes:
taking the opposite number of the length of the run length code where the pixel point is located as an independent variable of an exponential function taking a natural constant as a base number; acquiring an energy average value of an area where a pixel point is located; calculating a normalization result of the total energy value of the pixel point and the absolute value of the difference value of the energy mean value; taking the product of the normalization result and the calculation result of the exponential function as the knife illumination perception change of the pixel point.
Preferably, the consistency of the horizontal gradient of the knife switch of each pixel point is obtained according to the gradient amplitude variation of the pixel point in the horizontal direction in the neighborhood window of the pixel point, and the expression is:
in the method, in the process of the invention,knife lateral gradient uniformity for center pixel,/->Horizontal gradient magnitude for window center pixel, < >>、/>Left side and right side of the horizontal direction of the center pixel respectively +.>Gradient magnitude of each unit pixel.
Preferably, the obtaining the difference of the texture characteristics of the knife of the pixel according to the consistency of the knife transverse gradient and the consistency of the knife longitudinal gradient of the pixel specifically includes:
acquiring an entropy value of a window where the pixel point is located through a gray level co-occurrence matrix; calculating Euclidean distance between the consistency of the transverse gradient of the guillotine and the consistency of the longitudinal gradient of the guillotine; taking the product of the Euclidean distance and the entropy value as the difference of the texture characteristics of the knife of the pixel point.
Preferably, the growth criteria for obtaining the region growth according to the switch blade illumination perception change of the pixel point and the switch blade texture feature difference specifically include:
setting a growth threshold; and taking the product of the illumination perception change of the guillotine and the difference of the texture characteristics of the guillotine as a growth criterion of regional growth.
Preferably, the identifying of the abnormal state of the isolating switch is completed according to the knife defect area, and the specific method comprises the following steps:
acquiring the number of pixels in a defect area, and setting a defect pixel threshold; when the number of the pixels in the defect area is larger than or equal to the threshold value of the defective pixels, judging that the state of the isolating switch is abnormal; and when the number of the pixels in the defect area is smaller than the threshold value of the pixels in the defect area, judging that the isolating switch is in a normal state.
The invention has at least the following beneficial effects:
the invention provides a method for identifying the state of an isolating switch, which comprises the steps of firstly, calculating the difference of total energy values obtained by wavelet transformation, combining the total energy values to obtain the run length of a run length code where a pixel point is located, constructing the illumination perception change of a knife, and accurately capturing defects caused by uneven illumination or shadow in an image, thereby improving the accuracy of defect analysis. Secondly, the texture characteristic difference of the knife combines the texture characteristic and the gradient amplitude change rule, more information can be provided when defects are analyzed, the defects caused by the texture change in the image can be effectively captured, and the abnormal texture of the knife can be independently analyzed from the bright and dark areas of the knife, so that the property and the degree of the defects can be more comprehensively known, and the change of texture details caused by different illumination degrees is eliminated. And finally, the defect of the shadow part can be accurately segmented by analyzing the growth criterion of the improved region growth through the index, the characteristic difference between the shadow part and the normal part can be eliminated, the defect region at the light-dark junction is completely segmented, and the robustness and the accuracy of the algorithm are improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for identifying the status of an isolating switch according to the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description refers to specific implementation, structure, characteristics and effects of a method for identifying the state of a disconnecting switch according to the invention, which are provided by the invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the method for identifying the state of the isolating switch provided by the invention with reference to the accompanying drawings.
The embodiment of the invention provides a method for identifying the state of an isolating switch.
Specifically, referring to fig. 1, the following method for identifying the state of an isolating switch is provided, and the method includes the following steps:
step S001: and acquiring an image of the isolating switch, and preprocessing the acquired image.
When the CCD industrial camera is used for collecting the image of the isolating switch in a short distance, the photographing angle needs to be adjusted when the image is collected, so that the knife blade and the connecting part of the isolating switch can be clearly photographed. After the image acquisition is completed, carrying out graying treatment on the acquired image to obtain a gray image, carrying out denoising treatment on the gray image by using a median filtering algorithm to remove most of noise influence, and then enhancing the denoised image by using histogram equalization to complete preprocessing on the acquired image. It should be noted that, the denoising and enhancement processing implementation of the image can select the processing method according to the actual situation, the median filtering algorithm and the histogram equalization are the prior known techniques, and will not be described in detail in this embodiment.
Step S002: constructing a knife switch illumination perception change through energy distribution difference; and extracting entropy values of texture features by combining the gradient amplitude variation and the gray level co-occurrence matrix to construct the difference of the texture features of the guillotine.
Firstly, a knife and a region of a connecting part are acquired by utilizing a semantic segmentation algorithm, the knife and the region of the connecting part are set as ROI regions, and an acquisition implementer of the ROI regions can select the segmentation algorithm according to actual conditions, in the embodiment, a semantic segmentation model based on a Convolutional Neural Network (CNN) is adopted to segment a preprocessed image, a label of a pixel level is created, the knife region and the pixels of the connecting part are marked as 1, and the pixels of the background or other regions are marked as 0. And then, using a Linknet image segmentation grid structure as a semantic segmentation model, carrying out model training optimization and cross entropy loss function guide model training by an Adam algorithm, classifying the preprocessed image at a pixel level, and outputting a knife and a connecting part region extracted after semantic segmentation processing by training and optimizing a semantic segmentation model based on a convolutional neural network. The region of CNN output is set as ROI region, and the subsequent analysis is performed on the ROI region.
Since the blade portion is connected to the post at both ends, the connection portion typically involves some components or fixtures to ensure a secure connection and reliable operation. Firstly, K-means clustering is used for an ROI region, in the ROI region, three color categories are contained in the whole ROI region due to shielding of a connecting part fixing device and components, namely a knife region under no shielding, a knife region under a shadow part and a connecting part, so that the number K=3 of clustering clusters is set, K pixel points in a data set are randomly selected as initial clustering centers for clustering, and segmentation of the knife region, the shadow region and the connecting part is completed. According to the gray features of the three parts, the region with the largest gray average value in the three regions is recorded as a normal knife switch region, the region with the smallest gray average value is secondly a connecting region, and the region with the smallest gray average value is recorded as a shadow knife switch region. The Sobel operator is then used to calculate the gradient magnitude for each pixel point within the ROI region. It should be noted that, the K-means clustering and Sobel operator are known techniques in the prior art, and are not described in detail in this embodiment.
Since the blade defects in component shadows may be affected by the surrounding environment, such as projection of surrounding components, these additional shadows and projections interfere with the detection and analysis of defects, so that the position and shape of defects in the image in this case are different from those in normal illumination. Using Daubechies wavelet basis function to perform one-layer two-dimensional wavelet transformation on the ROI area, dividing the image into four sub-bands, namely LL (low frequency), LH (horizontal high frequency), HL (vertical high frequency) and HH (diagonal high frequency), wherein the decomposed corresponding coefficient matrixes are respectively as follows:let the image size be +.>The four coefficient matrices are all +.>Wherein->As a round-up function. Calculating pixels in the region of the ROI>Total energy value in the above three high frequency sub-bands +.>：
In the method, in the process of the invention,representing pixel dot +.>Total energy in high frequency subband, < >>Indicate->The coefficient matrix of the high-frequency sub-band +.>Line->Values of the columns, representing the different frequency components of the ROI area at the position +.>A contribution to the above.
After the total energy value of each pixel on each high-frequency subband in the ROI area is obtained by calculating the position contribution mean value of three high-frequency subbands of the pixel point, the total energy mean value calculation is respectively carried out on the normal guillotine area and the shadow guillotine area, and the total energy value of the pixels in the guillotine area under the shadow part is recordedThe energy mean value isThe overall energy mean of the blade region under normal illumination is +.>。
In the acquired image, defects appear as anomalies in color, texture, or brightness, which can lead to energy differences between pixels. In some cases, the pixel where the defect is located is relatively different from most of the surrounding pixels, and this difference appears more prominent in the energy calculation of the entire image. Taking the total energy value of each pixel as input instead of the gray value of the pixel, constructing a gray scale run-length matrix, and setting pixel pointsRun length of the run length code is。
The overall energy value of the pixel under the high frequency component corresponds to details in the image, and in case of illumination changes the high frequency component may highlight local changes in different areas, e.g. more intense high frequency components may be generated in areas where the illumination changes more. Constructing pixel points according to the total energy value of the pixels and the run length of the run length code of the pixelIs a knife switch light perception change->：
In the method, in the process of the invention,is pixel dot +.>Is a switch blade light perception change->Is pixel dot +.>Is>Is the total energy mean value of the area where the pixel point is located, < > and is the pixel point>Run length encoded for the run length of the pixel, < >>Is a normalization function.
It should be noted that, the overall energy average of the region where the pixel point is located is divided into two cases: if the pixel point is in the shadow knife region, thenThe method comprises the steps of carrying out a first treatment on the surface of the If the pixel point is in the normal knife region, +.>。/>The absolute value of the difference between the total energy value of the pixel and the total energy mean value of the region where the pixel is located is reflected, and the difference between the defective pixel and the normal pixel can be amplified by the energy value obtained by the analysis, so that the greater the value is, the greater the possibility that the pixel is the defective pixel is represented, the +.>The larger; while run length +.>Reflecting the length of the run length code of the pixel, since the knife surface is made of uniform and smooth metalThe material is made, the pixel energy values of the pixels under the same illumination under the normal condition are similar, so that the runlength of the pixels under the normal condition is long, namely, the longer the runlength is, the smaller the possibility that the pixel is a defective pixel is>Will be +.>Is decreased by an increase in (c).
Further, pixel points are usedFor the center, construct a size of +.>Is to say +.>The value practitioner of (a) can be set according to the actual situation, in this embodiment +.>The empirical value was taken to be 5. The inside of the window is analyzed by the pixel gradient magnitude obtained using the Sobel operator. Calculating the gradient amplitude value in the horizontal direction and the vertical direction of the window center, and calculating the consistency of the transverse gradient and the consistency of the longitudinal gradient of the switch blade in the window, wherein the expression of the consistency of the transverse gradient of the switch blade is as follows:
in the method, in the process of the invention,knife lateral gradient uniformity for center pixel,/->Horizontal gradient magnitude for window center pixel, < >>、/>Left side and right side of the horizontal direction of the center pixel respectively +.>Gradient magnitude of each unit pixel. />Smaller means that the gradient change of the central pixel is more similar to the gradient change trend of the surrounding pixels, and is more likely to be a smoother region; the more likely it is to be a textured area.
Obtaining blade longitudinal gradient consistency using the same method。
Taking gray values of pixels in a window as input to obtain gray co-occurrence matrix, and extracting entropy value from the gray co-occurrence matrixAnalyzing the difference of the knife texture characteristics of the center pixel of the window by combining the consistency of the knife transverse gradient and the consistency of the knife longitudinal gradient in the window>：
Wherein,for the difference of the texture characteristics of the knife of the central pixel of the window, F is the entropy value extracted by using a gray level co-occurrence matrix in the window, and +.>、/>Knife transversal gradient uniformity, knife longitudinal gradient uniformity, respectively, of the center pixel,>is a normalization function.
The entropy of the gray level co-occurrence matrix is a texture feature of the gray level co-occurrence matrix, which reflects the complexity of the texture in the image, its valueWhen the pixel is bigger, the representing image area has stronger texture characteristics, and the matching between gray values of different pixels is more diversified, so +.>Will be dependent on the entropy value>Is increased by an increase in (a); while a greater uniformity of the gradient of the central pixel in the horizontal direction and a greater uniformity of the gradient of the longitudinal direction of the blade mean that the gradient change of the central pixel is more different from the gradient change trend of the surrounding pixels, representing a region with more pronounced changes in texture, and a region with more pronounced changes in texture>And also increases.
Step S003: and acquiring a growth criterion of region growth, dividing the defect part of the knife switch, and completing identification of abnormal detection of the isolating switch.
The region growing algorithm is used for detecting the defects of the knife switch region of the isolating switch, however, due to the influences of the knife switch and the connecting part, the characteristic change difference near the demarcation region under the shadow and normal illumination is large, such as the characteristics of illumination, color, texture and the like, so that the accuracy of the algorithm is disturbed, the defects and the normal region cannot be accurately distinguished, the defect detection result is not reliable enough, the false alarm or missing alarm condition occurs, the maintenance cost is increased, and even safety accidents are caused. To avoid such consequences, the growth criteria for improving the region growth based on the pixel point's guillotine illumination perception change and the guillotine texture feature difference are:
in the method, in the process of the invention,switch light perception change for pixel point, < >>Is the difference in the texture characteristics of the blades of the pixel and is obtainable from the above analysis, the +.>And->The larger the pixel, the more likely it is to be a defective area, the more +.>For the growth threshold, it should be noted that the growth threshold implementer may choose according to the actual situation, and in this embodiment, the checked value is 0.7.
After the new growth criterion is set, initializing a seed point, calculating the difference between the illumination perception change of the guillotine and the texture characteristics of the guillotine of the pixel point, judging whether the pixel belongs to a defect area by using the new growth criterion, and if the pixel does not meet the new growth criterion, reselecting the seed point until the seed pointUntil that point. And then starting from the seed pixel point, searching adjacent pixel points on the 8 neighborhood according to the growth criterion, and judging whether the pixel points meet the new growth criterion. If a pixel point is judged to meet the new growth criterion, the pixel point is added into the area and marked as accessed, the new pixel point is searched and added continuously through iteration until the pixel point can not grow any more, and finally the grown area is the defect area of the knife switch.
Calculate the mostThe number of pixel points of the defect area is finally obtainedSetting a defect pixel threshold, and judging that the state of the isolating switch is abnormal when the number of pixels in the defect area is larger than or equal to the defect pixel threshold, and arranging a worker for field confirmation to replace the isolating switch in time; and when the number of the pixels in the defect area is smaller than the threshold value of the pixels in the defect area, judging that the isolating switch is in a normal state. It should be noted that, the threshold value implementation of the defective pixel point can be set by the operator according to the actual situation, and the threshold value selected in this embodiment is 200. Thus, the abnormal state identification of the isolating switch is completed.
In summary, the embodiment of the invention provides a method for identifying the state of the isolating switch, which firstly calculates the difference of the total energy value obtained by wavelet transformation, and combines the total energy value to obtain the run length of the run length code where the pixel point is located, so as to construct the illumination perception change of the knife, and accurately capture the defects caused by uneven illumination or shadow in the image, thereby improving the accuracy of defect analysis. Secondly, the texture characteristic difference of the knife combines the texture characteristic and the gradient amplitude change rule, more information can be provided when defects are analyzed, the defects caused by the texture change in the image can be effectively captured, and the abnormal texture of the knife can be independently analyzed from the bright and dark areas of the knife, so that the property and the degree of the defects can be more comprehensively known, and the change of texture details caused by different illumination degrees is eliminated. And finally, the defect of the shadow part can be accurately segmented by analyzing the growth criterion of the improved region growth through the index, the characteristic difference between the shadow part and the normal part can be eliminated, the defect region at the light-dark junction is completely segmented, and the robustness and the accuracy of the algorithm are improved.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and the same or similar parts of each embodiment are referred to each other, and each embodiment mainly describes differences from other embodiments.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; the technical solutions described in the foregoing embodiments are modified or some of the technical features are replaced equivalently, so that the essence of the corresponding technical solutions does not deviate from the scope of the technical solutions of the embodiments of the present application, and all the technical solutions are included in the protection scope of the present application.
Claims (10)
1. A method for identifying the state of an isolating switch, comprising the steps of:
acquiring a gray level image of the isolating switch;
acquiring a region of a knife and a connecting part according to a gray image, and marking the region as an ROI region; dividing the ROI area into a normal guillotine area, a shadow guillotine area and a connection area in an airspace; dividing the ROI area into a low-frequency sub-band and three high-frequency sub-bands in the frequency domain; obtaining coefficient matrixes corresponding to the high-frequency sub-bands; acquiring the total energy value of the pixel points in the ROI according to coefficient matrixes corresponding to the three high-frequency sub-bands; using the gray scale run matrix to construct a gray scale run matrix by taking the total energy values of the pixel points in three high-frequency sub-bands as input; acquiring the length of a run length code where a pixel point is located according to a gray scale run matrix; acquiring the switch blade illumination perception change of the pixel point according to the length of the run length code of the pixel point; acquiring the consistency of the horizontal gradient of the guillotine of the pixel point according to the amplitude variation of the horizontal gradient in the neighborhood window of the pixel point; acquiring longitudinal gradient consistency of a knife of a pixel point; acquiring the difference of the texture characteristics of the guillotine of the pixel point according to the consistency of the transverse gradient and the consistency of the longitudinal gradient of the guillotine of the pixel point;
acquiring a growth criterion of regional growth according to the switch blade illumination perception change of the pixel points and the switch blade texture characteristic difference; completing the segmentation of the knife switch defect area according to the growth criterion; and completing identification of the abnormal state of the isolating switch according to the knife defect area.
2. The method for identifying the state of a disconnecting switch according to claim 1, wherein the acquiring the regions of the knife and the connecting portion according to the gray level image comprises:
segmenting the image by using a semantic segmentation network, marking a knife region and a joint as 1, and marking other regions as 0; the semantic segmentation network then extracts the regions of the blades and the connected portions.
3. The method for identifying the state of the disconnecting switch according to claim 1, wherein the ROI area is spatially divided into a normal knife area, a shadow knife area and a connecting area, and the method comprises the following steps:
dividing the ROI into three areas by adopting a clustering algorithm;
and (3) marking the region with the largest gray average value in the three regions as a normal knife switch region, marking the second largest gray average value as a connecting region, and marking the region with the smallest gray average value as a shadow knife switch region.
4. The method for identifying the state of an isolating switch according to claim 1, wherein the ROI area is divided into a low frequency subband and three high frequency subbands in the frequency domain, specifically:
a layer of two-dimensional wavelet transform is performed on the ROI area using a wavelet basis function to divide the ROI area into a low frequency subband, a horizontal high frequency subband, a vertical high frequency subband, and a diagonal high frequency subband.
5. The method for identifying the state of a disconnecting switch according to claim 1, wherein the obtaining the total energy value of the pixel point in the ROI area according to the coefficient matrix corresponding to the three high-frequency sub-bands specifically comprises:
taking the values of the half of the horizontal coordinates and the vertical coordinates of the pixel points which are rounded upwards as the row and column values of the corresponding coefficient matrix elements; and taking the square average value of the element values of the corresponding coefficient matrix of the pixel points in the three high-frequency sub-bands as the total energy value of the pixel points.
6. The method for identifying the state of an isolating switch according to claim 1, wherein the step of obtaining the perceived change of the illumination of the knife of the pixel according to the length of the run length code where the pixel is located is specifically:
taking the opposite number of the length of the run length code where the pixel point is located as an independent variable of an exponential function taking a natural constant as a base number; acquiring an energy average value of an area where a pixel point is located; calculating a normalization result of the total energy value of the pixel point and the absolute value of the difference value of the energy mean value; taking the product of the normalization result and the calculation result of the exponential function as the knife illumination perception change of the pixel point.
7. The method for identifying the state of an isolating switch according to claim 1, wherein the expression is as follows, the consistency of the gradient of the knife transverse of each pixel point is obtained according to the gradient amplitude variation of the pixel point in the horizontal direction in the neighborhood window of the pixel point:
in the method, in the process of the invention,for the center pixel->Is of the knife transversal gradient consistency, +.>For the window center pixel +.>Is of the horizontal gradient amplitude of (2)，/>、/>Respectively the center pixel->Horizontal left side, right side->The gradient amplitude of each unit pixel, N is a value rounded down by half the edge length of the neighborhood window.
8. The method for identifying the state of the disconnecting switch according to claim 1, wherein the step of obtaining the difference of the texture characteristics of the knife of the pixel according to the consistency of the lateral gradient and the consistency of the longitudinal gradient of the knife of the pixel comprises the following steps:
acquiring an entropy value of a window where the pixel point is located through a gray level co-occurrence matrix; calculating Euclidean distance between the consistency of the transverse gradient of the guillotine and the consistency of the longitudinal gradient of the guillotine; taking the product of the Euclidean distance and the entropy value as the difference of the texture characteristics of the knife of the pixel point.
9. The method for identifying the state of an isolating switch according to claim 1, wherein the growth criteria for the growth of the region according to the perceived change of the illumination of the switch and the difference of the texture characteristics of the switch of the pixel point are specifically as follows:
setting a growth threshold; and taking the product of the illumination perception change of the guillotine and the difference of the texture characteristics of the guillotine as a growth criterion of regional growth.
10. The method for identifying the abnormal state of the isolating switch according to claim 1, wherein the identifying the abnormal state of the isolating switch is completed according to the defect area of the knife switch, and the method comprises the following steps:
acquiring the number of pixels in a defect area, and setting a defect pixel threshold; when the number of the pixels in the defect area is larger than or equal to the threshold value of the defective pixels, judging that the state of the isolating switch is abnormal; and when the number of the pixels in the defect area is smaller than the threshold value of the pixels in the defect area, judging that the isolating switch is in a normal state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311443231.5A CN117173703B (en) | 2023-11-02 | 2023-11-02 | Isolating switch state identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311443231.5A CN117173703B (en) | 2023-11-02 | 2023-11-02 | Isolating switch state identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117173703A CN117173703A (en) | 2023-12-05 |
| CN117173703B true CN117173703B (en) | 2024-01-16 |
Family
ID=88932048
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311443231.5A Active CN117173703B (en) | 2023-11-02 | 2023-11-02 | Isolating switch state identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117173703B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118570707B (en) * | 2024-08-01 | 2024-10-08 | 东方电子股份有限公司 | Isolation switch knife switch motion state analysis method and system based on image processing |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6804403B1 (en) * | 1998-07-15 | 2004-10-12 | Digital Accelerator Corporation | Region-based scalable image coding |
| CN105374021A (en) * | 2014-08-19 | 2016-03-02 | 南京理工大学 | Graphic text image segmentation method and system based on line cutting direction |
| CN109409190A (en) * | 2018-08-21 | 2019-03-01 | 南京理工大学 | Pedestrian detection method based on histogram of gradients and Canny edge detector |
| CN110781896A (en) * | 2019-10-17 | 2020-02-11 | 暨南大学 | Track garbage identification method, cleaning method, system and resource allocation method |
| CN113159035A (en) * | 2021-05-10 | 2021-07-23 | 北京世纪好未来教育科技有限公司 | Image processing method, device, equipment and storage medium |
| WO2022074643A1 (en) * | 2020-10-08 | 2022-04-14 | Edgy Bees Ltd. | Improving geo-registration using machine-learning based object identification |
| CN115082418A (en) * | 2022-07-14 | 2022-09-20 | 山东聊城富锋汽车部件有限公司 | Precise identification method for automobile parts |
| CN115082763A (en) * | 2022-08-22 | 2022-09-20 | 江苏安弘电气有限公司 | Isolation reversing switch state identification method and system |
| CN115294158A (en) * | 2022-09-26 | 2022-11-04 | 江苏万森绿建装配式建筑有限公司 | Hot continuous rolling strip steel image segmentation method based on machine vision |
| CN115909028A (en) * | 2022-11-16 | 2023-04-04 | 昭通学院 | High-voltage isolating switch state identification method based on gradient image fusion |
| CN116310360A (en) * | 2023-05-18 | 2023-06-23 | 实德电气集团有限公司 | Reactor surface defect detection method |
| CN116304778A (en) * | 2023-05-17 | 2023-06-23 | 温州华嘉电器有限公司 | Maintenance data processing method for miniature circuit breaker |
-
2023
- 2023-11-02 CN CN202311443231.5A patent/CN117173703B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6804403B1 (en) * | 1998-07-15 | 2004-10-12 | Digital Accelerator Corporation | Region-based scalable image coding |
| CN105374021A (en) * | 2014-08-19 | 2016-03-02 | 南京理工大学 | Graphic text image segmentation method and system based on line cutting direction |
| CN109409190A (en) * | 2018-08-21 | 2019-03-01 | 南京理工大学 | Pedestrian detection method based on histogram of gradients and Canny edge detector |
| CN110781896A (en) * | 2019-10-17 | 2020-02-11 | 暨南大学 | Track garbage identification method, cleaning method, system and resource allocation method |
| WO2022074643A1 (en) * | 2020-10-08 | 2022-04-14 | Edgy Bees Ltd. | Improving geo-registration using machine-learning based object identification |
| CN113159035A (en) * | 2021-05-10 | 2021-07-23 | 北京世纪好未来教育科技有限公司 | Image processing method, device, equipment and storage medium |
| CN115082418A (en) * | 2022-07-14 | 2022-09-20 | 山东聊城富锋汽车部件有限公司 | Precise identification method for automobile parts |
| CN115082763A (en) * | 2022-08-22 | 2022-09-20 | 江苏安弘电气有限公司 | Isolation reversing switch state identification method and system |
| CN115294158A (en) * | 2022-09-26 | 2022-11-04 | 江苏万森绿建装配式建筑有限公司 | Hot continuous rolling strip steel image segmentation method based on machine vision |
| CN115909028A (en) * | 2022-11-16 | 2023-04-04 | 昭通学院 | High-voltage isolating switch state identification method based on gradient image fusion |
| CN116304778A (en) * | 2023-05-17 | 2023-06-23 | 温州华嘉电器有限公司 | Maintenance data processing method for miniature circuit breaker |
| CN116310360A (en) * | 2023-05-18 | 2023-06-23 | 实德电气集团有限公司 | Reactor surface defect detection method |
Non-Patent Citations (2)
| Title |
|---|
| Computer aided mass segmentation in mammogram images using Grey wolf Optimized Region growing technique;Ashi Ashok，等;《2021 5th International Conference on Trends in Electronics and Informatics (ICOEI)》;第1082-1087页 * |
| 基于BP神经网络的高压隔离开关分合闸监测识别;刘子英;张靖;邓芳明;;电力系统保护与控制(第05期);第140-146页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117173703A (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020224458A1 (en) | Method for detecting corona discharge employing image processing | |
| CN117173703B (en) | Isolating switch state identification method | |
| Fu et al. | Combined thresholding and neural network approach for vein pattern extraction from leaf images | |
| CN113808138B (en) | Artificial intelligence-based wire and cable surface defect detection method | |
| CN111260616A (en) | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization | |
| CN107679495B (en) | Detection method for movable engineering vehicles around power transmission line | |
| CN109446982B (en) | AR glasses-based electric power cabinet pressing plate state identification method and system | |
| CN106023185A (en) | Power transmission equipment fault diagnosis method | |
| CN108563979B (en) | Method for judging rice blast disease conditions based on aerial farmland images | |
| Valliammal et al. | Hybrid image segmentation algorithm for leaf recognition and characterization | |
| CN111539293A (en) | Fruit tree disease diagnosis method and system | |
| CN115861987A (en) | Intelligent electric power fitting defect identification method for on-line monitoring of power transmission line | |
| CN110660065A (en) | Infrared fault detection and identification algorithm | |
| CN112115807A (en) | Transformer substation secondary equipment switching operation anti-error simulation method and system based on image recognition | |
| CN105741244A (en) | Method for removing shadows and halos under weak light through indoor polling robot | |
| CN114648714A (en) | YOLO-based workshop normative behavior monitoring method | |
| CN109682821B (en) | Citrus surface defect detection method based on multi-scale Gaussian function | |
| CN111192332B (en) | Smoke detection-based smoke machine control method and smoke machine | |
| CN107886493B (en) | Method for detecting conductor stranding defects of power transmission line | |
| CN111476784A (en) | Product surface defect online detection method based on image enhancement recognition technology | |
| CN116612178A (en) | Method for extracting grain shape of integrated rice spike grains | |
| CN114119634B (en) | Building automatic extraction method and system combining vegetation rejection and image feature consistency constraint | |
| Khan et al. | Shadow removal from digital images using multi-channel binarization and shadow matting | |
| CN111260589B (en) | Retinex-based power transmission line monitoring image defogging method | |
| CN115294440A (en) | Power transmission line identification method and device and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |