CN111402215A - Contact net insulator state detection method based on robust principal component analysis method - Google Patents
Contact net insulator state detection method based on robust principal component analysis method Download PDFInfo
- Publication number
- CN111402215A CN111402215A CN202010153833.7A CN202010153833A CN111402215A CN 111402215 A CN111402215 A CN 111402215A CN 202010153833 A CN202010153833 A CN 202010153833A CN 111402215 A CN111402215 A CN 111402215A
- Authority
- CN
- China
- Prior art keywords
- insulator
- image
- value
- state
- principal component
- 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.)
- Granted
Links
- 239000012212 insulator Substances 0.000 title claims abstract description 122
- 238000001514 detection method Methods 0.000 title claims abstract description 31
- 238000012847 principal component analysis method Methods 0.000 title claims abstract description 14
- 239000011159 matrix material Substances 0.000 claims abstract description 39
- 230000000007 visual effect Effects 0.000 claims abstract description 17
- 230000011218 segmentation Effects 0.000 claims abstract description 8
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 30
- 230000008569 process Effects 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000012887 quadratic function Methods 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 5
- 230000002411 adverse Effects 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 238000013139 quantization Methods 0.000 claims description 3
- 241001270131 Agaricus moelleri Species 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 5
- 230000002950 deficient Effects 0.000 abstract description 4
- 230000002596 correlated effect Effects 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 239000000725 suspension Substances 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 210000000707 wrist Anatomy 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/45—Analysis of texture based on statistical description of texture using co-occurrence matrix computation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/48—Extraction of image or video features by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
Abstract
The invention discloses a contact network insulator state detection method based on a robust principal component analysis method, which comprises the steps of establishing an insulator sample data set according to collected images of a contact network supporting and hanging device, and performing target detection and segmentation by adopting a Mask-RCNN convolutional neural network, so as to position and segment an insulator in the images; calculating the minimum external torque of the insulator according to the positioning result, detecting the inclination angle, and rotating the obtained picture according to the inclination angle to obtain a horizontal insulator image; cutting the collected insulator images piece by piece to obtain an insulator piece data set with a single fixed visual angle; carrying out foreground and background segmentation on the insulator sheet data set with a fixed visual angle; and extracting texture features of the separated foreground through a gray level co-occurrence matrix, extracting the texture features of the image by adopting energy and entropy, carrying out weighted summation according to whether the texture features are positively correlated or not, and setting a threshold value to identify the state of the insulator. The invention realizes the detection and rapid positioning of defective states such as insulator defect, dirt and the like.
Description
Technical Field
The invention relates to the technical field of intelligent detection of high-speed railway images, in particular to a contact network insulator state detection method based on a robust principal component analysis method.
Background
The rapid development of the high-speed railway puts higher requirements on the operation safety of a traction power supply system, and advanced detection technology and modern detection equipment are used for ensuring the improvement of the maintenance quality of the traction power supply system and are important means for realizing the state detection and state maintenance of the electrified railway. In an electrified railway power supply system, a cantilever supporting device mainly comprises an inclined cantilever, a horizontal cantilever (pull rod), a rod insulator and related parts. The rod insulator is used for suspending and supporting the inclined cantilever and the horizontal cantilever and keeping the contact wire to be electrically insulated from the grounding body. The inclined cantilever and the horizontal cantilever form a stable triangular structure, and the stable triangular structure provides supporting force for the carrier cable and is connected with the positioning device. Because the gravity load of the inclined cantilever and the horizontal cantilever needs to be borne and severe environmental conditions are faced, abnormal operation states such as insulator piece defect or dirt are generated inevitably during operation, and hidden dangers are brought to safe operation of the motor train unit, so that the fault of the insulator needs to be detected, and measures are taken to eliminate the hidden dangers. The 4C system technical specification promulgated by the original railway ministry comprises high-definition video monitoring of a suspension part and a wrist arm part of a contact network, and relates to fault detection of parts in a contact network supporting and suspension device based on a digital image processing technology.
At present, the detection method for the state defect of the contact network parts at home and abroad mainly comprises the following steps: manual detection, laser testing, eddy current, ultrasonic, and the like. The detection methods all achieve certain effects, but many methods have the problems of inaccurate measurement, high danger, complex operation, expensive and heavy equipment, heavy detection task, poor anti-interference capability and the like. The non-contact bow net detection technology based on the image processing technology can realize development of a bow net detection device which does not interfere driving safety, and the used equipment has strong expansibility, realizes automatic identification of bow net parameters and faults, and has numerous advantages.
The method comprises the steps of firstly extracting insulators by using L BP and Adaboost combined models, then calculating geometrical characteristics of insulator cracks by using a method of solving the area and the circumference of a connected domain, so as to realize the detection of the insulator cracks.
Disclosure of Invention
In view of the above problems, the present invention aims to provide a method for detecting the state of an insulator of a contact network based on a robust principal component analysis method, which can realize the detection and rapid positioning of defective states such as insulator incomplete and dirt. The technical scheme is as follows:
a contact network insulator state detection method based on a robust principal component analysis method comprises the following steps:
step A: acquiring images of a high-speed railway contact net supporting and hanging device;
and B: establishing a sample data set of four insulators with fixed visual angles of each strut, and adopting a Mask-RCNN convolutional neural network to perform target detection and segmentation, so as to position and segment the insulators in the images of the supporting and hanging device of the contact network;
and C: calculating the minimum external torque of the insulator according to the positioning result, detecting an inclination angle by adopting Hough transformation, and rotating the picture obtained in the step B according to the inclination angle to finally obtain a horizontal insulator image; then, cutting the acquired insulator images piece by utilizing the visual characteristics of the insulator to obtain an insulator piece data set with a single fixed visual angle;
step D: performing foreground and background segmentation on the insulator sheet data set with the fixed view angle by using a robust principal component analysis algorithm;
step E: and extracting texture features of the separated foreground through a gray level co-occurrence matrix, extracting the texture features of the image by adopting two features of energy and entropy, carrying out weighted summation according to whether positive correlation exists or not, and setting a threshold value to identify the state of the insulator.
Further, the specific process of step B is as follows:
step B1: performing convolution operation on an input image to obtain a characteristic diagram;
step B2: extracting the region of interest by the region suggestion network;
step B3: and classifying, positioning and segmenting each region of interest to generate a mask.
Furthermore, in the step C, the visual characteristics of the insulator are used to cut the acquired insulator image piece by piece, and a specific process of obtaining an insulator piece data set with a single fixed viewing angle is as follows:
step C1: carrying out graying processing on the global image, further calculating a gray value gradient, and positioning to obtain an edge coordinate value of the insulator piece by utilizing a gradient peak value generated by the alternation of white and black at the edge of the single insulator piece;
step C2: performing quadratic function fitting on the edges of the insulators by utilizing the coordinate values of the four edges of each insulator sheet;
step C3: and according to the fitting result, calculating a quadratic function vertex, namely the division position of the single insulator sheet, and cutting according to the vertex coordinate to obtain the image of the single insulator sheet.
Further, the specific process of step D is as follows:
step D1: stretching a single insulator sheet image with a fixed visual angle from a to b into a one-dimensional vector 1 to m, wherein m is a to b; sequentially arranging the n images to obtain a two-dimensional matrix of n x m;
step D3: and stretching a single row of the processed two-dimensional matrix to obtain a foreground image and a background image of a.
Further, the specific process of step E is as follows:
step E1: firstly, extracting a foreground image of an insulator sheet, and calculating texture characteristics after extraction;
step E2: gray level quantization: dividing the gray scale into 8 gray scales;
step E3: selecting texture feature extraction parameters;
step E4: computing a co-occurrence matrix;
step E5: and identifying a fault state: judging whether the insulator is in fault by adopting energy and entropy as judgment indexes:
the energy value and entropy value are expressed as:
wherein P (i, j) is a value of an element (i, j) in the gray level co-occurrence matrix;
step E6: weighting the calculated texture features, wherein the energy value and the bad state are in negative correlation, and taking a negative weight; the entropy value is in positive correlation with the adverse state, and a positive weight is taken; the texture feature values are written as:
f=Ent-ASM
and selecting a threshold value for fault detection after calculating the texture characteristic value.
The invention has the beneficial effects that:
1. the method provided by the invention can be used for detecting the incomplete insulator and poor dirt states of the high-speed rail contact net by a robust principal component analysis method and an image processing method, is not influenced by illumination intensity, shooting angle and distance, gives objective, real and accurate detection and analysis results, and overcomes the defects of the traditional manual detection method.
2. According to the structural characteristics of the insulator sheet, the gray value gradient peak value and the optical rule of the insulator sheet are skillfully combined, and the single insulator sheet is quickly and effectively segmented.
3. The method can effectively detect the faults of the defective insulator, dirt and other bad states of the high-speed rail contact network, has high correct detection rate, and simplifies the difficulty of fault detection.
Drawings
FIG. 1 is a block diagram of the processing procedure of the method of the present invention.
Fig. 2 is an image of the high-speed railway catenary supporting and suspending device collected on site.
FIG. 3 shows the insulator regions located by the Mask-RCNN convolutional neural network: (a) the picture is an original picture; (b) positioning the Mask-RCNN convolution neural network to an insulation subarea; (c) the obtained insulator picture is cut.
Fig. 4 is a diagram of the insulator transformation rotation through Hough.
Fig. 5 is a feature point extraction diagram calculated by using a gradient peak in the period extraction.
(a) The horizontal coordinate represents a certain line of pixel points in the horizontal direction of the image, and the vertical coordinate represents a certain line of pixel values in the horizontal direction of the image;
(b) the abscissa represents a certain line of pixel points in the horizontal direction of the image, and the ordinate represents the gradient of pixel values in a certain line in the horizontal direction of the image.
Fig. 6 is a cutting diagram of insulator period extraction and insulator piece fitting.
Fig. 7 is a single insulator image taken.
Fig. 8 is a comparison graph of the fault state of the insulator ((a) defect, (b) dirt) and (c) normal image processed by the robust principal component analysis method.
Fig. 9 is a comparison graph of the intercepted fault states of the insulator ((a) incomplete, (b) dirty) and (c) normal image foreground.
FIG. 10 is a calculated texture feature for a failure versus normal foreground image: (a) energy, (b) entropy.
Fig. 11 shows the weighted texture feature values.
Fig. 12 is a grayscale matrix a.
Detailed Description
The invention is described in further detail below with reference to the figures and specific embodiments. FIG. 1 is a block diagram of the processing procedure of the method of the present invention. Fig. 2 is a field-acquired image of a suspension device of a high-speed rail catenary, and the catenary insulator state detection method based on the robust principal component analysis method is characterized by comprising the following steps:
step A: a special comprehensive train inspection vehicle is adopted to image the high-speed railway contact net supporting and hanging device;
and B: and establishing a sample data set of each post insulator at four fixed visual angles, and detecting and segmenting an insulator target by adopting a Mask-RCNN convolutional neural network, so as to position and segment the insulator in the image of the contact net supporting and suspending device.
As shown in FIG. 3, the Mask-RCNN convolutional neural network can be used for accurately positioning the insulating sub-region and segmenting and extracting the insulating sub-region from the background. The specific process is as follows:
step a: firstly, convolution operation is carried out on an input image to obtain a characteristic diagram
Step b: extracting regions of interest (RoI) by Region Proposal Network (RPN, Region of interest Network)
Step c: and classifying, positioning and segmenting each RoI to generate a mask.
And C: and (4) rotating the image, as shown in fig. 4, calculating the minimum external moment of the insulator according to the positioning result of the step (B), detecting an inclination angle by adopting Hough transformation after the external moment is calculated, and rotating the picture acquired in the step (B) according to the inclination angle to finally obtain a horizontal insulator image. And cutting the acquired insulator images piece by utilizing the visual characteristics of the insulators to finally obtain an insulator piece data set with a single fixed visual angle.
The specific process is as follows:
step a: and C, calculating the minimum external moment of the insulator region, detecting the inclination angle by adopting Hough transformation after the external moment is calculated, and rotating the picture acquired in the step B according to the inclination angle to obtain a horizontal insulator image.
Step b: and carrying out graying processing on the global image, firstly displaying a gray value on a horizontal line of the insulator, calculating the gradient of the gray value, and extracting a gradient peak value. Due to the optical reflection characteristic of the insulator, the gradient of the gray value (the place of alternating bright and dark) is the edge of the insulator. The peak value is extracted to obtain the point coordinates of n insulator piece numbers (assuming that the number is n insulator pieces of one insulator).
As shown in fig. 5 (left image: abscissa represents a certain line of pixel points in the horizontal direction of the image, and ordinate represents a certain line of pixel values in the horizontal direction of the image; right image: abscissa represents a certain line of pixel points in the horizontal direction of the image, and ordinate represents a gradient of pixel values in the horizontal direction of the image). By using the visual characteristics of the insulator sheet, firstly displaying the insulator gray value on a horizontal straight line, then performing gradient calculation, and obtaining 9 characteristic points on the horizontal straight line by using the characteristics of the gradient peak value of the edge of the insulator sheet. And 4 straight lines are sequentially calculated to obtain 4 edge characteristic points of a single insulator sheet.
Step c: and calculating peak values on 4 horizontal straight lines to obtain n insulator sheets and 4 point coordinates of the edge of each insulator sheet, wherein the coordinates are 4 x n. A quadratic function fit was performed on the 4 point coordinates of each insulator sheet edge. And calculating a quadratic function vertex, wherein the function vertex is a single insulator sheet segmentation point.
As shown in fig. 6, a quadratic function is fitted according to 4 feature points of a single insulator piece, and a vertex of the function is calculated, where the vertex is a partition of the insulator piece. The division points of the 9 insulator pieces are calculated in sequence.
Step d: each insulator is divided into insulator pieces according to the individual insulator piece dividing points and stored. The clipping results are shown in fig. 7.
Step D: and C, performing foreground and background segmentation on the insulator sheet data set acquired in the step C by using a robust principal component analysis algorithm.
The specific process is as follows:
step a: the method comprises the steps of stretching a single insulator sheet image (a & ltb & gt) with a fixed visual angle into a one-dimensional vector (1 & ltm & gt, m & lta & gt & ltb & gt), sequentially arranging n images to obtain a two-dimensional matrix D with n & ltm & gt, and using an algorithm to decompose the matrix D into a low-rank matrix A (linear correlation among rows or columns due to certain internal structural information) and a sparse matrix E (containing noise and sparse). The above requirements can be written as the following optimization problem:
since the rank and L0 norm are optimally non-convex and non-smooth, the problem is generally transformed to solve a one
The relaxed convex optimization problem:
b, performing rank reduction on the two-dimensional matrix processed in the step a by using an Alternating Direction Method (ADM), namely an inaccurate Lagrange Multiplier method (inexact amplified L aggregate Multiplier; inexact A L M), and extracting a foreground and a background, wherein the target function is changed into the following steps:
the lagrange function is:
2.1) update Ak+1
2.2) update Ek+1
3) Update others
And c, respectively stretching the two-dimensional matrixes A (original image), D (background) and E (foreground) processed in the step b in a single row to obtain an original image, a background and a foreground comparison map of a and b. The comparison between the defective state and the normal state of the treated insulator is shown in fig. 8 and fig. 9.
And E, intercepting a foreground image of the insulator sheet, extracting texture features of the separated foreground through a Gray-level co-occurrence matrix (G L CM), extracting the texture features of the image by adopting two most commonly used features, wherein the texture features comprise energy, entropy and texture feature values calculated by the poor state and the normal state of the insulator are shown in figure 10, carrying out weighted summation according to whether the two most commonly used features are positively correlated or not, and setting a threshold value to identify the state of the insulator, wherein the weighted feature value is shown in figure 11.
The specific process is as follows:
step a: firstly, extracting a foreground image of the insulator sheet, and calculating texture characteristics after extraction.
Step b: and (5) gray level quantization. The gray levels in an image have 256 levels from 0 to 255. However, 256 gray levels are not required in calculating the gray co-occurrence matrix, and the amount of calculation is too large, so it is divided into 8 gray levels.
Step c: and selecting texture feature extraction parameters.
1) Step distance d: d is 1, namely the central pixel is directly compared with the adjacent pixel points;
2) selecting a direction: the direction selection for calculating the gray level co-occurrence matrix is four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees; and after the eigenvalues of the four direction matrixes are obtained, calculating the average value of the four eigenvalues to serve as a final eigenvalue co-occurrence matrix.
Step d: and calculating a symbiotic matrix. The gray level co-occurrence matrix is a square matrix, and the dimension is equal to the gray level of the image. And after the parameters are selected, carrying out gray level co-occurrence matrix calculation, wherein the value of an element (i, j) in the gray level co-occurrence matrix represents the occurrence frequency of two pixels, namely the gray level of one pixel in the image is i, the gray level of the other pixel in the image is j, the adjacent distance is d, and the direction is A. The following examples are given. A certain gray matrix a (size 3 × 3, gray level 2) is shown in fig. 12.
And (3) selecting the step distance d to be 1, and calculating a co-occurrence matrix of the matrix A in the 0-degree direction, wherein the matrix values (1 and 2) are counted according to the 0-degree direction (namely the horizontal direction from left to right), and at the moment, the total number of values meeting the statistical condition of the matrix values (1 and 2) is 4, so that the value of the element at the (1 and 2) position of the G L CM statistical matrix corresponding to the window is 4.
Step e: and identifying a fault state. The gray level co-occurrence matrix theory provides 14 characteristic values, but because the calculation amount of the gray level co-occurrence matrix is large, for simplicity, 2 most commonly used characteristics are adopted to extract the texture characteristics of the image: and the energy and the entropy are used as evaluation indexes to judge whether the insulator is in fault or not.
1) Energy, also known as angular second moment (angular second moment): the energy is the sum of the squares of the elements of the gray level co-occurrence matrix, also known as the angular second order distance. The method is a measure for uniform change of the texture gray level of an image, and reflects the uniform degree of the gray level distribution of the image and the thickness degree of the texture.
2) Entropy (entropy) is a measure of the amount of information an image has, texture information also belongs to the information of the image, and is a measure of randomness, and when all elements in the co-occurrence matrix have the maximum randomness and all values in the spatial co-occurrence matrix are almost equal, and the elements in the co-occurrence matrix are distributed dispersedly, the entropy is large. Which represents the degree of non-uniformity or complexity of the texture in the image. If the gray level co-occurrence matrix values are uniformly distributed, that is, the image is nearly random or the noise is large, the entropy has a large value.
Step f: weighting the calculated texture features, wherein the energy value and the bad state are in negative correlation, the value range is [0,1], and a negative weight is taken; the entropy value is positively correlated with the adverse state, the value range is [0,1], and the positive weight is taken; thus, the texture feature value (feature) can be written as:
f=Ent-ASM
and after the texture characteristic value is calculated, selecting a proper threshold value for fault detection.
Claims (5)
1. A contact network insulator state detection method based on a robust principal component analysis method is characterized by comprising the following steps:
step A: acquiring images of a high-speed railway contact net supporting and hanging device;
and B: establishing a sample data set of four insulators with fixed visual angles of each strut, and adopting a Mask-RCNN convolutional neural network to perform target detection and segmentation, so as to position and segment the insulators in the images of the supporting and hanging device of the contact network;
and C: calculating the minimum external torque of the insulator according to the positioning result, detecting an inclination angle by adopting Hough transformation, and rotating the picture obtained in the step B according to the inclination angle to finally obtain a horizontal insulator image; then, cutting the acquired insulator images piece by utilizing the visual characteristics of the insulator to obtain an insulator piece data set with a single fixed visual angle;
step D: performing foreground and background segmentation on the insulator sheet data set with the fixed view angle by using a robust principal component analysis algorithm;
step E: and extracting texture features of the separated foreground through a gray level co-occurrence matrix, extracting the texture features of the image by adopting two features of energy and entropy, carrying out weighted summation according to whether positive correlation exists or not, and setting a threshold value to identify the state of the insulator.
2. The method for detecting the state of the insulator of the overhead line system based on the robust principal component analysis method as claimed in claim 1, wherein the specific process of the step B is as follows:
step B1: performing convolution operation on an input image to obtain a characteristic diagram;
step B2: extracting the region of interest by the region suggestion network;
step B3: and classifying, positioning and segmenting each region of interest to generate a mask.
3. The method for detecting the state of the insulator of the overhead line system based on the robust principal component analysis method according to claim 1, wherein in the step C, the acquired insulator images are cut piece by using the visual characteristics of the insulators, and a specific process of obtaining an insulator piece data set with a single fixed view angle is as follows:
step C1: carrying out graying processing on the global image, further calculating a gray value gradient, and positioning to obtain an edge coordinate value of the insulator piece by utilizing a gradient peak value generated by the alternation of white and black at the edge of the single insulator piece;
step C2: performing quadratic function fitting on the edges of the insulators by utilizing the coordinate values of the four edges of each insulator sheet;
step C3: and according to the fitting result, calculating a quadratic function vertex, namely the division position of the single insulator sheet, and cutting according to the vertex coordinate to obtain the image of the single insulator sheet.
4. The method for detecting the state of the insulator of the overhead line system based on the robust principal component analysis method as claimed in claim 1, wherein the specific process of the step D is as follows:
step D1: stretching a single insulator sheet image with a fixed visual angle from a to b into a one-dimensional vector 1 to m, wherein m is a to b; sequentially arranging the n images to obtain a two-dimensional matrix of n x m;
step D3: and stretching a single row of the processed two-dimensional matrix to obtain a foreground image and a background image of a.
5. The method for detecting the state of the insulator of the overhead line system based on the robust principal component analysis method as claimed in claim 1, wherein the specific process of the step E is as follows:
step E1: firstly, extracting a foreground image of an insulator sheet, and calculating texture characteristics after extraction;
step E2: gray level quantization: dividing the gray scale into 8 gray scales;
step E3: selecting texture feature extraction parameters;
step E4: computing a co-occurrence matrix;
step E5: and identifying a fault state: judging whether the insulator is in fault by adopting energy and entropy as judgment indexes:
the energy value and entropy value are expressed as:
wherein P (i, j) is a value of an element (i, j) in the gray level co-occurrence matrix;
step E6: weighting the calculated texture features, wherein the energy value and the bad state are in negative correlation, and taking a negative weight; the entropy value is in positive correlation with the adverse state, and a positive weight is taken; the texture feature values are written as:
f=Ent-ASM
and selecting a threshold value for fault detection after calculating the texture characteristic value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010153833.7A CN111402215B (en) | 2020-03-07 | 2020-03-07 | Contact net insulator state detection method based on robust principal component analysis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010153833.7A CN111402215B (en) | 2020-03-07 | 2020-03-07 | Contact net insulator state detection method based on robust principal component analysis method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111402215A true CN111402215A (en) | 2020-07-10 |
CN111402215B CN111402215B (en) | 2022-04-29 |
Family
ID=71430614
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010153833.7A Expired - Fee Related CN111402215B (en) | 2020-03-07 | 2020-03-07 | Contact net insulator state detection method based on robust principal component analysis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111402215B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111310761A (en) * | 2020-03-06 | 2020-06-19 | 西南交通大学 | Contact net insulator detection method based on reconstruction and classification convolution self-coding network |
CN112634254A (en) * | 2020-12-29 | 2021-04-09 | 北京市商汤科技开发有限公司 | Insulator defect detection method and related device |
CN113158728A (en) * | 2020-12-31 | 2021-07-23 | 杭州图歌科技有限公司 | Parking space state detection method based on gray level co-occurrence matrix |
CN114414660A (en) * | 2022-03-18 | 2022-04-29 | 盐城工学院 | Method for identifying axle number and cracks of railway vehicle wheel set |
CN115546568A (en) * | 2022-12-01 | 2022-12-30 | 合肥中科类脑智能技术有限公司 | Insulator defect detection method, system, equipment and storage medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103886610A (en) * | 2014-04-05 | 2014-06-25 | 东北电力大学 | Image type defect detecting method for insulator |
CN104504394A (en) * | 2014-12-10 | 2015-04-08 | 哈尔滨工业大学深圳研究生院 | Dese population estimation method and system based on multi-feature fusion |
CN106960178A (en) * | 2017-02-23 | 2017-07-18 | 中国科学院自动化研究所 | The training method of insulator identification model and the identification of insulator and localization method |
CN108305259A (en) * | 2018-02-06 | 2018-07-20 | 西安交通大学 | A kind of abrasive type automatic identifying method of multi-texturing Fusion Features |
CN108564254A (en) * | 2018-03-15 | 2018-09-21 | 国网四川省电力公司绵阳供电公司 | Controller switching equipment status visualization platform based on big data |
US20190132343A1 (en) * | 2016-09-07 | 2019-05-02 | Patternex, Inc. | Method and system for generating synthetic feature vectors from real, labelled feature vectors in artificial intelligence training of a big data machine to defend |
CN110533642A (en) * | 2019-08-21 | 2019-12-03 | 深圳新视达视讯工程有限公司 | A kind of detection method of insulator damage |
CN110728283A (en) * | 2019-10-11 | 2020-01-24 | 高新兴科技集团股份有限公司 | License plate type identification method and device |
-
2020
- 2020-03-07 CN CN202010153833.7A patent/CN111402215B/en not_active Expired - Fee Related
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103886610A (en) * | 2014-04-05 | 2014-06-25 | 东北电力大学 | Image type defect detecting method for insulator |
CN104504394A (en) * | 2014-12-10 | 2015-04-08 | 哈尔滨工业大学深圳研究生院 | Dese population estimation method and system based on multi-feature fusion |
US20190132343A1 (en) * | 2016-09-07 | 2019-05-02 | Patternex, Inc. | Method and system for generating synthetic feature vectors from real, labelled feature vectors in artificial intelligence training of a big data machine to defend |
CN106960178A (en) * | 2017-02-23 | 2017-07-18 | 中国科学院自动化研究所 | The training method of insulator identification model and the identification of insulator and localization method |
CN108305259A (en) * | 2018-02-06 | 2018-07-20 | 西安交通大学 | A kind of abrasive type automatic identifying method of multi-texturing Fusion Features |
CN108564254A (en) * | 2018-03-15 | 2018-09-21 | 国网四川省电力公司绵阳供电公司 | Controller switching equipment status visualization platform based on big data |
CN110533642A (en) * | 2019-08-21 | 2019-12-03 | 深圳新视达视讯工程有限公司 | A kind of detection method of insulator damage |
CN110728283A (en) * | 2019-10-11 | 2020-01-24 | 高新兴科技集团股份有限公司 | License plate type identification method and device |
Non-Patent Citations (8)
Title |
---|
LEI MA: "Detection Method of Insulator Based on Faster R-CNN", 《2017 IEEE 7TH ANNUAL INTERNATIONAL CONFERENCE ON CYBER TECHNOLOGY IN AUTOMATION, CONTROL, AND INTELLIGENT SYSTEMS (CYBER)》 * |
QINGGANG WU: "A Texture Segmentation Algorithm Based on PCA and Global Minimization Active Contour Model for Aerial Insulator Images", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING ( VOLUME: 5, ISSUE: 5, OCT. 2012)》 * |
YIFAN LIAO: "Study on Evaluation Method of Insulator Surface Contamination Level Based on LIBS Technology and PCA Algorithm", 《2019 2ND INTERNATIONAL CONFERENCE ON ELECTRICAL MATERIALS AND POWER EQUIPMENT (ICEMPE)》 * |
刘志刚: "多小波理论及其在电力系统故障信号处理中的应用研究", 《中国博士学位论文全文数据库 (工程科技Ⅱ辑)》 * |
张桂南: "基于角点匹配与谱聚类的接触网绝缘子破损/夹杂异物故障检测", 《仪器仪表学报 学术论文》 * |
杨红梅: "基于快速鲁棒性特征匹配的电气化铁路绝缘子不良状态检测", 《电网技术》 * |
胡建平: "一种新的航拍玻璃绝缘子图像分割方法", 《东北电力大学学报》 * |
阳武: "基于航拍图像的绝缘子识别与状态检测方法研究", 《中国博士学位论文全文数据库 (工程科技Ⅱ辑)》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111310761A (en) * | 2020-03-06 | 2020-06-19 | 西南交通大学 | Contact net insulator detection method based on reconstruction and classification convolution self-coding network |
CN111310761B (en) * | 2020-03-06 | 2022-04-29 | 西南交通大学 | Contact net insulator detection method based on reconstruction and classification convolution self-coding network |
CN112634254A (en) * | 2020-12-29 | 2021-04-09 | 北京市商汤科技开发有限公司 | Insulator defect detection method and related device |
CN113158728A (en) * | 2020-12-31 | 2021-07-23 | 杭州图歌科技有限公司 | Parking space state detection method based on gray level co-occurrence matrix |
CN114414660A (en) * | 2022-03-18 | 2022-04-29 | 盐城工学院 | Method for identifying axle number and cracks of railway vehicle wheel set |
CN114414660B (en) * | 2022-03-18 | 2024-01-12 | 盐城工学院 | Axle number and crack identification method for railway vehicle wheel set |
CN115546568A (en) * | 2022-12-01 | 2022-12-30 | 合肥中科类脑智能技术有限公司 | Insulator defect detection method, system, equipment and storage medium |
CN115546568B (en) * | 2022-12-01 | 2023-03-10 | 合肥中科类脑智能技术有限公司 | Insulator defect detection method, system, equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN111402215B (en) | 2022-04-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111402215B (en) | Contact net insulator state detection method based on robust principal component analysis method | |
CN103442209B (en) | Video monitoring method of electric transmission line | |
WO2018028103A1 (en) | Unmanned aerial vehicle power line inspection method based on characteristics of human vision | |
CN104318582B (en) | Detection method for bad state of rotating double-lug component pin of high-speed rail contact network | |
CN109489724B (en) | Tunnel train safe operation environment comprehensive detection device and detection method | |
CN111242123B (en) | Power equipment fault diagnosis method based on infrared image | |
CN106372667B (en) | A kind of high iron catenary diagonal brace sleeve part screw defective mode detection method | |
CN104268505A (en) | Automatic cloth defect point detection and recognition device and method based on machine vision | |
CN111428748A (en) | Infrared image insulator recognition and detection method based on HOG characteristics and SVM | |
CN109191421B (en) | Visual detection method for pits on circumferential surface of cylindrical lithium battery | |
CN106485694B (en) | A kind of high iron catenary double-jacket tube connector six-sided nut based on cascade classifier falls off defective mode detection method | |
CN103729856B (en) | A kind of Fabric Defects Inspection detection method utilizing S-transformation signal extraction | |
CN114241364A (en) | Method for quickly calibrating foreign object target of overhead transmission line | |
CN108537170A (en) | A kind of power equipment firmware unmanned plane inspection pin missing detection method | |
CN110288571B (en) | High-speed rail contact net insulator abnormity detection method based on image processing | |
CN111462045B (en) | Method for detecting defects of catenary support assembly | |
CN109272484B (en) | Rainfall detection method based on video image | |
CN112884795A (en) | Power transmission line inspection foreground and background segmentation method based on multi-feature significance fusion | |
CN112991432A (en) | Icing shape identification method based on image processing | |
CN112541478A (en) | Insulator string stain detection method and system based on binocular camera | |
CN110276747B (en) | Insulator fault detection and fault rating method based on image analysis | |
CN116309407A (en) | Method for detecting abnormal state of railway contact net bolt | |
CN116228659A (en) | Visual detection method for oil leakage of EMS trolley | |
Li et al. | Vehicle seat detection based on improved RANSAC-SURF algorithm | |
CN112396572A (en) | Composite insulator double-light fusion method based on feature enhancement and Gaussian pyramid |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220429 |
|
CF01 | Termination of patent right due to non-payment of annual fee |