CN113487543A - Contact net arcing firing detection method, device, computer equipment and storage medium - Google Patents

Contact net arcing firing detection method, device, computer equipment and storage medium Download PDF

Info

Publication number
CN113487543A
CN113487543A CN202110666642.5A CN202110666642A CN113487543A CN 113487543 A CN113487543 A CN 113487543A CN 202110666642 A CN202110666642 A CN 202110666642A CN 113487543 A CN113487543 A CN 113487543A
Authority
CN
China
Prior art keywords
image
contact line
line
contact
arcing
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
Application number
CN202110666642.5A
Other languages
Chinese (zh)
Other versions
CN113487543B (en
Inventor
王瑞锋
张郭
王云龙
李想
张金鑫
刘颖强
高雄杰
冯中伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu Tangyuan Electric Co Ltd
Original Assignee
Chengdu Tangyuan Electric Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chengdu Tangyuan Electric Co Ltd filed Critical Chengdu Tangyuan Electric Co Ltd
Priority to CN202110666642.5A priority Critical patent/CN113487543B/en
Publication of CN113487543A publication Critical patent/CN113487543A/en
Application granted granted Critical
Publication of CN113487543B publication Critical patent/CN113487543B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/0008—Industrial image inspection checking presence/absence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/50—Photovoltaic [PV] energy

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a method and a device for detecting the arcing firing of a contact network, computer equipment and a storage medium, and relates to the technical field of contact network measurement in rail transit. The contact network arcing ignition detection method comprises an image acquisition step, a contact line positioning step, a wear detection step and an arcing ignition detection step, wherein in the arcing ignition detection step, whether each line is an effective line is judged according to the number of pixels of each line, which is smaller than a gray threshold, and whether arcing ignition exists in a current image is judged according to whether the number of effective lines is larger than the threshold or whether the number of continuous effective lines is larger than the threshold. The detection method can accurately detect the trace and the position of the contact line which are seriously burned by the arcing, and can provide important help and reference information for maintenance personnel.

Description

Contact net arcing firing detection method, device, computer equipment and storage medium
Technical Field
The invention relates to the technical field of rail transit contact network detection, in particular to a contact network arcing firing detection method, a contact network arcing firing detection device, computer equipment and a storage medium.
Background
The wear between the bow nets is classified into electrical wear and mechanical wear. Mechanical abrasion generally refers to abrasion caused by mutual friction between pantograph nets when a pantograph slides across the bottom surface of a contact line, and the mechanical abrasion generally belongs to normal abrasion; the electrical abrasion is mainly caused by the fact that the pantograph breaks away from the bottom surface of the contact line to generate arcing and ablate the pantograph and the surface of the contact line, and in most cases, the electrical abrasion is abnormal abrasion and should be avoided as much as possible.
According to actual operation and maintenance experience, electrical abrasion is often accompanied with phenomena such as arcing electric spark and the like, if the current to be disconnected exceeds 0.25-1A, and the voltage between two disconnected contact points exceeds 12-20V, a group of approximately cylindrical gas-electric arcs with extremely high temperature and strong light emission and capable of conducting electricity are usually generated in the gap between the contact points, and the gas-electric arcs can ablate and burn contact surfaces of an arch net, so that contact line materials become soft and are easy to abrade, and abrasion is more easily generated when subsequent vehicles pass through.
Therefore, in the subway overhead line system detection, if the trace and the position of the bus bar which is seriously burned by the arcing can be detected, extremely important help and reference information can be provided for maintenance personnel, so that the improvement and optimization of the pantograph-catenary relationship can be judged.
The chinese patent application with publication number CN111323683A discloses an arcing detection system, which comprises an arcing detection unit, an image acquisition unit, an image storage unit and an upper computer, wherein the arcing detection unit is used for detecting arcing of a contact net; the image acquisition unit is used for acquiring an arcing image of the contact network; the image storage unit is respectively connected with the arcing detection unit, the image acquisition unit and the upper computer, and is used for storing the arcing images acquired by the image acquisition unit and uploading the arcing images at the arcing occurrence time to the upper computer when the arcing detection unit detects arcing signals.
The above prior art needs to be equipped with an arcing detection unit and an image acquisition unit for arcing detection at the same time, which belongs to real-time detection of arcing, and still cannot detect and locate the traces and positions of arcing on the contact network, and cannot distinguish whether the existing abrasion on the contact network is mechanical abrasion or electrical abrasion.
Disclosure of Invention
In order to overcome the defects and shortcomings in the prior art, the invention provides a contact net arcing firing detection method, and aims to solve the problems of detection and positioning of electrical abrasion between bow nets and arcing firing in the prior art. The invention depends on the existing detection equipment on the running vehicle to carry out abrasion detection and arcing firing detection on the contact line in the contact network. The detection method can accurately detect the trace and the position of the contact line which are seriously burned by the arcing, and provides help and reference for contact line maintenance personnel.
In order to solve the problems in the prior art, the invention is realized by the following technical scheme:
the contact net arcing firing detection method comprises the following steps:
an image acquisition step, namely acquiring a rigid contact line image in the whole pull-out value range from the linear array camera;
a contact line positioning step, namely preprocessing the rigid contact line image, and extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image; positioning the position of the contact line according to the offset of the contact line relative to the boundary of the busbar;
a wear detection step, namely extracting a contact line boundary by taking the position of a contact line as a reference in the bus area image, searching a line segment which is the longest and continuously larger than a gray threshold in the contact line area and is the wear pixel length, and extracting a wear image;
the method comprises the following steps of (1) arc burning detection, namely searching line by line in a wear image, setting a left searching range and a right searching range in each line according to the position of a contact line, searching pixels of which the image gray level is smaller than a gray level threshold value in the searching ranges, and counting the number; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
Furthermore, in the arc ignition detection step, L is a contact line positioning vector obtained in the contact line positioning step, i.e., L ═ L0,…,lh,…,lH-1]，lhThe H-th row of contact lines is the abscissa position, I represents a linear array camera image, the size of the image I is W x H, namely W represents the image width, and H represents the image height; then each row sets a left and right search range according to the position of the contact line, which is expressed as:
according to the h-th row contact line position lhSetting left and right search ranges SRhThe setting method is as follows:
SRh＝[max(lh-offset,0),min(lh+offset,W)]
wherein, offset is a preset offset, max (-) is an operator for solving the maximum value, and min (-) is an operator for solving the minimum value;
in the search range SRhSearching pixels with image gray smaller than gray threshold value gt in the image, and counting the number ChThe gray threshold gt determination method is as follows:
Figure BDA0003117683340000021
wherein the content of the first and second substances,
Figure BDA0003117683340000022
the bus bar image obtained in the contact line positioning step is an image in the range of the bus bar in the original image I; fac2 represents the set coefficient, fac2 ∈ (1, 3)]；
Judgment ChWhether the h-th behavior is greater than a threshold ct or not, and if the h-th behavior is greater than the threshold ct, the h-th behavior is judged to be a valid line, namely
F＝[f0,…,fh,…,fH-1]T
Figure BDA0003117683340000031
Wherein F is a valid row flag vector, FhFor valid line marking, when fhWhen f is 1, it is effectivehWhen the value is 0, the operation is invalid;
the threshold ct is set in relation to the contact line width and the worn pixel length determined by the row, i.e.:
Figure BDA0003117683340000032
wherein offset is the search range offset; aLhWorn pixel length detected for row h; cons is constant, Cons is E [1,3 ]]；
The total number of valid lines and the maximum continuous valid line length, i.e.
Figure BDA0003117683340000033
Figure BDA0003117683340000034
Wherein svf is the total number of active rows; vcFiFor the ith active consecutive line, S is the current image active consecutive line set, vcFmaxvcF for the longest valid consecutive rowmaxL is the longest effective continuous line length, len (-) is the vector length operator;
judging whether the number of effective lines is greater than a threshold value vt or whether the number of continuous effective lines is greater than a threshold value vct, and if any condition meets the requirement, determining that arcing firing exists in the current image, namely
Figure BDA0003117683340000035
And if the resFlag is equal to 0, the current image is considered to have no arcing.
In the contact line positioning step, preprocessing an acquired rigid contact line image, specifically, solving a transverse gradient image E of an original image I; then, a binarized image B of the transverse gradient image E is obtained.
Furthermore, the following gradient operator is used for solving the transverse gradient image E:
Figure BDA0003117683340000036
then, the gradient image can be calculated as follows:
E(x,y)＝∑i∑j I(x+i,y+j)·A(i+1,j+1),i,j∈(-1,0,1)；
when the binarized image B is obtained, the binarized threshold bt is obtained by the following formula:
bt max (e) fac 0; wherein max (·) is an operator for solving the maximum value, fac0 is a manual setting coefficient, fac0 belongs to [0.5,1 ]) is; the binarized image B can be calculated by the following equation:
Figure BDA0003117683340000041
the method comprises the steps of extracting the boundary information of the busbar from a preprocessed image, specifically, searching connected domains according to a binary image, tracking curves of all image heights through each connected domain, screening out a curve with the maximum gray average value from all the tracked curves, and selecting a curve with the minimum difference variance of the horizontal coordinates of the curve under the condition that the gray average value is maximum if two or more curves have the same gray average value, wherein the curve can be used for representing the shape of the busbar in the image; based on the curve, move to both sides at fixed intervalsCalculating the moved curve pixel mean value, if the continuous curve pixel mean values are less than the threshold value
Figure BDA0003117683340000042
The bus boundary is considered to be searched.
The positioning of the position of the contact line according to the offset of the contact line relative to the boundary of the bus bar specifically means that the bus bar image is intercepted from the original gray level image I according to the boundary curve of the bus bar
Figure BDA0003117683340000043
Then, the obtained value is obtained
Figure BDA0003117683340000044
Corresponding transverse gradient image
Figure BDA0003117683340000045
Obtaining bus bar image
Figure BDA0003117683340000046
Longitudinal gray scale projection of
Figure BDA0003117683340000047
And bus bar image
Figure BDA0003117683340000048
Corresponding transverse gradient image
Figure BDA0003117683340000049
Longitudinal gray scale projection of
Figure BDA00031176833400000410
In that
Figure BDA00031176833400000411
Moving to the right with a fixed window, calculating the sum in each window; obtaining a sum vector after all sliding windows
Figure BDA00031176833400000412
Obtaining
Figure BDA00031176833400000413
The maximum value is obtained, and the column index is obtained; then on the bus image
Figure BDA00031176833400000414
Longitudinal gray scale projection of
Figure BDA00031176833400000415
Respectively searching out minimum values from left to right in a certain range by taking the obtained column index as a reference to obtain two minimum value indexes, and finally taking the average value of the two minimum value indexes as the offset of the contact line; and then calculating the position of the contact line in the original image I according to the left boundary of the busbar and the offset of the contact line.
In the abrasion detection step, a contact line position is taken as a reference to extract a contact line boundary in a busbar area image, specifically, a contact line curve obtained by positioning in the contact line positioning step is taken as a reference, the contact line curve is moved towards two sides, a moved curve pixel mean value is calculated, and if a plurality of continuous curve pixel mean values are smaller than a threshold value, the contact line boundary is considered; searching the maximum pixel value in the contact line boundary searching range, setting a gray threshold according to the maximum value, and searching a line segment continuously larger than the gray threshold in the searching range, wherein the line segment is a wear pixel length line segment.
The invention also provides a device for detecting the burning of the contact net due to arcing, which can accurately detect the trace and the position of the contact net due to severe burning due to arcing and provide important help and reference information for maintenance personnel.
The contact network arcing firing detection device comprises an image acquisition module, a contact line positioning module, a contact line abrasion detection module and an arcing firing detection module;
the image acquisition module comprises a light source and a linear array camera, the light source is used for supplementing light to a contact line area, the linear array camera is used for acquiring an original image of a rigid contact line in the whole pulling value range, and the original image of the contact line is input into the contact line positioning module;
the contact line positioning module is used for preprocessing an input contact line original image, extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image, and positioning the position of a contact line according to the offset of the contact line relative to the bus bar boundary; the contact line positioning module sends the bus area image and the contact line position to the contact line abrasion detection module;
the contact line abrasion detection module extracts a contact line boundary by taking a contact line position as a reference in a bus area image to obtain a contact line area, and searches a line segment which is the longest continuous line segment larger than a gray threshold value in the contact line area, namely the abrasion pixel length, to obtain an abrasion image; the contact line abrasion detection module sends the abrasion image to an arcing firing detection module;
the arcing firing detection module searches the abrasion image line by line, each line sets a left searching range and a right searching range according to the position of a contact line, pixels of which the image gray level is smaller than a gray level threshold value are searched in the searching ranges, and the number of the pixels is counted; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
Furthermore, in the arc burning detection module, L is a contact line positioning vector obtained in the contact line positioning step, i.e., L ═ L0,…,lh,…,lH-1]，lhThe H-th row of contact lines is the abscissa position, I represents a linear array camera image, the size of the image I is W x H, namely W represents the image width, and H represents the image height; then each row sets a left and right search range according to the position of the contact line, which is expressed as:
according to the h-th row contact line position lhSetting left and right search ranges SRhThe setting method is as follows:
SRh＝[max(lh-offset,0),min(lh+offset,W)]
wherein, offset is a preset offset, max (-) is an operator for solving the maximum value, and min (-) is an operator for solving the minimum value;
in the search range SRhSearching pixels with image gray smaller than gray threshold value gt in the image, and counting the number ChThe gray threshold gt determination method is as follows:
Figure BDA0003117683340000061
wherein the content of the first and second substances,
Figure BDA0003117683340000062
the bus bar image obtained in the contact line positioning step is an image in the range of the bus bar in the original image I; fac2 represents the set coefficient, fac2 ∈ (1, 3)]；
Judgment ChWhether the h-th behavior is greater than a threshold ct or not, and if the h-th behavior is greater than the threshold ct, the h-th behavior is judged to be a valid line, namely
F＝[f0,…,fh,…,fH-1]T
Figure BDA0003117683340000063
Wherein F is a valid row flag vector, FhFor valid line marking, when fhWhen f is 1, it is effectivehWhen the value is 0, the operation is invalid;
the threshold ct is set in relation to the contact line width and the worn pixel length determined by the row, i.e.:
Figure BDA0003117683340000064
wherein offset is the search range offset; aLhWorn pixel length detected for row h; cons is constant, Cons is E [1,3 ]]；
The total number of valid lines and the maximum continuous valid line length, i.e.
Figure BDA0003117683340000065
Figure BDA0003117683340000066
Wherein svf is the total number of active rows; vcFiFor the ith active consecutive line, S is the current image active consecutive line set, vcFmaxvcF for the longest valid consecutive rowmaxL is the longest effective continuous line length, len (-) is the vector length operator;
judging whether the number of effective lines is greater than a threshold value vt or whether the number of continuous effective lines is greater than a threshold value vct, and if any condition meets the requirement, determining that arcing firing exists in the current image, namely
Figure BDA0003117683340000067
And if the resFlag is equal to 0, the current image is considered to have no arcing.
In the contact line positioning module, preprocessing an input contact line original image specifically means that a transverse gradient image E of an original image I is obtained; then, a binarized image B of the transverse gradient image E is obtained.
Furthermore, the following gradient operator is used for solving the transverse gradient image E:
Figure BDA0003117683340000071
then, the gradient image can be calculated as follows:
E(x,y)＝∑i∑j I(x+i,y+j)·A(i+1,j+1),i,j∈(-1,0,1)；
when the binarized image B is obtained, the binarized threshold bt is obtained by the following formula:
bt max (e) fac 0; where max (. cndot.) is the operator for the maximum, fac0 ∈ [0.5, 1); the binarized image B can be calculated by the following equation:
Figure BDA0003117683340000072
the method comprises the steps of extracting the boundary information of the busbar from a preprocessed image, specifically, searching connected domains according to a binary image, tracking curves of all image heights through each connected domain, screening out a curve with the maximum gray average value from all the tracked curves, and selecting a curve with the minimum difference variance of the horizontal coordinates of the curve under the condition that the gray average value is maximum if two or more curves have the same gray average value, wherein the curve can be used for representing the shape of the busbar in the image; moving to both sides at fixed intervals by taking the curve as a reference, calculating the moved curve pixel mean value, and if the mean values of a plurality of continuous curve pixels are smaller than a threshold value
Figure BDA0003117683340000073
The bus boundary is considered to be searched.
The positioning of the position of the contact line according to the offset of the contact line relative to the boundary of the bus bar specifically means that the bus bar image is intercepted from the original gray level image I according to the boundary curve of the bus bar
Figure BDA0003117683340000074
Then, the obtained value is obtained
Figure BDA0003117683340000075
Corresponding transverse gradient image
Figure BDA0003117683340000076
Obtaining bus bar image
Figure BDA00031176833400000717
Longitudinal gray scale projection of
Figure BDA0003117683340000077
And bus bar image
Figure BDA0003117683340000078
Corresponding transverse gradient image
Figure BDA0003117683340000079
Longitudinal gray scale projection of
Figure BDA00031176833400000711
In that
Figure BDA00031176833400000712
Moving to the right with a fixed window, calculating the sum in each window; obtaining a sum vector after all sliding windows
Figure BDA00031176833400000713
Obtaining
Figure BDA00031176833400000714
The maximum value is obtained, and the column index is obtained; then on the bus image
Figure BDA00031176833400000715
Longitudinal gray scale projection of
Figure BDA00031176833400000716
Respectively searching out minimum values from left to right in a certain range by taking the obtained column index as a reference to obtain two minimum value indexes, and finally taking the average value of the two minimum value indexes as the offset of the contact line; and then calculating the position of the contact line in the original image I according to the left boundary of the busbar and the offset of the contact line.
In the contact line abrasion detection module, a contact line position is taken as a reference to extract a contact line boundary in a busbar area image, specifically, a contact line curve obtained by positioning in a contact line positioning module is taken as a reference, the contact line curve is moved towards two sides, a moved curve pixel mean value is calculated, and if a plurality of continuous curve pixel mean values are smaller than a threshold value, the contact line boundary is considered; searching the maximum pixel value in the contact line boundary searching range, setting a gray threshold according to the maximum value, and searching a line segment continuously larger than the gray threshold in the searching range, wherein the line segment is a wear pixel length line segment.
Compared with the prior art, the beneficial technical effects brought by the invention are as follows:
1. compared with the prior art, the contact line arcing firing detection method provided by the invention runs by depending on the existing equipment, and can realize arcing firing detection positioning of the contact line without additionally adding hardware equipment. Moreover, the arcing ignition detection method provided by the invention can identify the abrasion of the surface of the rigid contact line in the contact network as electrical abrasion or mechanical abrasion, and can accurately identify the arcing ignition point. The invention can identify the worn contact line and the contact line which is just subjected to arcing electric spark, does not increase hardware equipment, and can finish detection by depending on the existing equipment.
2. The arcing detection method judges whether the line is an effective line or not according to the number of pixels of each line smaller than the gray threshold, and judges whether arcing firing exists in the current image or not according to whether the effective line number is larger than the threshold or whether the continuous effective line number is larger than the threshold. The method can accurately identify the burning point of the arcing, has high identification precision, can accurately detect the trace and the position of the contact line which are seriously burnt by the arcing, and can provide important help and reference information for maintenance personnel.
3. In the invention, the contact line positioning method converts rigid contact line identification with high difficulty and higher environmental complexity into contact line busbar boundary identification with single environment, simplicity and easiness in identification, and obtains the position of the contact line based on the offset positioning of the contact line busbar boundary and the contact line, thereby completing the positioning of the contact line with high precision.
4. The abrasion detection method can overcome various interferences caused by complex and various contact network environments to a great extent and improve the abrasion detection precision.
Drawings
FIG. 1 is a flow chart of the method for detecting the arcing firing of a contact network of the present invention;
FIG. 2 is a schematic structural view of an arc ignition detection device of a contact network of the present invention;
FIG. 3 is an original image acquired by a line camera;
FIG. 4 is a graph traced to characterize the shape of the buss bar;
FIG. 5 is a graph of identified bus left and right boundaries;
FIG. 6 is a cut-away image of a bus bar;
FIG. 7 is a longitudinal gray scale projection image of a buss bar image;
FIG. 8 is an image of contact line offset in a reflow bar image;
FIG. 9 is an image of contact line location in an original image;
FIG. 10 is an image of the approximate range of the identified contact line;
FIG. 11 is an image of an identified worn area;
fig. 12 is an image of an identified arcing burn trace.
Detailed Description
The technical scheme of the invention is further elaborated in the following by combining the drawings in the specification.
Example 1
Referring to the attached figure 1 of the specification, the embodiment discloses a contact network arcing ignition detection method, which comprises the following steps:
an image acquisition step, namely acquiring a rigid contact line image in the whole pull-out value range by adopting a linear array camera;
a contact line positioning step, namely preprocessing the rigid contact line image, extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image, and positioning the position of a contact line according to the offset of the contact line relative to the bus bar boundary;
a wear detection step, namely extracting a contact line boundary by taking the position of a contact line as a reference in the bus area image, searching a line segment which is the longest and continuously larger than a gray threshold in the contact line area and is the wear pixel length, and extracting a wear image;
the method comprises the following steps of (1) arc burning detection, namely searching line by line in a wear image, setting a left searching range and a right searching range in each line according to the position of a contact line, searching pixels of which the image gray level is smaller than a gray level threshold value in the searching ranges, and counting the number; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
As an implementation mode of the invention, the arc burning step is mainly used for detecting the arc burning trace on the contact line, and comprises five main steps: determining the search range of each line, searching target pixels, judging effective lines, counting the effective lines and the maximum continuous effective line length, and judging the burning trace of the arcing, wherein the specific implementation mode is as follows:
let L be the contact line location vector obtained in the contact line location step, i.e., L ═ L0,…,lh,…,lH-1]，lhThe H-th row of contact lines is the abscissa position, I represents a linear array camera image, the size of the image I is W x H, namely W represents the image width, and H represents the image height; then each row sets a left and right search range according to the position of the contact line, which is expressed as:
step 1: according to the h-th row contact line position lhSetting left and right search ranges SRhThe setting method is as follows:
SRh＝[max(lh-offset,0),min(lh+offset,W)]
wherein, offset is a preset offset, max (-) is an operator for solving the maximum value, and min (-) is an operator for solving the minimum value;
step 2: in the search range SRhSearching pixels with image gray smaller than gray threshold value gt in the image, and counting the number ChThe gray threshold gt determination method is as follows:
Figure BDA0003117683340000101
wherein the content of the first and second substances,
Figure BDA0003117683340000102
positioning the contact lineThe obtained bus bar image is an image in the range of the bus bar in the original image I; fac2 represents the set coefficient, fac2 ∈ (1, 3)]；
Step 3: judgment ChWhether the h-th behavior is greater than a threshold ct or not, and if the h-th behavior is greater than the threshold ct, the h-th behavior is judged to be a valid line, namely
F＝[f0,…,fh,…,fH-1]T
Figure BDA0003117683340000103
Wherein F is a valid row flag vector, FhFor valid line marking, when fhWhen f is 1, it is effectivehWhen the value is 0, the operation is invalid;
the threshold ct is set in relation to the contact line width and the worn pixel length determined by the row, i.e.:
Figure BDA0003117683340000104
wherein offset is a search range offset, which has been set in Step 1; aLhWorn pixel length detected for row h; cons is constant, Cons is E [1,3 ]]Configuring according to actual conditions;
step 4: the total number of valid lines and the maximum continuous valid line length, i.e.
Figure BDA0003117683340000105
Assume the valid row flag vector is as follows:
d＝[1,0,0,0,1,1,0,1,0,0,1,1,1,1,0]T，
the effective continuous rows have
vcF0＝[F4,F5]
vcF1＝[F10,F11,F12,F13]
Therefore, the maximum valid consecutive line length, i.e. the maximum valid consecutive line length, is:
vcFmax＝bcF2
vcFmaxL＝len(vcF2)
in general terms, the amount of the solvent to be used,
Figure BDA0003117683340000106
wherein svf is the total number of active rows; vcFiFor the ith active consecutive line, s is the current image active consecutive line set, vcFmaxvcF for the longest valid consecutive rowmaxL is the longest effective continuous line length, len (-) is the vector length operator;
step5, judging whether the number of effective lines is greater than a threshold value vt or whether the number of continuous effective lines is greater than a threshold value vct, and if any condition meets the requirement, determining that arcing ignition exists in the current image, namely the current image is burnt
Figure BDA0003117683340000111
And if the resFlag is equal to 0, the current image is considered to have no arcing.
Further, as an implementation manner of this embodiment, in the step of positioning the contact line, the acquired rigid contact line image is preprocessed, specifically, a transverse gradient image E of the original image I is obtained; then, a binarized image B of the transverse gradient image E is obtained.
When a transverse gradient image E is obtained, the following gradient operators are adopted:
Figure BDA0003117683340000112
then, the gradient image can be calculated as follows:
E(x,y)＝∑i∑j I(x+i,y+j)·A(i+1,j+1),i,j∈(-1,0,1)；
when the binarized image B is obtained, the binarized threshold bt is obtained by the following formula:
bt max (e) fac 0; where max (·) is the operator to find the maximum, fac0 is the manual setting coefficient, and the referable range is [0.5,1 ]; the binarized image B can be calculated by the following equation:
Figure BDA0003117683340000113
extracting bus boundary information from the preprocessed image, specifically, searching a connected domain according to a binary image; in this embodiment, an eight-connected domain is searched according to the binary image B, in this embodiment, the acquired image connected domain API of opencv is directly called: findContours (·), assuming that N connected domains are obtained by the binary image B in total, the connected domain set is recorded as:
Con＝[con0,…,conn,…,conN](ii) a Wherein, connIs the nth connected domain point set;
and tracking curves of all image heights through each connected domain in the following specific mode:
respectively obtaining the maximum and minimum points of the vertical coordinate of each connected domain, and respectively recording the points as the maximum and minimum points
Figure BDA0003117683340000114
Tracing a line segment upwards from the maximum point to the minimum point, or tracing a line segment downwards from the minimum point to the maximum point to downwards in the connected domain connMiddle trace is an example:
firstly, determining a transverse search range according to the abscissa of the minimum point:
Figure BDA0003117683340000121
where min (-) is the operator for the minimum,
Figure BDA0003117683340000122
represents a connected domain connThe abscissa and the offset of the middle and minimum point are manually set search offset and are set according to actual conditions;
at the minimum point
Figure BDA0003117683340000123
In the corresponding row, and searching for the range in that row
Figure BDA0003117683340000124
Inner search gray value greater than threshold
Figure BDA0003117683340000125
The threshold is determined by:
Figure BDA0003117683340000126
wherein fac1 is a manual setting coefficient, and can refer to a range [0.5,1 ];
finally, taking the center of the line segment with the longest continuous length larger than the threshold value as a connected domain connThe abscissa of the curve traced in this row, i.e.
Suppose that the pixel values in the line search range are:
Figure BDA0003117683340000127
setting the corresponding gray level threshold of the line
Figure BDA0003117683340000128
The line segments continuously greater than the threshold value have
Figure BDA0003117683340000129
Figure BDA00031176833400001210
The longest line segment continuously greater than the threshold is therefore
Figure BDA00031176833400001211
Finally, the coordinates of the curve traced by the line are line segments with the longest length continuously larger than the threshold value
Figure BDA00031176833400001212
Of the center point of (i.e. coordinates of
Figure BDA00031176833400001213
Wherein the content of the first and second substances,
Figure BDA00031176833400001214
is represented in the connected domain connMiddle minimum point
Figure BDA00031176833400001215
Coordinates of the curves traced in the corresponding row, Pt (-) representing
Figure BDA00031176833400001216
Pixel coordinates of the point in (1) in the image I;
when tracing connected domain connWhen the curve coordinate of the ith row is determined, the search range of the current row is determined according to the curve coordinate tracked by the previous row, and generally,
Figure BDA0003117683340000131
wherein the content of the first and second substances,
Figure BDA0003117683340000132
is represented in the connected domain connMiddle (i-Curve abscissa, SR, traced by line 1iRepresents a connected domain connThe horizontal search range of the ith row;
then searching the longest continuous larger than the gray threshold value gt of the current lineiLine segment of
Figure BDA0003117683340000133
Finally will be
Figure BDA0003117683340000134
The coordinates of the center point are used as the coordinates of the curve traced by the current line, i.e. in general,
Figure BDA0003117683340000135
wherein, the threshold value gtiThe general determination of (c) is as follows:
Figure BDA0003117683340000136
so far, each connected domain respectively tracks an incomplete curve, then the upper end and the lower end of each curve respectively begin to extend upwards and downwards to the whole image height range, and when the curve extends upwards, the search range is determined by adopting the following mode:
Figure BDA0003117683340000137
when extended upward, the search range is determined as follows:
Figure BDA0003117683340000138
finally, N full height curves are traced:
Cur＝[Cur0,…,Curn,…,CurN]。
screening out a curve with the maximum gray level mean value from all the tracked curves, if two or more curves have the same gray level mean value, selecting a curve with the minimum difference variance of the horizontal coordinates of the curve under the condition of the maximum gray level mean value, wherein the curve can be used as a shape of a representation busbar in an image; the method comprises the following specific steps:
let ave be [ ave ]0,…,aven,…,aveN](ii) a Where ave represents the curve gray level mean vector, avenDenotes the nth connected domain connThe pixel mean of the tracked curve is calculated in the following way:
Figure BDA0003117683340000139
let diff be [ diff ]0,…,diffn,…,diffN]；
diffn＝[diffn0,…,diffnh,…,diffn(H-step)](ii) a Wherein diff represents a difference matrix of the abscissa of the curve, diffnAnd (3) representing the difference component of the horizontal coordinate of the nth curve, H is the image height, step is the difference step length, and the specific difference and calculation mode is as follows:
Figure BDA0003117683340000141
wherein | is an absolute operator;
then calculating the variance of the difference vector of the horizontal coordinates of each curve,
var＝[var0,…,varn,…,varN](ii) a Wherein var represents the amount of difference between the horizontal coordinates of the curve, varnThe difference variance of the horizontal coordinates of the nth curve is represented, and the specific calculation mode is as follows:
Figure BDA0003117683340000142
where mean (-) represents the operator that calculates the mean;
then screening out the maximum gray mean value and the minimum horizontal coordinate difference varianceCurve characterizing the shape of the busbar: curshape＝minvar(maxave(Cur))。
Taking the curve as a reference, moving towards two sides at fixed intervals gap, calculating the moved curve pixel mean value, and if the mean values of a plurality of continuous curve pixels are smaller than a threshold value
Figure BDA0003117683340000143
The bus boundary is considered to be searched; specifically, assume that the curve pixel mean after moving to the right k-th time is less than the threshold
Figure BDA0003117683340000144
Namely, it is
Figure BDA0003117683340000145
Wherein the content of the first and second substances,
Figure BDA0003117683340000146
all abscissas in the bus bar shape curve are added with k.gap, namely, the k gap is shifted to the right;
Figure BDA0003117683340000147
the calculation method of (c) is as follows:
Figure BDA0003117683340000148
wherein fac2 is a manual setting coefficient, reference range [0.5,2](ii) a And the subsequently moved M curve pixel mean values are all smaller than the threshold value
Figure BDA0003117683340000149
Namely, it is
Figure BDA00031176833400001410
m∈[1,2,…,m,…,M]Then the current curve is determined to be the bus right boundary, i.e.
Figure BDA00031176833400001411
Wherein HLrRepresents the busbar right boundary curve;
similarly, a left boundary curve of the busbar can be obtained after moving the busbar to the left for a certain number of times, and the condition is satisfied after the busbar is supposed to move for q times, that is
Figure BDA00031176833400001412
The positioning of the position of the contact line according to the offset of the contact line relative to the boundary of the bus bar specifically means that the bus bar image is intercepted from the original gray level image I according to the boundary curve of the bus bar
Figure BDA00031176833400001413
Then, the obtained value is obtained
Figure BDA00031176833400001414
Corresponding transverse gradient image
Figure BDA00031176833400001415
Obtaining bus bar image
Figure BDA00031176833400001416
Longitudinal gray scale projection of
Figure BDA00031176833400001417
And bus bar image
Figure BDA00031176833400001418
Corresponding transverse gradient image
Figure BDA00031176833400001419
Longitudinal gray scale projection of
Figure BDA00031176833400001420
In that
Figure BDA00031176833400001421
Moving to the right with a fixed window, calculating the sum in each window; obtaining a sum vector after all sliding windows
Figure BDA00031176833400001422
Obtaining
Figure BDA00031176833400001423
The maximum value is obtained, and the column index is obtained; then on the bus image
Figure BDA00031176833400001424
Longitudinal gray scale projection of
Figure BDA00031176833400001425
Respectively searching out minimum values from left to right in a certain range by taking the obtained column index as a reference to obtain two minimum value indexes, and finally taking the average value of the two minimum value indexes as the offset of the contact line; and then calculating the position of the contact line in the original image I according to the left boundary of the busbar and the offset of the contact line.
Specifically, a bus image is intercepted from an original gray image I according to a bus boundary curve
Figure BDA0003117683340000151
Then, the obtained value is obtained
Figure BDA0003117683340000152
Corresponding transverse gradient image
Figure BDA0003117683340000153
The obtaining method is the same as that of obtaining the corresponding transverse gradient image E from the original gray level image I; obtaining
Figure BDA0003117683340000154
And
Figure BDA0003117683340000155
longitudinal gray scale projection, in particular, of
Figure BDA0003117683340000156
Wherein
Figure BDA0003117683340000157
Is composed of
Figure BDA0003117683340000158
The longitudinal gray projection vector of (a),
Figure BDA0003117683340000159
is composed of
Figure BDA00031176833400001510
The width of the image of (a) is,
Figure BDA00031176833400001511
is composed of
Figure BDA00031176833400001512
The projection value in the x-th column is calculated as follows:
Figure BDA00031176833400001513
can be calculated by the same principle
Figure BDA00031176833400001514
Longitudinal gray scale projection of, i.e.
Figure BDA00031176833400001515
Then, at
Figure BDA00031176833400001516
Sliding right with a fixed window, setting the window length according to the actual situation, and calculating the sum in each window, i.e.
Figure BDA00031176833400001517
Figure BDA00031176833400001518
Wherein
Figure BDA00031176833400001519
For the sum vector obtained after all sliding windows,
Figure BDA00031176833400001520
is centered in the window
Figure BDA00031176833400001521
The sum calculated at the xth element; then, the obtained value is obtained
Figure BDA00031176833400001522
Medium maximum value, and obtain its column index, i.e.
Figure BDA00031176833400001523
For convenience, let maxIdx be
Figure BDA00031176833400001524
In that
Figure BDA00031176833400001525
The corresponding column index; then, at
Figure BDA00031176833400001526
In a certain range with maxIdx as a reference, respectively searching out minimum values from left to right to obtain two minimum value indexes, and finally taking the index average value of the two minimum values as the offset of the contact line, namely los＝(dl+dr) 2; wherein losOffset of contact line with respect to left boundary of bus bar, dlIs composed of
Figure BDA00031176833400001527
The minimum index of the left search based on maxIdx, drIs composed of
Figure BDA00031176833400001528
The minimum value of the medium is searched rightwards by taking maxIdx as a referenceIndexing;
the position of a contact line in the original image I is obtained according to the left boundary of the busbar and the offset of the contact line, namely the abscissa vector of the contact line is set as L ═ L0,…,lh,…,lH-1]Then, then
Figure BDA00031176833400001529
In the abrasion detection step, a contact line position is taken as a reference to extract a contact line boundary in a busbar area image, specifically, a contact line curve obtained by positioning in the contact line positioning step is taken as a reference, the contact line curve is moved towards two sides, a moved curve pixel mean value is calculated, and if a plurality of continuous curve pixel mean values are smaller than a threshold value, the contact line boundary is considered; searching the maximum pixel value in the contact line boundary searching range, setting a gray threshold according to the maximum value, and searching a line segment continuously larger than the gray threshold in the searching range, wherein the line segment is a wear pixel length line segment.
Example 2
As another preferred embodiment of the present invention, referring to fig. 2 in the specification, the present embodiment discloses a detecting device for contact network arcing ignition, which can accurately detect the trace and position of the contact network that is seriously ignited by arcing, and can provide important help and reference information for maintenance personnel.
The contact network arcing firing detection device comprises an image acquisition module, a contact line positioning module, a contact line abrasion detection module and an arcing firing detection module;
the image acquisition module comprises a light source and a linear array camera, the light source is used for supplementing light to a contact line area, the linear array camera is used for acquiring an original image of a rigid contact line in the whole pull-out value range and inputting the original image of the contact line into the contact line positioning module;
the contact line positioning module is used for preprocessing an input contact line original image, extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image, and positioning the position of a contact line according to the offset of the contact line relative to the bus bar boundary; the contact line positioning module sends the bus area image and the contact line position to the contact line abrasion detection module;
the contact line abrasion detection module extracts a contact line boundary by taking a contact line position as a reference in a bus area image to obtain a contact line area, and searches a line segment which is the longest continuous line segment larger than a gray threshold value in the contact line area, namely the abrasion pixel length, to obtain an abrasion image; the contact line abrasion detection module sends the abrasion image to an arcing firing detection module;
the arcing firing detection module searches the abrasion image line by line, each line sets a left searching range and a right searching range according to the position of a contact line, pixels of which the image gray level is smaller than a gray level threshold value are searched in the searching ranges, and the number of the pixels is counted; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
Furthermore, in the arc burning detection module, L is a contact line positioning vector obtained in the contact line positioning step, i.e., L ═ L0,…,lh,…,lH-1]，lhThe H-th row of contact lines is the abscissa position, I represents a linear array camera image, the size of the image I is W x H, namely W represents the image width, and H represents the image height; then each row sets a left and right search range according to the position of the contact line, which is expressed as:
according to the h-th row contact line position lhSetting left and right search ranges SRhThe setting method is as follows:
SRh＝[max(lh-offset,0),min(lh+offset,W)]
wherein, offset is a preset offset, max (-) is an operator for solving the maximum value, and min (-) is an operator for solving the minimum value;
in the search range SRhSearching pixels with image gray smaller than gray threshold value gt in the image, and counting the number ChThe gray threshold gt determination method is as follows:
Figure BDA0003117683340000171
wherein the content of the first and second substances,
Figure BDA0003117683340000172
the bus bar image obtained in the contact line positioning step is an image in the range of the bus bar in the original image I; fac2 represents the set coefficient, fac2 ∈ (1, 3)]；
Judgment ChWhether the h-th behavior is greater than a threshold ct or not, and if the h-th behavior is greater than the threshold ct, the h-th behavior is judged to be a valid line, namely
F＝[f0,…,fh,…,fH-1]T
Figure BDA0003117683340000173
Wherein F is a valid row flag vector, FhFor valid line marking, when fhWhen f is 1, it is effectivehWhen the value is 0, the operation is invalid;
the threshold ct is set in relation to the contact line width and the worn pixel length determined by the row, i.e.:
Figure BDA0003117683340000174
wherein offset is the search range offset; aLhWorn pixel length detected for row h; cons is constant, Cons is E [1,3 ]]；
The total number of valid lines and the maximum continuous valid line length, i.e.
Figure BDA0003117683340000175
Figure BDA0003117683340000176
Wherein svf is the total number of active rows; vcFiFor the ith active consecutive line, S is the current image active consecutive line set, vcFmaxvcF for the longest valid consecutive rowmaxL is the longest effective continuous line length, len (-) is the vector length operator;
judging whether the number of effective lines is greater than a threshold value vt or whether the number of continuous effective lines is greater than a threshold value vct, and if any condition meets the requirement, determining that arcing firing exists in the current image, namely
Figure BDA0003117683340000181
And if the resFlag is equal to 0, the current image is considered to have no arcing.
In the contact line positioning module, preprocessing an input contact line original image specifically means that a transverse gradient image E of an original image I is obtained; then, a binarized image B of the transverse gradient image E is obtained.
Furthermore, the following gradient operator is used for solving the transverse gradient image E:
Figure BDA0003117683340000182
then, the gradient image can be calculated as follows:
E(x,y)＝∑i∑j I(x+i,y+j)·A(i+1,j+1),i,j∈(-1,0,1)；
when the binarized image B is obtained, the binarized threshold bt is obtained by the following formula:
bt max (e) fac 0; where max (·) is the operator to find the maximum, fac0 is the manual setting coefficient, and the referable range is [0.5,1 ]; the binarized image B can be calculated by the following equation:
Figure BDA0003117683340000183
the method comprises the steps of extracting the boundary information of the busbar from a preprocessed image, specifically, searching connected domains according to a binary image, tracking curves of all image heights through each connected domain, screening out a curve with the maximum gray average value from all the tracked curves, and selecting a curve with the minimum difference variance of the horizontal coordinates of the curve under the condition that the gray average value is maximum if two or more curves have the same gray average value, wherein the curve can be used for representing the shape of the busbar in the image; moving to both sides at fixed intervals by taking the curve as a reference, calculating the moved curve pixel mean value, and if the mean values of a plurality of continuous curve pixels are smaller than a threshold value
Figure BDA0003117683340000184
The bus boundary is considered to be searched.
The positioning of the position of the contact line according to the offset of the contact line relative to the boundary of the bus bar specifically means that the bus bar image is intercepted from the original gray level image I according to the boundary curve of the bus bar
Figure BDA0003117683340000185
Then, the obtained value is obtained
Figure BDA0003117683340000186
Corresponding transverse gradient image
Figure BDA0003117683340000187
Obtaining bus bar image
Figure BDA0003117683340000188
Longitudinal gray scale projection of
Figure BDA0003117683340000189
And bus bar image
Figure BDA00031176833400001810
Corresponding transverse gradient image
Figure BDA00031176833400001811
Longitudinal gray scale projection ofShadow
Figure BDA00031176833400001812
In that
Figure BDA00031176833400001813
Moving to the right with a fixed window, calculating the sum in each window; obtaining a sum vector after all sliding windows
Figure BDA00031176833400001814
Obtaining
Figure BDA00031176833400001815
The maximum value is obtained, and the column index is obtained; then on the bus image
Figure BDA00031176833400001816
Longitudinal gray scale projection of
Figure BDA00031176833400001817
Respectively searching out minimum values from left to right in a certain range by taking the obtained column index as a reference to obtain two minimum value indexes, and finally taking the average value of the two minimum value indexes as the offset of the contact line; and then calculating the position of the contact line in the original image I according to the left boundary of the busbar and the offset of the contact line.
In the contact line abrasion detection module, a contact line position is taken as a reference to extract a contact line boundary in a busbar area image, specifically, a contact line curve obtained by positioning in a contact line positioning module is taken as a reference, the contact line curve is moved towards two sides, a moved curve pixel mean value is calculated, and if a plurality of continuous curve pixel mean values are smaller than a threshold value, the contact line boundary is considered; searching the maximum pixel value in the contact line boundary searching range, setting a gray threshold according to the maximum value, and searching a line segment continuously larger than the gray threshold in the searching range, wherein the line segment is a wear pixel length line segment.
Example 3
As another preferred embodiment of the invention, referring to the attached drawings 3-12 in the specification, the embodiment discloses a contact netAn arc burning detection method, wherein an image I acquired by a line camera is shown in fig. 3, and W × H is 8192 × 1000; traced curve Cur representing the shape of the busbarshapeAs shown in fig. 4; further, the respective identified left and right bus bar boundaries are shown in fig. 5; clipping bus images according to bus boundaries
Figure BDA0003117683340000191
As shown in fig. 6; bus image
Figure BDA0003117683340000192
The longitudinal gray scale projection image of (a) is shown in fig. 7; the contact line offset l thus identifiedosAs shown in fig. 8; the contact lines in the original image I are shown in fig. 9; further, the approximate contact line range identified is shown in FIG. 10; the worn area identified from the contact line left and right boundaries is shown in FIG. 11; finally, the identified arc burn trace results are shown in fig. 12.
Example 4
In order to achieve the above object, according to another aspect of the present application, there is also provided a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the above cyclic neural network-based catenary failure prediction method when executing the computer program.
The processor may be a Central Processing Unit (CPU) in this embodiment. The Processor may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or a combination thereof.
The memory, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and units, such as the corresponding program units in the above-described method embodiments of the present invention. The processor executes various functional applications of the processor and the processing of the work data by executing the non-transitory software programs, instructions and modules stored in the memory, that is, the method in the above method embodiment is realized.
The memory may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor, and the like. Further, the memory may include high speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory located remotely from the processor, and such remote memory may be coupled to the processor via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more units are stored in the memory and, when executed by the processor, perform the method of embodiment 1 above.
Example 5
As another preferred embodiment of the present invention, this embodiment discloses a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps of embodiment 1 above.

Claims (10)

1. The contact net arcing ignition detection method is characterized by comprising the following steps:
an image acquisition step, namely acquiring a rigid contact line image in the whole pull-out value range from the linear array camera;
a contact line positioning step, namely preprocessing the rigid contact line image, and extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image; positioning the position of the contact line according to the offset of the contact line relative to the boundary of the busbar;
a wear detection step, namely extracting a contact line boundary by taking the position of a contact line as a reference in the bus area image, searching a line segment which is the longest and continuously larger than a gray threshold in the contact line area and is the wear pixel length, and extracting a wear image;
the method comprises the following steps of (1) arc burning detection, namely searching line by line in a wear image, setting a left searching range and a right searching range in each line according to the position of a contact line, searching pixels of which the image gray level is smaller than a gray level threshold value in the searching ranges, and counting the number; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
2. The contact network arcing ignition detection method of claim 1, wherein: in the step of detecting arcing firing, L is set as a contact line positioning vector obtained in the step of positioning a contact line, i.e., L ═ L0，…，lh，…，lH-1]，lhThe H-th row of contact lines is the abscissa position, I represents a linear array camera image, the size of the image I is W x H, namely W represents the image width, and H represents the image height; then each row sets a left and right search range according to the position of the contact line, which is expressed as:
setting a left and right search range SR according to the h-th row contact line position lhhThe setting method is as follows:
SRh＝[max(lh-offset，0)，min(lh+offset，W)]
wherein, offset is a preset offset, max (-) is an operator for solving the maximum value, and min (-) is an operator for solving the minimum value;
in the search range SRhSearching pixels with image gray smaller than gray threshold value gt in the image, and counting the number ChThe gray threshold gt determination method is as follows:
Figure FDA0003117683330000011
wherein the content of the first and second substances,
Figure FDA0003117683330000012
the bus bar image obtained in the contact line positioning step is an image in the range of the bus bar in the original image I; fac2 represents the set coefficient, fac2 ∈ (1, 3)]；
Judging whether Ch is greater than a threshold value ct or not, and if so, judging the h-th behavior valid line, namely
F＝[f0，…，fh，…，fH-1]T
Figure FDA0003117683330000021
Wherein F is a valid row flag vector, FhFor valid line marking, when fhWhen f is 1, it is effectivehWhen the value is 0, the operation is invalid;
the threshold ct is set in relation to the contact line width and the worn pixel length determined by the row, i.e.:
Figure FDA0003117683330000022
wherein offset is the search range offset; aLhWorn pixel length detected for row h; cons is constant, Cons is E [1,3 ]]；
The total number of valid lines and the maximum continuous valid line length, i.e.
Figure FDA0003117683330000023
Figure FDA0003117683330000024
Wherein svf is the total number of active rows; vcFiFor the ith active consecutive line, S is the current image active consecutive line set, vcFmaxvcF for the longest valid consecutive rowmaxL is the longest effective continuous line length, len (-) is the vector length operator;
judging whether the number of effective lines is greater than a threshold value vt or whether the number of continuous effective lines is greater than a threshold value vct, and if any condition meets the requirement, determining that arcing firing exists in the current image, namely
Figure FDA0003117683330000025
And if the resFlag is equal to 0, the current image is considered to have no arcing.
3. The contact network arcing ignition detection method of claim 1, wherein: in the contact line positioning step, preprocessing an acquired rigid contact line image, specifically, solving a transverse gradient image E of an original image I; then, a binarized image B of the transverse gradient image E is obtained.
4. The contact network arcing ignition detection method of claim 3, wherein: when a transverse gradient image E is obtained, the following gradient operators are adopted:
Figure FDA0003117683330000026
then, the gradient image can be calculated as follows:
E(x，y)＝∑i∑jI(x+i，y+j)·A(i+1，j+1)，i，j∈(-1，0，1)；
when the binarized image B is obtained, the binarized threshold bt is obtained by the following formula:
bt max (e) fac 0; wherein max (·) is the operator for finding the maximum value, fac0 is the manual setting coefficient, fac0 belongs to [0.5,1 ]; the binarized image B can be calculated by the following equation:
Figure FDA0003117683330000031
5. the contact network arcing ignition detection method of claim 1, wherein: the method comprises the steps of extracting the boundary information of the busbar from a preprocessed image, specifically, searching connected domains according to a binary image, tracking curves of all image heights through each connected domain, screening out a curve with the maximum gray average value from all the tracked curves, and selecting a curve with the minimum difference variance of the horizontal coordinates of the curve under the condition that the gray average value is maximum if two or more curves have the same gray average value, wherein the curve can be used for representing the shape of the busbar in the image; moving to both sides at fixed intervals by taking the curve as a reference, calculating the moved curve pixel mean value, and if the mean values of a plurality of continuous curve pixels are smaller than a threshold value
Figure FDA0003117683330000032
The bus boundary is considered to be searched.
6. The contact network arcing ignition detection method of claim 1, wherein: the positioning of the position of the contact line according to the offset of the contact line relative to the boundary of the bus bar specifically means that the bus bar image is intercepted from the original gray level image I according to the boundary curve of the bus bar
Figure FDA0003117683330000033
Then, the obtained value is obtained
Figure FDA0003117683330000034
Corresponding transverse gradient image
Figure FDA0003117683330000035
Obtaining bus bar image
Figure FDA0003117683330000036
Longitudinal gray scale projection of
Figure FDA0003117683330000037
And bus bar image
Figure FDA0003117683330000038
Corresponding transverse gradient image
Figure FDA0003117683330000039
Longitudinal gray scale projection of
Figure FDA00031176833300000310
In that
Figure FDA00031176833300000311
Moving to the right with a fixed window, calculating the sum in each window; obtaining a sum vector after all sliding windows
Figure FDA00031176833300000312
Obtaining
Figure FDA00031176833300000313
The maximum value is obtained, and the column index is obtained; then on the bus image
Figure FDA00031176833300000314
Longitudinal gray scale projection of
Figure FDA00031176833300000315
Respectively searching out minimum values from left to right in a certain range by taking the obtained column index as a reference to obtain two minimum value indexes, and finally taking the average value of the two minimum value indexes as the offset of the contact line; and then calculating the position of the contact line in the original image I according to the left boundary of the busbar and the offset of the contact line.
7. The contact network arcing ignition detection method of claim 1, wherein: in the abrasion detection step, a contact line position is taken as a reference to extract a contact line boundary in a busbar area image, specifically, a contact line curve obtained by positioning in the contact line positioning step is taken as a reference, the contact line curve is moved towards two sides, a moved curve pixel mean value is calculated, and if a plurality of continuous curve pixel mean values are smaller than a threshold value, the contact line boundary is considered; searching the maximum pixel value in the contact line boundary searching range, setting a gray threshold according to the maximum value, and searching a line segment continuously larger than the gray threshold in the searching range, wherein the line segment is a wear pixel length line segment.
8. Contact net arcing firing detection device, its characterized in that: the device comprises an image acquisition module, a contact line positioning module, a contact line abrasion detection module and an arcing firing detection module;
the image acquisition module comprises a light source and a linear array camera, the light source is used for supplementing light to a contact line area, the linear array camera is used for acquiring an original image of a rigid contact line in the whole pull-out value range and inputting the original image of the contact line into the contact line positioning module;
the contact line positioning module is used for preprocessing an input contact line original image, extracting bus bar boundary information from the preprocessed image to obtain a bus bar area image, and positioning the position of a contact line according to the offset of the contact line relative to the bus bar boundary; the contact line positioning module sends the bus area image and the contact line position to the contact line abrasion detection module;
the contact line abrasion detection module extracts a contact line boundary by taking a contact line position as a reference in a bus area image to obtain a contact line area, and searches a line segment which is the longest continuous line segment larger than a gray threshold value in the contact line area, namely the abrasion pixel length, to obtain an abrasion image; the contact line abrasion detection module sends the abrasion image to an arcing firing detection module;
the arcing firing detection module searches the abrasion image line by line, each line sets a left searching range and a right searching range according to the position of a contact line, pixels of which the image gray level is smaller than a gray level threshold value are searched in the searching ranges, and the number of the pixels is counted; judging whether the number of pixels smaller than a preset gray threshold is larger than a threshold, and if so, judging the pixels to be effective lines; and counting the total number of the effective lines and the maximum continuous effective line length, and judging whether arcing firing exists in the current image according to whether the total number of the effective lines is greater than a preset threshold or not or whether the maximum continuous effective line length is greater than the preset threshold or not.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor performing the steps of the method of any of the preceding claims 1-7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of the preceding claims 1 to 7.
CN202110666642.5A 2021-06-16 2021-06-16 Contact net arcing burning detection method, device, computer equipment and storage medium Active CN113487543B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110666642.5A CN113487543B (en) 2021-06-16 2021-06-16 Contact net arcing burning detection method, device, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110666642.5A CN113487543B (en) 2021-06-16 2021-06-16 Contact net arcing burning detection method, device, computer equipment and storage medium

Publications (2)

Publication Number Publication Date
CN113487543A true CN113487543A (en) 2021-10-08
CN113487543B CN113487543B (en) 2023-08-04

Family

ID=77935145

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110666642.5A Active CN113487543B (en) 2021-06-16 2021-06-16 Contact net arcing burning detection method, device, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN113487543B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114384072A (en) * 2021-11-30 2022-04-22 杭州申昊科技股份有限公司 Abrasion detection method and system for track inspection rigid contact net
CN116309280A (en) * 2022-12-16 2023-06-23 上海药明康德新药开发有限公司 Lymphocyte labeling method and system
CN116958099A (en) * 2023-07-27 2023-10-27 微牌科技(浙江)有限公司 Cable abrasion detection method, system, device and computer equipment
CN116958099B (en) * 2023-07-27 2024-05-24 微牌科技(浙江)有限公司 Cable abrasion detection method, system, device and computer equipment

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008089523A (en) * 2006-10-05 2008-04-17 Meidensha Corp Wear measurement device for trolley wire
US20100016720A1 (en) * 2006-09-20 2010-01-21 Shimadzu Corporation Ultrasonograph
CN102202182A (en) * 2011-04-29 2011-09-28 北京工业大学 Device and method for acquiring high dynamic range images by adopting linear array charge coupled device (CCD)
CN103440657A (en) * 2013-08-27 2013-12-11 武汉大学 Method for online screening cracks of road
CN105652154A (en) * 2016-01-25 2016-06-08 成都国铁电气设备有限公司 Safety monitoring analysis system for contact net running state
CN109001597A (en) * 2017-10-30 2018-12-14 北京华开领航科技有限责任公司 A kind of electric spark determines method and device
CN109269416A (en) * 2017-07-17 2019-01-25 成都唐源电气股份有限公司 A kind of contact line conducting wire measurement of wear method and device
CN110717900A (en) * 2019-09-27 2020-01-21 南京理工大学 Pantograph abrasion detection method based on improved Canny edge detection algorithm
US20200175909A1 (en) * 2018-11-30 2020-06-04 Beijing Xiaomi Mobile Software Co., Ltd. Screen display method and screen display device
CN111323683A (en) * 2020-03-23 2020-06-23 中铁电气化局集团有限公司 Arcing detection system, arcing detection method and device and computer equipment
CN111339797A (en) * 2020-02-25 2020-06-26 福州符号信息科技有限公司 Decoding method and terminal capable of accurately identifying damaged one-dimensional bar code

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100016720A1 (en) * 2006-09-20 2010-01-21 Shimadzu Corporation Ultrasonograph
JP2008089523A (en) * 2006-10-05 2008-04-17 Meidensha Corp Wear measurement device for trolley wire
CN102202182A (en) * 2011-04-29 2011-09-28 北京工业大学 Device and method for acquiring high dynamic range images by adopting linear array charge coupled device (CCD)
CN103440657A (en) * 2013-08-27 2013-12-11 武汉大学 Method for online screening cracks of road
CN105652154A (en) * 2016-01-25 2016-06-08 成都国铁电气设备有限公司 Safety monitoring analysis system for contact net running state
CN109269416A (en) * 2017-07-17 2019-01-25 成都唐源电气股份有限公司 A kind of contact line conducting wire measurement of wear method and device
CN109001597A (en) * 2017-10-30 2018-12-14 北京华开领航科技有限责任公司 A kind of electric spark determines method and device
US20200175909A1 (en) * 2018-11-30 2020-06-04 Beijing Xiaomi Mobile Software Co., Ltd. Screen display method and screen display device
CN110717900A (en) * 2019-09-27 2020-01-21 南京理工大学 Pantograph abrasion detection method based on improved Canny edge detection algorithm
CN111339797A (en) * 2020-02-25 2020-06-26 福州符号信息科技有限公司 Decoding method and terminal capable of accurately identifying damaged one-dimensional bar code
CN111323683A (en) * 2020-03-23 2020-06-23 中铁电气化局集团有限公司 Arcing detection system, arcing detection method and device and computer equipment

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
G. SILVA RAMOS等: "Prototype of a Rail Wear Measuring System Using Digital Image Processing", 《IEEE LATIN AMERICA TRANSACTIONS》, vol. 17, no. 4, pages 582 - 589, XP011754906, DOI: 10.1109/TLA.2019.8891882 *
徐清霞等: "一种钢轨磨耗测量的新方法", 《上海工程技术大学学报》, vol. 27, no. 3, pages 278 - 282 *
校美玲: "基于单目视觉的轮对磨耗检测方法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, no. 3, pages 033 - 757 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114384072A (en) * 2021-11-30 2022-04-22 杭州申昊科技股份有限公司 Abrasion detection method and system for track inspection rigid contact net
CN116309280A (en) * 2022-12-16 2023-06-23 上海药明康德新药开发有限公司 Lymphocyte labeling method and system
CN116958099A (en) * 2023-07-27 2023-10-27 微牌科技(浙江)有限公司 Cable abrasion detection method, system, device and computer equipment
CN116958099B (en) * 2023-07-27 2024-05-24 微牌科技(浙江)有限公司 Cable abrasion detection method, system, device and computer equipment

Also Published As

Publication number Publication date
CN113487543B (en) 2023-08-04

Similar Documents

Publication Publication Date Title
CN107014294B (en) Contact net geometric parameter detection method and system based on infrared image
CN113487543A (en) Contact net arcing firing detection method, device, computer equipment and storage medium
US5590217A (en) Vehicle activity measuring apparatus
CN107292870B (en) Track plug pin fault detection method and system based on image alignment and detection network model
US5555312A (en) Automobile apparatus for road lane and vehicle ahead detection and ranging
CN104866823B (en) A kind of moving vehicles detection and tracking method based on monocular vision
CN106780565B (en) Multi-student sitting-up detection method based on optical flow and k-means clustering
Landi et al. Hough transform and thermo-vision for monitoring pantograph-catenary system
CN110450706B (en) Self-adaptive high beam control system and image processing algorithm
JP5136504B2 (en) Object identification device
US20200394415A1 (en) Object detection apparatus, prediction model generation apparatus, object detection method, and program
Aydin A new approach based on firefly algorithm for vision-based railway overhead inspection system
CN111539436B (en) Rail fastener positioning method based on straight template matching
CN107730454B (en) Linear-tracking-based method for removing horizontal stripe noise in linear array image
CN110084830B (en) Video moving object detection and tracking method
CN111238365A (en) Subway train distance measurement and positioning method and system based on stereoscopic vision
CN108226890A (en) Based on the time to the airport foreign matter radar detecting method of statistics
CN108917621B (en) Pantograph slide plate upper edge single-pixel tracking detection method
CN103150550B (en) A kind of road pedestrian event detection method based on gripper path analysis
CN109325963B (en) SVM-based three-dimensional trajectory classification method for bus passengers
CN111323683B (en) Arcing detection system, arcing detection method, arcing detection device and computer equipment
JP4459162B2 (en) Velocity measuring device, method and program
CN110322508B (en) Auxiliary positioning method based on computer vision
Li et al. Rear lamp based vehicle detection and tracking for complex traffic conditions
CN110634122A (en) Fastener elastic strip displacement detection method

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