CN102288614B - Method for detecting pantograph crack fault based on curvelet domain moving parallel window - Google Patents
Method for detecting pantograph crack fault based on curvelet domain moving parallel window Download PDFInfo
- Publication number
- CN102288614B CN102288614B CN2011101302503A CN201110130250A CN102288614B CN 102288614 B CN102288614 B CN 102288614B CN 2011101302503 A CN2011101302503 A CN 2011101302503A CN 201110130250 A CN201110130250 A CN 201110130250A CN 102288614 B CN102288614 B CN 102288614B
- Authority
- CN
- China
- Prior art keywords
- pantograph
- image
- slide plate
- crack
- curvelet
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a method for detecting a pantograph crack fault based on a curvelet domain moving parallel window, and the method comprises the following steps of: acquiring pantograph images through image pickup systems along a train track, performing a series of preprocessing operations on the acquired images to obtain slipper images; performing second-generation curvelet transform on the slipper images, converting images in a spatial domain into a curvelet decomposition coefficient matrix in a curvelet domain, analyzing, classifying and threshold-processing the curvelet decomposition coefficient matrix by using a moving parallel window method, filtering off curvelet coefficients corresponding to all the inherent image elements, except cracks, on the slipper images, and finally, performing inverse transform on the processed curvelet decomposition coefficient matrix so that slipper crack images can be extracted to implement the detection and recognition for pantograph slipper crack fault. Aiming at the singularity characteristic of pantograph slipper crack lines, the method provided by the invention can be used for effectively capturing the linear characteristics of the crack images by means of curvelet transform, and precisely and effectively extracting crack characteristics to implement the detection for the pantograph crack fault.
Description
Technical field
The present invention relates to the high-speed railway checkout equipment and make field, especially pantograph pan method for crack
Background technology
High-speed railway all adopts electric propulsion, and electric propulsion is a track traffic mode of traction of knowing at present the highest, the energy-conserving and environment-protective of energy utilization rate.Therefore, no matter all newly-built track traffics at present are high speed, heavy duty, or urban track traffic, almost all adopt electric propulsion.Electric locomotive draws electric current from contact net under the high-speed cruising condition, how must guarantee to be flowed uninterrupted and absolute reliable in service guarantees that it is one of gordian technique of electric railway that the high-speed cruising train has good current carrying quality.
Pantograph is electric locomotive draws electric current from contact net a device; It directly contacts with contact line conducting wire; Obtain electric current from contact line conducting wire; The confession locomotive uses, and pantograph cooperates the electric energy of undertaking jointly obtaining from the traction net directly to flow to the vital task that electric locomotive uses with osculatory.Because have complicated mechanics and electric reciprocal effect between pantograph and the contact net, its failure rate is higher always.The pantograph and catenary fault more complicated, the damage of pantograph is sometimes inseparable with the fault of contact net, and the fault of pantograph possibly cause the damage of contact net, and that contact net goes wrong is also unfavorable to pantograph.So, through real-time monitoring, can in time find the pantograph fault to the pantograph state, can judge the state of contact net through the ruuning situation of pantograph simultaneously, thereby help the timely maintenance of pantograph and contact net, reduce the generation of pantograph and catenary fault.
Present existing pantograph fault detection technique mainly contains manual detection mode, the online detection mode of fixed point and based on the detection RM of Flame Image Process etc., realizes the pantograph status detection through servicing units such as sensor, infrored equipment, video cameras.But existing research to the pantograph fault concentrates on the detection that abrasion are transfinited to slide plate, and the research that the sled surface defective mode is detected (like crack detection) is less, and the slide plate crack fault very easily causes more serious accident between pantograph.
Document " current situation of the other monitoring system in Minoru Ogasawara. rail truck road. external rolling stock technology, Yu Huijuan translates .2000,5:1-3 " introduced the pantograph pan bar wearing and tearing self-operated measuring unit of Developed by JR East.This system is made up of ultrasonic sensor, video camera, illuminating lamp, trigger switch.This device sets up ultrasonic sensor along the railway, when locomotive through the time, ultrasonic sensor sends ultrasound wave, arrives on the measured object after reflection turns back on the sensor through air transfer.According to the time of ultrasonic transmission and velocity of wave at that time, can calculate the thickness of slide plate bar.
Document " Li Zongzhi. the research of Dynamic Monitoring Device of Pantographs. Southwest Jiaotong University's master thesis; 2004 " introduced a pantograph dynamic monitoring system; This system can 24 hours carries out real time monitoring from level, vertical direction to the train pantograph going; Through manual analysis, judge the abnormal information of bow net system then, and in time report to the police monitoring image.The pantograph picture that the pantograph automatic monitoring device of introducing in the document " Otsuki Yasutake; Kosaka shuiehi; Naka shigeki.Development of a pantograph automatic monitor.JREA.1999,42 (7): 26266-26269 " has taked robot calculator that ccd video camera is taken carries out analyzing and processing.This device is made up of optical system, picture processing system and display system three parts.Optical system is by being erected at photoelectric sensor on the overhead crossbeam of osculatory, not formed by light fixture that intensity of illumination influences and the parts that ccd video camera is housed.Picture processing system is that the optical system various piece is controlled and ccd video camera is clapped the device that the pantograph picture is handled, and this part energy measurement slide plate bar thickness, the identification pantograph is unusual and the improper abrasion of slide plate bar.
Document " L.G.C.Hamey; T.Watkins; S.W.T.Yen.Pancam:In-Service Inspection of Locomotive Pantographs.9th Biennial Conference of the Australian Pattern Recognition Society on Digital Image Computing Techniques and Applications, 2007:493-499 " has been introduced a kind of Pancam motorcycle pantograph on-line monitoring system.It adopts the pictures different disposal route according to the architectural feature of pantograph to the pantograph various piece, has realized robotization fully.This system uses two CCD cameras to catch the side view and the vertical view of running status pantograph respectively, and side view is used to detect the wearing and tearing of carbon slide plate bar, and vertical view is used to analyze the damage of bow angle part.
Document " Feng Qian. the research of motorcycle pantograph image intelligent processing and detecting. Southwest Jiaotong University's master thesis; 2008 " introduced motorcycle pantograph wireless video monitoring system based on Flame Image Process, this system mainly to the pantograph of TSG3-630/25 type single arm pantograph picture tilt, draw runner is lost, the draw runner abrasion are transfinited and the draw runner crack fault is discerned.It is through using the dynamic image of camera pantograph; The appliance computer intelligent identification technology is discerned the various faults of motorcycle pantograph automatically; Use wireless transmitting system that motorcycle pantograph is carried out remote monitoring, and realize the long-distance on-line of each control point by Ethernet.
Summary of the invention
In view of the above deficiency of prior art, the object of the present invention is to provide a kind of detection method of pantograph pan crack fault, this method is to adopt image processing techniques that the pantograph image is handled, analyzed, and goes out existing of pantograph slider crackle with Real time identification.Its core is that the pantograph pan image transitions of spatial domain is analyzed to bent wave zone, through the processing that the Qu Bofen to the pantograph pan image separates matrix of coefficients, effectively extracts slide plate crack fault characteristic.Qu Bofen separates matrix of coefficients for slide plate crackle image; Employing is distinguished crackle image and the intrinsic image of other slide plate based on the pantograph pan Identification of Cracks algorithm that bent wave zone moves parallel window; Then the various characteristics of image of slide plate are classified, handled, extract the crackle characteristic.
The objective of the invention is to realize through following means:
A kind ofly move the pantograph crack fault detection method of parallel window, gather the pantograph image, comprise following means through the camera system that train rail is along the line based on bent wave zone:
(1), pantograph IMAQ and image pre-service
CCD (Charge-coupled Device) industrial camera is at the image of the vertical shooting of pantograph workplace; Merge obtaining image, the pretreatment operation of enhancing, rim detection and slide plate image interception, obtain the slide plate image;
(2), the slide plate image of (1) institute intercepting is carried out bent wave conversion of two generations, be that the Qu Bofen of bent wave zone separates matrix of coefficients with the image transitions of spatial domain; Pseudo-fault signatures such as the scratch in the threshold process filtering image of overall coefficient, screw;
(3), use mobile parallel window method that (2) data are handled; The corresponding bent wave system number of the intrinsic pictorial element of on the filtering slide plate image except that crackle other; At last; Qu Bofen after handling is separated matrix of coefficients carry out inverse transformation, extract slide plate crackle image, realize the detection identification of pantograph pan crack fault.
Compared with prior art, adopt the beneficial effect of method of the present invention to be:
1, the pantograph crack fault detection method that the present invention adopted is based on image processing techniques, is to be detected object with the pantograph image, through the crack fault characteristic is extracted in the processing of collection in worksite picture, realizes the purpose that detects.
2, the present invention is directed to the characteristic that pantograph pan crackle image has the line singularity; In image processing techniques, used the bent wave conversion that the line singularity is had the good representation ability; Convert the picture of spatial domain into the matrix of coefficients of bent wave zone; Analysis through Qu Bofen being separated matrix of coefficients is classified to the various elements in the slide plate image with processing, extracts the difference of bent wave system number of crackle image and the bent wave system number of the intrinsic image of other slide plate, realizes the identification to crack fault.
3, the present invention adopts image fusion technology to remedy the dynamic fuzzy and the uneven defective of exposure of collection in worksite image, and the pantograph picture of taking the same video camera of adjacent moment merges, thereby obtains complete relatively picture clearly.
As stated, the method that the present invention adopts has been directed against the characteristics of pantograph pan fault line singularity, uses bent wave conversion effectively to catch the linear feature of crackle image, can accurately, effectively extract the crackle characteristic, realizes the detection of pantograph crack fault.
Description of drawings
Fig. 1 powder metallurgy slide plate synoptic diagram.
Fig. 2 is pantograph pan characteristics of image figure.
Fig. 3 moves parallel window schematic diagram.
Fig. 4 pantograph pan extracts the result.
Fig. 5 extracts figure as a result for the pantograph pan crackle.
Fig. 6 overexposure slide plate image.
Fig. 7 pantograph pan crackle extracts the result.
Fig. 8 pantograph pan figure.
Fig. 9 pantograph pan crackle extracts figure
Embodiment
Below in conjunction with accompanying drawing embodiment of the present invention is done further to detail.
Fig. 1 is image pretreatment process figure of the present invention.
A, the pre-service of pantograph image
A. pantograph image co-registration
For the different pantograph picture of two width of cloth that same sensor adjacent moment obtains, employing is carried out fusion treatment based on the Image Fusion of bent wave conversion of two generations to it, to obtain relative image than complete display.Based on the pantograph Image Fusion of bent wave conversion be at bent wave zone to the fusion that the Qu Bofen of image separates coefficient, adopt following fusion criterion:
(1) low frequency part adopts the weighted mean criterion, promptly
If
is the l yardstick; Position coordinates is (i on the k direction; J) low frequency coefficient, then:
In the formula, the low frequency coefficient in the different source images of
expression on the same position.
(2) HFS adopts big Laplce's energy and criterion, promptly
If I
L, k(i, j) expression l yardstick, on the k direction position coordinates be (then improved Laplce's energy (ML) with improved Laplce's energy and (SML branch) is not for i, high frequency coefficient j):
ML
l，k(i，j)＝|2I
l，k(i，j)-I
l，k(i-step，j)-I
l，k(i+step，j)|+|2I
l，k(i，j)-I
l，k(i，j-step)-I
l，k(i，j+step)|
(2)
In the formula, step representes the variable spacing between the coefficient of dissociation, gets step=1 among the present invention.
In the formula, P and Q represent to move the size (2P+1) * (2Q+1) of window.
Use I
1 L, k(i, j), I
2 L, k(i, j) and I
F L, k(i j) representes that respectively the Qu Bofen on source images and the fused images same position separates coefficient, SML
1 L, k(i, j) and SML
2 L, k(i j) representes I respectively
1 L, k(i, j) and I
2 L, k(so, fusion rule is for i, SML value j):
Image co-registration can remedy the fuzzy or uneven problem of making public of collection in worksite Image Dynamic to a certain extent, improves the utilization factor of captured picture.
B. pantograph figure image intensifying
Image enhancement technique is had living space, and the territory strengthens and frequency field strengthens two big types; Histogram-equalized image enhancement algorithms and the frequency field that the present invention adopts spatial domain respectively based on wavelet transformation and based on the algorithm for image enhancement of bent wave conversion to the pantograph original image strengthen, denoising, improve the sharpness and the contrast of image.
Basic step based on the image denoising of small echo (Qu Bo) conversion, enhancing is following:
(1) (x y) carries out small echo (Qu Bo) direct transform with image f;
(2) calculate the threshold value λ of small echo (Qu Bo) coefficient according to iconic model;
(3) at small echo (Qu Bo) transform domain coefficient of dissociation is carried out threshold process;
(4) carry out small echo (Qu Bo) inverse transformation, reconstructed image.
C. the pantograph pan Edge extraction is technological
The pantograph pan along continuous straight runs is fixed on (angular error is not more than 5 °) on the pantograph support, two slide plate equal in length, stationkeeping.Utilize this characteristic, can carry out rim detection, only keep the edge of horizontal direction, with the edge filtering of other direction the pantograph image.So just can from the pantograph image, clearly locate position, slide plate edge so that to the slide plate parts of images cut apart, intercepting.This algorithm picks has the better extract effect, Sobel operator and Canny operator that edge fogization is lighter carry out the slide plate edge extracting to the pantograph image respectively, only keeps the edge (being the slide plate horizontal boundary) of horizontal direction in the algorithm.
D. pantograph pan image interception algorithm
Rim detection has obtained the horizontal boundary of pantograph pan, according to the exact position of horizontal boundary in the pantograph image, can locate the concrete zone at pantograph pan place, carries out the intercepting of slide plate.The concrete practice is: at first the edge image that obtains is carried out 0～180 ° of Radon conversion in the scope, can find to have two maximum value as Radon territory horizontal ordinate θ during 90 ° of left and right sides; About the horizontal direction of the corresponding spatial domain of this direction, because alignment error, slide plate is not to be fixed on the pantograph support in strict accordance with horizontal direction; There is certain error; But error generally is not more than 5 °, and promptly the direction at slide plate edge maybe be between-5 °～5 °, so in θ=85 °～95 ° of scopes, edge image is carried out the Radon conversion once more; Find out the corresponding concrete angle of maximum value; Then edge image is carried out the Radon conversion of this angle, we can obtain the angle and the exact position of slide plate like this, and intercepting goes out the slide plate image.
B, pantograph pan crackle extractive technique
According to the difference of material, slide plate is divided into pure carbon slide plate, powder metallurgy slide plate and impregnating metal slide plate.The present invention is a detected object with the powder metallurgy slide plate.The powder metallurgy slide plate belongs to the Metal Substrate slide plate, can be divided into two kinds of copper base and iron-baseds according to the difference of agglomerated material, and iron-based is applicable to the steel aluminum conductor, and the copper base is applicable to copper conductor.The physical strength of powder metallurgy slide plate is high, good toughness, and anti-attrition property is good, and shock resistance is strong, and the phenomenon that fracture takes place is fewer, and resistance is very little, helps electric locomotive and is flowed.Powder metallurgy slide plate such as Fig. 1.
The pantograph pan characteristics of image is as shown in Figure 2: 1. on the powder metallurgy slide plate, there is parallel seams in metal slider; 2. the screw of fixed glides; 3. the long-term a large amount of cuts that use the back to exist; The crackle that 4. possibly exist.
The powder metallurgy slide plate is generally connected to form by the short draw runner that 5 block sizes equate.All there is a seam between per two draw runners, therefore, has 4 seams on the slide plate.These 4 is parallel on direction, is 3 π/4 directions, and the spacing distance of adjacent two seams equates.For fixed glides, screw has been installed on draw runner, the screw area is less and be circular.
For preventing the excessive wear of pantograph same position, contact net becomes " Z " font to arrange.Pantograph is in running status for a long time, and exists contact pressure between the osculatory, therefore can form scratch on the pantograph pan surface, and scratch generally is strip and distributes, direction basically identical in certain zone.
After wavelength-division was separated to Fig. 2 pantograph pan image march, all kinds of image coefficient characteristic distribution were as shown in table 1 on the slide plate:
Table 1 pantograph pan coefficients statistics characteristic
Can know that from table 1 Qu Bofen that scratch is corresponding separates coefficient sets and will be distributed in π/4～3 π/4 direction matrixes, screw is because it is shaped as circle, and Qu Bofen separates coefficient and is distributed in all angles matrix.Though it is wider that background, screw, cut coefficient distribute, coefficient scope and crackle and seam have than big difference, can classify through setting global threshold.Seam and crackle are comparatively similar on coefficient distributes, and 4 corresponding coefficients of seam are present in the same angle matrix, and also approximately equal of the interval of corresponding big coefficient region, need special classification processing.
Through pseudo-fault signatures such as the scratch in can the filtering image to the threshold process of overall coefficient, screw, remaining crackle characteristic and parallel seams characteristic.Like this, the fault signature that is included in the slide plate image just can classify as three types: crackle, parallel seams and other pseudo-fault signature.On direction, be all 3 π/4 according to parallel seams; The characteristic that spacing equates on the position; The present invention proposes based on the method that moves parallel window (Translational Parallel Window) the intrinsic parallel seams of slide plate image is handled, extract the crackle characteristic.
The size of supposing slide plate image to be detected is M * N, and this image march wavelength-division is separated, and the decomposition number of plies is n, and the n-1 layer is made up of l direction matrix, l=(l
1, l
2, K, l
l) represent l different decomposition direction respectively.If the l in the corresponding Qu Bofen dematrix of 3 π in the slide plate image/4 direction parallel seams
k(the individual direction matrix of 1≤k≤l), then this matrix can be expressed as C{n-1}{k}, and establishing this matrix size is M ' * N '.
The coefficients statistics characteristic of table 1 shows: the parallel seams direction is 3 π/4, and is more fixing, and coefficient value is bigger, is higher than background and other slide plate inherent feature coefficient values, shows as tangible line singularity characteristic.Therefore, after process Qu Bofen separated, { among the n-1}{k}, the parallel seams correspondence had the banded parallel zone of big coefficient value, and the width of belt-like zone and adjacent spaces are all about equally in Matrix C.Correspondingly, in the Qu Bofen dematrix, the vergence direction of the banded parallel zone of crack site also is approximately 3 π/4.The characteristic parallel with equal at interval according to the direction of belt-like zone utilized and moved parallel window thought, extracts the parallel seams characteristic in the slide plate.
Rectangular window as shown in Figure 3, that the parallelogram that structure is 3 π/4 by three directions is formed.The horizontal width of banded parallelogram window is (2p+1), and vertical height is the line number M ' of matrix, and the spacing of adjacent parallel quadrilateral window is d.The matrix element that dash area is corresponding in the rectangular window puts 1, and remainder puts 0.
In the formula (5), (2p+1) be the horizontal length of side of parallelogram, d is the tetragonal spacing of adjacent parallel.
Using this rectangular window that Qu Bofen is separated coefficient direction matrix handles:
C′{n-1}{k}＝A
ij·|C{n-1}{k}| (6)
In the formula (6), { n-1}{k} separates the coefficient absolute value for the matrix after filtering through window, the element value size of handling in back three banded parallel zones for the Qu Bofen of correspondence position to C ', and extra-regional element value then is 0.
Can know by above-mentioned seam signature analysis; { there are four parallel belt-like zones of concentrating big coefficient value among the n-1}{k} in Matrix C; Therefore; Can just can effectively distinguish the crackle that intrinsic parallel seams of slide plate and contingent cause according to these accurate positions that belt-like zone matrix element absolute value is bigger and characteristics that adjacent spacing is identical are oriented belt-like zone.
Parallel window is separated the beginning of angle matrix first row from Qu Bofen, with the step-length of step=1 successively to right translation, and to the rectangular window after each translation ask window energy with (SPW), will obtain an element and be the one-dimensional vector of each window energy with size.
Obviously, when rectangular window moves to three banded parallel windows that this window comprises when just coincideing with parallel seams, this energy value is the element maximum value of one-dimensional vector obviously greater than the energy value under other situation.Because there are 4 parallel seams in general powder metallurgy slide plate, so the higher maximum value of window energy can occur twice, obtains two extreme value SPW (j
1) and SPW (j
2).Like this,, just can locate the accurate position of parallel seams in the slide plate image fast, the slide plate parallel seams is carried out filtering according to detected maximum value position.
If there be the crackle parallel in the slide plate with the seam direction; Since exist simultaneously three parallel and the possibility crackle that spacing equates with the parallel seams spacing is very little; Even there is the crackle of 3 π/4 directions; General also have only one at most, and the SPW value of single crackle in rectangular window compare much for a short time with containing 3 SPW values under the parallel seams situation simultaneously, can not be mistaken for seam.Therefore, no matter whether direction of check is parallel with the seam direction, can both discern.
After image march wave system numbers such as the scratch on the slide plate, screw, parallel seams are handled,, just think that there is crack fault in slide plate if still there is the singularity characteristics of wire or approximate wire on the slide plate image.
Embodiment
Following the present invention combines instance to elaborate:
Pantograph front sled cut phenomenon is more serious, has a vertical crack in the middle of the slide plate of rear end, and it is as shown in Figure 4 that its slide plate extracts the result; The image size is 47 * 1508 * 3; Because there is crackle in slide plate itself, add reasons such as out of true is installed, cause the short draw runner on the slide plate that distortion and displacement have taken place; Include a part of background image in the feasible slide plate image that extracts, the last lower limb of slide plate is not in strict accordance with horizontal direction yet.It is as shown in Figure 5 that crackle extracts experimental result.Like Fig. 6, select two width of cloth to exist the picture of light exposure problems to carry out the checking of algorithm.Figure has a vertical crack in the middle of (a), has a vertical crack in the middle of the figure (b), has a crackle with the seam parallel direction in the middle of two seams in left side.Its crackle extract experimental result such as Fig. 7 (a) and (b) correspondence: if the existence crackle parallel on the slide plate with the seam direction; Owing to exist three the parallel and possibilities crackle that spacing equates with the parallel seams spacing not exist basically simultaneously; Even there is the crackle of 135 ° of directions; General also have only one at most, and the SPW value of single crackle in rectangular window compare much for a short time with containing 3 SPW values under the parallel seams situation simultaneously, can not be mistaken for seam.Therefore, no matter whether direction of check is parallel with the seam direction, can both discern.
Choose the pantograph pan image that a width of cloth contains with seam direction parallel crack and carry out the checking of algorithm, as shown in Figure 8, this slide plate image contains two crackles, and one is vertical direction, is positioned at the central authorities of slide plate, and another direction is identical with seam, is positioned at the slide plate left side.It is as shown in Figure 9 that crackle extracts the result.
No matter can know whether there be the crackle parallel with seam the slide plate from algorithm examples, the present invention can both accurately detect and locate crackle.
Claims (2)
1. one kind is moved the pantograph crack fault detection method of parallel window based on bent wave zone, gathers the pantograph image through the camera system that train rail is along the line, comprises following steps:
(1), pantograph IMAQ and image pre-service
Gather the CCD industrial camera at the vertical pantograph image of taking of pantograph workplace; Merge obtaining image, the pretreatment operation of enhancing, rim detection and slide plate image interception, obtain the slide plate image;
(2), the slide plate image of step (1) institute intercepting is carried out bent wave conversion of two generations, be that the Qu Bofen of bent wave zone separates matrix of coefficients with the image transitions of spatial domain; Scratch in the threshold process filtering image of overall coefficient, the pseudo-fault signature of screw;
(3), use mobile parallel window method that step (2) data are handled; The bent wave system number that the intrinsic pictorial element of on the filtering slide plate image except that crackle other is corresponding, that is: the rectangular window formed of the structure parallelogram that is 3 π/4 by three directions, the horizontal width of banded parallelogram window is (2p+1); Vertical height is the line number M ' of matrix; The spacing of adjacent parallel quadrilateral window is d, and the matrix element that dash area is corresponding in the rectangular window puts 1, and remainder puts 0;
In the formula (1), (2p+1) be the horizontal length of side of parallelogram, d is the tetragonal spacing of adjacent parallel;
Using this rectangular window that Qu Bofen is separated coefficient direction matrix handles:
C′{n-1}{k}＝A
ij·|C{n-1}{k}| （2）
In the formula (2), { n-1}{k} separates the coefficient absolute value for the matrix after filtering through window, the element value size of handling in back three banded parallel zones for the Qu Bofen of correspondence position to C ', and extra-regional element value then is 0; N is the decomposition number of plies that slide plate image march to be detected wavelength-division is separated;
According to these accurate positions that belt-like zone matrix element absolute value is bigger and characteristics that adjacent spacing is identical are oriented belt-like zone, the crackle that causes with intrinsic parallel seams of effective differentiation slide plate and contingent;
At last, the Qu Bofen after handling is separated matrix of coefficients carry out inverse transformation, extract slide plate crackle image, realize the detection identification of pantograph pan crack fault.
2. the pantograph crack fault detection method that moves parallel window based on bent wave zone according to claim 1; It is characterized in that; Said image co-registration is the different pantograph picture of two width of cloth that is obtained by same CCD camera adjacent moment, and employing is carried out fusion treatment based on the Image Fusion of bent wave conversion of two generations to it; Its fusion criterion: low frequency part adopts the weighted mean criterion, and HFS adopts big Laplce's energy and criterion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011101302503A CN102288614B (en) | 2011-05-19 | 2011-05-19 | Method for detecting pantograph crack fault based on curvelet domain moving parallel window |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011101302503A CN102288614B (en) | 2011-05-19 | 2011-05-19 | Method for detecting pantograph crack fault based on curvelet domain moving parallel window |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102288614A CN102288614A (en) | 2011-12-21 |
| CN102288614B true CN102288614B (en) | 2012-11-07 |
Family
ID=45335210
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011101302503A Expired - Fee Related CN102288614B (en) | 2011-05-19 | 2011-05-19 | Method for detecting pantograph crack fault based on curvelet domain moving parallel window |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102288614B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102749335B (en) * | 2012-05-31 | 2014-11-05 | 西南交通大学 | Insulator breakage fault detection method based on second-generation curvelet coefficient morphology band energy method |
| CN103310438B (en) * | 2013-04-16 | 2015-10-14 | 西南交通大学 | Based on the contact net rod insulator fault detection method of Harris angle point and image difference |
| CN103247044B (en) * | 2013-04-16 | 2015-08-19 | 西南交通大学 | Based on the defective mode detection method of high ferro contact net insulator curve-like and point-like singularity characteristics |
| CN107742288A (en) * | 2017-10-11 | 2018-02-27 | 成都国铁电气设备有限公司 | A kind of pantograph beats bow defect inspection method and system |
| CN107991318B (en) * | 2017-12-01 | 2024-01-05 | 西南交通大学 | High-precision detection device |
| CN109490244B (en) * | 2018-11-13 | 2021-05-14 | 华东理工大学 | Terahertz technology-based thermal barrier coating parallel crack monitoring method |
| CN109801248A (en) * | 2018-12-18 | 2019-05-24 | 重庆邮电大学 | One New Image fusion method based on non-lower sampling shear transformation |
| CN110695992A (en) * | 2019-09-27 | 2020-01-17 | 中国铁路昆明局集团有限公司昆明机务段 | Intelligent robot for overhauling roof of electric locomotive |
| CN113324864B (en) * | 2020-02-28 | 2022-09-20 | 南京理工大学 | Pantograph carbon slide plate abrasion detection method based on deep learning target detection |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5659013B2 (en) * | 2007-08-06 | 2015-01-28 | オーリゾン オペレイションズ リミテッド | System for monitoring pantograph damage and wear |
| CN100529702C (en) * | 2007-12-20 | 2009-08-19 | 成都主导科技有限责任公司 | On-line checkout equipment for contact pressure of motorcycle pantograph |
| CN101487873B (en) * | 2008-01-15 | 2011-04-06 | 中国神华能源股份有限公司 | Wireless video automatic monitoring system for pantograph of locomotive |
-
2011
- 2011-05-19 CN CN2011101302503A patent/CN102288614B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN102288614A (en) | 2011-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102288614B (en) | Method for detecting pantograph crack fault based on curvelet domain moving parallel window | |
| CN102129564B (en) | Contact network failure detection and diagnosis method based on unmanned aerial vehicle | |
| Banić et al. | Intelligent machine vision based railway infrastructure inspection and monitoring using UAV | |
| CN104567684B (en) | A kind of contact net geometric parameter detection method and device | |
| CN108759670B (en) | Contact line abrasion dynamic detection device based on non-contact detection technology | |
| CN107576666A (en) | A kind of double light spectrum image-forming rails and fastener method for detecting abnormality | |
| CN202793389U (en) | On-line routing inspection device for high-speed rail overhead line system | |
| CN103400379B (en) | The detection of a kind of icing based on image and thickness method of estimation | |
| CN105203552A (en) | 360-degree tread image detecting system and method | |
| CN106679567A (en) | Contact net and strut geometric parameter detecting measuring system based on binocular stereoscopic vision | |
| CN103076338B (en) | Rapid fuzzy matching method for testing bad conditions of high-speed rail catenary rod type insulators | |
| CN104318582A (en) | Detection method for bad state of rotating double-lug component pin of high-speed rail contact network on basis of image invariance target positioning | |
| CN105258727B (en) | Brake block backboard automatic detection device based on machine vision | |
| CN102175692A (en) | System and method for detecting defects of fabric gray cloth quickly | |
| CN109269416B (en) | Contact net lead abrasion measuring method and device | |
| CN109269415B (en) | Contact net lead abrasion measuring method and device based on multi-path camera | |
| CN103196372A (en) | Optical imaging detection method of electrified railway overhead line system supporting device | |
| CN106778439B (en) | A kind of lithium battery batch scan code system and method based on image procossing | |
| CN102682602A (en) | Collecting method for road traffic parameters based on video technique | |
| CN108830822B (en) | Pantograph carbon skateboard abrasion identification method based on improved Canny operator | |
| CN103010258A (en) | System and method for detecting cracks of fasteners of high-speed rails and subways | |
| CN105321173B (en) | A kind of train crossing cable cleat automatic defect detection method based on machine vision | |
| CN111561967A (en) | Real-time online detection method and system for pantograph-catenary operation state | |
| KR101527079B1 (en) | Area Catenary contact force measurement device with camera and measurement method | |
| CN108759671B (en) | Dynamic detection method for contact line abrasion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121107 Termination date: 20150519 |
|
| EXPY | Termination of patent right or utility model |