CN104608799B - Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss - Google Patents
Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss Download PDFInfo
- Publication number
- CN104608799B CN104608799B CN201410756876.9A CN201410756876A CN104608799B CN 104608799 B CN104608799 B CN 104608799B CN 201410756876 A CN201410756876 A CN 201410756876A CN 104608799 B CN104608799 B CN 104608799B
- Authority
- CN
- China
- Prior art keywords
- damage
- tread
- image
- wheel
- infrared
- 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
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/12—Measuring or surveying wheel-rims
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Pathology (AREA)
- Quality & Reliability (AREA)
- Theoretical Computer Science (AREA)
- Mechanical Engineering (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
Abstract
A kind of to be based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss, its step is:(1) image is obtained;(2) Image semantic classification;(3) damage Preliminary detection;(4) damage and be accurately positioned:Near-infrared image result carries out data fusion with gray level image result, to tread damage labelling and positioning, obtains the texture and area that damage through Morphological scale-space, the length that Morphological scale-space obtains crackle is carried out to tread crack;(5) judgement for damaging:Extract and select the feature in tread damage region, design and train a BP neural network, and defect is classified using BP neural network, Railway wheelset tread damage information is recorded by terminal processing units, can provide reference for the analysis of the ruuning situation of wheel, locomotive and circuit.The present invention has the advantages that detection comprehensive, high degree of automation, detection speed be fast, accuracy of detection is high.
Description
Technical field
The invention belongs to field of optical measuring technologies, more particularly to a kind of Railway wheelset tread damage on-line checking and identification
Method.
Background technology
Wheel is the important part out of shape of train, and the safe operation to train plays critical effect.And train wheel
Running environment it is extremely severe, except with rail, the shock of track switch and friction in addition to, may also suffer rainwater, greasy dirt, caustic etc.
Erosion, such wheel operation a period of time after often produce a certain degree of damage and disappearance, especially train is tight
During anxious braking, as the high temperature that friction is produced can make wheel tread scratch, stripping, rhegma occur and fall the phenomenons such as block, these systems
Referred to as tread damage, train tread damage are a parameters for being most difficult in Railway wheelset on-line checking measure.Common wheel pedal
Surface damage includes that flat sliding, shelled tread, tread grind heap, lose circle (local dent) etc..Due to train in the process of moving, take turns
Tread is contacted with rail, thus wheel tread damage location be difficult positioning, while train operationally rail vibration to survey
Amount brings very big error.
At present using static detection method more than the detection of country's tread damage:After train is put in storage using soon, hammer strike,
The method of sound is listened manually to be investigated, as the impact of anthropic factor, guilty culprit position, environmental condition etc. is difficult to discovery in time
Wheel tread failure.There is scientific research personnel to propose a kind of wheel tread defect on-line checking and identifying system, it utilizes machine vision
Or the method for photoelectric sensor, detecting system is fixedly mounted on the rail of train driving, when train enters measured zone,
Replaced human eye dynamic detection defect and judged by visual system, and defect is classified.These methods on the one hand can be with
The Financial cost and human cost of vehicle wheel maintenance are greatlyd save, on the other hand can also ensure reliability and the repetition for detecting simultaneously
Property.The technical scheme of this respect includes:Chinese invention patent application " know by a kind of tire tread defects based on image photographic form
Not " (number of patent application:201210475411.7) the utilization image photographic form tire tread defects scratch for proposing and stripping inspection
Survey method;Chinese invention patent application " the all-round tread surface defect dynamic on-line monitoring device of train wheel " (number of patent application:
201010575920.8) the utilization red LED light source for proposing and line array CCD collection image tire tread defects scratch and stripping inspection
Survey method.Applicant also once Chinese invention patent application " the contactless device for dynamically detecting of tire tread defects and
Its detection method " (number of patent application:201110361994.6) in propose a kind of to utilize laser displacement sensor and CCD shooting skills
Art flat sliding and stripping detection method.
But, scratch and stripping of the current detection scheme mainly for wheel tread, and for wheel tread grinds heap, loses circle
And online measuring technique and system report situations such as scratch and stripping mixing are very few;Simultaneously with generally target detection and
Recognition method is different, as wheel tread damages the high speed of on-line checking, high-precision requirement so that bullet train wheel is to stepping on
Surface damage detection and recognition methodss are faced with following subject matter:
(1) image disruption factor is more.At high speeds, detection means and system produce shifting such as the vibration of train
Dynamic or low-angle rotation, and the factor such as site environment, illumination variation and weather image may be introduced noise, fuzzy band and
Shadow region so that general detection method and technology are easily interfered.
(2) topical railway wheel problems include that flat sliding, shelled tread, tread grind heap, lose circle (local dent) etc., and wheel
Tread damage forms front tread and has crackle appearance, while tread is often disturbed by factors such as greasy dirt, dusts, these ginsengs
Several textural characteristics are not quite similar, therefore single detection means and method can not be efficiently identified.
(3) detection speed and required precision.As speed of the Railway wheelset tread damage on-line checking with recognition methodss is wanted
Ask, general Target detection and identification method can not meet online high speed and high-precision requirement simultaneously.
The content of the invention
The technical problem to be solved in the present invention is that:For the technical problem that prior art is present, the present invention provides one
Plant simple principle, high degree of automation, detection speed is fast, accuracy of detection is high, the Railway wheelset tread damage of multiparameter detection exists
Line is detected and recognition methodss.
To solve above-mentioned technical problem, the present invention is employed the following technical solutions:
A kind of to be based on many vision Railway wheelset tread damage on-line checkings and recognition methodss, its step is:
(1) image is obtained:Wheel tread gray level image is obtained using black and white camera, it is red near using near ir laser
Outer camera obtains wheel tread line image and line image is spliced into a width panoramic picture;
(2) Image semantic classification:Noise jamming removal, image binaryzation and image block are carried out including to image;
(3) damage Preliminary detection:Morphological scale-space is carried out to near-infrared image, pseudo- damage is rejected, labelling damage position is obtained
Obtain lesion depths and lose circle;
(4) damage and be accurately positioned:Near-infrared image result carries out data fusion with gray level image result, to stepping on
Surface damage labelling and positioning, obtain the texture and area that damage through Morphological scale-space, carry out Morphological scale-space to tread crack
Obtain the length of crackle;
(5) judgement for damaging:The feature of tread crack, Nian Dui, scratch and stripping area is extracted and selected, is designed and is trained
One BP neural network, and defect is classified using BP neural network, Railway wheelset is recorded by terminal processing units step on
Face scratch, stripping, mistake circle, stone roller heap, crackle equivalent damage information, not only contributing to technical staff is carried out to wheel when train is stopped
Targetedly recheck, moreover it is possible to which the ruuning situation analysis for wheel, locomotive and circuit provides reference.
Compared with prior art, it is an advantage of the current invention that:
(1) detection and recognition speed of the invention are fast, it is ensured that the rate request of Railway wheelset tread damage on-line checking.
The present invention is different from general detection method, and directly entire image is processed, but extracts wheel tread area using piecemeal
The algorithm in domain, so as to reach unless impact of the wheel tread region to testing result and the purpose of saving algrithm time.
(2) accuracy of detection of the invention is high.The complex shape of topical railway wheel problems, different types of defect gray feature is not
Equally, while Jing is commonly present the phenomenon of defect mixing, and tread surface is often subject to the false defects such as greasy dirt, dust, therefore not
General single, same type Target detection and identification is same as, therefore how effectively all defect on image to be detected simultaneously
Out, it is a difficult point.Near-infrared image testing result and black white image testing result are dexterously combined by the present invention, from
And detect the various damages of wheel tread well simultaneously and reject false defect, solve this problem.
(3) voting algorithm in information fusion is present invention employs so as to be further ensured that the accuracy of testing result, profit
With near-infrared image and the features and complementarity of black white image, processed by "or" and "AND", damaged according to wheel tread
The characteristics of hindering, Comprehensive Evaluation testing result, it is possible to achieve the comprehensive assessment of topical railway wheel problems.
Description of the drawings
Fig. 1 is the structural representation of Railway wheelset tread damage on-line checking and identifying device.
Fig. 2 is the schematic diagram of Railway wheelset tread damage on-line checking and identifying device in embodiment.
Fig. 3 is the overall procedure schematic diagram of the present invention.
Fig. 4 is Railway wheelset tread damage on-line checking and identifying device in-site installation schematic diagram in embodiment.
Fig. 5 is that embodiment Railway wheelset tread damage on-line checking gathers schematic diagram with identifying device near-infrared image.
Fig. 6 is the flow chart of black and white cell picture process in embodiment.
Fig. 7 is the flow chart of near-infrared cell picture process in embodiment.
Fig. 8 is information fusion wheel tread non-destructive tests algorithm flow chart in embodiment.
Specific embodiment
The present invention is described in further details below with reference to Figure of description and specific embodiment.
As depicted in figs. 1 and 2, the present invention includes terminal processing units 3, also including the orientation sensing for being arranged on rail side
Device 1, alignment sensor 1 are connected with terminal processing units 3, and terminal processing units 3 are imaged with the near-infrared for being arranged on rail both sides
Unit 2 is connected with black and white image unit 4, and near-infrared image unit 2 is made up of near infrared camera 22 and near ir laser 21.
When train is walked, rail is all arranged in pairs, and the rail side arranges alignment sensor 1 and refers to the steel being arranged in pairs
In rail, wherein a side of a rail is provided with alignment sensor 1, the rail both sides arrange 2 He of near-infrared image unit
Black and white image unit 4 refers to that the lateral surface of two rail 20 is equipped with near-infrared image unit 2, and medial surface is equipped with black and white shooting
Unit 4, for any of which bar rail, near-infrared image unit 2 and black and white image unit 4 are located at the side of rail 20.
The alignment sensor 1 is two current vortex sensors or two photoswitches, and the near-infrared image unit 2 is
Refer near infrared camera 22 and near ir laser 21.The travel speed of train is detected by alignment sensor 1, by near-infrared
Image unit 2 and black and white image unit 4 shoot the image that wheel tread is damaged, and transfer data to terminal processing units 3.
As shown in figure 3, the present invention based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss,
Its flow process is:
1. detection is started:When wheel 19 enters detection working region, alignment sensor 1 outputs signals to terminal processing units
3, trigger is provided by terminal processing units 3, start near-infrared image unit 2 and black and white image unit 4 is started working and gathered
19 tread image of wheel.
2. image is obtained:Near-infrared image unit 2 obtains wheel tread line image, and black and white image unit 4 obtains wheel pedal
Face circumference gray level image.
3. Image semantic classification:Due to the interference of the factors such as the live natural light of detection, ambient light, draw to the image for collecting
Enter noise, the present invention filters the noise in image using a kind of median filtering algorithm based on template operation.Using based on template
The interference of 2 the gathered noise in image of median filter method removal step of operation.
4. Morphological scale-space:4 shooting image of black and white image unit is sent to terminal processing units 3, Jing image procossings and stricture of vagina
Reason analysis judges crackle and defect area size in tread damage.2 shooting image Jing light thinning of near-infrared image unit is calculated
Method process, by N main feed lines image mosaic into a width tread circumference panoramic picture, is combined algorithm based on small echo and Optimal-threshold segmentation
Extract and damage target area, the false defects such as dust, greasy dirt are rejected according to textural characteristics, with not damaged tread circumference image or fortune
Calculate, calculate the size for obtaining defect.
5. tread damage classification:Nearly 2 shooting image result of infrared photography unit and 4 shooting figure of black and white image unit
As result carries out information fusion, according to characteristics of image recognizer, the faces such as the line features such as crackle, stripping and scratch are distinguished
Shape feature, classifies to tread damage.
6. terminal processing units 3 transmit signals to locating alarm device 5 according to classification situation, and locating alarm device record has
The wheel of damage, and send alarm signal.
The near-infrared image unit 2 and black and white image unit 4 of 20 both sides of rail, the tread for recording two sidecar wheels 19 respectively are damaged
Wound.Alignment sensor 1, near-infrared image unit 2 and black and white image unit 4 can be bonded on rail 20, it is also possible to by card
Block is fixed on the outside of rail.
Limited by near-infrared image unit 2 and 4 working range of black and white image unit, two neighboring shooting in the present embodiment
The distance of system is the 1/4 of wheel circumference, and each camera system is made up of near-infrared image unit 2 and black and white image unit 4, point
Not An Zhuan rail 20 both sides.As shown in figure 4, the circumference of wheel 19 is divided into 3 parts, respectively AB, BC and CA.As selected
Select near-infrared image unit 2 and black and white image unit 4 coverage it is wider in the case of, can by between camera system away from
From be set to wheel circumference 1/3 or 1/2 or equal to whole girth with a distance from, to meet the shooting car that camera system can be complete
Wheel tread circumference is defined.
As shown in Figure 4 and Figure 5, when wheel 19 is crossed, the near ir laser in near-infrared image unit 2 is oblique
The formation tread contour line on 19 tread of wheel is penetrated, with the movement of wheel 19, near ir laser oblique fire is stepped in wheel 19
Contour line on face forms a series of scan lines on tread and crosses up to wheel 19, and near infrared camera records laser line generator in car
Take turns a series of contour lines for being formed on 19 treads and be transferred to terminal processing units 3, Jing algorithm process is spliced into a width wheel to 19
Tread panorama sketch.Black and white image unit is recorded 19 tread image of wheel and is transferred to terminal processing units 3.
As shown in fig. 6, black and white image unit 4 shoots 19 tread image processing flow of wheel being:
After the gray level image of shooting is transferred to terminal processing units 3, piecemeal is carried out to image first.Due to 19 tread of wheel
High light reflectivity can be formed with Outboard Sections, image both sides and the mid portion of acquisition form larger gray scale difference, therefore using figure
As tread region and exterior lateral area are separated by the algorithm of piecemeal, binary conversion treatment is carried out using different threshold values respectively.Then enter
Row Morphological scale-space extracts target area, identifies crackle, scratch, grinds the treads such as heap damage according to textural characteristics and line feature
Wound.
As shown in fig. 7, near-infrared image unit 2 shoots 19 tread image processing flow of wheel being:
Binary conversion treatment is carried out near infrared camera shooting image, then using thinning algorithm Extracting contour centrage,
By the N width image mosaic after process into a width wheel tread panoramic picture.Hinder the position that image comparison searches damage with standard non-destructive
Put, detect that wheel loses circle degree to 19 treads, Morphological scale-space is carried out to damage field, false defect is rejected according to textural characteristics.
Maximum difference between two pixel of statistical picture column direction, (Δ h), as max, (Δ h) exceedes to be calculated depth capacity max of damage
During 0.2mm, it is judged to tread damage.
As shown in figure 8, information fusion wheel tread non-destructive tests algorithm flow is:
Nearly infrared photography unit 2 and 4 result of black and white image unit carry out information fusion by voting method.It is nearly red
" having " damage field in outer image unit testing result is done with black and white image unit result (regardless of whether having damage)
Inclusive-OR operation;"None" damage field nearly in infrared photography unit testing result and black and white image unit result (nothing again
By whether having damage) AND operation is done, fusion results carry out Morphological scale-space respectively twice, extract tread damage feature, design
And a BP neural network is trained, and defect is classified using BP neural network, to the crackle on 19 tread of wheel, lose
Circle, stone roller heap, scratch and stripping damage field carry out fixation and recognition, and are calculated the attrition value of correlation.Terminal processing units 3
Locating alarm device 5 is transmitted signals to according to classification situation and respective attrition value, locating alarm device record has the wheel of damage, and
Send alarm signal.
Claims (3)
1. a kind of to be based on many vision Railway wheelset tread damage on-line checkings and recognition methodss, its step is:
(1) image is obtained:Wheel tread gray level image is obtained using black and white camera, using near ir laser and near-infrared phase
Machine obtains wheel tread line image and line image is spliced into a width panoramic picture;
(2) Image semantic classification:Noise jamming removal, image binaryzation and image block are carried out including to image;
(3) damage Preliminary detection:Morphological scale-space is carried out to near-infrared image, pseudo- damage is rejected, labelling damage position is damaged
Hinder depth and lose circle;
(4) damage and be accurately positioned:Near-infrared image result carries out data fusion with gray level image result, and tread is damaged
Hinder labelling and positioning, the texture and area that damage are obtained through Morphological scale-space, Morphological scale-space acquisition is carried out to tread crack
The length of crackle;
(5) judgement for damaging:The feature of tread crack, Nian Dui, scratch and stripping area is extracted and selected, one is designed and train
BP neural network, and defect is classified using BP neural network, Railway wheelset tread is recorded by terminal processing units wipe
Wound, stripping, mistake circle, stone roller heap, crackle equivalent damage information, terminal processing units transmit signals to locating alarming dress according to classification situation
Put, locating alarm device record has the wheel of damage, and sends alarm signal.
2. according to claim 1 based on many vision Railway wheelset tread damage on-line checkings and recognition methodss, its feature
It is the concrete steps of step (3):
Binary conversion treatment is carried out near infrared camera shooting image, then using thinning algorithm Extracting contour centrage, will place
N width image mosaic after reason hinders position that image comparison search damage, inspection with standard non-destructive into a width wheel tread panoramic picture
Measure wheel tread and lose circle degree, Morphological scale-space is carried out to damage field, false defect, statistical picture is rejected according to textural characteristics
Maximum difference between two pixel of column direction, be calculated damage depth capacity max (Δ h), when max (when Δ h) is more than 0.2mm,
It is judged to tread damage.
3. according to claim 1 based on many vision Railway wheelset tread damage on-line checkings and recognition methodss, its feature
It is the concrete steps of step (4):
Nearly infrared photography unit and black and white image unit result carry out information fusion, nearly infrared photography by voting method
" having " damage field in unit testing result does inclusive-OR operation with black and white image unit result;Nearly infrared photography list again
"None" damage field in first testing result does AND operation with black and white image unit result, and fusion results are entered respectively twice
Row Morphological scale-space, extracts tread damage feature.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410756876.9A CN104608799B (en) | 2014-12-12 | 2014-12-12 | Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410756876.9A CN104608799B (en) | 2014-12-12 | 2014-12-12 | Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104608799A CN104608799A (en) | 2015-05-13 |
CN104608799B true CN104608799B (en) | 2017-03-29 |
Family
ID=53143516
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410756876.9A Expired - Fee Related CN104608799B (en) | 2014-12-12 | 2014-12-12 | Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104608799B (en) |
Families Citing this family (46)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105203552A (en) * | 2015-09-18 | 2015-12-30 | 南京拓控信息科技有限公司 | 360-degree tread image detecting system and method |
CN105139413A (en) * | 2015-09-25 | 2015-12-09 | 深圳大学 | Railway connecting piece quality detection method based on image characteristic extraction and detection system |
CN105301016B (en) * | 2015-10-13 | 2018-12-21 | 东莞市诺丽电子科技有限公司 | Wheel tread flat image detecting method |
CN105374038B (en) * | 2015-11-11 | 2016-09-07 | 石家庄铁道大学 | Vehicular double camera wheel drive track displacement image detecting system and method |
CN105501250B (en) * | 2015-11-25 | 2017-09-05 | 广西大学 | Railway wheelset flat sliding method for diagnosing faults based on vehicle-mounted detection apparatus |
CN105548196B (en) * | 2015-12-07 | 2018-12-04 | 郑州轻工业学院 | The method and apparatus of the online non-destructive testing of hard alloy top hammer |
CN106683076B (en) * | 2016-11-24 | 2019-08-09 | 南京航空航天大学 | The method of train wheel tread damage detection based on textural characteristics cluster |
CN106650831B (en) * | 2017-01-10 | 2020-04-28 | 北京邮电大学 | Damage detection method and device based on feature fusion |
CN106940318B (en) * | 2017-02-23 | 2019-04-09 | 大连理工大学 | A kind of fibre reinforced composites processing effect evaluation method |
CN107024480A (en) * | 2017-04-12 | 2017-08-08 | 浙江硕和机器人科技有限公司 | A kind of stereoscopic image acquisition device |
US10445871B2 (en) * | 2017-05-22 | 2019-10-15 | General Electric Company | Image analysis neural network systems |
CN107958452B (en) * | 2017-11-20 | 2021-08-13 | 中车株洲电力机车有限公司 | Method, device and system for monitoring running attitude of vehicle wheel set |
DE102017222964A1 (en) * | 2017-12-15 | 2019-06-19 | Zf Friedrichshafen Ag | Method and system for damage detection of a component |
CN108520274B (en) * | 2018-03-27 | 2022-03-11 | 天津大学 | High-reflectivity surface defect detection method based on image processing and neural network classification |
CN108876802B (en) * | 2018-04-20 | 2022-04-15 | 北京交通大学 | Device and method for automatically positioning fluorescent magnetic powder defects of wheel |
CN108871431A (en) * | 2018-05-21 | 2018-11-23 | 北京京天威科技发展有限公司 | Wheel is to dynamic detection system and method |
CN108921845A (en) * | 2018-07-10 | 2018-11-30 | 深圳大学 | Wheel tread rehabilitation plan generation method and device |
CN108961252B (en) * | 2018-07-27 | 2021-06-08 | Oppo(重庆)智能科技有限公司 | Method for detecting logo paste failure, electronic device and computer readable storage medium |
CN109060828A (en) * | 2018-08-23 | 2018-12-21 | 南京拓控信息科技股份有限公司 | A kind of locomotive wheel thread defect image detecting system |
CN109685124A (en) * | 2018-12-14 | 2019-04-26 | 斑马网络技术有限公司 | Road disease recognition methods neural network based and device |
CN109591846A (en) * | 2018-12-20 | 2019-04-09 | 成都唐源电气股份有限公司 | A kind of wheel tread online test method |
CN111353515B (en) * | 2018-12-21 | 2024-01-26 | 湖南工业大学 | Multi-scale classification-based train wheel set tread damage classification and identification method |
CN109978821A (en) * | 2019-02-03 | 2019-07-05 | 湖南工业大学 | A kind of wheel wear monitoring method based on image recognition |
CN110223266A (en) * | 2019-03-08 | 2019-09-10 | 湖南工业大学 | A kind of Railway wheelset tread damage method for diagnosing faults based on depth convolutional neural networks |
CN110261139A (en) * | 2019-06-12 | 2019-09-20 | 中国神华能源股份有限公司 | Wheel tread flat recognition methods and identification device |
CN110254468B (en) * | 2019-06-20 | 2021-12-03 | 吉林大学 | Intelligent online detection device and detection method for track surface defects |
CN110217266A (en) * | 2019-06-21 | 2019-09-10 | 中国神华能源股份有限公司 | Anomalous wheel identifying system and method under train braking state |
CN110276297B (en) * | 2019-06-21 | 2021-08-06 | 中国神华能源股份有限公司 | Train tread wear detection system and method |
CN110567966A (en) * | 2019-07-31 | 2019-12-13 | 众宏(上海)自动化股份有限公司 | Gear detection method based on line laser scanner |
CN110816588B (en) * | 2019-12-04 | 2021-02-05 | 中车株洲电力机车有限公司 | Locomotive wheel out-of-roundness detection method, device, equipment and system |
CN111023998A (en) * | 2019-12-11 | 2020-04-17 | 神华铁路货车运输有限责任公司 | Tread damage measuring device, method, apparatus and storage medium |
CN111429422A (en) * | 2020-03-19 | 2020-07-17 | 中国工程物理研究院激光聚变研究中心 | Laser near-field state analysis method and device based on deep learning |
CN111524112B (en) * | 2020-04-17 | 2023-04-07 | 中冶赛迪信息技术(重庆)有限公司 | Steel chasing identification method, system, equipment and medium |
CN112129777A (en) * | 2020-10-19 | 2020-12-25 | 北京市地铁运营有限公司 | Rail train maintains detection device based on computer vision |
CN112785547A (en) * | 2020-11-11 | 2021-05-11 | 平高集团有限公司 | Electrode contact surface damage identification method and system |
CN112508883A (en) * | 2020-11-24 | 2021-03-16 | 南京航空航天大学 | Self-adaptive defect identification method for train wheel tread taper hole detection |
CN112712552A (en) * | 2020-12-29 | 2021-04-27 | 哈尔滨市科佳通用机电股份有限公司 | Fault detection method for vehicle tread scratch |
CN113418926B (en) * | 2021-06-02 | 2022-02-18 | 南京航空航天大学 | Wheel set tread damage detection method based on laser strip center line positioning extraction algorithm |
CN113450530A (en) * | 2021-06-03 | 2021-09-28 | 河北华电石家庄热电有限公司 | Safety early warning system based on intelligent video analysis algorithm |
CN114572273B (en) * | 2022-03-15 | 2024-04-30 | 南京拓控信息科技股份有限公司 | Railway vehicle wheel set tread 3D image detection method |
CN114749342B (en) * | 2022-04-20 | 2023-09-26 | 华南理工大学 | Lithium battery pole piece coating defect identification method, device and medium |
DE102022204607A1 (en) | 2022-05-11 | 2023-07-13 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Apparatus, monitoring system and method for monitoring a rail vehicle for thermal damage |
CN115879343B (en) * | 2022-12-14 | 2023-08-01 | 西南交通大学 | Ballastless track self-compacting concrete void two-stage identification method |
CN116109638B (en) * | 2023-04-13 | 2023-07-04 | 中铁四局集团有限公司 | Rail break detection method and system |
CN117197538B (en) * | 2023-08-16 | 2024-05-14 | 哈尔滨工业大学 | Bayesian convolution neural network structure apparent damage identification method based on Gaussian distribution weight sampling |
CN117173183B (en) * | 2023-11-03 | 2024-01-26 | 山西阳光三极科技股份有限公司 | Railway wagon fault detection method based on artificial intelligence |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4932784A (en) * | 1986-10-13 | 1990-06-12 | Caltronic A/S | Apparatus for track-based detection of the wheel profile of passing railway wheels |
CN1065431A (en) * | 1992-04-06 | 1992-10-21 | 西南交通大学 | The wheel tread for railway rolling stock damage check |
CN102060036A (en) * | 2010-12-14 | 2011-05-18 | 成都主导科技有限责任公司 | System for detecting rapid train wheel pair dynamically |
CN102431576A (en) * | 2011-10-13 | 2012-05-02 | 成都主导科技有限责任公司 | Fault dynamic detecting and data processing method and system of wheel set |
CN102501887A (en) * | 2011-11-16 | 2012-06-20 | 郑州轻工业学院 | Non-contact dynamic detection device and detection method for tire tread defects |
-
2014
- 2014-12-12 CN CN201410756876.9A patent/CN104608799B/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4932784A (en) * | 1986-10-13 | 1990-06-12 | Caltronic A/S | Apparatus for track-based detection of the wheel profile of passing railway wheels |
CN1065431A (en) * | 1992-04-06 | 1992-10-21 | 西南交通大学 | The wheel tread for railway rolling stock damage check |
CN102060036A (en) * | 2010-12-14 | 2011-05-18 | 成都主导科技有限责任公司 | System for detecting rapid train wheel pair dynamically |
CN102431576A (en) * | 2011-10-13 | 2012-05-02 | 成都主导科技有限责任公司 | Fault dynamic detecting and data processing method and system of wheel set |
WO2013053140A1 (en) * | 2011-10-13 | 2013-04-18 | 成都主导科技有限责任公司 | Data processing method and system for dynamic wheelset failure detection |
CN102501887A (en) * | 2011-11-16 | 2012-06-20 | 郑州轻工业学院 | Non-contact dynamic detection device and detection method for tire tread defects |
Also Published As
Publication number | Publication date |
---|---|
CN104608799A (en) | 2015-05-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104608799B (en) | Based on information fusion technology Railway wheelset tread damage on-line checking and recognition methodss | |
Liu et al. | A review of applications of visual inspection technology based on image processing in the railway industry | |
AU2018224073B2 (en) | Broken wheel detection system | |
CN100449259C (en) | On-line detection method and device for vehicle wheel set diameter | |
CN100449264C (en) | On-line detection method and device for thread defect of vehicle wheel set | |
CN102788803B (en) | On-line high-speed detection system of train wheel set tread and wheel rim and detection method thereof | |
CN109230351A (en) | A kind of recognition methods that belt conveyor is operating abnormally | |
CN103322936A (en) | Wheel set size on-line monitoring method | |
US20140232852A1 (en) | Optical device and method for inspecting tires | |
CN201233223Y (en) | Lumber automatic monitor system | |
CN102060037A (en) | Online dynamic detection device of all-periphery contour dimension of locomotive wheel pair | |
KR101602376B1 (en) | A train faulty monitoring system | |
CN106192634A (en) | A kind of railroad track elastic bar fastener condition automatic detection device and method | |
CN107576666A (en) | A kind of double light spectrum image-forming rails and fastener method for detecting abnormality | |
CN110047070B (en) | Method and system for identifying rail wear degree | |
CN102608126A (en) | On-line detection method and device for surface defects of high-temperature continuously cast bloom | |
CN1843822A (en) | Dynamic detection system of train wheel pair and detection method thereof | |
CN102854191A (en) | Real-time visual detection and identification method for high speed rail surface defect | |
CN106274977A (en) | The direct current Magnetic Flux Leakage Testing System of a kind of trigger-type drainage pattern and method thereof | |
CN103837087B (en) | Pantograph automatic testing method based on active shape model | |
CN110862033A (en) | Intelligent early warning detection method applied to coal mine inclined shaft winch | |
CN103307982A (en) | On-line monitoring device for dimensions of wheel sets | |
CN111295564A (en) | Optical wheel evaluation | |
CN202548069U (en) | On-line detection device for surface defects of high temperature continuous casting billet | |
CN104280397A (en) | Eddy current-video comprehensive detection evaluation method of in-service steel rail tread scale cracks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170329 Termination date: 20171212 |
|
CF01 | Termination of patent right due to non-payment of annual fee |