CN104200464B - A kind of detection method and system for identifying that train is abnormal - Google Patents
A kind of detection method and system for identifying that train is abnormal Download PDFInfo
- Publication number
- CN104200464B CN104200464B CN201410406535.9A CN201410406535A CN104200464B CN 104200464 B CN104200464 B CN 104200464B CN 201410406535 A CN201410406535 A CN 201410406535A CN 104200464 B CN104200464 B CN 104200464B
- Authority
- CN
- China
- Prior art keywords
- train
- depth information
- depth
- full car
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention discloses a kind of detection method and system for identifying that train is abnormal, belong to the technical field of image automatic identification.Recognition methods comprises the following steps:Gather the full car depth information of train and reference train to be detected, the full car depth information of train to be detected and reference train is matched to identify faulty components by vehicle license number, then matches the depth informations of faulty components with the depth information with reference to part to determine failed part.Identifying system includes:Gather the acquisition module of the full car depth information of train to be detected, reference train data obtaining module and failed part detection module.It is this to utilize method of the depth information to crossing the carry out three-dimensional image information collection of car, overcome the limitation of conventional two-dimensional image information visual performance, so as to facilitate testing staff to be quickly found out actual fault point, system rate of false alarm is reduced, improves the overhaul efficiency of testing staff.
Description
Technical field
The invention discloses a kind of detection method and system for identifying that train is abnormal, the technology for belonging to image automatic identification is led
Domain.
Background technology
The full car parts abnormality detection generally use of traditional train is manually detected, and this mode needs train to enter the station
Or storage when detected, due to train (including lorry, car, the EMUs and other types of track train) dwell time compared with
Short, along with the parts on train are numerous and complicated, maintainer is difficult to pay close attention to each zero in a short period of time
The normal condition of part.Especially, exist between train bottom and the parts at top and block and vision dead zone be present, this both enters one
The abnormal false dismissal probability of step increase Train Parts, also further reduces abnormality detection efficiency and the degree of accuracy.
In addition, the train most of the time is in running status, track train is in the process of running because of debris impact, relative fortune
Dynamic, impact, vibrations etc. can cause the wearing of parts, abnormal the problems such as releasing, and after train enters the station or complete operation milimeter number
Just overhauled after storage, this, which can cause train bottom parts exception occur, to be found and overhauled in time, most give at last
The operation of track traffic causes great potential safety hazard.
The content of the invention
The technical problems to be solved by the invention are the deficiencies for above-mentioned background technology, there is provided one kind identification train is different
Normal detection method and system.
The present invention adopts the following technical scheme that for achieving the above object:
A kind of detection method for identifying that train is abnormal, comprises the following steps:
Full the car depth information and vehicle vehicle number information of train to be detected are gathered,
Match to obtain reference train by the vehicle vehicle number information of the train to be detected, and obtain the complete of the reference train
Car depth information,
Reference train and the geometric element of train different depth hierarchical data to be detected are extracted respectively, are intended in units of component
The geometric element of each depth hierarchical data is closed, obtains the full car geometric properties standard profile letter that train to be detected contains depth information
Breath figure and reference train contain the full car geometric properties distributed intelligence figure of depth information, by matching rain model car to be detected
Number information is with reference train vehicle vehicle number information to contrast full the car geometric properties distributed intelligence figure and reference columns of train to be detected
Module sets corresponding to the unmatched hum pattern of depth information are failure by the full car geometric properties standard profile hum pattern of car
Component, and,
Extract and handle the geometric element of train different depth hierarchical data to obtain the full car depth in units of part
Information, the depth information of faulty components is matched with the depth information with reference to part, enters one from the depth information of faulty components
Step extracts abnormal depth information, and the part further extracted corresponding to obtained abnormal depth information is set as into failed part;
Full car depth information described in above-mentioned steps is train bottom depth information, train sidepiece depth information and train top
At least one of portion's depth information.
The further prioritization scheme of the detection method abnormal as the identification train, the full car depth information of reference train
For:
The full car depth information of the normal train pre-set,
In the full car depth information of the same train passed through with the current sample time closest moment,
In multigroup full car depth information of the same train passed through with current sample time adjacent to the moment, and
Above-mentioned full car depth information is merged or statistical analysis after at least one of obtained full car depth information.
The further prioritization scheme of the detection method abnormal as the identification train, it is right for reference train with reference to part
At least one of corresponding part in the part and default parts library answered.
The further prioritization scheme of the detection method abnormal as the identification train, extracts and handles train different depth
The geometric element of hierarchical data is to obtain the full car depth information in units of part, by the depth information of faulty components and reference
The depth information matching of part, abnormal depth information is further extracted from the depth information of faulty components, will further be extracted
In the step of obtained part corresponding to abnormal depth information is set as failed part, failure zero is set as follows
Part:
The geometric element of the faulty components different depth hierarchical data is extracted, each depth level is fitted in units of part
The geometric element of data, obtain the faulty components geometric properties distributed intelligence figure containing depth information;
Geometric element of the reference train in different depth hierarchical data is extracted, is fitted in units of the reference part
The geometric element of each depth hierarchical data, obtains matching with the faulty components and the reference component geometry containing depth information
Characteristic standard distributed intelligence figure;
The faulty components geometric properties distributed intelligence figure and the reference component geometric properties standard profile hum pattern ratio
To rear, the part corresponding to the unmatched hum pattern of depth information is set as failed part.
Further, the acquisition method of the full car depth information of train to be detected be laser triangulation, binocular telemetry and
At least one of pulse ranging method.
Further, identify that the abnormal detection method of train is additionally included in and will further extract obtained abnormal depth information
After the step of corresponding part is set as failed part, the step of to failed part alarm indication.
A kind of detecting system for identifying that train is abnormal, including:
Acquisition module, for gathering the full car depth information of train to be detected and the vehicle vehicle number information of train to be detected;
Reference train data obtaining module, match to obtain reference train by the vehicle vehicle number information of the train to be detected,
And obtain the full car depth information of the reference train;And
Failed part detection module, the geometry member of reference train and train different depth hierarchical data to be detected is extracted respectively
Element, the geometric element of each depth hierarchical data is fitted in units of component, obtains the full car that train to be detected contains depth information
Geometric properties standard profile hum pattern and reference train contain the full car geometric properties distributed intelligence figure of depth information, by
With rain model vehicle number information to be detected and reference train vehicle vehicle number information to contrast the full car geometric properties of train to be detected
The full car geometric properties standard profile hum pattern of distributed intelligence figure and reference train, the unmatched hum pattern institute of depth information is right
The module sets answered are faulty components, extract and handle the geometric element of train different depth hierarchical data with obtain using part as
The full car depth information of unit, the depth information of faulty components is matched with the depth information with reference to part, from faulty components
Abnormal depth information is further extracted in depth information, the part further extracted corresponding to obtained abnormal depth information is set
It is set to failed part.
Further, the abnormal detecting system of above-mentioned identification train also includes the report for failed part described in alarm indication
Alert display module.
The full car depth information for some groups of trains to be detected that the present invention is gathered by multi-angle, and by the row to be detected
The vehicle vehicle number information of car matches to obtain reference train, and obtains the full car depth information of the reference train, and process is to be detected
The matching of the full car depth information of full the car depth information and reference train of train, so as to detect failed part.This utilization
Method of the depth information to crossing the carry out three-dimensional image information collection of car, can be in real time to important on the full car of track train
Part and/or part carry out abnormality identification, so as to change traditional work pattern, drastically increase China Express Railway
Monitoring capability safe for operation.In addition, the method for this three-dimensional image information collection, additionally it is possible to overcome conventional two-dimensional image information to regard
Feel the limitation of function, the alarm of some non-faulting points such as water stain dust can be shielded, so as to facilitate testing staff to be quickly found out reality
Border trouble point, system rate of false alarm is reduced, and improve the overhaul efficiency of testing staff.
Brief description of the drawings
Fig. 1 is the flow chart of the abnormal detection method of the first identification train;
Fig. 2 is the flow chart of the abnormal detection method of second of identification train.
Embodiment
The technical scheme of invention is described in detail below in conjunction with the accompanying drawings.
The flow chart of the present invention is as shown in Figure 1.Here arrange for two example two to illustrate the present invention.
Specific embodiment one:
Step 1, the full car depth information of train to be detected is gathered:To carrying out real-time detection, control system by train speed
System adjusts IMAQ pulse mode according to real-time speed, embedded timing generator circuit solution in modularization binocular acquisition subsystem
Acquisition pulse is analysed, and timing synchronization is carried out to the collection sequential of each camera in module, it is ensured that three-dimensional data is in time and sky
Between on meet image co-registration specification.The acquisition method of the full car depth information of above-mentioned train to be detected can be pulse ranging method or
Laser triangulation, but it is also not necessarily limited to the above method.
Step 2, vehicle license number index is established:The vehicle and license number of each car are obtained by AEI car number identification systems, is used
In matching reference train, the full car depth information of reference train is obtained.
Wherein, the full car depth information of above-mentioned reference train can be the full car depth letter of the normal train pre-set
Breath, such as the full car depth information determined with a train when using first;In addition, the full car depth information of reference train is also
Can be the full car depth information in the same train passed through with the current sample time closest moment, or with currently adopting
The multigroup full car depth information for the same train that the sample moment passes through adjacent to the moment, or above-mentioned full car depth information is melted
The full car depth information obtained after conjunction or statistical analysis.
Step 3, the full car depth information of train to be detected is matched with the full car depth information of reference train, detects failure
Part:I.e. by the way that the full car depth information of train to be detected is matched with the full car depth information of reference train, from full car depth
Abnormal depth information is extracted in information, part corresponding to the abnormal depth information of extraction is defined as failed part.Wherein, it is above-mentioned
Part with reference to corresponding to part can be reference train, can also be part corresponding in default parts library.
In order to improve the operation speed that the full car depth information of train to be detected matches with the full car depth information of reference train
Above-mentioned steps can also be refined as following steps by degree:
Step A, the full car depth information of train to be detected is matched with the full car depth information of reference train, it is deep from full car
Abnormal depth information is extracted in degree information, it (is that component corresponding to the abnormal depth information of extraction is defined as into faulty components
One-level positioning parts result):
Step A-1, pre-processed by cluster, denoising etc., extract the profile of reference train different depth hierarchical data, used
The geometric elements such as line, rectangle, the geometric element of each depth hierarchical data is fitted in units of component, is obtained containing depth information
Full car geometric properties standard profile hum pattern;
Step A-2, the geometric element of train different depth hierarchical data to be detected is extracted, each depth is fitted in units of component
The geometric element of hierarchical data is spent, obtains the full car geometric properties distributed intelligence figure containing depth information;
Step A-3, by the matching of rain model vehicle number information to be detected and reference train vehicle vehicle number information, it will contain
The full car geometric properties distributed intelligence figure of depth information closes with the full car geometric properties standard profile hum pattern containing depth information
Connection gets up;
Step A-4, match the associated full car geometric properties standard profile hum pattern containing depth information, contain depth
The full car geometric properties distributed intelligence figure of information, is defined as failure group by component corresponding to the unmatched hum pattern of depth information
Part.
Step B, the depth information of faulty components is matched with the depth information with reference to part, believed from the depth of faulty components
Abnormal depth information is further extracted in breath, the part further extracted corresponding to obtained abnormal depth information is defined as therefore
Hinder part (being second level positioning parts result):
Step B-1, the geometric element of faulty components different depth hierarchical data is extracted, each depth is fitted in units of part
The geometric element of hierarchical data, obtain the trouble unit geometric properties distributed intelligence figure containing depth information;
Step B-2, geometric element of the reference train in different depth hierarchical data is extracted, be fitted in units of reference part
The geometric element of each depth hierarchical data, obtains matching with faulty components and the reference component geometric properties containing depth information
Standard profile hum pattern;
Step B-3, matching faulty components geometric properties distributed intelligence figure, reference component geometric properties standard profile information
Figure, is defined as failed part by the part corresponding to the unmatched hum pattern of depth information.
Deep anomalies are detected using geometric element fit approach, can be substantially reduced for the imaging of multipart labyrinth
The operand of matching detection is also higher to the tolerance of acquisition noise;Simultaneously because the matching is related to vehicle number information, Mei Yiche
Information only matches with the historical data of specific license number, also reduces interference caused by otherness between vehicle.
Specific embodiment two:
Step 1, camera array is laid out by the way of middle linear array, two sides battle array in train direction of advance, and passed through
Camera array gathers full car three-dimensional image information and middle linear array, two sides battle array information in real time, passes through different-time exposure skill during collection
Art controls the triggering frequency of laser rays, so as to obtain some three-dimensional image informations of multi-angle collection, and from above-mentioned 3-D view
The full car depth information of train to be detected is extracted in information.
Wherein, the acquisition method of the full car depth information of above-mentioned train to be detected can be but be not limited to pulse ranging method,
At least one of binocular telemetry and laser triangulation.
Step 2, vehicle license number index is established, matches to obtain reference train by the vehicle vehicle number information of train to be detected, and
Obtain the full car depth information of the reference train.
Step 3, the depth for the train to be detected under each angle being matched according to the method for step 3 in specific embodiment one is believed
Breath, the full car depth information of reference train obtain the recognition result of the failed part of each angle.
Step 4, the depth information that faulty components are extracted after each angle trouble unit recognition result is merged, according still further to specific reality
Apply the refinement step classification positioning failed part of the method for step 4 in example one.
On the basis of specific embodiment one, the present embodiment is matched using multi-angle depth information carries out faulty components, event
Hinder the classification positioning of part, some non-faulting points such as water stain dust can be shielded, and failed part matching result can be improved
Accuracy, so as to facilitate testing staff to be quickly found out actual fault point, system rate of false alarm is reduced, and improve the maintenance of testing staff
Efficiency.In addition, the three-dimensional image information of above-mentioned multi-angle image extraction can be effectively improved the depth letter of single imaging acquisition
The defects of breath can cover smaller depth information.
There is the method for matching block depth information in specific embodiment one, the step 4 of specific embodiment two:According to vehicle license number
Faulty components geometric properties distributed intelligence figure, the component geometric properties standard profile hum pattern of index matching vehicle license number association,
This realizes the matching between specific data;Vehicle license number, matching faulty components geometric properties distributed intelligence figure, Suo Youke are not limited
The component geometric properties distributed intelligence figure of energy type, to adapt to the needs of train component exchange.Present invention additionally comprises to failure group
The step of part, failed part carry out alarm indication, to feed back abnormal information in time to operating personnel.
In addition, present invention further teaches the identification train exception for carrying the abnormal detection method of above-mentioned identification train
Detecting system, the system mainly include:For gathering the full car depth information of train to be detected and the vehicle car of train to be detected
The acquisition module of number information;Match to obtain reference train by the vehicle vehicle number information of train to be detected, and obtain reference train
The reference train data obtaining module of full car depth information;And by full the car depth information and reference train of train to be detected
Full car depth information matching, detect the failed part detection module of failed part.In preferred scheme, said system also includes being used for
The alarm display module of failed part described in alarm indication.
Claims (8)
1. a kind of detection method for identifying that train is abnormal, it is characterised in that comprise the following steps:
Full the car depth information and vehicle vehicle number information of train to be detected are gathered,
Match to obtain reference train by the vehicle vehicle number information of the train to be detected, and the full car for obtaining the reference train is deep
Spend information,
Reference train and the geometric element of train different depth hierarchical data to be detected are extracted respectively, are fitted in units of component each
The geometric element of depth hierarchical data, obtain the full car geometric properties standard profile hum pattern that train to be detected contains depth information
And reference train contains the full car geometric properties distributed intelligence figure of depth information, believed by matching rain model license number to be detected
Cease with reference train vehicle vehicle number information to contrast full the car geometric properties distributed intelligence figure and reference train of train to be detected
Module sets corresponding to the unmatched hum pattern of depth information are failure group by full car geometric properties standard profile hum pattern
Part, and,
Extract and handle the geometric element of train different depth hierarchical data to obtain the full car depth information in units of part,
The depth information of faulty components is matched with the depth information with reference to part, further extracted from the depth information of faulty components
Abnormal depth information, the part further extracted corresponding to obtained abnormal depth information is set as failed part;
Full car depth information described in above-mentioned steps is deep at the top of train bottom depth information, train sidepiece depth information and train
Spend at least one of information.
2. the abnormal detection method of train is identified according to claim 1, it is characterised in that the full car of the reference train is deep
Spending information is:
The full car depth information of the normal train pre-set,
In the full car depth information of the same train passed through with the current sample time closest moment,
In multigroup full car depth information of the same train passed through with current sample time adjacent to the moment, and
Above-mentioned full car depth information is merged or statistical analysis after at least one of obtained full car depth information.
3. the abnormal detection method of train is identified according to claim 1, it is characterised in that the reference part is reference columns
At least one of corresponding part in part and default parts library corresponding to car.
4. the abnormal detection method of train is identified according to claim 1, it is characterised in that:It is described to extract and handle train not
With the geometric element of depth hierarchical data to obtain the full car depth information in units of part, by the depth information of faulty components
Matched with the depth information with reference to part, abnormal depth information is further extracted from the depth information of faulty components, one will be entered
In the step of part corresponding to abnormal depth information that step extraction obtains is set as failed part, setting event as follows
Hinder part:
The geometric element of the faulty components different depth hierarchical data is extracted, each depth hierarchical data is fitted in units of part
Geometric element, obtain the faulty components geometric properties distributed intelligence figure containing depth information;
Geometric element of the reference train in different depth hierarchical data is extracted, each depth is fitted in units of the reference part
The geometric element of hierarchical data is spent, obtains matching with the faulty components and the reference component geometric properties containing depth information
Standard profile hum pattern;
After the faulty components geometric properties distributed intelligence figure compares with the reference component geometric properties standard profile hum pattern,
Part corresponding to the unmatched hum pattern of depth information is set as failed part.
5. the abnormal detection method of train is identified according to any one in claim 1,2,3 and 4, it is characterised in that institute
State the full car depth information of train to be detected acquisition method be laser triangulation, binocular telemetry and pulse ranging method in extremely
Few one kind.
6. according to identifying the abnormal detection method of train described in claim 1, it is characterised in that will further extraction obtain it is different
After the step of part corresponding to normal depth information is set as failed part, including the step of to failed part alarm indication.
A kind of 7. detecting system for identifying that train is abnormal, it is characterised in that including:
Acquisition module, for gathering the full car depth information of train to be detected and the vehicle vehicle number information of train to be detected;
Reference train data obtaining module, match to obtain reference train by the vehicle vehicle number information of the train to be detected, and obtain
Take the full car depth information of the reference train;And
Failed part detection module, reference train and the geometric element of train different depth hierarchical data to be detected are extracted respectively,
The geometric element of each depth hierarchical data is fitted in units of component, obtains the full car geometry that train to be detected contains depth information
Characteristic standard distributed intelligence figure and reference train contain the full car geometric properties distributed intelligence figure of depth information, are treated by matching
Detection rain model vehicle number information is distributed with reference train vehicle vehicle number information with contrasting the full car geometric properties of train to be detected
The full car geometric properties standard profile hum pattern of hum pattern and reference train, by corresponding to the unmatched hum pattern of depth information
Module sets are faulty components, extract and handle the geometric element of train different depth hierarchical data to obtain in units of part
Full car depth information, the depth information of faulty components is matched with reference to the depth information of part, from the depth of faulty components
Abnormal depth information is further extracted in information, the part further extracted corresponding to obtained abnormal depth information is set as
Failed part.
8. the abnormal detecting system of identification train according to claim 7, it is characterised in that also include being used for alarm indication
The alarm display module of the failed part.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410406535.9A CN104200464B (en) | 2014-08-19 | 2014-08-19 | A kind of detection method and system for identifying that train is abnormal |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410406535.9A CN104200464B (en) | 2014-08-19 | 2014-08-19 | A kind of detection method and system for identifying that train is abnormal |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104200464A CN104200464A (en) | 2014-12-10 |
CN104200464B true CN104200464B (en) | 2018-03-09 |
Family
ID=52085750
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410406535.9A Active CN104200464B (en) | 2014-08-19 | 2014-08-19 | A kind of detection method and system for identifying that train is abnormal |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104200464B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104567725A (en) * | 2014-12-17 | 2015-04-29 | 苏州华兴致远电子科技有限公司 | Vehicle operation fault detection system and method |
KR101927648B1 (en) * | 2014-12-17 | 2018-12-10 | 쑤저우 뉴 비전 사이언스 앤드 테크놀러지 컴퍼니 리미티드 | Vehicle operation fault detection system and method |
CN104567726B (en) * | 2014-12-17 | 2017-12-08 | 苏州华兴致远电子科技有限公司 | Vehicle operation troubles detecting system and method |
CN104809427B (en) * | 2014-12-31 | 2019-02-15 | 朱金良 | Failure automatic identification detection system and detection method based on 3D information |
CN106468914B (en) * | 2015-08-21 | 2023-04-28 | 苏州华兴致远电子科技有限公司 | Train maintenance method and system |
CN106557773B (en) * | 2015-09-25 | 2020-04-24 | 北京宝沃汽车有限公司 | Vehicle fault diagnosis method and system |
CN105243381B (en) * | 2015-11-25 | 2019-04-19 | 朱金良 | Failure automatic identification detection system and method based on 3D information |
CN106940884B (en) * | 2015-12-15 | 2020-11-24 | 北京康拓红外技术股份有限公司 | Motor train unit operation fault image detection system containing depth information and detection method |
CN106198559A (en) * | 2016-08-31 | 2016-12-07 | 成都铁安科技有限责任公司 | A kind of dynamic detection system of roof of train image |
CN107220609A (en) * | 2017-05-22 | 2017-09-29 | 哈尔滨市科佳通用机电股份有限公司 | A kind of motor-car failure automatic detection method |
CN107697098B (en) * | 2017-09-29 | 2018-08-03 | 金辉 | A kind of compartment gradient in-site measurement platform |
CN108711152B (en) * | 2018-05-23 | 2021-08-13 | 深圳市阡丘越科技有限公司 | Image analysis method and device based on AI technology and user terminal |
CN110580748A (en) * | 2018-06-08 | 2019-12-17 | 苏交科集团股份有限公司 | Daily inspection and evaluation system and method for load-bearing support based on image processing |
CN110095061B (en) * | 2019-03-31 | 2020-07-14 | 唐山百川智能机器股份有限公司 | Vehicle form and position detection system and method based on contour scanning |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102323070A (en) * | 2011-06-10 | 2012-01-18 | 北京华兴致远科技发展有限公司 | Method and system for detecting abnormality of train |
CN103077526A (en) * | 2013-02-01 | 2013-05-01 | 苏州华兴致远电子科技有限公司 | Train abnormality detection method and system with deep detection function |
CN202952978U (en) * | 2012-11-06 | 2013-05-29 | 苏州华兴致远电子科技有限公司 | Train part information collector |
CN103512762A (en) * | 2012-06-29 | 2014-01-15 | 北京华兴致远科技发展有限公司 | Image processing method and device and train fault detection system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8855404B2 (en) * | 2012-08-27 | 2014-10-07 | The Boeing Company | Methods and systems for inspecting a workpiece |
-
2014
- 2014-08-19 CN CN201410406535.9A patent/CN104200464B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102323070A (en) * | 2011-06-10 | 2012-01-18 | 北京华兴致远科技发展有限公司 | Method and system for detecting abnormality of train |
CN103512762A (en) * | 2012-06-29 | 2014-01-15 | 北京华兴致远科技发展有限公司 | Image processing method and device and train fault detection system |
CN202952978U (en) * | 2012-11-06 | 2013-05-29 | 苏州华兴致远电子科技有限公司 | Train part information collector |
CN103077526A (en) * | 2013-02-01 | 2013-05-01 | 苏州华兴致远电子科技有限公司 | Train abnormality detection method and system with deep detection function |
Also Published As
Publication number | Publication date |
---|---|
CN104200464A (en) | 2014-12-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104200464B (en) | A kind of detection method and system for identifying that train is abnormal | |
CN102759347B (en) | Online in-process quality control device and method for high-speed rail contact networks and composed high-speed rail contact network detection system thereof | |
CN102323070B (en) | Method and system for detecting abnormality of train | |
CN108491758B (en) | Track detection method and robot | |
CN103512762B (en) | Image processing method, device and train failure detection system | |
CN110136447A (en) | Lane change of driving a vehicle detects and method for distinguishing is known in illegal lane change | |
CN105115605A (en) | Track train infrared detection system and detection method | |
CN105844222A (en) | System and method for front vehicle collision early warning based on visual sense | |
CN109801284A (en) | A kind of high iron catenary insulator breakdown detection method based on deep learning | |
CN105067287A (en) | Motor train unit running fault dynamic image detecting system and method | |
CN105243381B (en) | Failure automatic identification detection system and method based on 3D information | |
CN103770780A (en) | Vehicle active safety system alarm shielding device | |
CN111754483A (en) | Method and system for identifying abnormal equipment based on thermal imager | |
CN102854191A (en) | Real-time visual detection and identification method for high speed rail surface defect | |
CN104567725A (en) | Vehicle operation fault detection system and method | |
CN1664877A (en) | Airplane berth plane type automatic identification and indication system | |
CN105528894B (en) | The method and apparatus of O&M detection is carried out to monitoring frontend based on trace information | |
CN105241679B (en) | A kind of hidden fault detection method of EMUs | |
CN104210500A (en) | Overhead lines suspension state detecting and monitoring device and working method thereof | |
CN110455214A (en) | A kind of abrasion of pantograph pan condition monitoring system and method | |
CN104316782A (en) | Pantograph monitoring method and system | |
KR20190024447A (en) | Real-time line defect detection system | |
CN207850304U (en) | A kind of electrification railway contact net detecting system | |
CN205737579U (en) | Train operation composite image information detection device | |
CN104320613A (en) | Monitoring method, monitoring system and positioning system for maintenance operation of motor car |
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 |