CN109325519A - Fault recognition method and device - Google Patents
Fault recognition method and device Download PDFInfo
- Publication number
- CN109325519A CN109325519A CN201810950509.0A CN201810950509A CN109325519A CN 109325519 A CN109325519 A CN 109325519A CN 201810950509 A CN201810950509 A CN 201810950509A CN 109325519 A CN109325519 A CN 109325519A
- Authority
- CN
- China
- Prior art keywords
- picture
- unit cell
- element unit
- picture element
- strong classifier
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present invention provides a kind of fault recognition method and device.It include the first preset number picture element unit cell in the picture to be identified the described method includes: obtaining the picture to be identified of target position；According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；It wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier is made of the cascade of the second preset number Weak Classifier.The present invention can accurately and rapidly carry out fault identification, and entire fault identification process efficiency is higher, cost is relatively low, and operating worker can be replaced to realize automation fault detection, missing inspection and erroneous detection caused by artificial detection is effectively reduced, improves the degree of automation of fault detection.
Description
Technical field
The present embodiments relate to field of fault detection more particularly to a kind of fault recognition methods and device.
Background technique
With the progress of current high-speed rail EMU level, the Fast Construction and operating mileage of high-speed rail EMU grow at top speed,
Railway transportation equipment maintenance cost is rapidly increasing.The operation of high-speed rail EMU is related to extensive stream of people's transport, and safety inspection is
The most important thing, once it goes wrong, it will become major hidden danger.
Currently, the EMU maintenance shop of substantial amounts has had been established in high-speed rail EMU operating system, with standard tool car
Between for, inner space area is larger, and 4 maintenance tracks of standard configuration are overhauled simultaneously for 8 groups of trains, equipped with vehicle by every track
Top, compartment, three layers of vehicle bottom platform, convenient working personnel take action up and down.Under normal conditions, motor train unit train needs to overhaul daily,
For example the train of operation on daytime, maintenance can only carry out at night, next day starts the operation of outbound successively.Current overhaul of train-set is still
Artificial progress is relied primarily on, operating personnel's night by flashlight lighting, checks the state at each position of motor train unit train, such as
One column are about on 200 meters of motor train unit train, and primary basic maintenance repair, operating personnel usually requires to check more than 30,000 bolts
Whether break down, and the whether faulty loosening of artificial judgment bolt occurs, and checks the paint graticule in every failure.Huge
Maintenance amount and the height of service pit are generally lower than human height, and operating personnel is caused to need work of bending over.High-strength work
Intensity and uncomfortable operating attitude, cause the detection process of entire failure cumbersome, take a long time, and erroneous detection and false dismissal probability are high.
Summary of the invention
The embodiment of the present invention provides a kind of fault recognition method and device, to solve in the prior art, high-speed rail EMU
The detection process of the failure of train is cumbersome, takes a long time, the problem that erroneous detection and false dismissal probability are high.
On the one hand, the embodiment of the present invention provides a kind of fault recognition method, which comprises
The picture to be identified of target position is obtained, includes the first preset number picture element unit cell in the picture to be identified；
According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；
Wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier by
The cascade of second preset number Weak Classifier is constituted.
On the other hand, the embodiment of the present invention provides a kind of fault identification device, and described device includes:
Picture obtains module, includes first pre- in the picture to be identified for obtaining the picture to be identified of target position
If number picture element unit cell；
Fault identification module, for the strong classifier group according to the target position pre-established, to described to be identified
Picture carries out fault identification；It wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, often
A strong classifier is made of the cascade of the second preset number Weak Classifier.
On the other hand, the embodiment of the invention also provides a kind of electronic equipment, including memory, processor, bus and
The computer program that can be run on a memory and on a processor is stored, the processor is realized above-mentioned when executing described program
Step in fault recognition method.
In another aspect, being stored thereon with the embodiment of the invention also provides a kind of non-transient computer readable storage medium
Computer program realizes the step in above-mentioned fault recognition method when described program is executed by processor.
Fault recognition method and device provided in an embodiment of the present invention obtain the picture to be identified of target position, to be identified
It include multiple picture element unit cells in picture, according to the strong classifier group of the target position pre-established, to the figure to be identified
Each picture element unit cell of piece carries out fault identification, since the error rate of strong classifier is infinitely low, if picture element unit cell is by corresponding strong
Classifier successful classification is then abort situation at the corresponding position of the picture element unit cell.Wherein, picture to be identified can be set by automation
Standby automatic shooting, is taken on site without manually going；Fault identification is carried out to picture element unit cell based on strong classifier, it can be accurate, quick
Ground carries out fault identification, and entire fault identification process efficiency is higher, cost is relatively low, and operating worker can be replaced to realize automation failure
Detection, is effectively reduced missing inspection and erroneous detection caused by artificial detection, improves the degree of automation of fault detection.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is the flow diagram of fault recognition method provided in an embodiment of the present invention；
Fig. 2 is the structural schematic diagram of fault identification device provided in an embodiment of the present invention；
Fig. 3 is the structural schematic diagram of electronic equipment provided in an embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
Fig. 1 shows a kind of flow diagram of fault recognition method provided in an embodiment of the present invention.
As shown in Figure 1, fault recognition method provided in an embodiment of the present invention, the method specifically includes the following steps:
Step 101, the picture to be identified of target position is obtained, includes that the first preset number is schemed in the picture to be identified
Blade unit.
Wherein, target position can overhaul any position in position with motor train unit train, be with the chassis of motor train unit train
Example, picture to be identified can be picture of the shooting from chassis.And shooting work can be automatically snapped by automation equipment, without artificial
It goes to be taken on site.It is subsequent by picture segmentation to be identified at the first preset number picture element unit cell in order to improve fault identification precision
Fault identification is carried out individually for each picture element unit cell.
Step 102, according to the strong classifier group of the target position pre-established, event is carried out to the picture to be identified
Barrier identification；It wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each described strong point
Class device is made of the cascade of the second preset number Weak Classifier.
Specifically, strong classifier group includes distinguishing one-to-one strong classifier with each picture element unit cell, and strong classifier is
It is when picture element unit cell to be identified is by strong classifier successful classification, then described for the classifier of correct identification failure picture rapidly
The corresponding position of picture element unit cell is malfunction.Strong classifier is made of the cascade of the second preset number Weak Classifier, cascade
Process is combined each Weak Classifier with certain weight.And Weak Classifier is by the event of the corresponding position of multiple picture element unit cells
Barrier picture is obtained by successive ignition training, and the second preset number Weak Classifier is formed one strong classification with certain weight
Device may be implemented in whole picture picture to be identified using strong classifier and quickly search failure, as a kind of self-adaptive enhancement algorithm, reason
It can reach unlimited low by upper error rate.
Specifically, picture to be identified for one, carries out picture segmentation to it first, obtains the first preset number picture
Unit；Strong classifier group according to the target position again, respectively classifies to each picture element unit cell, due to strong classifier
Error rate is infinitely low, if some picture element unit cell by successful classification, is abort situation at the corresponding position of the picture element unit cell.
In the above embodiment of the present invention, the picture to be identified of target position is obtained, includes multiple pictures in picture to be identified
Unit, according to the strong classifier group of the target position pre-established, to each picture element unit cell of the picture to be identified into
Row fault identification, since the error rate of strong classifier is infinitely low, if picture element unit cell is somebody's turn to do by corresponding strong classifier successful classification
It is abort situation at the corresponding position of picture element unit cell.Wherein, picture to be identified can be automatically snapped by automation equipment, without artificial
It goes to be taken on site；Fault identification is carried out to picture element unit cell based on strong classifier, can accurately and rapidly carry out fault identification, entirely
Fault identification process efficiency is higher, cost is relatively low, and operating worker can be replaced to realize automation fault detection, artificial inspection is effectively reduced
Missing inspection and erroneous detection caused by survey improve the degree of automation of fault detection.The present invention solves in the prior art, high-speed rail EMU
The detection process of the failure of train is cumbersome, takes a long time, the problem that erroneous detection and false dismissal probability are high.
Optionally, in the embodiment of the present invention, it is described obtain target position picture to be identified the step of before, the method
Include:
Construct the strong classifier of each picture element unit cell；
The step of strong classifier of each picture element unit cell of building, comprising:
The first step obtains the sample image of the picture element unit cell；Wherein, the sample image includes positive sample and negative sample
This；The positive sample is the malfunction of the picture element unit cell corresponding position, and the negative sample is that the picture element unit cell corresponds to position
The non-faulting state set；
Second step obtains the Weak Classifier of the picture element unit cell according to the sample image；
Second preset number Weak Classifier is cascaded, constitutes the strong classifier of the picture element unit cell by third step.
Specifically, the process for constructing the strong classifier of each picture element unit cell mainly includes three steps.
The first step obtains the sample image of the picture element unit cell, can obtain a large amount of sample image, to improve strong classifier
Recognition accuracy.
Wherein, the sample image includes positive sample and negative sample；The positive sample is that the picture element unit cell corresponds to position
The malfunction set, the negative sample are the non-faulting state of the picture element unit cell corresponding position；
Second step during establishing Weak Classifier, is primarily based on positive and negative sample training, is each according to sample image
Initial weight is arranged in sample, and building filtering algorithm is iterated；During iteration, if sample is correctly classified, reducing should
The weight of sample, otherwise improves weight, weight and optimal filtering algorithm after thus obtaining one group of Sample Refreshment.
Third step cascades the second preset number Weak Classifier, constitutes strong classifier, and entire cascade includes several layers,
Each layer is the Weak Classifier Jing Guo above-mentioned training；To each layer of setting threshold value, different mistakes is obtained by adjusting threshold size
Inspection rate.Layer more rearward, the Weak Classifier number for forming strong classifier is more, and classifier is more complicated, and classification performance is also stronger.Inspection
When survey, picture element unit cell only can just be judged as being malfunction by all layers, if any layer is judged as non-event wherein
Barrier state is abandoned.
Optionally, described according to the sample image in the embodiment of the present invention, obtain the Weak Classifier of the picture element unit cell
The step of, comprising:
According to the first preset algorithm, the integrogram of the sample image is calculated, and extracts the haar feature of the integrogram；
According to the second preset algorithm, the training of third present count contents is carried out to the haar feature, obtains the picture list
The Weak Classifier of member.
Wherein, after obtaining the corresponding a large amount of positive negative samples of picture element unit cell, sample image is integrated to obtain integrogram；
Extract integrogram haar feature, wherein haar feature be it is a kind of reflect image grey scale change, pixel sub-module seek difference
A kind of feature.It is divided into edge feature, linear character, central feature and diagonal line feature, with two kinds of rectangle frame combinations of black and white
At feature templates, the characteristic value of this template is indicated with pixel in feature templates.After the haar feature for extracting integrogram,
The training of third present count contents is carried out to haar feature, obtains Weak Classifier.
Optionally, in the embodiment of the present invention, the strong classifier group for the target position that the basis pre-establishes, to institute
State the step of picture to be identified carries out fault identification, comprising:
Obtain the strong classifier group of the target position pre-established；
The determining strong classifier with each strong classifier group, corresponding picture element unit cell；
Classified by the strong classifier to the corresponding picture element unit cell:
If classifying successfully, the corresponding position of the picture element unit cell is malfunction；
Otherwise, the corresponding position of the picture element unit cell is non-faulting state.
Wherein, the strong classifier group according to the target position, respectively classifies to each picture element unit cell, due to dividing by force
The error rate of class device is infinitely low, if some picture element unit cell by successful classification, is fault bit at the corresponding position of the picture element unit cell
It sets.
Optionally, in the embodiment of the present invention, it is described obtain target position picture to be identified the step of, comprising:
Obtain the video file of target position；
It intercepts in the video file, the picture to be identified including the target position.
Wherein, the process for obtaining picture to be identified, the video file by obtaining target position realize, can when each
Between shot under the conditions of section and different weather respectively motor train unit train target position multitude of video, it includes described for intercepting in video
The picture of target position is as sample, and wherein defect center is located at picture center；Wherein, the picture of malfunction is as positive sample
This, the picture of non-faulting state is as negative sample.Largely dynamic positive and negative samples are established, and give the setting of each sample initial power
Weight calculates the haar feature of each sample using integrogram, constructs Weak Classifier；If sample is correctly classified, weight is reduced,
Otherwise its weight is aggravated, weight and optimal Weak Classifier after thus obtaining one group of Sample Refreshment.
Then continue to classify by sample and updated weight, until iteration n times, obtain n trained weak point
N Weak Classifier is formed a strong classifier with certain weight by class device, is realized in entire image quickly using strong classifier
Search failure.
Video can be automatically snapped by automation equipment, be taken on site without manually going, and the automation journey of fault detection is improved
Degree.
Optionally, in the embodiment of the present invention, the target position is the chassis position of motor train unit train；
The step of video file for obtaining target position, comprising:
The video acquisition device of predeterminated position on vehicle RGV, the video of photographic subjects position are guided by being mounted on rail
File；Wherein, the sliding rail sliding that the RGV is arranged along the maintenance slot of the motor train unit train.
Further, target position is the chassis position of motor train unit train, then obtains the mistake of the video file of target position
Journey can be realized by being mounted on rail guidance vehicle (Rail Guided Vehicle, RGV), be arranged in RGV predeterminated position and regard
Frequency acquisition device, and the sliding rail sliding that RGV is arranged along the maintenance slot of the motor train unit train is set, in this way, in RGV sliding
In the process, the video file of chassis position can be shot.
As a specific example, it can be laid with track in motor train unit train and set up RGV.Since service work must be at night
It carries out, six shaft mechanical arms can be installed on RGV, and video acquisition device and light source are erected to the terminal position of mechanical arm,
In the vehicle bottom chassis image that video acquisition device takes, carry out that the detection of failure mainly utilizes is the form of failure, and
The characteristics such as position stationarity feature and intensity profile that failure occurs in the picture.The type of motor train unit train can be first analyzed before detection
Number parameter, the length of each compartment, obtain the relative position of each failure to be detected.Under light source irradiation condition, pass through
Video acquisition device is sent to each location of fault to be detected by the cooperation of RGV and mechanical arm, by the side for extracting characteristic point
Formula comes the position of dynamic corrections video acquisition device, and calculates spatial positional information, the posture of dynamic adjusting machine tool arm, guarantees
Video acquisition device is moved to specified position, it is ensured that the image for collecting normal place, by fault identification technology in standard
Failure is searched in the image of position, the image corrected after then being handled by deformation pattern geometric correction method, analysis event
The form of barrier simultaneously judges whether current failure occurs, and result is finally fed back to maintenance monitoring center staff.
If it was found that failure, first extracts the profile of failure, is marked with special color, and save this picture, in picture list
Defeated out of order quantity in member, and result and the physical fault quantity of the position in preset database are compared, one
Denier finds that quantity is inconsistent, is write results on picture to be saved immediately, and number of faults and specific location are passed through hair
It send to maintenance monitoring center.
Optionally, in the embodiment of the present invention, the video acquisition device includes respectively corresponding with each picture element unit cell
Grid.
Wherein, grid is set for video acquisition device, main purpose is in order to divide picture element unit cell, followed by collecting
Image on carry out geometric correction, method particularly includes: by image, each pixel is split first, with each pixel of 80*80
For point, according to 80*80 mesh segmentation at 80*80 picture element unit cell, and four sides and the original size of each grid are calculated
Then offset calculates the offset on four vertex of the relatively former grid of pixel in each network one by one, grid
Offset is equal with the offset of pixel in each grid, extrapolates the image pixel after correction by this corresponding relationship
Point position.By the geometric correction algorithm of deformation pattern, the position error of RGV and mechanical arm are eliminated, it is ensured that there is height every time
Discrimination.
In the above embodiment of the present invention, the picture to be identified of target position is obtained, includes multiple pictures in picture to be identified
Unit, according to the strong classifier group of the target position pre-established, to each picture element unit cell of the picture to be identified into
Row fault identification, since the error rate of strong classifier is infinitely low, if picture element unit cell is somebody's turn to do by corresponding strong classifier successful classification
It is abort situation at the corresponding position of picture element unit cell.Wherein, picture to be identified can be automatically snapped by automation equipment, without artificial
It goes to be taken on site；Fault identification is carried out to picture element unit cell based on strong classifier, can accurately and rapidly carry out fault identification, entirely
Fault identification process efficiency is higher, cost is relatively low, and operating worker can be replaced to realize automation fault detection, artificial inspection is effectively reduced
Missing inspection and erroneous detection caused by survey improve the degree of automation of fault detection.
Fault recognition method provided in an embodiment of the present invention is described above, introduces the present invention below in conjunction with attached drawing and implements
The fault identification device that example provides.
Referring to fig. 2, the embodiment of the invention provides a kind of fault identification device, described device includes:
Picture obtains module 201, includes first in the picture to be identified for obtaining the picture to be identified of target position
Preset number picture element unit cell.
Wherein, target position can overhaul any position in position with motor train unit train, be with the chassis of motor train unit train
Example, picture to be identified can be picture of the shooting from chassis.And shooting work can be automatically snapped by automation equipment, without artificial
It goes to be taken on site.It is subsequent by picture segmentation to be identified at the first preset number picture element unit cell in order to improve fault identification precision
Fault identification is carried out individually for each picture element unit cell.
Fault identification module 202, for the strong classifier group according to the target position pre-established, to described wait know
Other picture carries out fault identification；It wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group,
Each strong classifier is made of the cascade of the second preset number Weak Classifier.
Specifically, strong classifier group includes distinguishing one-to-one strong classifier with each picture element unit cell, and strong classifier is
It is when picture element unit cell to be identified is by strong classifier successful classification, then described for the classifier of correct identification failure picture rapidly
The corresponding position of picture element unit cell is malfunction.Strong classifier is made of the cascade of the second preset number Weak Classifier, cascade
Process is combined each Weak Classifier with certain weight.And Weak Classifier is by the event of the corresponding position of multiple picture element unit cells
Barrier picture is obtained by successive ignition training, and the second preset number Weak Classifier is formed one strong classification with certain weight
Device may be implemented in whole picture picture to be identified using strong classifier and quickly search failure, as a kind of self-adaptive enhancement algorithm, reason
It can reach unlimited low by upper error rate.
Specifically, picture to be identified for one, carries out picture segmentation to it first, obtains the first preset number picture
Unit；Strong classifier group according to the target position again, respectively classifies to each picture element unit cell, due to strong classifier
Error rate is infinitely low, if some picture element unit cell by successful classification, is abort situation at the corresponding position of the picture element unit cell.
Optionally, in the embodiment of the present invention, described device includes:
First building module, for constructing the strong classifier of each picture element unit cell；
Described first, which constructs module, includes:
Sample acquisition submodule, for obtaining the sample image of the picture element unit cell；Wherein, the sample image includes just
Sample and negative sample；The positive sample is the malfunction of the picture element unit cell corresponding position, and the negative sample is the figure
The non-faulting state of blade unit corresponding position；
First building submodule, for obtaining the Weak Classifier of the picture element unit cell according to the sample image；
Submodule is cascaded, for cascading the second preset number Weak Classifier, constitutes the strong classification of the picture element unit cell
Device.
Optionally, in the embodiment of the present invention, the first building submodule is used for:
According to the first preset algorithm, the integrogram of the sample image is calculated, and extracts the haar feature of the integrogram；
According to the second preset algorithm, the training of third present count contents is carried out to the haar feature, obtains the picture list
The Weak Classifier of member.
Optionally, in the embodiment of the present invention, the picture obtains module 201 and includes:
Video acquisition submodule, for obtaining the video file of target position；
Intercept submodule, for intercepting in the video file, the picture to be identified including the target position.
Optionally, in the embodiment of the present invention, the target position is the chassis position of motor train unit train；
The acquisition video acquisition submodule is used for:
The video acquisition device of predeterminated position on vehicle RGV, the video of photographic subjects position are guided by being mounted on rail
File；Wherein, the sliding rail sliding that the RGV is arranged along the maintenance slot of the motor train unit train.
Optionally, in the embodiment of the present invention, the video acquisition device includes respectively corresponding with each picture element unit cell
Grid.
Optionally, in the embodiment of the present invention, the fault identification module 202 is used for:
Obtain the strong classifier group of the target position pre-established；
The determining strong classifier with each strong classifier group, corresponding picture element unit cell；
Classified by the strong classifier to the corresponding picture element unit cell:
If classifying successfully, the corresponding position of the picture element unit cell is malfunction；
Otherwise, the corresponding position of the picture element unit cell is non-faulting state.
In the above embodiment of the present invention, the picture to be identified that module 201 obtains target position is obtained by picture, it is to be identified
Include multiple picture element unit cells in picture, fault identification module 202 according to the strong classifier group of the target position pre-established,
Fault identification is carried out to each picture element unit cell of the picture to be identified, since the error rate of strong classifier is infinitely low, if picture
Unit is then abort situation at the corresponding position of the picture element unit cell by corresponding strong classifier successful classification.Wherein, figure to be identified
Piece can be automatically snapped by automation equipment, be taken on site without manually going；Failure knowledge is carried out to picture element unit cell based on strong classifier
Not, fault identification can be accurately and rapidly carried out, entire fault identification process efficiency is higher, cost is relatively low, can replace operating worker
It realizes automation fault detection, missing inspection and erroneous detection caused by artificial detection is effectively reduced, improves the degree of automation of fault detection.
Fig. 3 shows the structural schematic diagram of a kind of electronic equipment of further embodiment of this invention offer.
Referring to Fig. 3, electronic equipment provided in an embodiment of the present invention, the electronic equipment include memory (memory) 31,
Processor (processor) 32, bus 33 and it is stored in the computer program that can be run on memory 31 and on a processor.
Wherein, the memory 31, processor 32 complete mutual communication by the bus 33.
The processor 32 is used to call the program instruction in the memory 31, realizes when executing described program as schemed
1 method.
In another embodiment, following method is realized when the processor executes described program:
The picture to be identified of target position is obtained, includes the first preset number picture element unit cell in the picture to be identified；
According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；
Wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier by
The cascade of second preset number Weak Classifier is constituted.
Electronic equipment provided in an embodiment of the present invention can be used for executing the corresponding program of method of above method embodiment,
This implementation repeats no more.
Electronic equipment provided in an embodiment of the present invention obtains the picture to be identified of target position, includes in picture to be identified
Multiple picture element unit cells, according to the strong classifier group of the target position pre-established, to each figure of the picture to be identified
Blade unit carries out fault identification, since the error rate of strong classifier is infinitely low, if picture element unit cell is by the success of corresponding strong classifier
Classification is then abort situation at the corresponding position of the picture element unit cell.Wherein, picture to be identified can be clapped automatically by automation equipment
It takes the photograph, is taken on site without manually going；Fault identification is carried out to picture element unit cell based on strong classifier, can accurately and rapidly carry out event
Barrier identification, entire fault identification process efficiency is higher, cost is relatively low, and operating worker can be replaced to realize automation fault detection, had
Effect reduces missing inspection and erroneous detection caused by artificial detection, improves the degree of automation of fault detection.
A kind of non-transient computer readable storage medium that further embodiment of this invention provides, the non-transient computer can
It reads to be stored with computer program on storage medium, realize when described program is executed by processor such as the step of Fig. 1.
In another embodiment, following method is realized when described program is executed by processor:
The picture to be identified of target position is obtained, includes the first preset number picture element unit cell in the picture to be identified；
According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；
Wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier by
The cascade of second preset number Weak Classifier is constituted.
Non-transient computer readable storage medium provided in an embodiment of the present invention, realization when described program is executed by processor
The method of above method embodiment, this implementation repeat no more.
Non-transient computer readable storage medium provided in an embodiment of the present invention obtains the picture to be identified of target position,
Include multiple picture element unit cells in picture to be identified, according to the strong classifier group of the target position pre-established, to it is described to
Identify that each picture element unit cell of picture carries out fault identification, since the error rate of strong classifier is infinitely low, if picture element unit cell is right
The strong classifier successful classification answered then is abort situation at the corresponding position of the picture element unit cell.Wherein, picture to be identified can be by certainly
The automatic shooting of dynamicization equipment, is taken on site without manually going；Fault identification is carried out to picture element unit cell based on strong classifier, it can be quasi-
Really, fault identification is rapidly carried out, entire fault identification process efficiency is higher, cost is relatively low, operating worker can be replaced to realize certainly
Missing inspection and erroneous detection caused by artificial detection is effectively reduced in dynamicization fault detection, improves the degree of automation of fault detection.
Further embodiment of this invention discloses a kind of computer program product, and the computer program product is non-including being stored in
Computer program in transitory computer readable storage medium, the computer program include program instruction, when described program refers to
When order is computer-executed, computer is able to carry out method provided by above-mentioned each method embodiment, for example,
The picture to be identified of target position is obtained, includes the first preset number picture element unit cell in the picture to be identified；
According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；
Wherein, include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier by
The cascade of second preset number Weak Classifier is constituted.
The apparatus embodiments described above are merely exemplary, wherein described, unit can as illustrated by the separation member
It is physically separated with being or may not be, component shown as a unit may or may not be physics list
Member, it can it is in one place, or may be distributed over multiple network units.It can be selected according to the actual needs
In some or all of the modules achieve the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creativeness
Labour in the case where, it can understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It realizes by means of software and necessary general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be embodied in the form of software products in other words, should
Computer software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Method described in certain parts of example or embodiment.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of fault recognition method characterized by comprising
The picture to be identified of target position is obtained, includes the first preset number picture element unit cell in the picture to be identified；
According to the strong classifier group of the target position pre-established, fault identification is carried out to the picture to be identified；Wherein,
It include strong classifier corresponding with each picture element unit cell in the strong classifier group, each strong classifier is pre- by second
If the cascade of number Weak Classifier is constituted.
2. the method according to claim 1, wherein it is described obtain target position picture to be identified the step of it
Before, which comprises
Construct the strong classifier of each picture element unit cell；
The step of strong classifier of each picture element unit cell of building, comprising:
Obtain the sample image of the picture element unit cell；Wherein, the sample image includes positive sample and negative sample；The positive sample
This is the malfunction of the picture element unit cell corresponding position, and the negative sample is the non-faulting shape of the picture element unit cell corresponding position
State；
According to the sample image, the Weak Classifier of the picture element unit cell is obtained；
Second preset number Weak Classifier is cascaded, the strong classifier of the picture element unit cell is constituted.
3. according to the method described in claim 2, obtaining the picture list it is characterized in that, described according to the sample image
The step of Weak Classifier of member, comprising:
According to the first preset algorithm, the integrogram of the sample image is calculated, and extracts the haar feature of the integrogram；
According to the second preset algorithm, the training of third present count contents is carried out to the haar feature, obtains the picture element unit cell
Weak Classifier.
4. the method according to claim 1, wherein it is described obtain target position picture to be identified the step of,
Include:
Obtain the video file of target position；
It intercepts in the video file, the picture to be identified including the target position.
5. the method according to claim 1, wherein the target position is the chassis position of motor train unit train；
The step of video file for obtaining target position, comprising:
The video acquisition device of predeterminated position on vehicle RGV, the video file of photographic subjects position are guided by being mounted on rail；
Wherein, the sliding rail sliding that the RGV is arranged along the maintenance slot of the motor train unit train.
6. according to the method described in claim 5, it is characterized in that, the video acquisition device includes and each picture list
The corresponding grid of member.
7. the method according to claim 1, wherein strong point of the target position that the basis pre-establishes
Class device group, the step of fault identification is carried out to the picture to be identified, comprising:
Obtain the strong classifier group of the target position pre-established；
The determining strong classifier with each strong classifier group, corresponding picture element unit cell；
Classified by the strong classifier to the corresponding picture element unit cell:
If classifying successfully, the corresponding position of the picture element unit cell is malfunction；
Otherwise, the corresponding position of the picture element unit cell is non-faulting state.
8. a kind of fault identification device, which is characterized in that described device includes:
Picture obtains module, includes the first present count in the picture to be identified for obtaining the picture to be identified of target position
Mesh picture element unit cell；
Fault identification module, for the strong classifier group according to the target position pre-established, to the picture to be identified
Carry out fault identification；It wherein, include strong classifier corresponding with each picture element unit cell, Mei Gesuo in the strong classifier group
Strong classifier is stated to be made of the cascade of the second preset number Weak Classifier.
9. a kind of electronic equipment, which is characterized in that on a memory and can be including memory, processor, bus and storage
The computer program run on processor, the processor are realized when executing described program such as any one of claims 1 to 7 institute
The step in fault recognition method stated.
10. a kind of non-transient computer readable storage medium, is stored thereon with computer program, it is characterised in that: described program
The step in the fault recognition method as described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810950509.0A CN109325519A (en) | 2018-08-20 | 2018-08-20 | Fault recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810950509.0A CN109325519A (en) | 2018-08-20 | 2018-08-20 | Fault recognition method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109325519A true CN109325519A (en) | 2019-02-12 |
Family
ID=65264335
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810950509.0A Pending CN109325519A (en) | 2018-08-20 | 2018-08-20 | Fault recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109325519A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101867699A (en) * | 2010-05-25 | 2010-10-20 | 中国科学技术大学 | Real-time tracking method of nonspecific target based on partitioning |
| CN102024149A (en) * | 2009-09-18 | 2011-04-20 | 北京中星微电子有限公司 | Method of object detection and training method of classifier in hierarchical object detector |
| CN102519972A (en) * | 2011-12-10 | 2012-06-27 | 山东明佳包装检测科技有限公司 | Detection method of PET bottle cap and liquid level |
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
| CN108022235A (en) * | 2017-11-23 | 2018-05-11 | 中国科学院自动化研究所 | High-voltage power transmission tower critical component defect identification method |
-
2018
- 2018-08-20 CN CN201810950509.0A patent/CN109325519A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024149A (en) * | 2009-09-18 | 2011-04-20 | 北京中星微电子有限公司 | Method of object detection and training method of classifier in hierarchical object detector |
| CN101867699A (en) * | 2010-05-25 | 2010-10-20 | 中国科学技术大学 | Real-time tracking method of nonspecific target based on partitioning |
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| CN102519972A (en) * | 2011-12-10 | 2012-06-27 | 山东明佳包装检测科技有限公司 | Detection method of PET bottle cap and liquid level |
| CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
| CN108022235A (en) * | 2017-11-23 | 2018-05-11 | 中国科学院自动化研究所 | High-voltage power transmission tower critical component defect identification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110703800A (en) | Unmanned aerial vehicle-based intelligent identification method and system for electric power facilities | |
| KR102017870B1 (en) | Real-time line defect detection system | |
| EP3196853A1 (en) | Machine vision-based method and system for aircraft docking guidance and aircraft type identification | |
| CN110728225B (en) | High-speed face searching method for attendance checking | |
| CN108846331B (en) | Video identification method for judging whether screw fastener of motor train unit chassis falls off or not | |
| CN102902974B (en) | Image based method for identifying railway overhead-contact system bolt support identifying information | |
| CN107958201A (en) | A kind of intelligent checking system and method for vehicle annual test insurance policy form | |
| CN105302151A (en) | Aircraft docking guidance and type recognition system and method | |
| CN111080611A (en) | Railway wagon bolster spring fracture fault image identification method | |
| CN104599249A (en) | Cable rope bearing bridge deck vehicle load distribution real-time detection method | |
| CN107886047A (en) | A kind of detecting system and method for vehicle annual test survey report | |
| CN106447699B (en) | High iron catenary object detecting and tracking method based on Kalman filtering | |
| CN113947731A (en) | Foreign matter identification method and system based on contact net safety inspection | |
| CN115810134B (en) | Image acquisition quality inspection method, system and device for vehicle insurance anti-fraud | |
| CN111339927A (en) | Intelligent work state identification system for personnel in electric power business hall | |
| CN114633774A (en) | Rail transit fault detection system based on artificial intelligence | |
| CN113591643A (en) | Underground vehicle station entering and exiting detection system and method based on computer vision | |
| CN113033443A (en) | Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method | |
| CN109325519A (en) | Fault recognition method and device | |
| CN114442658B (en) | Automatic inspection system for power transmission and distribution line unmanned aerial vehicle and operation method thereof | |
| CN115272284A (en) | Power transmission line defect identification method based on image quality evaluation | |
| CN110516551B (en) | Vision-based line patrol position deviation identification system and method and unmanned aerial vehicle | |
| CN113763326A (en) | Pantograph detection method based on Mask scaling R-CNN network | |
| CN115035412B (en) | Method for identifying foreign matters of overhead contact system | |
| Andonovski et al. | Computer vision system for cabin door detection and location |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190212 |