CN105719305A - Assembly falloff defect identification method and system of overhead contact system - Google Patents
Assembly falloff defect identification method and system of overhead contact system Download PDFInfo
- Publication number
- CN105719305A CN105719305A CN201610049171.2A CN201610049171A CN105719305A CN 105719305 A CN105719305 A CN 105719305A CN 201610049171 A CN201610049171 A CN 201610049171A CN 105719305 A CN105719305 A CN 105719305A
- Authority
- CN
- China
- Prior art keywords
- target
- target connector
- assembly
- comes
- connector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an assembly falloff defect identification method and system of an overhead contact system. Assembly template images slide in different scales of an image to be detected to search a target assembly, a position area image of the target assembly is matched in the image to be detected, and the assembly template images include template images of different assemblies in the overhead contact system; an area of a target connecting piece is positioned in the position area image according to an image edge analysis algorithm; according to the structural relation between the target assembly and the target connecting piece, the target connecting piece is segmented from the position area image in a corresponding proportion; and a gray scale histogram and a gradient feature of the target connecting piece are obtained, whether a suspected falloff defect exists the target connecting piece is determined according to the gray scale histogram of the target connecting piece, whether the suspected falloff defect is true is determined according to the gradient feature of the target connecting piece, the target connecting piece is determined to have the assembly falloff defect if the suspected falloff defect is true, and the target connecting piece is determined to be normal if the suspected falloff defect is false. Thus, the assembly falloff defect can be identified accurately.
Description
Technical field
The present invention relates to contact net field, particularly relate to assembly in contact net and come off defect identification method and system.
Background technology
Contact net is in electric railway, along the erection of rail overhead "the" shape, takes the high voltage transmission line of stream for pantograph.Contact net is the main truss of railway electrification project, is the transmission line of electricity of the specific form powered to track vehicle of downline overhead erection.Generally by contact suspension, support that device, positioner, a few part of pillar and basis form.
Contact net employs and substantial amounts of fixes relevant support, localizer, location bar etc. as screw, bolt etc. connect assembly, and this type of connects assembly once there is the situation loosening or coming off, and contact net safe operation is had very big potential danger.
Existing detection mode is then check corresponding video image by human eye, due to being continuously increased along with amount of video, human eye checks that such as screw, bolt etc. connect whether assembly exists the defect that comes off, and not only expend substantial amounts of manpower, it is also possible to because the fatigue test of people causes missing inspection.What therefore assembly came off defect automatically identifies have very strong practical value.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, confirm that assembly comes off the inconvenience of defect for artificial analysis video frame by frame, it is proposed that in a kind of contact net, the assembly assembly in defect identification method and contact net that comes off comes off defect recognition system.
It is an object of the invention to be achieved through the following technical solutions:
(1) in contact net, assembly comes off defect identification method, and each assembly in contact net is connected to form by connector, and described method is used for detecting whether this connector exists the defect that comes off, and said method comprising the steps of:
S1, obtains image to be detected；
S2, positioning component: utilize component template image slip scan target element on the different scale of image to be detected, image to be detected matches the band of position image of target element, described component template image includes the template image of each assembly in contact net；
S3, positioning link: orient target connector region in the area image of position according to edge analysis algorithm；
S4, cutting connection fitting: the structural relation according to target element Yu target connector, in the area image of position, it is partitioned into target connector according to corresponding ratio；
S5, feature analysis: obtain grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
Further, described positioning component step S2 includes following sub-step:
S201, formation component subtemplate image: in component template image object, intercept the subimage of component template image object as assembly subtemplate image object_sub；
S202, multiple dimensioned location target element: carry out up-sampling and down-sampling on n the yardstick of component template image object, obtain the component template image object of 2n+1 different scalek, n the yardstick of assembly subtemplate image object_sub carries out up-sampling and down-sampling, obtains the assembly subtemplate image object_sub of 2n+1 different scalek, wherein, k ∈ [1,2n+1]；
S203, extracts the gradient magnitude of image: computation module template image objectk, assembly subtemplate image object_subkAmplitude Amp is obtained with the gradient magnitude of image to be detectedk, amplitude Ampk_subWith amplitude Ampdetect；
S204, the band of position of location target element: by amplitude AmpkWith amplitude Ampk_subRespectively at amplitude AmpdetectIn carry out template matching, obtain the optimum network for location of target element under current scale as LockWith network for location as Lock_sub,
S205, calculates the similarity of network for location picture and template image:
If network for location is as Lock_subIt is that network for location is as LockSubimage, namelyThen extract component template image object under current scale respectivelykHOG feature and network for location as LockHOG feature, and calculate the Euclidean distance dist between two HOG featuresk；
If network for location is as Lock_subIt not that network for location is as LockSubimage, namelyThen component template image object under definition current scalekWith network for location as LockEuclidean distance distkIt is a maximum MaxValue, namely
S206, it is determined that the band of position image of target element: from 2n+1 Euclidean distance distkThe Euclidean distance dist that middle selection is minimumk, and it is compared with threshold value MaxTh, if this minimum Euclidean distance distk> MaxTh, then judge that this network for location is as LockInvalid, if this minimum Euclidean distance distk< MaxTh, then judge this location target LockBand of position image for this target element.
Further, described positioning link step S3 includes following sub-step:
S301, obtains the sobel edge graph of band of position image；
S302, the horizontal edge rectangular histogram of statistics sobel edge graph and vertical edge rectangular histogram；
S303, compare with threshold value Th1 according to horizontal edge rectangular histogram, first time orients coboundary and the lower boundary in target connector region, compare with threshold value Th2 according to vertical edge rectangular histogram, first time orients left margin and the right margin in target connector region, obtains the target connector region of first time location；
S304, the positioning result according to S303, it is repeated once or repeatedly step S301-S303, obtains the target connector region after multiple bearing.
Further, described step S303 includes following sub-step:
S3031, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the coboundary top in target connector region: from horizontal edge rectangular histogram HistRpInitial value start to compare with its average AvgR successively, until pth row HistRpDuring with the ratio of average AvgR more than threshold value Th1, then make the coboundary top=p in target connector region；
S3032, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the lower boundary bottom in target connector region: from horizontal edge rectangular histogram HistRpEnd value start to compare with its maximum MaxR successively, until pth row HistRpMore than average AvgR and HistRpWith the ratio of maximum MaxR more than threshold value Th0, then make the lower boundary bottom=p in target connector region；
S3033, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the left margin left in target connector region: from vertical edge rectangular histogram HistCqInitial value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the left margin left=q in target connector region；
S3034, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the right margin right in target connector region: from vertical edge rectangular histogram HistCqEnd value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the right margin right=q in target connector region；
S3035, according to coboundary top, lower boundary bottom, left margin left and right margin right, the target connector region I that first time orientsfirst。
Further, described positioning link step S3 also includes sub-step S305: get rid of the positioning result of mistake in the target connector region after multiple bearing.
Further, described structural relation includes relative position relation and area proportionate relationship.
Further, judge described in feature analysis step S5 whether target connector exists the doubtful defect that comes off and include following sub-step: calculate the grey level histogram of each target connector, and judge whether the maximum gray level of the grey level histogram of each target connector identical or whether difference between same grey level is less than threshold value Th3, if, then judge that this target connector exists the doubtful defect that comes off, and record connector come off mark flag=1, otherwise judging that this target connector is absent from the doubtful defect that comes off, record connector comes off and indicates flag=0.
Further, determine described in feature analysis step S5 whether the doubtful defect that comes off is true, including following sub-step:
S501, calculates horizontal gradient and the vertical gradient of each target connector respectively, calculates the argument θ of each pixel；
S502, adds up the distribution situation of argument θ in whole target connector image, argument rectangular histogram is normalized, obtains argument features vector；
S503, the mark flag=1 if connector comes off, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if having one group of mean square deviation more than threshold value Th4, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect；
S504, the mark flag=0 if connector comes off, then whether the argument principal direction comparing all target connectors is identical；
If it is identical, it is determined that whether the doubtful defect that comes off is false, it is determined that this target connector is normal；
If differing, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if there being one group of mean square deviation more than threshold value Th5 and to contain one group of mean square deviation less than threshold value Th6, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect.
(2) in contact net, assembly comes off defect recognition system, applying method as defined above, described system includes the image collection module to be detected being sequentially connected with, multiple dimensioned assembly locating module, connector locating module, connector segmentation module and connector characteristics analysis module.
Described image collection module to be detected is for the image to be detected of securing component.
Described multiple dimensioned assembly locating module is for utilizing component template image slip scan target element on the different scale of image to be detected, matching the band of position image of target element in image to be detected, described component template image includes the template image of each assembly in contact net.
Described connector locating module for orienting target connector region according to edge analysis algorithm in the area image of position.
Described connector segmentation module, for the structural relation according to target element Yu target connector, is partitioned into target connector according to corresponding ratio in the area image of position.
Described connector characteristics analysis module is for obtaining grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
The invention has the beneficial effects as follows:
1) present invention can be suitably applied to detect the various connectors of each assembly, only need to for different assemblies and different connector, adopt different parameters, its principle is essentially identical, the connector on each assembly and each assembly thereof can be oriented accurately, by technology such as multiple dimensioned location, template matching and HOG feature analysiss, be effectively improved assembly and come off the accuracy identified, reduce false drop rate, also can get rid of the environmental disturbances such as background, grove and tunnel.
2) realize assembly by the present invention to come off the Intelligent Measurement of defect, detect image exist come off defect time, send the defect that comes off to report to the police, and export this existence and come off the picture of defect, the picture gone out through Intelligent Recognition of the present invention only need to be carried out manual confirmation and analysis by staff, shorten staff greatly to check the time of video, improve work efficiency.
Accompanying drawing explanation
Fig. 1 is that in contact net of the present invention, assembly comes off the schematic flow sheet of defect identification method；
Fig. 2 is that in contact net of the present invention, assembly comes off the system block diagram of defect recognition system.
Detailed description of the invention
Below in conjunction with accompanying drawing, technical scheme is described in further detail, but protection scope of the present invention is not limited to the following stated.
(1) in contact net, assembly comes off defect identification method
As shown in Figure 1, This embodiment describes assembly in a kind of contact net to come off defect identification method, each assembly in contact net is connected to form by connector, and described method is used for detecting whether this connector exists the defect that comes off, and described connector can include screw, nut, bolt etc..Present invention can be suitably applied to detect the various connectors of each assembly, only for different assemblies and different connector, need to adopt different parameters, its principle is essentially identical.
Method proposed by the invention comprises the following steps:
S1, obtains image to be detected.
S2, positioning component: utilize component template image slip scan target element on the different scale of image to be detected, image to be detected matches the band of position image of target element, described component template image includes the template image of each assembly in contact net.
S3, positioning link: orient target connector region in the area image of position according to edge analysis algorithm.
S4, cutting connection fitting: the structural relation according to target element Yu target connector, in the area image of position, it is partitioned into target connector according to corresponding ratio.
S5, feature analysis: obtain grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
Further, described positioning component step S2 includes following sub-step:
S201, formation component subtemplate image: in component template image object, intercept the subimage of component template image object as assembly subtemplate image object_sub.
S202, multiple dimensioned location target element: carry out up-sampling and down-sampling on n the yardstick of component template image object, obtain the component template image object of 2n+1 different scalek, n the yardstick of assembly subtemplate image object_sub carries out up-sampling and down-sampling, obtains the assembly subtemplate image object_sub of 2n+1 different scalek, wherein, k ∈ [1,2n+1].
Wherein, up-sampling multiplying power is fup, down-sampling multiplying power is fdown。
S203, extracts the gradient magnitude of image: computation module template image objectk, assembly subtemplate image object_subkWith the gradient magnitude Amp of image to be detected, obtain amplitude Ampk, amplitude Ampk_subWith amplitude Ampdetect。
S204, the band of position of location target element: by amplitude AmpkWith amplitude Ampk_subRespectively at amplitude AmpdetectIn carry out template matching, obtain the optimum network for location of target element under current scale as LockWith network for location as Lock_sub。
Wherein, the matching way of described template matching includes difference of two squares coupling, standard deviation coupling, relevant matches, standard relevant matches, correlation coefficient matching method and canonical correlation coefficient coupling etc..
S205, calculates the similarity of network for location picture and template image:
If network for location is as Lock_subIt is that network for location is as LockSubimage, namelyThen extract component template image object under current scale respectivelykHOG feature and network for location as LockHOG feature, and calculate the Euclidean distance dist between two HOG featuresk。
If network for location is as Lock_subIt not that network for location is as LockSubimage, namelyThen component template image object under definition current scalekWith network for location as LockEuclidean distance distkIt is a maximum MaxValue, namelyWherein, maximum MaxValue is theoretically infinitely-great positive number, the positive number of approach infinity can be adopted to describe, such as MaxValue=10000.
S206, it is determined that the band of position image of target element: from 2n+1 Euclidean distance distkThe Euclidean distance dist that middle selection is minimumk, and it is compared with threshold value MaxTh, if this minimum Euclidean distance distk> MaxTh, then judge that this network for location is as LockInvalid, if this minimum Euclidean distance distk< MaxTh, then judge this location target LockBand of position image for this target element.
In the present invention, threshold value can be adjusted correspondingly according to different assemblies.
Further, described positioning link step S3 includes following sub-step:
S301, obtains the sobel edge graph of band of position image.
S302, the horizontal edge rectangular histogram of statistics sobel edge graph and vertical edge rectangular histogram.
S303, compare with threshold value Th1 according to horizontal edge rectangular histogram, first time orients coboundary and the lower boundary in target connector region, compare with threshold value Th2 according to vertical edge rectangular histogram, first time orients left margin and the right margin in target connector region, obtains the target connector region of first time location.
S304, the positioning result according to S303, it is repeated once or repeatedly step S301-S303, obtains the target connector region after multiple bearing.
Further, described positioning link step S3 also includes sub-step S305: get rid of the positioning result of mistake in the target connector region after multiple bearing.
Further, described step S303 includes following sub-step:
S3031, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the coboundary top in target connector region: from horizontal edge rectangular histogram HistRpInitial value start to compare with its average AvgR successively, until pth row HistRpDuring with the ratio of average AvgR more than threshold value Th1, then make the coboundary top=p in target connector region；
S3032, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the lower boundary bottom in target connector region: from horizontal edge rectangular histogram HistRpEnd value start to compare with its maximum MaxR successively, until pth row HistRpMore than average AvgR and HistRpDuring with the ratio of maximum MaxR more than threshold value Th0, then make the lower boundary bottom=p in target connector region；
S3033, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the left margin left in target connector region: from vertical edge rectangular histogram HistCqInitial value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the left margin left=q in target connector region；
S3034, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the right margin right in target connector region: from vertical edge rectangular histogram HistCqEnd value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the right margin right=q in target connector region；
S3035, according to coboundary top, lower boundary bottom, left margin left and right margin right, the target connector region I that first time orientsfirst。
Obtaining target connector region IfirstAfter, perform step S304, general, only target connector can need to be performed twice at location, namely when step S304, only repeat a S301-S303.Border top, bottom, left, the right of location are truncated to the target connector area image I of second time locationsceond.At the target connector area image I obtaining second time locationsceondAfter, can also carry out the positioning result operation removing target connector mistake.
Further, the structural relation described in step S4 can include relative position relation and area proportionate relationship.Such as, for screw assembly, owing to being identical according to the size of each screw in screw assembly, and the position of screw is relatively-stationary in assembly.Therefore the segmentation of screw can be split according to this characteristic.Concrete grammar is as follows:
1. according to screw proportion in assembly, can calculating the height Height*0.5-offset of single screw, wherein Height is IsceondPicture altitude.Owing to the scale visual of screw is square, namely the width of known screw is Height*0.5-offset, wherein takes offset=4.
2. according to screw relative position relation in assembly, divisible corresponding single screw is gone out accordingly.Screw as being distributed for triangle divisible go out Atria apex screw.
Further, judge described in feature analysis step S5 whether target connector exists the doubtful defect that comes off and include following sub-step: calculate the grey level histogram of each target connector, and judge whether the maximum gray level of the grey level histogram of each target connector identical or whether difference between same grey level is less than threshold value Th3, if meeting, then judge that this target connector exists the doubtful defect that comes off, and record connector come off mark flag=1, if being unsatisfactory for, then record connector comes off and indicates flag=0.
Further, determine described in feature analysis step S5 whether the doubtful defect that comes off is true, including following sub-step:
S501, calculates horizontal gradient and the vertical gradient of each target connector respectively, calculates the argument θ of each pixel.
S503, adds up the distribution situation of argument θ in whole target connector image, argument rectangular histogram is normalized, obtains argument features vector.
S504, the mark flag=1 if connector comes off, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if having one group of mean square deviation more than threshold value Th4, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect.
S505, the mark flag=0 if connector comes off, then whether the argument principal direction comparing all target connectors is identical.
If it is identical, it is determined that whether the doubtful defect that comes off is false, it is determined that this target connector is normal.
If differing, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if there being one group of mean square deviation more than threshold value Th5 and to contain one group of mean square deviation less than threshold value Th6, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect.
Experiment proves that, the present invention can effectively identify assembly in the full background image of different gray scales and come off defect, gets rid of the interference such as background, tunnel, grove, multiple dimensioned location objective result is good, missing inspection is few, flase drop is less, and wherein, it is also ideal that screw is accurately positioned result.
(2) in contact net, assembly comes off defect recognition system
As shown in Figure 2, This embodiment describes assembly in a kind of contact net to come off defect recognition system, applying method as defined above, described system includes the image collection module to be detected being sequentially connected with, multiple dimensioned assembly locating module, connector locating module, connector segmentation module and connector characteristics analysis module.
Described image collection module to be detected is for the image to be detected of securing component.
Described multiple dimensioned assembly locating module is for utilizing component template image slip scan target element on the different scale of image to be detected, matching the band of position image of target element in image to be detected, described component template image includes the template image of each assembly in contact net.
Described connector locating module for orienting target connector region according to edge analysis algorithm in the area image of position.
Described connector segmentation module, for the structural relation according to target element Yu target connector, is partitioned into target connector according to corresponding ratio in the area image of position.
Described connector characteristics analysis module is for obtaining grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
Defect identification method and the system of coming off according to assembly in the contact net of the present invention is described in an illustrative manner above with reference to accompanying drawing.But; skilled artisan would appreciate that; assembly in the contact net that the invention described above is proposed is come off defect identification method and system; various improvement can also be made on without departing from the basis of present invention; or wherein portion of techniques feature is carried out equivalent replacement; all within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention.Therefore, protection scope of the present invention should be determined by the content of appending claims.
Claims (9)
1. in contact net, assembly comes off defect identification method, and each assembly in contact net is connected to form by connector, and described method is used for detecting whether this connector exists the defect that comes off, it is characterised in that said method comprising the steps of:
S1, obtains image to be detected；
S2, positioning component: utilize component template image slip scan target element on the different scale of image to be detected, image to be detected matches the band of position image of target element, described component template image includes the template image of each assembly in contact net；
S3, positioning link: orient target connector region in the area image of position according to edge analysis algorithm；
S4, cutting connection fitting: the structural relation according to target element Yu target connector, in the area image of position, it is partitioned into target connector according to corresponding ratio；
S5, feature analysis: obtain grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
2. in contact net according to claim 1, assembly comes off defect identification method, it is characterised in that described positioning component step S2 includes following sub-step:
S201, formation component subtemplate image: in component template image object, intercept the subimage of component template image object as assembly subtemplate image object_sub；
S202, multiple dimensioned location target element: carry out up-sampling and down-sampling on n the yardstick of component template image object, obtain the component template image object of 2n+1 different scalek, n the yardstick of assembly subtemplate image object_sub carries out up-sampling and down-sampling, obtains the assembly subtemplate image object_sub of 2n+1 different scalek, wherein, k ∈ [1,2n+1]；
S203, extracts the gradient magnitude of image: computation module template image objectk, assembly subtemplate image object_subkAmplitude Amp is obtained with the gradient magnitude of image to be detectedk, amplitude Ampk_subWith amplitude Ampdetect；
S204, the band of position of location target element: by amplitude AmpkWith amplitude Ampk_subRespectively at amplitude AmpdetectIn carry out template matching, obtain the optimum network for location of target element under current scale as LockWith network for location as Lock_sub,
S205, calculates the similarity of network for location picture and template image:
If network for location is as Lock_subIt is that network for location is as LockSubimage, namelyThen extract component template image object under current scale respectivelykHOG feature and network for location as LockHOG feature, and calculate the Euclidean distance dist between two HOG featuresk；
If network for location is as Lock_subIt not that network for location is as LockSubimage, namelyThen component template image object under definition current scalekWith network for location as LockEuclidean distance distkIt is maximum MaxValue, i.e. a diskT=MaxVa；
S206, it is determined that the band of position image of target element: from 2n+1 Euclidean distance distkThe Euclidean distance dist that middle selection is minimumk, and it is compared with threshold value MaxTh, if this minimum Euclidean distance distk> MaxTh, then judge that this network for location is as LockInvalid, if this minimum Euclidean distance distk< MaxTh, then judge this location target LockBand of position image for this target element.
3. in contact net according to claim 1, assembly comes off defect identification method, it is characterised in that described positioning link step S3 includes following sub-step:
S301, obtains the sobel edge graph of band of position image；
S302, the horizontal edge rectangular histogram of statistics sobel edge graph and vertical edge rectangular histogram；
S303, compare with threshold value Th1 according to horizontal edge rectangular histogram, first time orients coboundary and the lower boundary in target connector region, compare with threshold value Th2 according to vertical edge rectangular histogram, first time orients left margin and the right margin in target connector region, obtains the target connector region of first time location；
S304, the positioning result according to S303, it is repeated once or repeatedly step S301-S303, obtains the target connector region after multiple bearing.
4. in contact net according to claim 3, assembly comes off defect identification method, it is characterised in that described step S303 includes following sub-step:
S3031, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the coboundary top in target connector region: from horizontal edge rectangular histogram HistRpInitial value start to compare with its average AvgR successively, until pth row HistRpDuring with the ratio of average AvgR more than threshold value Th1, then make the coboundary top=p in target connector region；
S3032, according to horizontal edge rectangular histogram HistRpCompare with threshold value Th1, orient the lower boundary bottom in target connector region: from horizontal edge rectangular histogram HistRpEnd value start to compare with its maximum MaxR successively, until pth row HistRpMore than average AvgR and HistRpDuring with the ratio of maximum MaxR more than threshold value Th0, then make the lower boundary bottom=p in target connector region；
S3033, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the left margin left in target connector region: from vertical edge rectangular histogram HistCqInitial value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the left margin left=q in target connector region；
S3034, according to vertical edge rectangular histogram HistCqCompare with threshold value Th2, orient the right margin right in target connector region: from vertical edge rectangular histogram HistCqEnd value start to compare with threshold value Th2 successively, as q row HistCqDuring more than threshold value Th2, then make the right margin right=q in target connector region；
S3035, according to coboundary top, lower boundary bottom, left margin left and right margin right, the target connector region I that first time orientsfirst。
5. in contact net according to claim 3, assembly comes off defect identification method, it is characterised in that described positioning link step S3 also includes sub-step S305: get rid of the positioning result of mistake in the target connector region after multiple bearing.
6. in contact net according to claim 1, assembly comes off defect identification method, it is characterised in that: described structural relation includes relative position relation and area proportionate relationship.
7. in contact net according to claim 1, assembly comes off defect identification method, it is characterized in that, judge described in feature analysis step S5 whether target connector exists the doubtful defect that comes off and include following sub-step: calculate the grey level histogram of each target connector, calculate the grey level histogram of each target connector, and judge whether the maximum gray level of the grey level histogram of each target connector identical or whether difference between same grey level is less than threshold value Th3, if, then judge that this target connector exists the doubtful defect that comes off, and record connector come off mark flag=1, otherwise judge that this target connector is absent from the doubtful defect that comes off, record connector comes off and indicates flag=0.
8. in contact net according to claim 1, assembly comes off defect identification method, it is characterised in that determine described in feature analysis step S5 whether the doubtful defect that comes off is true, including following sub-step:
S501, calculates horizontal gradient and the vertical gradient of each target connector respectively, calculates the argument θ of each pixel；
S502, adds up the distribution situation of argument θ in whole target connector image, argument rectangular histogram is normalized, obtains argument features vector；
S503, the mark flag=1 if connector comes off, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if having one group of mean square deviation more than threshold value Th4, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect；
S504, the mark flag=0 if connector comes off, then whether the argument principal direction comparing all target connectors is identical；
If it is identical, it is determined that whether the doubtful defect that comes off is false, it is determined that this target connector is normal；
If differing, then calculate in all target connectors that this target connector region is partitioned into, the mean square deviation of every argument features vector between target connector between two, if there being one group of mean square deviation more than threshold value Th5 and to contain one group of mean square deviation less than threshold value Th6, then determine whether the doubtful defect that comes off is true, it is determined that this target connector exists assembly and comes off defect.
9. in contact net, assembly comes off defect recognition system, apply method as described in any one in claim 1-8, it is characterised in that: described system includes the image collection module to be detected being sequentially connected with, multiple dimensioned assembly locating module, connector locating module, connector segmentation module and connector characteristics analysis module；
Described image collection module to be detected is for the image to be detected of securing component；
Described multiple dimensioned assembly locating module is for utilizing component template image slip scan target element on the different scale of image to be detected, matching the band of position image of target element in image to be detected, described component template image includes the template image of each assembly in contact net；
Described connector locating module for orienting target connector region according to edge analysis algorithm in the area image of position；
Described connector segmentation module, for the structural relation according to target element Yu target connector, is partitioned into target connector according to corresponding ratio in the area image of position；
Described connector characteristics analysis module is for obtaining grey level histogram and the Gradient Features of target connector, grey level histogram according to target connector judges whether target connector exists the doubtful defect that comes off, Gradient Features further according to target connector determines whether this doubtful defect that comes off is true, if very then judging that this target connector exists assembly and comes off defect, if vacation then judges that this target connector is normal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610049171.2A CN105719305B (en) | 2016-01-25 | 2016-01-25 | Component falls off defect identification method and system in contact net |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610049171.2A CN105719305B (en) | 2016-01-25 | 2016-01-25 | Component falls off defect identification method and system in contact net |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105719305A true CN105719305A (en) | 2016-06-29 |
| CN105719305B CN105719305B (en) | 2018-10-19 |
Family
ID=56154081
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610049171.2A Active CN105719305B (en) | 2016-01-25 | 2016-01-25 | Component falls off defect identification method and system in contact net |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105719305B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106504238A (en) * | 2016-10-31 | 2017-03-15 | 成都交大光芒科技股份有限公司 | Railway contact line defect inspection method based on image procossing and convolutional neural networks |
| CN106570857A (en) * | 2016-09-11 | 2017-04-19 | 西南交通大学 | High-speed railway overhead contact system lateral conductor fixation hook nut falling bad state detection method based on HOG features |
| CN107230205A (en) * | 2017-05-27 | 2017-10-03 | 国网上海市电力公司 | A kind of transmission line of electricity bolt detection method based on convolutional neural networks |
| CN107588809A (en) * | 2017-10-24 | 2018-01-16 | 中国矿业大学 | A kind of frictional high-strength bolts in rod member connecting node come off recognition methods |
| CN107680091A (en) * | 2017-10-12 | 2018-02-09 | 成都思晗科技股份有限公司 | A kind of stockbridge damper of transmission line of electricity comes off defect inspection method |
| CN109389566A (en) * | 2018-10-19 | 2019-02-26 | 辽宁奇辉电子系统工程有限公司 | Subway height adjusting valve fastening nut defective mode detection method based on boundary characteristic |
| CN110059631A (en) * | 2019-04-19 | 2019-07-26 | 中铁第一勘察设计院集团有限公司 | The contactless monitoring defect identification method of contact net |
| CN110610516A (en) * | 2019-08-26 | 2019-12-24 | 湖南大学 | Railway fastener nut center positioning method |
| CN110910373A (en) * | 2019-11-25 | 2020-03-24 | 西南交通大学 | Identification method of orthotropic steel bridge deck fatigue crack detection image |
| CN111754508A (en) * | 2020-07-06 | 2020-10-09 | 中兴飞流信息科技有限公司 | Contact net limiting nail-stopping allowance state detection system |
| CN112285111A (en) * | 2019-07-09 | 2021-01-29 | 株洲变流技术国家工程研究中心有限公司 | Pantograph front carbon sliding plate defect detection method, device, system and medium |
| CN112365481A (en) * | 2020-11-13 | 2021-02-12 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting bolt loss of cross beam assembly based on image processing |
| CN113192071A (en) * | 2021-06-29 | 2021-07-30 | 南方电网数字电网研究院有限公司 | Insulator binding wire missing identification method and device and computer equipment |
| CN116385353A (en) * | 2023-02-10 | 2023-07-04 | 南通大学 | Camera module abnormality detection method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567998A (en) * | 2012-01-06 | 2012-07-11 | 西安理工大学 | Head-shoulder sequence image segmentation method based on double-pattern matching and edge thinning |
| CN103400393A (en) * | 2013-08-21 | 2013-11-20 | 中科创达软件股份有限公司 | Image matching method and system |
| CN104318582A (en) * | 2014-11-14 | 2015-01-28 | 西南交通大学 | Detection method for bad state of rotating double-lug component pin of high-speed rail contact network on basis of image invariance target positioning |
| 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 |
-
2016
- 2016-01-25 CN CN201610049171.2A patent/CN105719305B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567998A (en) * | 2012-01-06 | 2012-07-11 | 西安理工大学 | Head-shoulder sequence image segmentation method based on double-pattern matching and edge thinning |
| CN103400393A (en) * | 2013-08-21 | 2013-11-20 | 中科创达软件股份有限公司 | Image matching method and system |
| CN104318582A (en) * | 2014-11-14 | 2015-01-28 | 西南交通大学 | Detection method for bad state of rotating double-lug component pin of high-speed rail contact network on basis of image invariance target positioning |
| 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 |
Non-Patent Citations (3)
| Title |
|---|
| HAO FENG 等: "Automatic Fastener Classification and Defect Detection in Vision-Based Railway Inspection Systems", 《IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》 * |
| 张春春 等: "基于图像处理的接触网定位线夹目标检测", 《铁道科学与工程学报》 * |
| 李东明: "车载铁路路障智能图像检测技术研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570857A (en) * | 2016-09-11 | 2017-04-19 | 西南交通大学 | High-speed railway overhead contact system lateral conductor fixation hook nut falling bad state detection method based on HOG features |
| CN106504238A (en) * | 2016-10-31 | 2017-03-15 | 成都交大光芒科技股份有限公司 | Railway contact line defect inspection method based on image procossing and convolutional neural networks |
| CN107230205A (en) * | 2017-05-27 | 2017-10-03 | 国网上海市电力公司 | A kind of transmission line of electricity bolt detection method based on convolutional neural networks |
| CN107680091B (en) * | 2017-10-12 | 2020-11-17 | 成都思晗科技股份有限公司 | Method for detecting falling defect of vibration damper of power transmission line |
| CN107680091A (en) * | 2017-10-12 | 2018-02-09 | 成都思晗科技股份有限公司 | A kind of stockbridge damper of transmission line of electricity comes off defect inspection method |
| CN107588809A (en) * | 2017-10-24 | 2018-01-16 | 中国矿业大学 | A kind of frictional high-strength bolts in rod member connecting node come off recognition methods |
| CN107588809B (en) * | 2017-10-24 | 2019-07-30 | 中国矿业大学 | A kind of frictional high-strength bolts in rod piece connecting node fall off recognition methods |
| CN109389566A (en) * | 2018-10-19 | 2019-02-26 | 辽宁奇辉电子系统工程有限公司 | Subway height adjusting valve fastening nut defective mode detection method based on boundary characteristic |
| CN109389566B (en) * | 2018-10-19 | 2022-01-11 | 辽宁奇辉电子系统工程有限公司 | Method for detecting bad state of fastening nut of subway height adjusting valve based on boundary characteristics |
| CN110059631A (en) * | 2019-04-19 | 2019-07-26 | 中铁第一勘察设计院集团有限公司 | The contactless monitoring defect identification method of contact net |
| CN112285111A (en) * | 2019-07-09 | 2021-01-29 | 株洲变流技术国家工程研究中心有限公司 | Pantograph front carbon sliding plate defect detection method, device, system and medium |
| CN110610516A (en) * | 2019-08-26 | 2019-12-24 | 湖南大学 | Railway fastener nut center positioning method |
| CN110610516B (en) * | 2019-08-26 | 2023-03-24 | 湖南大学 | Railway fastener nut center positioning method |
| CN110910373A (en) * | 2019-11-25 | 2020-03-24 | 西南交通大学 | Identification method of orthotropic steel bridge deck fatigue crack detection image |
| CN111754508A (en) * | 2020-07-06 | 2020-10-09 | 中兴飞流信息科技有限公司 | Contact net limiting nail-stopping allowance state detection system |
| CN112365481A (en) * | 2020-11-13 | 2021-02-12 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting bolt loss of cross beam assembly based on image processing |
| CN113192071A (en) * | 2021-06-29 | 2021-07-30 | 南方电网数字电网研究院有限公司 | Insulator binding wire missing identification method and device and computer equipment |
| CN116385353A (en) * | 2023-02-10 | 2023-07-04 | 南通大学 | Camera module abnormality detection method |
| CN116385353B (en) * | 2023-02-10 | 2024-01-30 | 南通大学 | Camera module abnormality detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105719305B (en) | 2018-10-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105719305A (en) | Assembly falloff defect identification method and system of overhead contact system | |
| CN105652154B (en) | Contact Running State security auditing system | |
| CN106919978B (en) | Method for identifying and detecting parts of high-speed rail contact net supporting device | |
| CN104318582B (en) | Detection method for bad state of rotating double-lug component pin of high-speed rail contact network | |
| CN106960179B (en) | Rail line Environmental security intelligent monitoring method and device | |
| CN102980896B (en) | High ferro overhead contact line device auricle fracture detection method | |
| CN109118479A (en) | Defects of insulator identification positioning device and method based on capsule network | |
| CN105404867B (en) | A kind of substation isolating-switch state identification method of view-based access control model | |
| CN103196372B (en) | A kind of optical imagery detection method of electrification railway contact net supportive device | |
| CN106679567A (en) | Contact net and strut geometric parameter detecting measuring system based on binocular stereoscopic vision | |
| CN102737252B (en) | Method for detecting faults caused by foreign body pollution between electrified railway insulator plates based on affine invariant moment | |
| CN106485694B (en) | A kind of high iron catenary double-jacket tube connector six-sided nut based on cascade classifier falls off defective mode detection method | |
| CN108846331B (en) | Video identification method for judging whether screw fastener of motor train unit chassis falls off or not | |
| CN111311567A (en) | Method for identifying fastener and steel rail diseases of track line image | |
| Gui et al. | Automated defect detection and visualization for the robotic airport runway inspection | |
| CN107154034B (en) | State detection method and system for stay wire positioning hook of high-speed rail contact network | |
| CN105354827A (en) | Method and system for intelligently identifying clamp nut shedding in catenary image | |
| CN108537170A (en) | A kind of power equipment firmware unmanned plane inspection pin missing detection method | |
| CN105574515A (en) | Pedestrian re-identification method in zero-lap vision field | |
| CN106340019A (en) | Method for detecting adverse state of high-speed rail overhead line system inclined cable fixing hook | |
| Zhao et al. | Image-based comprehensive maintenance and inspection method for bridges using deep learning | |
| CN112508911A (en) | Rail joint touch net suspension support component crack detection system based on inspection robot and detection method thereof | |
| CN110147748A (en) | A kind of mobile robot obstacle recognition method based on road-edge detection | |
| Premachandra et al. | Image based automatic road surface crack detection for achieving smooth driving on deformed roads | |
| CN104217206A (en) | Real-time attendance counting method based on high-definition videos |
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 |