CN110084171A - A kind of subway roof detection device for foreign matter and detection method - Google Patents
A kind of subway roof detection device for foreign matter and detection method Download PDFInfo
- Publication number
- CN110084171A CN110084171A CN201910327152.5A CN201910327152A CN110084171A CN 110084171 A CN110084171 A CN 110084171A CN 201910327152 A CN201910327152 A CN 201910327152A CN 110084171 A CN110084171 A CN 110084171A
- Authority
- CN
- China
- Prior art keywords
- image
- photoelectric sensor
- subway
- foreign matter
- plc
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 72
- 230000008878 coupling Effects 0.000 claims abstract description 48
- 238000010168 coupling process Methods 0.000 claims abstract description 48
- 238000005859 coupling reaction Methods 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 26
- 238000004321 preservation Methods 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 5
- 239000000284 extract Substances 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 10
- 238000004378 air conditioning Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 8
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims description 6
- 230000000717 retained effect Effects 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000013519 translation Methods 0.000 claims description 6
- 238000003708 edge detection Methods 0.000 claims description 4
- 101100328884 Caenorhabditis elegans sqt-3 gene Proteins 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 238000009792 diffusion process Methods 0.000 claims description 3
- 238000005530 etching Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000003706 image smoothing Methods 0.000 claims description 3
- 229910052742 iron Inorganic materials 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000001960 triggered effect Effects 0.000 claims description 2
- 238000002271 resection Methods 0.000 claims 1
- 238000012423 maintenance Methods 0.000 abstract description 5
- 239000000126 substance Substances 0.000 abstract 1
- 238000007689 inspection Methods 0.000 description 7
- 230000005622 photoelectricity Effects 0.000 description 6
- 238000012546 transfer Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 230000014759 maintenance of location Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000005693 optoelectronics Effects 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 239000013065 commercial product Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
-
- 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
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30268—Vehicle interior
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
A kind of subway roof detection device for foreign matter and detection method, on the basis of subway Train number recognition, propose a kind of subway roof foreign matter detecting method, devise a kind of subway roof detection device for foreign matter, image processing module is spliced to roof image and is distinguished key area and non-critical areas, non-critical areas, which is extracted, to be used to carry out foreign bodies detection to non-critical areas with detection module, key area, which is extracted, is used for key area foreign bodies detection and the region foreign bodies detection that is flexible coupling with detection module, information preservation and display module are for receiving and showing foreign substance information, realize subway roof foreign matter automatic detection, foreign matter existing for subway roof can be found in time, it alarms subway operation maintenance personnel, processing subway roof foreign matter in time, during solving subway circulation, roof foreign matter safety problem.
Description
Technical field
The present invention relates to target detections, Computer Vision Recognition field, more particularly to a kind of subway roof foreign bodies detection
Device and detection method.
Background technique
When subway circulation, there may be foreign matter at the top of subway carriage, foreign matter is forgotten when including worker's maintenance in roof
Tool, small part, the sundries such as cement block fallen down in driving conditions.In subway driving process, if deposited at the top of subway carriage
Car body top normal configuration component will be collided, traffic safety is endangered due to effect of inertia in foreign matter.Existing subway roof
Foreign matter detecting method is to pass through artificial substep, by section car body top inspection in maintenance section after metro operation.Existing people
Work checks subway roof foreign matter detecting method, and subway roof foreign matter can not be detected during operation, can not accomplish to protect in real time
Demonstrate,prove the safety of railcar；Manual inspection method there are missing inspection or detects the problems such as not comprehensive, simultaneously since human factor is big
Huge human cost is wasted, working efficiency is low.
Summary of the invention
The present invention for the technical problems in the prior art, on the basis of subway Train number recognition, inventive concept
Image registration, edge detection, the image procossings technique such as picture difference propose a kind of subway roof foreign matter detecting method, design
A kind of subway roof detection device for foreign matter, realizes subway roof foreign matter automatic detection, can find that subway roof is deposited in time
Foreign matter, alarm subway operation maintenance personnel, in time handle subway roof foreign matter, during solving subway circulation, vehicle
Push up foreign matter safety problem.
To realize the present invention purpose use technical solution first is that:
A kind of subway roof foreign matter detecting method, it is characterized in that:
(1) acquisition subway serves as a fill-in evidence, stores to database:
The route whole subway type roof key area template data is acquired, after the template and detection of acquisition
There are the storages of the image information of foreign matter into the database module of image capturing system, and key area template data includes subway
License number mark, air-conditioning template image and pantograph template image, foreign matter image information is by there are the license numbers of the subway of foreign matter
Coach number and foreign matter image composition where mark, foreign matter, are stored in the database module of image capturing system；
(2) subway Train number recognition:
During subway advances, when the wheel of subway triggers wheel detector, wheel detector transmits trigger signal
To PLC, signal is transmitted to Train number recognition host, Train number recognition host command Train number recognition antenna search by PLC, and license number is known
Other antenna receives the license number label signal of subway, and the license number label signal for receiving subway is returned to identification master by Train number recognition antenna
Machine carries out Train number recognition；
(3) Image Acquisition:
During subway advances, when subway first segment compartment triggers first photoelectric sensor, first photoelectric sensing
Trigger signal is passed to PLC by device, and signal is transmitted to image capturing system by PLC, and image capturing system instruction high definition industry is taken the photograph
Camera shoots first photo；When subway first segment compartment triggers second photoelectric sensor, second photoelectric sensor will
Trigger signal passes to PLC, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera
Shoot second photo；When subway first segment compartment triggers third photoelectric sensor, third photoelectric sensor will be triggered
Signal passes to PLC, and signal is transmitted to image capturing system by PLC, and image capturing system instructs the shooting of high definition industrial camera
Third photo；When subway first segment compartment triggers the 4th photoelectric sensor, the 4th photoelectric sensor is by trigger signal
PLC is passed to, signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera shooting the 4th
Open photo；When subway first segment compartment triggers the 5th photoelectric sensor, the 5th photoelectric sensor transmits trigger signal
To PLC, signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera to shoot the 5th Zhang Zhao
Piece；When subway first segment compartment triggers the 6th photoelectric sensor, the 6th photoelectric sensor passes to trigger signal
Signal is transmitted to image capturing system by PLC, PLC, and image capturing system instruction high definition industrial camera shoots the 6th photo；
When subway first segment compartment triggers the 7th photoelectric sensor, trigger signal is passed to PLC by the 7th photoelectric sensor,
Signal is transmitted to image capturing system by PLC, and image capturing system instruction high definition industrial camera shoots the 7th photo；It is local
When iron first segment compartment triggers the 8th optoelectronic switch, trigger signal is passed to PLC by the 8th optoelectronic switch, and PLC is by signal
It is transmitted to image capturing system, image capturing system instruction high definition industrial camera shoots the 8th photo；Subway first segment vehicle
The whole triggering photoelectric sensor in compartment, by the image of photoelectric sensor control shooting, per between adjacent two relative to subway
There is the lap of 0.5m on top, since compartment junction is lower than compartment, all photosensor resets, the second section of subway compartment
Start to trigger photoelectric sensor, high definition industrial camera shoots photo, and respectively section compartment sequentially triggers photoelectric sensor, high definition thereafter
Industrial camera shoots photo, and the photo of shooting is temporarily stored to image capturing system；
(4) image procossing:
Subway totally 6 section compartment is set, first image in first segment compartment is headstock, does not have to detection；Final section compartment
Last image be headstock, do not have to detection；Last photo in each section compartment is to be flexible coupling, and does not need to splice；Cause
There was only 2-7 images in first segment compartment and Section of six compartment, the second section compartment to Section of five compartment in this part to be spliced together
1-7 images, during photograph taking, may be by key area: air-conditioning, pantograph separately take two images
On, therefore, two adjacent images are spliced, to restore key area, image capture module is passed over 6 × 8
Photo, handles respectively according to coach number, 8 images of same coach number is numbered, number is respectively 1,2 ..., 8, so
Two adjacent images are spliced afterwards, i.e., first segment and Section six are (2,3), (3,4), (4,5), (5,6), (6,7) totally 5
Secondary splicing, the second section are that (1,2), (2,3), (3,4), (4,5), (5,6), (6,7) are spliced two-by-two to Section five, and splicing is successful
With the 8th image in first segment compartment to Section of five compartment be sent to key area extract and detection, splice it is unsuccessful be sent to it is non-
Key area extracts and detection；
(5) non-critical areas is extracted and is detected:
Since possibility described in step (4) is by key area: air-conditioning, pantograph separately take on two images,
That is there is key area on an image a possibility that, never splice in successful image, finds out these keys
The image in region is sent to key area and extracts and detect:
(5.1) image that will carry out non-critical areas foreign bodies detection carries out image smoothing using 5 × 5 Gaussian filter,
To reduce influence of the noise to edge detection, filtering mode is as follows:
Wherein, P is original image, and P ' is filtered image；
(5.2) K is definedxAnd KyTwo 3 × 3 gradient operators are as shown by the following formula:
KxGradient operator can extract lines vertical in image, calculate the horizontal gradient component of image, KyGradient is calculated
Son can extract lines horizontal in image, calculate in image the gradient component G on both horizontally and verticallyxAnd GyFormula is such as
Under:
Gx=P′*Kx, Gy=P ' * Ky
According to both horizontally and vertically gradient component GxAnd Gy, the gradient value G and direction θ of image are calculated, then according to every
Gradient direction at a point is compared the gradient value of former and later two points and current point in this direction, is retained centered on current point
Maximum gradient value in three points, off-peak gradient value are then suppressed and are set to 0, and the boundary finally obtained is all not only thin but also bright
Line.Gradient value and gradient direction calculation are as follows:
θ=arctan (Gy, Gx)
(5.3) the image P " generated after non-greatest gradient value inhibits has been carried out, can still have a large amount of noise, will scheme
The pixel of picture P " is provided with two threshold values, and small threshold value is known as threshold value lower bound, and big threshold value is known as the threshold value upper bound, when in image P "
When pixel value is greater than the threshold value upper bound, which is fully retained, referred to as strong boundary；When pixel value is less than threshold value lower bound, the picture
Vegetarian refreshments is not boundary, is completely inhibited；When pixel value is between the threshold value upper bound and threshold value lower bound, the pixel is as time
Select boundary, referred to as weak boundary；
Next, the pixel value and extension direction according to strong boundary screen weak boundary, it is connected to strong boundary weak
Boundary is considered as boundary, retains pixel and inhibits with the disconnected weak boundary in strong boundary not as boundary to pixel, this
When, obtain more visible image border；
Script parts of images edge should belong to a line, and after above-mentioned edge extracting, a line becomes multiple lines
Section, therefore we expand image border using 5 × 5 region, with reach multiple line segments are connected it is in alignment
Purpose will lead to connected region and become more, for convenience of the calculating in connection region, use 5 × 5 since by expanding, lines are thicker
Region carries out etching operation to the image after expansion；
(5.4) it is searched according to unconnected pixels, from a pixel until shape by the relationship being connected between pixel
At a ring, it is a connected region, finds out connected region all in image, and according to the maximum width of connected region
It is limited with maximum height, to realize that the screening to connected region calculates separately theirs to the connected region screened
Area, since the non-critical areas of foreign is nearly free from connected region, we are according to the connection area being set in advance
Domain area threshold carries out foreign bodies detection, when area is greater than preset threshold value, it is believed that and the connected region is foreign matter, and
The connection region is labeled in the picture with square-shaped frame, if there is foreign matter, then by foreign matter image and there are foreign matters
Vehicle, car information pass to information preservation and display module, otherwise, " foreign " information are sent to information preservation and display
Module；
(6) key area extraction and detection module:
(6.1) characteristic point of key area and key area template is extracted respectively using sift operator:
Key area is matched with the characteristic point of template extraction, determines matched characteristic point, by calculating all
With the Euclidean distance information between point, maximum distance (max_dist) and the minimum range between matched characteristic point are calculated
(min_dist), matching characteristic point is filtered according to maximum distance and minimum range, retains distance and is less than 0.3*max_
The matching characteristic point of dist；
(6.2) from filtered matching characteristic point, according to the one-to-one relationship of Feature Points Matching, matching image is obtained
The index of characteristic point calculates corresponding perspective transformation matrix, and key area image is carried out matrix change using perspective transformation matrix
It changes, is mapped to the characteristic point on key area image in the corresponding characteristic point of key area template image, completes matching for image
Quasi- operation, then carries out image difference operation for transformed image and template, i.e., after matched characteristic point being corresponded to, with
All pixels point in template image carries out the reducing of image；
(6.3) largest connected region is asked to differentiated image, and according to the maximum width and maximum height of connected region
It is limited, to realize that the screening to connected region calculates separately their area, then to the connected region screened
Foreign bodies detection is carried out using the mode of threshold value, when area is greater than preset threshold value, it is believed that the connected region is different
Object, and the connection region is labeled in the picture with square-shaped frame；
(7) be flexible coupling region progress foreign bodies detection:
(7.1) image that the presence received from image processing module is flexible coupling, wide and high respectively col and row,
Both sides be flexible coupling there are small part plane domain, our region foreign matter detecting method that is flexible coupling only carries out foreign matter inspection to being flexible coupling
It surveys, therefore we need to be flexible coupling region to being cut into.Since the position of photoelectric sensor is fixed, what is taken every time is soft
The position of join domain also determines, therefore, image is flexible coupling region both sides according to the known regional location that is flexible coupling by we
Plane domain excision；
(7.2) area image that is flexible coupling being cut into is carried out carrying out first three step of foreign bodies detection such as non-critical areas image
The edge detecting operation of process just will appear a large amount of due to being flexible coupling region based on vertical line, when only there is foreign matter
Horizontal line section, therefore we carry out Hough straight-line detection to the region that is flexible coupling, and to enhance the quantity of vertical direction line segment, reduce water
The quantity of flat direction line segment；
(7.3) edge image that is flexible coupling being made of horizontal line section and vertical segment has been obtained, has been used in the picture
The pixel frame (width of pixel frame is 30) of square carries out pixels statistics, this needs the double-deck circulation to realize:
When the pixel value counted in square-shaped frame is greater than preset threshold value, it is believed where there is foreign matters, and terminate
This statistics, if pixel until traversing completion completely in square-shaped frame, the pixel number summation greater than 0 is still less than threshold value, then
Think that there is no foreign matters in the square-shaped frame, carry out translation for square-shaped frame and continue searching.During entire translation search, such as
Not there is foreign matter in fruit, square-shaped frame will traverse completely the region that is entirely flexible coupling, and also need one in entire ergodic process
A double-deck circulation is completed:
It is being flexible coupling in edge image, according to the width and height of square-shaped frame, square-shaped frame is being translated, until most
When the top left co-ordinate of the latter frame is (row-30, col-30) and bottom right angular coordinate is (row-1, col-1), traversal is completed,
And return to the information for the region foreign that is flexible coupling.If finding foreign matter in ergodic process, terminate traversal, there are foreign matters for return
Information；
(8) information preservation and display module will receive non-critical areas extraction and detection module and key area extracts and inspection
Surveying module whether there is the information of foreign matter, if there is foreign matter, then the image, foreign matter place coach number and train of foreign matter will be present
License number identification information is transmitted to database module, stores data into the historical record tables of data in database, meanwhile, it will send out
The information such as existing foreign matter image are shown on system foreground, and prompt the warning note of " it was found that foreign matter, is please cleared up as early as possible ", if do not had
It is found foreign matter, then prompts " foreign "；
So far, the roof foreign bodies detection of the column subway is completed, and the information that will test completion passes to Train number recognition module
And image capture module.
To realize the present invention purpose use technical solution second is that:
A kind of subway roof detection device for foreign matter comprising: Train number recognition host, Train number recognition antenna, license number label,
It is characterized in, it further include: wheel detector, image capturing system, PLC, the first photoelectric sensor, the second photoelectric sensor,
Three photoelectric sensors, the 4th photoelectric sensor, the 5th photoelectric sensor, the 6th photoelectric sensor, the 7th photoelectric sensor, the 8th
Photoelectric sensor, high definition industrial camera, the wheel detector are arranged on subway approach track, the wheel-sensors
Device is electrically connected with PLC, and the Train number recognition host is electrically connected with PLC, the Train number recognition host and image capturing system
Communication connection, on the wheel detector track front, successively set up the first photoelectric sensor, the second photoelectric sensor,
Third photoelectric sensor, the 4th photoelectric sensor, the 5th photoelectric sensor, the 6th photoelectric sensor, the 7th photoelectric sensor,
Eight photoelectric sensors, between the wheel detector and the first photoelectric sensor, distance the first photoelectric sensor 2.5m, rail
Right above road, it is higher than subway carriage junction and high definition industrial camera, the high definition industrial camera and Image Acquisition is set
System electrical connection, first photoelectric sensor, the second photoelectric sensor, third photoelectric sensor, the 4th photoelectric sensor,
5th photoelectric sensor, the 6th photoelectric sensor, the 7th photoelectric sensor, the 8th photoelectric sensor are electrically connected with PLC respectively, institute
The PLC stated is electrically connected with image capturing system.
The photoelectric sensor is laser diffusion photoelectric sensor.
A kind of subway roof detection device for foreign matter of the present invention and detection method have the beneficial effect that:
1, a kind of application of subway roof foreign matter detecting method realizes subway roof foreign matter automatic detection, reduces
Human cost improves detection efficiency, ensure that detection quality；
2, a kind of subway roof foreign matter detecting method can find that subway roof has the maintenance tool lost or falls in time
The foreign matters such as stone, security risk is excluded, and solve the image shot during subway operation, in image mosaic reduction process
In, there are identification mistake or unrecognized problems caused by larger deformation.
Detailed description of the invention
Fig. 1 is a kind of subway roof detection device for foreign matter installation diagram；
Fig. 2 is a kind of flow chart of subway roof foreign matter detecting method；
In figure: 1. train electronic tags, 2. wheel detectors, 3. first photoelectric sensors, 4. second photoelectric sensors, 5.
Third photoelectric sensor, 6. the 4th photoelectric sensors, 7. the 5th photoelectric sensors, 8. the 6th photoelectric sensors, 9. the 7th photoelectricity
Sensor, 10. the 8th photoelectric sensors, 11. high definition industrial cameras, 12. tracks, 13. subways.
Specific embodiment
Below in conjunction with attached drawing 1- attached drawing 2 and specific embodiment, invention is further described in detail, described herein
Specific embodiment only to explain the present invention, be not intended to limit the present invention.
Referring to attached drawing 1, in the master control room of setting Train number recognition host, image capturing system and PLC, the vehicle are configured
Wheel sensor 2 is arranged on 13 approach track 12 of subway, and the wheel detector 2 is electrically connected with PLC, the Train number recognition
Host is electrically connected with PLC, and the PLC is electrically connected with image capturing system, the Train number recognition host and Image Acquisition system
System communication connection, front, successively sets up the first photoelectric sensor 3, the second photoelectricity on 2 track of setting wheel detector
Sensor 4, third photoelectric sensor 5, the 4th photoelectric sensor 6, the 5th photoelectric sensor 7, the 6th photoelectric sensor the 8, the 7th
Photoelectric sensor 9, the 8th photoelectric sensor 10, between the wheel detector 2 and the first photoelectric sensor 3, distance
One photoelectric sensor 2.5m, 12 surface of track, is higher than subway carriage junction and high definition industrial camera 11, the height is arranged
Clear industrial camera 11 is electrically connected with image capturing system, first photoelectric sensor 3, the second photoelectric sensor 4, third
Photoelectric sensor 5, the 4th photoelectric sensor 6, the 5th photoelectric sensor 7, the 6th photoelectric sensor 8, the 7th photoelectric sensor 9,
8th photoelectric sensor 10 is electrically connected with PLC respectively.
Reference attached drawing 2,
(1) acquisition subway serves as a fill-in evidence, stores to database:
The route whole subway type roof key area template data is acquired, after the template and detection of acquisition
There are the storages of the image information of foreign matter into the database module of image capturing system, and key area template data includes subway
License number mark, air-conditioning template image and pantograph template image, foreign matter image information is by there are the license numbers of the subway of foreign matter
Coach number and foreign matter image composition where mark, foreign matter, are stored in the database module of image capturing system；
(2) subway Train number recognition:
During subway 13 advances, when the wheel of subway 13 triggers wheel detector 2, wheel detector 2 believes triggering
Number passing to PLC, signal is transmitted to Train number recognition host by PLC, Train number recognition host command Train number recognition antenna search,
Train number recognition antenna receives the license number label signal 1 of subway, and Train number recognition antenna returns the license number label signal 1 for receiving subway
Train number recognition is carried out to identification host；
(3) Image Acquisition:
During subway 13 advances, when 13 first segment compartment of subway triggers first photoelectric sensor 3, first photoelectricity
Trigger signal is passed to PLC by sensor 3, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition work
Industry video camera 11 shoots first photo；When 13 first segment compartment of subway triggers second photoelectric sensor 4, second photoelectricity
Trigger signal is passed to PLC by sensor 4, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition work
Industry video camera 11 shoots second photo；When 13 first segment compartment of subway triggers third photoelectric sensor 5, third photoelectricity
Trigger signal is passed to PLC by sensor 5, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition work
Industry video camera shoots third photo；When 13 first segment compartment of subway triggers the 4th photoelectric sensor 6, the 4th photoelectric transfer
Trigger signal is passed to PLC by sensor 6, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industry
Video camera 11 shoots the 4th photo；When 13 first segment compartment of subway triggers the 5th photoelectric sensor 7, the 5th photoelectric transfer
Trigger signal is passed to PLC by sensor 7, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industry
Video camera 11 shoots the 5th photo；When 13 first segment compartment of subway triggers the 6th photoelectric sensor 8, the 6th photoelectric transfer
Trigger signal is passed to PLC by sensor 8, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industry
Video camera 11 shoots the 6th photo；When 13 first segment compartment of subway triggers the 7th photoelectric sensor 9, the 7th photoelectric transfer
Trigger signal is passed to PLC by sensor 9, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industry
Video camera 11 shoots the 7th photo；When 13 first segment compartment of subway triggers the 8th photoelectric sensor 10, the 8th photoelectricity
Trigger signal is passed to PLC by sensor 10, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition
Industrial camera 11 shoots the 8th photo；13 first segment compartment of subway all triggers photoelectric sensor, takes the photograph by high definition industry
The image of the control shooting of camera 11, per the lap for having 0.5m between adjacent two relative to subway roof, due to compartment
Place be flexible coupling lower than compartment, all photosensor resets, 13 second section compartment of subway starts to trigger photoelectric sensor, high definition work
Industry video camera 11 shoots photo, and respectively section compartment sequentially triggers photoelectric sensor thereafter, and high definition industrial camera 11 shoots photo, and
The photo of shooting is temporarily stored to image capturing system；
(4) image procossing:
Subway totally 6 section compartment is set, first image in first segment compartment is headstock, does not have to detection；Final section compartment
Last image be headstock, do not have to detection；Last photo in each section compartment is to be flexible coupling, and does not need to splice；Cause
There was only the 2-7 in first segment compartment and Section of six compartment images, the second section compartment to Section of five compartment in this part to be spliced together
1-7 images, during photograph taking, may be by key area: air-conditioning, pantograph be separately taken on two images,
Therefore, two adjacent images are spliced, to restore key area, 6 × 8 photographs that image capture module is passed over
Piece is handled respectively according to coach number, and 8 images of same coach number are numbered, number be respectively 1,2 ..., 8, then will
Two adjacent images are spliced, i.e., first segment and Section six are (2,3), (3,4), (4,5), (5,6), (6,7) totally 5 spellings
It connects, the second section is that (1,2), (2,3), (3,4), (4,5), (5,6), (6,7) are spliced two-by-two to Section five, and splicing is successfully and the
8th image in one section compartment to Section of five compartment be sent to key area extract and detection, splice it is unsuccessful be sent to it is non-key
Extracted region and detection；
(5) non-critical areas is extracted and is detected:
Since possibility described in step (4) is by key area: air-conditioning, pantograph separately take on two images,
That is there is key area on an image a possibility that, never splice in successful image, finds out these keys
The image in region is sent to key area and extracts and detect:
(5.1) image that will carry out non-critical areas foreign bodies detection carries out image smoothing using 5 × 5 Gaussian filter,
To reduce influence of the noise to edge detection, filtering mode is as follows:
Wherein, P is original image, and P ' is filtered image；
(5.2) K is definedxAnd KyTwo 3 × 3 gradient operators are as shown by the following formula:
KxGradient operator can extract lines vertical in image, calculate the horizontal gradient component of image, KyGradient is calculated
Son can extract lines horizontal in image, calculate in image the gradient component G on both horizontally and verticallyxAnd GyFormula is such as
Under:
Gx=P ' * Kx, Gy=P ' * Ky
According to both horizontally and vertically gradient component GxAnd Gy, the gradient value G and direction θ of image are calculated, then according to every
Gradient direction at a point is compared the gradient value of former and later two points and current point in this direction, is retained centered on current point
Maximum gradient value in three points, off-peak gradient value are then suppressed and are set to 0, and the boundary finally obtained is all not only thin but also bright
Line.Gradient value and gradient direction calculation are as follows:
θ=arctan (Gy, Gx)
(5.3) the image P " generated after non-greatest gradient value inhibits has been carried out, can still have a large amount of noise, will scheme
The pixel of picture P " is provided with two threshold values, and small threshold value is known as threshold value lower bound, and big threshold value is known as the threshold value upper bound, when in image P "
When pixel value is greater than the threshold value upper bound, which is fully retained, referred to as strong boundary；When pixel value is less than threshold value lower bound, the picture
Vegetarian refreshments is not boundary, is completely inhibited；When pixel value is between the threshold value upper bound and threshold value lower bound, the pixel is as time
Select boundary, referred to as weak boundary；
Next, the pixel value and extension direction according to strong boundary screen weak boundary, it is connected to strong boundary weak
Boundary is considered as boundary, retains pixel and inhibits with the disconnected weak boundary in strong boundary not as boundary to pixel, this
When, obtain more visible image border；
Script parts of images edge should belong to a line, and after above-mentioned edge extracting, a line becomes multiple lines
Section, therefore we expand image border using 5 × 5 region, with reach multiple line segments are connected it is in alignment
Purpose will lead to connected region and become more, for convenience of the calculating in connection region, use 5 × 5 since by expanding, lines are thicker
Region carries out etching operation to the image after expansion；
(5.4) it is searched according to unconnected pixels, from a pixel until shape by the relationship being connected between pixel
At a ring, it is a connected region, finds out connected region all in image, and according to the maximum width of connected region
It is limited with maximum height, to realize that the screening to connected region calculates separately theirs to the connected region screened
Area, since the non-critical areas of foreign is nearly free from connected region, we are according to the connection area being set in advance
Domain area threshold carries out foreign bodies detection, when area is greater than preset threshold value, it is believed that and the connected region is foreign matter, and
The connection region is labeled in the picture with square-shaped frame, if there is foreign matter, then by foreign matter image and there are foreign matters
Vehicle, car information pass to information preservation and display module, otherwise, " foreign " information are sent to information preservation and display
Module；
(6) key area extraction and detection module:
(6.1) characteristic point of key area and key area template is extracted respectively using sift operator:
Key area is matched with the characteristic point of template extraction, determines matched characteristic point, by calculating all
With the Euclidean distance information between point, maximum distance (max_dist) and the minimum range between matched characteristic point are calculated
(min_dist), matching characteristic point is filtered according to maximum distance and minimum range, retains distance and is less than 0.3*max_
The matching characteristic point of dist；
(6.2) from filtered matching characteristic point, according to the one-to-one relationship of Feature Points Matching, matching image is obtained
The index of characteristic point calculates corresponding perspective transformation matrix, and key area image is carried out matrix change using perspective transformation matrix
It changes, is mapped to the characteristic point on key area image in the corresponding characteristic point of key area template image, completes matching for image
Quasi- operation, then carries out image difference operation for transformed image and template, i.e., after matched characteristic point being corresponded to, with
All pixels point in template image carries out the reducing of image；
(6.3) largest connected region is asked to differentiated image, and according to the maximum width and maximum height of connected region
It is limited, to realize that the screening to connected region calculates separately their area, then to the connected region screened
Foreign bodies detection is carried out using the mode of threshold value, when area is greater than preset threshold value, it is believed that the connected region is different
Object, and the connection region is labeled in the picture with square-shaped frame；
(7) be flexible coupling region progress foreign bodies detection:
(7.1) image that the presence received from image processing module is flexible coupling, wide and high respectively col and row,
Both sides be flexible coupling there are small part plane domain, our region foreign matter detecting method that is flexible coupling only carries out foreign matter inspection to being flexible coupling
It surveys, therefore we need to be flexible coupling region to being cut into.Since the position of photoelectric sensor is fixed, what is taken every time is soft
The position of join domain also determines, therefore, image is flexible coupling region both sides according to the known regional location that is flexible coupling by we
Plane domain excision；
(7.2) area image that is flexible coupling being cut into is carried out carrying out first three step of foreign bodies detection such as non-critical areas image
The edge detecting operation of process just will appear a large amount of due to being flexible coupling region based on vertical line, when only there is foreign matter
Horizontal line section, therefore we carry out Hough straight-line detection to the region that is flexible coupling, and to enhance the quantity of vertical direction line segment, reduce water
The quantity of flat direction line segment；
(7.3) edge image that is flexible coupling being made of horizontal line section and vertical segment has been obtained, has been used in the picture
The pixel frame (width of pixel frame is 30) of square carries out pixels statistics, this needs the double-deck circulation to realize:
When the pixel value counted in square-shaped frame is greater than preset threshold value, it is believed where there is foreign matters, and terminate
This statistics, if pixel until traversing completion completely in square-shaped frame, the pixel number summation greater than 0 is still less than threshold value, then
Think that there is no foreign matters in the square-shaped frame, carry out translation for square-shaped frame and continue searching.During entire translation search, such as
Not there is foreign matter in fruit, square-shaped frame will traverse completely the region that is entirely flexible coupling, and also need one in entire ergodic process
A double-deck circulation is completed:
It is being flexible coupling in edge image, according to the width and height of square-shaped frame, square-shaped frame is being translated, until most
When the top left co-ordinate of the latter frame is (row-30, col-30) and bottom right angular coordinate is (row-1, col-1), traversal is completed,
And return to the information for the region foreign that is flexible coupling.If finding foreign matter in ergodic process, terminate traversal, there are foreign matters for return
Information；
(8) information preservation and display module will receive non-critical areas extraction and detection module and key area extracts and inspection
Surveying module whether there is the information of foreign matter, if there is foreign matter, then the image, foreign matter place coach number and train of foreign matter will be present
License number identification information is transmitted to database module, stores data into the historical record tables of data in database, meanwhile, it will send out
The information such as existing foreign matter image are shown on system foreground, and prompt the warning note of " it was found that foreign matter, is please cleared up as early as possible ", if do not had
It is found foreign matter, then prompts " foreign "；
So far, the roof foreign bodies detection of the column subway is completed, and the information that will test completion passes to Train number recognition module
And image capture module.
It is used in this preferred embodiment, photoelectric sensor is sharp Xiang LXDM-31 laser diffusion photoelectric sensor, high definition
Industrial camera uses 3.0 DFK 33UX287 colour industrial camera of The Imaging Source USB, and wheel detector is
CG3 active vehicle wheels sensor, PLC are the CP1E N-type programmable controller of Omron, and Train number recognition host, ground identifies day
Line, electronic tag are passive electronic label, are commercial product.
The above is only preferred embodiment of the invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
This is considered as protection scope of the present invention.
Claims (3)
1. a kind of subway roof foreign matter detecting method, it is characterized in that:
(1) acquisition subway serves as a fill-in evidence, stores to database:
The route whole subway type roof key area template data is acquired, will be existed after the template and detection of acquisition
The image information of foreign matter is stored into the database module of image capturing system, and key area template data includes the vehicle of subway
Number mark, air-conditioning template image and pantograph template image, foreign matter image information by there are the license number of the subway of foreign matter mark,
Coach number and foreign matter image composition where foreign matter, are stored in the database module of image capturing system；
(2) subway Train number recognition:
During subway advances, when the wheel of subway triggers wheel detector, wheel detector passes to trigger signal
Signal is transmitted to Train number recognition host, Train number recognition host command Train number recognition antenna search, Train number recognition by PLC, PLC
Antenna receives the license number label signal of subway, and the license number label signal for receiving subway is returned to identification host by Train number recognition antenna
Carry out Train number recognition；
(3) Image Acquisition:
During subway advances, when subway first segment compartment triggers first photoelectric sensor, first photoelectric sensor will
Trigger signal passes to PLC, and signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera
Shoot first photo；When subway first segment compartment triggers second photoelectric sensor, second photoelectric sensor will be triggered
Signal passes to PLC, and signal is transmitted to image capturing system by PLC, and image capturing system instructs the shooting of high definition industrial camera
Second photo；When subway first segment compartment triggers third photoelectric sensor, third photoelectric sensor is by trigger signal
PLC is passed to, signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera to shoot third
Open photo；When subway first segment compartment triggers the 4th photoelectric sensor, the 4th photoelectric sensor transmits trigger signal
To PLC, signal is transmitted to image capturing system by PLC, and image capturing system instructs high definition industrial camera to shoot the 4th Zhang Zhao
Piece；When subway first segment compartment triggers the 5th photoelectric sensor, the 5th photoelectric sensor passes to trigger signal
Signal is transmitted to image capturing system by PLC, PLC, and image capturing system instruction high definition industrial camera shoots the 5th photo；
When subway first segment compartment triggers the 6th photoelectric sensor, trigger signal is passed to PLC by the 6th photoelectric sensor,
Signal is transmitted to image capturing system by PLC, and image capturing system instruction high definition industrial camera shoots the 6th photo；It is local
When iron first segment compartment triggers the 7th photoelectric sensor, trigger signal is passed to PLC by the 7th photoelectric sensor, and PLC will
Signal is transmitted to image capturing system, and image capturing system instruction high definition industrial camera shoots the 7th photo；When subway
When one section compartment triggers the 8th photoelectric sensor, trigger signal is passed to PLC by the 8th photoelectric sensor, and PLC is by signal
It is transmitted to image capturing system, image capturing system instruction high definition industrial camera shoots the 8th photo；Subway first segment vehicle
The whole triggering photoelectric sensor in compartment, by the image of high definition industrial camera control shooting, per between adjacent two relative to ground
There is the lap of 0.5m on iron car top, since compartment junction is lower than compartment, all photosensor resets, the second section of subway
Compartment starts to trigger photoelectric sensor, and high definition industrial camera shoots photo, and respectively section compartment sequentially triggers photoelectric sensor thereafter,
High definition industrial camera shoots photo, and the photo of shooting is temporarily stored to image capturing system；
(4) image procossing:
Subway totally 6 section compartment is set, first image in first segment compartment is headstock, does not have to detection；Final section compartment is most
Latter image is headstock, does not have to detection；Last photo in each section compartment is to be flexible coupling, and does not need to splice；Therefore it needs
The part to be spliced only has 2-7 images in first segment compartment and Section of six compartment, the second section compartment to Section of five compartment
1-7 images, during photograph taking, may be by key area: air-conditioning, pantograph be separately taken on two images,
Therefore, two adjacent images are spliced, to restore key area, 6 × 8 photographs that image capture module is passed over
Piece is handled respectively according to coach number, and 8 images of same coach number are numbered, number be respectively 1,2 ..., 8, then will
Two adjacent images are spliced, i.e., first segment and Section six are (2,3), (3,4), (4,5), (5,6), (6,7) totally 5 spellings
It connects, the second section is that (1,2), (2,3), (3,4), (4,5), (5,6), (6,7) are spliced two-by-two to Section five, and splicing is successfully and the
8th image in one section compartment to Section of five compartment be sent to key area extract and detection, splice it is unsuccessful be sent to it is non-key
Extracted region and detection；
(5) non-critical areas is extracted and is detected:
Since possibility described in step (4) is by key area: air-conditioning, pantograph separately take on two images, that is,
It says with key area on an image a possibility that, never splices in successful image, find out these key areas
Image, be sent to key area extract and detection:
(5.1) image that will carry out non-critical areas foreign bodies detection carries out image smoothing using 5 × 5 Gaussian filter, with drop
Influence of the low noise to edge detection, filtering mode are as follows:
Wherein, P is original image, and P ' is filtered image；
(5.2) K is definedxAnd KyTwo 3 × 3 gradient operators are as shown by the following formula:
KxGradient operator can extract lines vertical in image, calculate the horizontal gradient component of image, KyGradient operator energy
It enough extracts lines horizontal in image, calculates in image the gradient component G on both horizontally and verticallyxAnd GyFormula is as follows:
Gx=P ' * Kx, Gy=P ' * Ky
According to both horizontally and vertically gradient component GxAnd Gy, the gradient value G and direction θ of image are calculated, then according to each point
The gradient direction at place is compared the gradient value of former and later two points and current point in this direction, is retained three centered on current point
Maximum gradient value in point, off-peak gradient value are then suppressed and are set to 0, and the boundary finally obtained is all not only thin but also bright
Line.Gradient value and gradient direction calculation are as follows:
θ=arctan (Gy, Gx)
(5.3) the image P " generated after non-greatest gradient value inhibits has been carried out, can still have a large amount of noise, by image P "
Pixel be provided with two threshold values, small threshold value is known as threshold value lower bound, and big threshold value is known as the threshold value upper bound, the pixel in image P "
When value is greater than the threshold value upper bound, which is fully retained, referred to as strong boundary；When pixel value is less than threshold value lower bound, the pixel
It is not boundary, is completely inhibited；When pixel value is between the threshold value upper bound and threshold value lower bound, the pixel is as candidate side
Boundary, referred to as weak boundary；
Next, the pixel value and extension direction according to strong boundary screen weak boundary, the weak boundary being connected to strong boundary
It is considered as boundary, retains pixel and pixel is inhibited with the disconnected weak boundary in strong boundary not as boundary, at this point,
More visible image border is arrived；
Script parts of images edge should belong to a line, and after above-mentioned edge extracting, a line becomes multiple line segments, because
We expand image border using 5 × 5 region for this, in alignment to achieve the purpose that connect on multiple line segments,
Since by expanding, lines are thicker, it will lead to connected region and become more, for convenience of the calculating in connection region, use 5 × 5 region
Etching operation is carried out to the image after expansion；
(5.4) it is searched from a pixel according to unconnected pixels by the relationship being connected between pixel, until forming one
A ring is a connected region, finds out connected region all in image, and according to the maximum width of connected region and most
Big height is limited, to realize that the screening to connected region calculates separately their face to the connected region screened
Product, since the non-critical areas of foreign is nearly free from connected region, we are according to the connection region being set in advance
Area threshold carries out foreign bodies detection, when area is greater than preset threshold value, it is believed that the connected region is foreign matter, and will
The connection region is labeled in the picture with square-shaped frame, if there is foreign matter, then by foreign matter image and there are the vehicles of foreign matter
, car information pass to information preservation and display module, otherwise, by " foreign " information be sent to information preservation and display mould
Block；
(6) key area extraction and detection module:
(6.1) characteristic point of key area and key area template is extracted respectively using sift operator:
Key area is matched with the characteristic point of template extraction, determines matched characteristic point, by calculating all match points
Between Euclidean distance information, calculate maximum distance (max_dist) and the minimum range (min_ between matched characteristic point
Dist), matching characteristic point is filtered according to maximum distance and minimum range, retains that distance is less than 0.3*max_dist
With characteristic point；
(6.2) from filtered matching characteristic point, according to the one-to-one relationship of Feature Points Matching, matching image feature is obtained
The index of point, calculates corresponding perspective transformation matrix, and key area image is carried out matrixing using perspective transformation matrix, is made
Characteristic point on key area image is mapped in the corresponding characteristic point of key area template image, completes the registration behaviour of image
Make, transformed image and template is then subjected to image difference operation, i.e., after corresponding to matched characteristic point, with template
All pixels point in image carries out the reducing of image；
(6.3) largest connected region is asked to differentiated image, and is carried out according to the maximum width of connected region and maximum height
Limitation, to realize that the screening to connected region calculates separately their area, then use to the connected region screened
The mode of threshold value carries out foreign bodies detection, when area is greater than preset threshold value, it is believed that and the connected region is foreign matter, and
The connection region is labeled in the picture with square-shaped frame；
(7) be flexible coupling region progress foreign bodies detection:
(7.1) image that the presence received from image processing module is flexible coupling, wide and high respectively col and row, in soft company
Both sides are connect there are small part plane domain, our region foreign matter detecting method that is flexible coupling only carries out foreign bodies detection to being flexible coupling,
Therefore we need to be flexible coupling region to being cut into.Since the position of photoelectric sensor is fixed, the soft company that takes every time
The position for connecing region also determines, therefore, image is flexible coupling the flat of region both sides according to the known regional location that is flexible coupling by we
Face Regional resection；
(7.2) area image that is flexible coupling being cut into is carried out such as three-step process before non-critical areas image progress foreign bodies detection
Edge detecting operation just will appear a large amount of level when only there is foreign matter due to being flexible coupling region based on vertical line
Line segment, therefore we carry out Hough straight-line detection to the region that is flexible coupling, and to enhance the quantity of vertical direction line segment, reduce level side
To the quantity of line segment；
(7.3) edge image that is flexible coupling being made of horizontal line section and vertical segment has been obtained, in the picture using pros
The pixel frame (width of pixel frame is 30) of shape carries out pixels statistics, this needs the double-deck circulation to realize:
When the pixel value counted in square-shaped frame is greater than preset threshold value, it is believed where there is foreign matters, and terminate this
Statistics, if pixel until traversing completion completely in square-shaped frame, the pixel number summation greater than 0 is still less than threshold value, then it is assumed that
Foreign matter is not present in the square-shaped frame, square-shaped frame is subjected to translation and is continued searching.During entire translation search, if do not had
Have there are foreign matter, square-shaped frame will traverse completely the region that is entirely flexible coupling, and one pair is also needed in entire ergodic process
Layer circulation is completed:
It is being flexible coupling in edge image, according to the width and height of square-shaped frame, square-shaped frame is being translated, to the last one
When the top left co-ordinate of a frame is (row-30, col-30) and bottom right angular coordinate is (row-1, col-1), traversal is completed, and is returned
Ease back the information of join domain foreign.If finding foreign matter in ergodic process, terminate traversal, there are the letters of foreign matter for return
Breath；
(8) information preservation and display module will receive non-critical areas extraction and extract and detection mould with detection module and key area
Block whether there is the information of foreign matter, if there is foreign matter, then the image, foreign matter place coach number and train license number of foreign matter will be present
Identification information is transmitted to database module, stores data into the historical record tables of data in database, simultaneously, it may be found that
The information such as foreign matter image are shown on system foreground, and prompt the warning note of " it was found that foreign matter, is please cleared up as early as possible ", if do not sent out
Existing foreign matter, then prompt " foreign "；
So far, the roof foreign bodies detection of the column subway is completed, and the information that will test completion passes to Train number recognition module and figure
As acquisition module.
2. a kind of subway roof foreign bodies detection involved in a kind of subway roof foreign matter detecting method according to claim 1
Device comprising: Train number recognition host, Train number recognition antenna, license number label, characterized in that it further include: wheel detector,
Image capturing system, PLC, the first photoelectric sensor, the second photoelectric sensor, third photoelectric sensor, the 4th photoelectric sensor,
5th photoelectric sensor, the 6th photoelectric sensor, the 7th photoelectric sensor, the 8th photoelectric sensor, high definition industrial camera, institute
The wheel detector stated is arranged on subway approach track, and the wheel detector is electrically connected with PLC, the Train number recognition
Host is electrically connected with PLC, the Train number recognition host and image capturing system communication connection, in the wheel detector rail
On road front, successively set up the first photoelectric sensor, the second photoelectric sensor, third photoelectric sensor, the 4th photoelectric sensor,
5th photoelectric sensor, the 6th photoelectric sensor, the 7th photoelectric sensor, the 8th photoelectric sensor, in the wheel-sensors
Between device and the first photoelectric sensor, distance the first photoelectric sensor 2.5m, track surface is set higher than subway carriage junction
High definition industrial camera is set, the high definition industrial camera is electrically connected with image capturing system, first photoelectric sensing
Device, the second photoelectric sensor, third photoelectric sensor, the 4th photoelectric sensor, the 5th photoelectric sensor, the 6th photoelectric sensing
Device, the 7th photoelectric sensor, the 8th photoelectric sensor are electrically connected with PLC respectively, and the PLC is electrically connected with image capturing system
It connects.
3. a kind of subway roof detection device for foreign matter according to claim 2, characterized in that the photoelectric sensor is
Laser diffusion photoelectric sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910327152.5A CN110084171B (en) | 2019-04-23 | 2019-04-23 | Detection device and detection method for foreign matters on top of subway train |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910327152.5A CN110084171B (en) | 2019-04-23 | 2019-04-23 | Detection device and detection method for foreign matters on top of subway train |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110084171A true CN110084171A (en) | 2019-08-02 |
| CN110084171B CN110084171B (en) | 2023-03-31 |
Family
ID=67416105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910327152.5A Active CN110084171B (en) | 2019-04-23 | 2019-04-23 | Detection device and detection method for foreign matters on top of subway train |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110084171B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458126A (en) * | 2019-08-16 | 2019-11-15 | 上海仁童电子科技有限公司 | A kind of pantograph state monitoring method and device |
| CN111608522A (en) * | 2020-03-30 | 2020-09-01 | 南京邮电大学 | Sensor system for detecting platform door obstacle and detection method thereof |
| CN115988413A (en) * | 2022-12-21 | 2023-04-18 | 北京工业职业技术学院 | Train operation supervision platform based on sensing network |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005022629A (en) * | 2003-06-12 | 2005-01-27 | Nifco Inc | Moving device |
| CN103984961A (en) * | 2014-05-30 | 2014-08-13 | 成都西物信安智能***有限公司 | Image detection method for detecting foreign matter at bottom of vehicle |
| CN104567729A (en) * | 2014-12-31 | 2015-04-29 | 苏州华兴致远电子科技有限公司 | Monitoring method and device for operating state of pantograph |
| US20150189140A1 (en) * | 2009-02-23 | 2015-07-02 | Gary Edwin Sutton | Curved sensor array camera |
| WO2015123035A1 (en) * | 2014-02-17 | 2015-08-20 | General Electric Company | Video system and method for data communication |
| CN104881861A (en) * | 2015-03-11 | 2015-09-02 | 西南交通大学 | High-speed rail contact net suspension device failure state detecting method based on primitive classification |
| EP2942230A1 (en) * | 2014-05-09 | 2015-11-11 | DB Systemtechnik GmbH | Method for measuring the uplift of electrical contact lines on rail lines |
| CN105158257A (en) * | 2015-05-21 | 2015-12-16 | 苏州华兴致远电子科技有限公司 | Sliding plate measurement method and device |
| ITUB20155405A1 (en) * | 2015-10-20 | 2016-01-20 | Luca Maritano | RECLAIMING ROPE SYSTEM, THE RISALITY |
| CN206074832U (en) * | 2016-08-02 | 2017-04-05 | 易讯科技股份有限公司 | A kind of railcar roof pantograph foreign matter detection system |
| CN108981811A (en) * | 2018-07-18 | 2018-12-11 | 东莞市诺丽电子科技有限公司 | A kind of train pantograph on-line detecting system |
-
2019
- 2019-04-23 CN CN201910327152.5A patent/CN110084171B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005022629A (en) * | 2003-06-12 | 2005-01-27 | Nifco Inc | Moving device |
| US20150189140A1 (en) * | 2009-02-23 | 2015-07-02 | Gary Edwin Sutton | Curved sensor array camera |
| WO2015123035A1 (en) * | 2014-02-17 | 2015-08-20 | General Electric Company | Video system and method for data communication |
| EP2942230A1 (en) * | 2014-05-09 | 2015-11-11 | DB Systemtechnik GmbH | Method for measuring the uplift of electrical contact lines on rail lines |
| CN103984961A (en) * | 2014-05-30 | 2014-08-13 | 成都西物信安智能***有限公司 | Image detection method for detecting foreign matter at bottom of vehicle |
| CN104567729A (en) * | 2014-12-31 | 2015-04-29 | 苏州华兴致远电子科技有限公司 | Monitoring method and device for operating state of pantograph |
| CN104881861A (en) * | 2015-03-11 | 2015-09-02 | 西南交通大学 | High-speed rail contact net suspension device failure state detecting method based on primitive classification |
| CN105158257A (en) * | 2015-05-21 | 2015-12-16 | 苏州华兴致远电子科技有限公司 | Sliding plate measurement method and device |
| ITUB20155405A1 (en) * | 2015-10-20 | 2016-01-20 | Luca Maritano | RECLAIMING ROPE SYSTEM, THE RISALITY |
| CN206074832U (en) * | 2016-08-02 | 2017-04-05 | 易讯科技股份有限公司 | A kind of railcar roof pantograph foreign matter detection system |
| CN108981811A (en) * | 2018-07-18 | 2018-12-11 | 东莞市诺丽电子科技有限公司 | A kind of train pantograph on-line detecting system |
Non-Patent Citations (2)
| Title |
|---|
| CHEN, W等: "Image object detection in monitoring of pantograph slippers", 《JOURNAL OF SOUTHWEST JIAOTONG UNIVERSITY》 * |
| 曹蔡旻: "机车受电弓及其车顶状态检测***", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458126A (en) * | 2019-08-16 | 2019-11-15 | 上海仁童电子科技有限公司 | A kind of pantograph state monitoring method and device |
| CN111608522A (en) * | 2020-03-30 | 2020-09-01 | 南京邮电大学 | Sensor system for detecting platform door obstacle and detection method thereof |
| CN115988413A (en) * | 2022-12-21 | 2023-04-18 | 北京工业职业技术学院 | Train operation supervision platform based on sensing network |
| CN115988413B (en) * | 2022-12-21 | 2024-05-07 | 北京工业职业技术学院 | Train operation supervision platform based on sensor network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110084171B (en) | 2023-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109878552B (en) | Foreign matter monitoring devices between track traffic platform door and train based on machine vision | |
| CN109460709B (en) | RTG visual barrier detection method based on RGB and D information fusion | |
| CN106203239B (en) | Information processing method, device and system for container tallying | |
| CN110084171A (en) | A kind of subway roof detection device for foreign matter and detection method | |
| EP1796043B1 (en) | Object detection | |
| CN106067023B (en) | Container number and truck number identification system and method based on image processing | |
| CN104933409B (en) | A kind of parking stall recognition methods based on panoramic picture dotted line feature | |
| CN101030256B (en) | Method and apparatus for cutting vehicle image | |
| Broggi et al. | Self-calibration of a stereo vision system for automotive applications | |
| CN103077526B (en) | There is train method for detecting abnormality and the system of depth detection function | |
| CN102509090B (en) | A kind of vehicle feature recognition device based on public safety video image in sky net engineering | |
| CN112329747B (en) | Vehicle parameter detection method based on video identification and deep learning and related device | |
| CN101527046A (en) | Motion detection method, device and system | |
| CN105893953A (en) | Method and system for detecting two license plates of one vehicle | |
| CN111738203B (en) | Platform pedestrian line crossing detection method based on automatic line marking area | |
| CN107578012A (en) | A kind of drive assist system based on clustering algorithm selection sensitizing range | |
| KR101375567B1 (en) | Partial image extractor and its partial image extracting method | |
| CN105956619A (en) | Container lockhole coarse positioning and tracking method | |
| CN112364707A (en) | System and method for over-the-horizon sensing of intelligent vehicle on complex road conditions | |
| CN101369312B (en) | Method and equipment for detecting intersection in image | |
| CN111855667A (en) | Novel intelligent train inspection system and detection method suitable for metro vehicle | |
| CN107358236A (en) | A kind of number-plate number identifying system and method based on camera device | |
| CN115546738A (en) | Rail foreign matter detection method | |
| US7116814B2 (en) | Image-based container defects detector | |
| CN102706883A (en) | System and method for recognizing holes in paved waterproof board of tunnel |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |