CN112200483A - Automatic inspection system and automatic inspection method for subway trackside equipment - Google Patents
Automatic inspection system and automatic inspection method for subway trackside equipment Download PDFInfo
- Publication number
- CN112200483A CN112200483A CN202011139155.5A CN202011139155A CN112200483A CN 112200483 A CN112200483 A CN 112200483A CN 202011139155 A CN202011139155 A CN 202011139155A CN 112200483 A CN112200483 A CN 112200483A
- Authority
- CN
- China
- Prior art keywords
- equipment
- subway
- camera
- inspection
- automatic inspection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000007689 inspection Methods 0.000 title claims abstract description 91
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000001514 detection method Methods 0.000 claims abstract description 28
- 230000000007 visual effect Effects 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 15
- 230000002159 abnormal effect Effects 0.000 claims description 13
- 238000005259 measurement Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 239000013589 supplement Substances 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 239000000725 suspension Substances 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 2
- 238000000926 separation method Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000012423 maintenance Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000007635 classification algorithm Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 235000008331 Pinus X rigitaeda Nutrition 0.000 description 1
- 235000011613 Pinus brutia Nutrition 0.000 description 1
- 241000018646 Pinus brutia Species 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
- G06Q10/063114—Status monitoring or status determination for a person or group
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C1/00—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people
- G07C1/20—Checking timed patrols, e.g. of watchman
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
Abstract
An automatic inspection system and an automatic inspection method for equipment beside a subway rail are characterized in that an inspection vehicle is provided with an image acquisition module, a positioning module and a fault detection module, the image acquisition module acquires image data of the equipment beside the subway rail, the positioning module acquires real-time position information of the inspection vehicle, and the fault detection module identifies and analyzes the equipment beside the subway rail according to the image data and the corresponding position information during image shooting, acquires the position of the fault equipment and determines the fault type. The invention realizes the full-automatic inspection of the equipment beside the subway rail, improves the identification precision and efficiency, has good flexibility and expandability, reduces the labor cost and ensures the operation safety.
Description
Technical Field
The invention relates to an automatic inspection system and an automatic inspection method for equipment beside a subway rail.
Background
Urban rail transit is an industry with extremely high safety requirements. The safe and reliable equipment is the infrastructure for normal operation and safety accident avoidance of the urban rail transit passenger transport system. In order to ensure that the equipment is in a good operation state, the inspection of the trackside equipment is one of the daily tasks of safe operation and maintenance. Generally, after the operation is stopped at night, the inspection work is limited to work time of only two to three hours, the detection of the whole line network is completed, the time is short, the task is heavy, and the inspection work becomes a difficult point of rail transit rail maintenance work.
Some of current trackside equipment system of patrolling and examining concentrate on high-speed railway more, through install the high dynamic camera of a plurality of different visual angles on patrolling and examining the train, carry out image data acquisition to trackside equipment to combine the kilometer post to calculate the roughly position of train through the time of image acquisition, conveniently trace back trouble place of occurrence. And analyzing and processing the acquired images, and screening out the existing equipment faults.
The existing related papers and patent reports have the same and different overall ideas and system frameworks, but the existing technical means are relatively single in the most important image analysis. Most of the images captured at present are compared with original normal state images at the same position in a database, the comparison comprises image pixel subtraction and local gray value mutation detection, the comparison accuracy of the pure images is difficult to guarantee, accurate positioning of shooting positions is required firstly, image registration is required before comparison, the corresponding positions of equipment in the images are ensured to be relatively overlapped when the pixels are subtracted, and meanwhile, accurate judgment is difficult to achieve through the change of pixel values unless the appearance of the equipment is obviously changed. The technical means have low intelligent degree, low accuracy and fussy image preprocessing, and particularly, the accurate positioning of the shooting position is difficult to realize. The image recognition module adopts a data processing mode of software automatic recognition and manual confirmation, and to a certain extent, the inspection system only partially reduces the manual workload and does not completely realize automatic equipment inspection with full automation and high reliability.
Disclosure of Invention
The invention aims to provide an automatic inspection system and an automatic inspection method for subway trackside equipment, which are used for realizing full-automatic inspection of the subway trackside equipment, improving the identification precision and efficiency, having good flexibility and expandability, reducing the labor cost and ensuring the operation safety.
In order to achieve the aim, the invention provides an automatic inspection system for subway trackside equipment, which comprises an image acquisition module, a positioning module and a fault detection module, wherein the image acquisition module, the positioning module and the fault detection module are arranged on an inspection vehicle;
the image acquisition module acquires image data of equipment beside a subway rail;
the positioning module acquires real-time position information of the inspection vehicle;
and the fault detection module identifies and analyzes the subway trackside equipment according to the image data and the corresponding position information during image shooting, and acquires the fault position and specifies the fault type.
The image acquisition module comprises: a plurality of cameras, and a plurality of fill-in lights;
each side surface of the train is provided with at least one camera for acquiring image information of equipment positioned at two sides of the subway tunnel;
the front part of the train is provided with at least one camera for acquiring image information of equipment positioned in a front visual angle of the train in the subway tunnel;
the bottom of the train is at least provided with a camera for acquiring image information of equipment in the subway track;
the light filling lamp sets up in the not enough inspection car both sides of light and bottom for set up in inspection car both sides and bottom the camera provides sufficient light.
The camera arranged on the side face of the train adopts a ball machine.
The positioning module includes: the inertia measurement unit is arranged on the inspection vehicle, the kilometer post information on the wall of the subway tunnel is recorded in real time through the video acquired by the camera, and the inertia measurement unit is combined to accumulate the displacement of the inspection vehicle to obtain the specific position of the inspection vehicle.
The invention also provides an automatic inspection method for the equipment beside the subway rail, which comprises the following steps:
pre-marking trackside equipment in a subway tunnel, and performing deep classification learning and identification on normal state image data and fault state image data of the trackside equipment in the subway tunnel through a neural network in advance;
in the process that the inspection vehicle runs along the subway track, all cameras in the image acquisition module are matched with the light supplement lamp to acquire images of all trackside equipment arranged in the subway tunnel in real time, and meanwhile, the positioning module records the current position of the inspection vehicle in real time and correspondingly stores real-time image data and real-time position information;
the fault detection module utilizes an image recognition algorithm to recognize and position the image data of the trackside equipment collected by the image collection module, performs characteristic analysis on the positioned trackside equipment, judges whether the equipment is normal or has a fault, and determines the fault type and the position of the fault equipment.
The method for inspecting the interval cable comprises the following steps:
the visual angle of a camera on the inspection machine is perpendicular to the wall surface to shoot the cable on the wall of the subway tunnel;
detecting a straight line, namely a suspension cable, in the image by using Hough transformation in an OpenCV library;
and calculating the slope of the straight line, and determining that the straight line is abnormal if the slope exceeds a set threshold.
The method for inspecting the door of the trackside equipment comprises the following steps:
the method comprises the following steps that a specific color is marked at a side gap of an equipment box in advance, and when the door of the equipment box is opened, the marks are staggered to present an abnormal state;
a camera on the inspection machine shoots the side surface of the equipment box;
the equipment box is accurately positioned by training the appearance of the equipment box through deep learning, and a detection area is reduced;
constructing a color mask corresponding to the marking color through HSV color space segmentation to realize extraction of the marking form;
and performing two-classification training on the marking forms when the equipment box door is closed and opened, and finally judging the opening and closing conditions of the equipment box door by identifying the marking states.
The method for inspecting the terminal of the trackside equipment comprises the following steps:
a camera on the inspection machine shoots equipment on a wall;
constructing a yellow shade to filter out the yellow label by utilizing the original yellow label at the terminal end and dividing HSV color space, and eliminating the interference of a complex background;
calling cv2.findContours () algorithm to mark connected regions in the image and counting;
and comparing the terminal number with the terminal number recorded in the database, wherein if the terminal number is inconsistent with the terminal number, the terminal number is abnormal.
The method for inspecting the cover of the junction box comprises the following steps:
the visual angle of a camera on the inspection machine is aligned with the junction box beside the rail to shoot;
and directly performing two classification training on the normal state and the separation state of the junction box cover plate, and finally realizing accurate identification.
The method for inspecting the cover plate of the annunciator comprises the following steps:
the visual angle of a camera on the inspection machine is aligned with a signal machine beside the rail to shoot;
and directly performing two-classification training on the two states to finally realize accurate recognition.
The method for inspecting the beacon screw comprises the following steps:
a camera on the inspection machine shoots beacons on the track;
constructing a red shade and filtering out a mark form by utilizing the existing red mark on the screw and dividing HSV color space;
and performing two-class training on the marked forms of the loosening and tightening states of the screw, and finally judging the state of the screw by identifying the marked states.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention makes up the vacancy of the automatic inspection system beside the rail aiming at the subway, the subway tunnel environment is closed, the manual work is very hard, and the labor intensity of the staff can be greatly reduced.
2. The invention makes qualitative change aiming at the image recognition algorithm, combines and uses the OpenCV algorithm to detect aiming at the special form performance under the abnormal state of the equipment, and integrates the neural network algorithm, so that the system is more intelligent, and the recognition accuracy and reliability are greatly improved aiming at different fault matching special algorithms.
3. The invention greatly increases the types of fault identification, covers common equipment faults, does not omit faults, can increase functions according to the requirements of users and is integrated into the system. Besides the state anomalies of a plurality of devices listed by the invention, other types of faults and anomalies which can be observed through the appearance of the devices can be identified through the technical means provided by the invention, and different classification algorithms or corresponding identification models can be respectively adopted or trained aiming at specific devices and specific faults. Therefore, the invention has good flexibility and expandability.
4. The existing technology can not automatically identify the type of equipment, the corresponding cameras need to be installed for different equipment for capturing, and the cost of the cameras is very high according to the requirement of the inspection types of the equipment in the subway. The video identification model used by the invention can identify various devices in the same picture, respectively cuts the picture according to the device classification, and then carries out further device state detection, thereby greatly reducing the number of required cameras and reducing the cost. Meanwhile, aiming at a certain type of equipment, a model is adopted for identification processing, so that the accuracy and reliability of an identification result are improved.
Drawings
Fig. 1 is a schematic structural diagram of an automatic inspection system for subway trackside equipment provided by the invention.
Fig. 2 is a schematic structural diagram of an image acquisition module according to an embodiment of the present invention.
FIG. 3 is a schematic diagram of the detection of inter-zone cable hang-ups in an embodiment of the present invention.
Fig. 4 is a schematic diagram illustrating the detection of the disconnection of the connection terminal of the device according to the embodiment of the present invention.
Fig. 5 is a schematic diagram of detecting whether the annunciator cover is opened according to the embodiment of the present invention.
Fig. 6 is a schematic view illustrating detection of a screw tightening state according to an embodiment of the present invention.
Fig. 7 is a schematic diagram illustrating detection of a screw loosening state according to an embodiment of the present invention.
Detailed Description
The preferred embodiment of the present invention will be described in detail below with reference to fig. 1 to 7.
The existing trackside equipment inspection technology has the advantages that the automation and accuracy of identification are required to be improved, and the types of equipment fault identification are few due to the limitation of the image detection technology. The existing fault identification is as follows: the method comprises the steps of detecting the falling of a waterproof cap of the responder, changing the positions of a plug pin and a lead on the outer side of the rail web, changing the installation position of the interval equipment and deforming the physical appearance of the equipment facility.
In a subway, the number of devices is relatively large, the arrangement is more complex, and the number of potential fault types is more, which is listed as follows:
1. hanging the section cable: the interval cable is bound and placed on the bracket in a straight line and is horizontal to the ground in a normal state. Can cause the cable to hang down because the support pine takes off under the actual situation, forget to bind reasons such as playback after the maintenance, if the cable rocks and enters into the rail and go on the district, then there is huge potential safety hazard.
2. Opening the door of the equipment: the box doors of some equipment boxes on the wall beside the rail are in open risks, and are in side-opening and upturning modes, so that the normal running of the subway can be influenced if the box doors intrude into a limit after being opened.
3. The equipment wiring terminal falls off: there are fixed quantity's binding post on the equipment box, artificial touching or long-time use can appear the terminal and drop or not hard up risk, lead to equipment unusual.
4. Loss of the terminal box cover: during the maintenance of the junction box, maintenance personnel may forget to return the cover plate, resulting in the exposure of the cable.
5. Opening a signal machine cover plate: the apron of semaphore probably because personnel are neglected, forgets the playback and pins after the construction, and the apron side direction is opened, invades the boundary limit, has certain risk.
6. Loosening the beacon screws: due to vibration of the rail, shaking of the equipment or nonstandard operation after construction, screws of the equipment may loosen, the equipment is loosened, and monitoring and alarming are needed.
As shown in fig. 1, the present invention provides an automatic inspection system for subway wayside equipment, which comprises an image acquisition module 1, a positioning module 2 and a fault detection module 3, wherein the image acquisition module 1, the positioning module 2 and the fault detection module 3 are arranged on an inspection vehicle. The image acquisition module 1 acquires image data of equipment beside a subway rail, the positioning module 2 acquires real-time position information of the inspection vehicle, and the fault detection module 3 identifies and analyzes the equipment beside the subway rail according to the position information corresponding to the image data to acquire a fault position and specify a fault type.
The image acquisition module 1 comprises: a plurality of cameras 101, and a plurality of fill lights 102. Each side surface of the train is provided with at least one camera 101 for collecting image information of cables, trackside equipment boxes and the like positioned at two sides of a subway tunnel; the front part of the train is at least provided with a camera 101 which is used for collecting image information of trackside equipment such as a junction box, a signal machine and the like in a subway tunnel within a front visual angle of the train; at least one camera 101 is arranged at the bottom of the train and used for collecting image information of equipment such as beacons and the like in the subway track. The light supplement lamp 102 is arranged on two sides and the bottom of the inspection vehicle with insufficient light, and is used for providing sufficient light for the cameras arranged on the two sides and the bottom of the inspection vehicle, so that clearer images can be obtained.
As shown in fig. 2, in one embodiment of the present invention, a total of four cameras are installed on the inspection vehicle, and the first camera 11 and the second camera 12 are installed and fixed on two sides of the inspection vehicle, with a view angle perpendicular to the wall of the subway tunnel, and are used for shooting cables, trackside equipment boxes, and the like. The first camera 11 and the second camera 12 are selected as ball machines in consideration of the length of the cameras, preventing the cameras from being protruded from the boundary. The third camera 13 is installed in the cab with a viewing angle facing the driving direction, and is used for shooting trackside equipment such as a junction box, a signal machine and the like. The fourth camera 14 is installed in a vacant area on the side of a coupler in the train, and has a view angle perpendicular to the track area, and is used for shooting equipment in a track such as a beacon. Cameras arranged on two sides of the inspection machine and cameras for shooting rail traveling areas are provided with light supplementing lamps, and a third camera 13 used for shooting the front of a train in a cab can directly use light of train headlights.
The positioning module 2 includes: an Inertial Measurement Unit (IMU)201 provided on the inspection vehicle. The kilometer post information on the wall of the subway tunnel is recorded in real time through video identification acquired by the camera 101, and the displacement of the inspection vehicle is accumulated by combining with an Inertia Measurement Unit (IMU)201, so that the specific position of the inspection vehicle is finally obtained.
In order to ensure the quality of the collected images, a high-dynamic and high-resolution camera is selected and matched with a light supplement lamp for use.
The invention provides an automatic inspection method for subway trackside equipment, which comprises the following steps:
and step S0, pre-marking the trackside equipment in the subway tunnel, and pre-performing deep classification learning identification on the normal state image data and the fault state image data of the trackside equipment in the subway tunnel through a neural network.
Step S1, in the process of driving the inspection vehicle along the subway track, each camera 101 in the image acquisition module 1 acquires real-time images of all trackside equipment arranged in the subway tunnel under the coordination of the light supplement lamp 102, and meanwhile, the positioning module 2 records the current position of the inspection vehicle in real time, correspondingly stores the real-time image data and the real-time position information, and is convenient for accurately finding the fault section position after the fault is identified and maintaining in time.
Step S2, the fault detection module 3 uses an image recognition algorithm to recognize and locate the image data of the trackside equipment collected by the image collection module 1, performs feature analysis on the located trackside equipment, determines that the equipment is normal or has a fault, and specifies the type of the fault and the position of the faulty equipment.
Example 1
Detecting whether the section cable hangs:
s1, shooting the cable on the wall of the subway tunnel by the visual angle of a camera on the inspection machine perpendicular to the wall;
step S2, detecting a straight line in the image, namely a suspension cable, by using Hough transformation in an OpenCV library;
step S3, calculating the slope of the straight line, and determining that the straight line is abnormal if the slope exceeds a set threshold;
as shown in fig. 3, the wires are marked on the image according to a line detection algorithm, the horizontal wires in a normal state are marked with green, and the wires with abnormal slope are marked with red.
Example 2
Detecting whether the equipment box door is opened:
step S1, marking the side gap of the equipment box with a specific color in advance, wherein the marks are staggered to present an abnormal state when the door of the equipment box is opened;
s2, shooting the side surface of the equipment box at a certain angle between the visual angle of a camera on the inspection machine and the wall surface of the subway tunnel; under normal conditions, the central visual angle of the camera and the train running direction form an included angle of 60 degrees and lean to the tunnel wall, and the shooting angle can be properly adjusted on site due to uncertain site conditions;
s3, accurately positioning the equipment box by deeply learning and training the appearance of the equipment box, and reducing a detection area;
step S4, constructing a color mask corresponding to the mark color through HSV color space segmentation to realize the extraction of the mark form;
and step S5, performing two-classification training on the marking forms when the equipment box door is closed and opened, and finally judging the opening and closing conditions of the equipment box door by identifying the marking states.
Example 3
Whether check out test set binding post drops:
s1, shooting equipment on the wall by a camera on the inspection machine, wherein the visual angle of the camera is vertical to the wall of the subway tunnel;
step S2, constructing a yellow shade to filter out yellow labels by utilizing the original yellow labels at the terminal ends of the terminals through HSV color space segmentation, and eliminating the interference of a complex background;
step S3, calling cv2.findcontours () algorithm to mark connected regions in the image (the label regions are regarded as independent connected regions) and count;
step S4, comparing the number of terminals with the number of terminals recorded in the database, wherein if the number of terminals is inconsistent with the number of terminals, the result is abnormal;
as shown in fig. 4, the system counts the terminals on the device, labels under the terminals are marked and the total is displayed on the screen.
Example 4
Detecting whether the terminal box cover is lost:
step S1, shooting the trackside junction box by aiming at the visual angle of a camera on the inspection machine; generally, the central visual angle of the camera and the running direction of the train form an included angle of 60-90 degrees and deflect to the tunnel wall;
and step S2, the cover of the junction box is yellow, has large volume and obvious characteristics, and directly performs two-classification training on the normal state and the separated state of the cover plate of the junction box to finally realize accurate identification.
Example 5
Whether the detection semaphore apron is opened:
step S1, shooting a signal machine beside a track by the visual angle of a camera on the inspection machine;
and step S2, the annunciator cover plate is positioned at the rear end, the cover plate is large in size and obvious in characteristics when opened and closed, two states are directly subjected to classification training, and accurate recognition is finally achieved.
As shown in fig. 5, the open/close state of the cover of the traffic signal is recognized, and the closed state is marked with green, and the rear cover is marked with red when opened.
Example 6
Detecting whether the beacon screw is loosened:
s1, installing a camera on the inspection machine on a support, and shooting the beacon at a visual angle perpendicular to the track;
step S2, constructing a red shade and filtering out the mark form by dividing HSV color space by using the existing red mark (if not, the existing red mark can be added manually) on the screw;
and step S3, performing two-classification training on the marked forms of the loosening and tightening states of the screw, and finally judging the state of the screw by identifying the marked states.
As shown in fig. 6 and 7, the tightness of the screw is successfully identified, fig. 6 shows the screw in the tightened state, the screw in fig. 7 is loosened, and the system is successfully detected and marked.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention makes up the vacancy of the automatic inspection system beside the rail aiming at the subway, the subway tunnel environment is closed, the manual work is very hard, and the labor intensity of the staff can be greatly reduced.
2. Most of the methods adopted by the conventional trackside equipment inspection and identification technology are as follows: the method is characterized in that a high-dynamic camera is used for shooting the same equipment at the same position at different time, then pixels among pictures are compared, and the abnormal change of the equipment is tried to be found out through the sudden change of local gray values in the images.
The invention makes qualitative change aiming at the image recognition algorithm, combines and uses the OpenCV algorithm to detect aiming at the special form performance under the abnormal state of the equipment, and integrates the neural network algorithm, so that the system is more intelligent, and the recognition accuracy and reliability are greatly improved aiming at different fault matching special algorithms.
3. The invention greatly increases the types of fault identification, covers common equipment faults, does not omit faults, can increase functions according to the requirements of users and is integrated into the system. Besides the state anomalies of a plurality of devices listed by the invention, other types of faults and anomalies which can be observed through the appearance of the devices can be identified through the technical means provided by the invention, and different classification algorithms or corresponding identification models can be respectively adopted or trained aiming at specific devices and specific faults. Therefore, the invention has good flexibility and expandability.
4. The existing technology can not automatically identify the type of equipment, the corresponding cameras need to be installed for different equipment for capturing, and the cost of the cameras is very high according to the requirement of the inspection types of the equipment in the subway. The video identification model used by the invention can identify various devices in the same picture, respectively cuts the picture according to the device classification, and then carries out further device state detection, thereby greatly reducing the number of required cameras and reducing the cost. Meanwhile, aiming at a certain type of equipment, a model is adopted for identification processing, so that the accuracy and reliability of an identification result are improved.
It should be noted that in the embodiments of the present invention, the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which is only for convenience of describing the embodiments, and do not indicate or imply that the referred device or element must have a specific orientation, be configured and operated in a specific orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
Claims (11)
1. An automatic inspection system for subway trackside equipment is characterized by comprising an image acquisition module, a positioning module and a fault detection module, wherein the image acquisition module, the positioning module and the fault detection module are arranged on an inspection vehicle;
the image acquisition module acquires image data of equipment beside a subway rail;
the positioning module acquires real-time position information of the inspection vehicle;
and the fault detection module identifies and analyzes the subway trackside equipment according to the image data and the corresponding position information during image shooting, and acquires the fault position and specifies the fault type.
2. The automatic inspection system for subway wayside equipment according to claim 1, wherein said image acquisition module comprises: a plurality of cameras, and a plurality of fill-in lights;
each side surface of the train is provided with at least one camera for acquiring image information of equipment positioned at two sides of the subway tunnel;
the front part of the train is provided with at least one camera for acquiring image information of equipment positioned in a front visual angle of the train in the subway tunnel;
the bottom of the train is at least provided with a camera for acquiring image information of equipment in the subway track;
the light filling lamp sets up in the not enough inspection car both sides of light and bottom for set up in inspection car both sides and bottom the camera provides sufficient light.
3. The automatic inspection system for subway wayside equipment according to claim 2, wherein the camera disposed on the side of the train is a ball machine.
4. The automatic inspection system for subway wayside equipment according to claim 1, wherein said positioning module comprises: the inertia measurement unit is arranged on the inspection vehicle, the kilometer post information on the wall of the subway tunnel is recorded in real time through the video acquired by the camera, and the inertia measurement unit is combined to accumulate the displacement of the inspection vehicle to obtain the specific position of the inspection vehicle.
5. An automatic inspection method for subway trackside equipment by using the automatic inspection system for subway trackside equipment according to any one of claims 1-4, characterized by comprising the following steps:
pre-marking trackside equipment in a subway tunnel, and performing deep classification learning and identification on normal state image data and fault state image data of the trackside equipment in the subway tunnel through a neural network in advance;
in the process that the inspection vehicle runs along the subway track, all cameras in the image acquisition module are matched with the light supplement lamp to acquire images of all trackside equipment arranged in the subway tunnel in real time, and meanwhile, the positioning module records the current position of the inspection vehicle in real time and correspondingly stores real-time image data and real-time position information;
the fault detection module utilizes an image recognition algorithm to recognize and position the image data of the trackside equipment collected by the image collection module, performs characteristic analysis on the positioned trackside equipment, judges whether the equipment is normal or has a fault, and determines the fault type and the position of the fault equipment.
6. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the method for inspecting the section cable comprises the following steps:
the visual angle of a camera on the inspection machine is perpendicular to the wall surface to shoot the cable on the wall of the subway tunnel;
detecting a straight line, namely a suspension cable, in the image by using Hough transformation in an OpenCV library;
and calculating the slope of the straight line, and determining that the straight line is abnormal if the slope exceeds a set threshold.
7. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the method for inspecting the door of the trackside equipment comprises the following steps:
the method comprises the following steps that a specific color is marked at a side gap of an equipment box in advance, and when the door of the equipment box is opened, the marks are staggered to present an abnormal state;
a camera on the inspection machine shoots the side surface of the equipment box;
the equipment box is accurately positioned by training the appearance of the equipment box through deep learning, and a detection area is reduced;
constructing a color mask corresponding to the marking color through HSV color space segmentation to realize extraction of the marking form;
and performing two-classification training on the marking forms when the equipment box door is closed and opened, and finally judging the opening and closing conditions of the equipment box door by identifying the marking states.
8. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the method for inspecting the trackside equipment wiring terminals comprises the following steps:
a camera on the inspection machine shoots equipment on a wall;
constructing a yellow shade to filter out the yellow label by utilizing the original yellow label at the terminal end and dividing HSV color space, and eliminating the interference of a complex background;
calling cv2.findContours () algorithm to mark connected regions in the image and counting;
and comparing the terminal number with the terminal number recorded in the database, wherein if the terminal number is inconsistent with the terminal number, the terminal number is abnormal.
9. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the method for inspecting the terminal box cover comprises the following steps:
the visual angle of a camera on the inspection machine is aligned with the junction box beside the rail to shoot;
and directly performing two classification training on the normal state and the separation state of the junction box cover plate, and finally realizing accurate identification.
10. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the inspection method for the annunciator cover plate comprises the following steps:
the visual angle of a camera on the inspection machine is aligned with a signal machine beside the rail to shoot;
and directly performing two-classification training on the two states to finally realize accurate recognition.
11. The automatic inspection method for the subway trackside equipment according to claim 5, wherein the method for inspecting the beacon screws comprises the following steps:
a camera on the inspection machine shoots beacons on the track;
constructing a red shade and filtering out a mark form by utilizing the existing red mark on the screw and dividing HSV color space;
and performing two-class training on the marked forms of the loosening and tightening states of the screw, and finally judging the state of the screw by identifying the marked states.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011139155.5A CN112200483A (en) | 2020-10-22 | 2020-10-22 | Automatic inspection system and automatic inspection method for subway trackside equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011139155.5A CN112200483A (en) | 2020-10-22 | 2020-10-22 | Automatic inspection system and automatic inspection method for subway trackside equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112200483A true CN112200483A (en) | 2021-01-08 |
Family
ID=74010832
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011139155.5A Pending CN112200483A (en) | 2020-10-22 | 2020-10-22 | Automatic inspection system and automatic inspection method for subway trackside equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112200483A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113823012A (en) * | 2021-09-13 | 2021-12-21 | 浙江众合科技股份有限公司 | BIM technology-based rail transit facility inspection method and device |
| CN114228778A (en) * | 2021-12-09 | 2022-03-25 | 上海市东方海事工程技术有限公司 | Inspection vehicle-based track three-vertical-face detection system and method |
| CN114627601A (en) * | 2022-02-23 | 2022-06-14 | 广东交通职业技术学院 | Subway passenger emergency escape guiding method, system, device and storage medium |
| CN116486503A (en) * | 2023-04-06 | 2023-07-25 | 中冶南方（武汉）自动化有限公司 | Rapid inspection method and system for comprehensive pipe rack |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719252A (en) * | 2016-01-19 | 2016-06-29 | 中国铁道科学研究院 | Real-time high dynamic range drawing method and device of railway patrol image |
| CN109658397A (en) * | 2018-12-12 | 2019-04-19 | 广州地铁集团有限公司 | A kind of rail polling method and system |
| CN109753929A (en) * | 2019-01-03 | 2019-05-14 | 华东交通大学 | A kind of united high-speed rail insulator inspection image-recognizing method of picture library |
| CN111208146A (en) * | 2018-11-22 | 2020-05-29 | 北京世纪东方通讯设备有限公司 | Tunnel cable detection system and detection method |
-
2020
- 2020-10-22 CN CN202011139155.5A patent/CN112200483A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719252A (en) * | 2016-01-19 | 2016-06-29 | 中国铁道科学研究院 | Real-time high dynamic range drawing method and device of railway patrol image |
| CN111208146A (en) * | 2018-11-22 | 2020-05-29 | 北京世纪东方通讯设备有限公司 | Tunnel cable detection system and detection method |
| CN109658397A (en) * | 2018-12-12 | 2019-04-19 | 广州地铁集团有限公司 | A kind of rail polling method and system |
| CN109753929A (en) * | 2019-01-03 | 2019-05-14 | 华东交通大学 | A kind of united high-speed rail insulator inspection image-recognizing method of picture library |
Non-Patent Citations (1)
| Title |
|---|
| 张凤羽: "基于计算机视觉的输电线路异常巡检图的识别和研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113823012A (en) * | 2021-09-13 | 2021-12-21 | 浙江众合科技股份有限公司 | BIM technology-based rail transit facility inspection method and device |
| CN114228778A (en) * | 2021-12-09 | 2022-03-25 | 上海市东方海事工程技术有限公司 | Inspection vehicle-based track three-vertical-face detection system and method |
| CN114627601A (en) * | 2022-02-23 | 2022-06-14 | 广东交通职业技术学院 | Subway passenger emergency escape guiding method, system, device and storage medium |
| CN116486503A (en) * | 2023-04-06 | 2023-07-25 | 中冶南方（武汉）自动化有限公司 | Rapid inspection method and system for comprehensive pipe rack |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112200483A (en) | Automatic inspection system and automatic inspection method for subway trackside equipment | |
| CN108491758B (en) | Track detection method and robot | |
| CN202947884U (en) | Motor train unit operation fault dynamic image detection system | |
| CN100507970C (en) | Red light overriding detection system and method based on digital video camera | |
| CN103186982B (en) | Intelligent access system for public security | |
| CN107123274A (en) | Double parking stall video detecting devices and method | |
| CN102129775A (en) | Method and system for obtaining evidence by capturing vehicles at traffic crossing under panoramic video detection | |
| CN102759347B (en) | Online in-process quality control device and method for high-speed rail contact networks and composed high-speed rail contact network detection system thereof | |
| CN105067287A (en) | Motor train unit running fault dynamic image detecting system and method | |
| CN201845435U (en) | Video-based vehicle red-light-running monitoring device | |
| KR100994159B1 (en) | Catenary furniture deformation video surveillance system | |
| CN107993470A (en) | Count down traffic signal lamp condition detection method and the monitoring system based on the method | |
| CN109979242A (en) | A kind of freeway tunnel lane change, overtake other vehicles prior-warning device and method | |
| CN111192462A (en) | Overrun information acquisition method, device and system and overrun detection equipment | |
| CN106683400A (en) | Method and a system for obtaining evidence by capturing vehicles at traffic crossing under panoramic video detection | |
| CN201037962Y (en) | Electronic pile examining system | |
| CN214039881U (en) | 360-degree image detection system for metro vehicle | |
| CN1725248A (en) | Charged monitoring system for road charging monitoring and management section and its monitoring method | |
| CN203812397U (en) | Urban road video monitor system | |
| CN115600124A (en) | Subway tunnel inspection system and inspection method | |
| CN205383993U (en) | Identification system is measured to vehicle license plate and profile | |
| JP5506537B2 (en) | A vehicle side light detection method provided in a traveling body, and a movable fence control device that opens and closes a door provided in the movable fence using the detection method. | |
| CN211335993U (en) | 360-degree fault image detection system for metro vehicle | |
| CN114999170A (en) | Intelligent analysis visualization system based on highway vehicle path data | |
| CN114179866A (en) | Track turnout opening direction safety detection system and detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210108 |
|
| RJ01 | Rejection of invention patent application after publication |