CN112241946A - Track fastener detection device based on neural network - Google Patents
Track fastener detection device based on neural network Download PDFInfo
- Publication number
- CN112241946A CN112241946A CN201910652711.XA CN201910652711A CN112241946A CN 112241946 A CN112241946 A CN 112241946A CN 201910652711 A CN201910652711 A CN 201910652711A CN 112241946 A CN112241946 A CN 112241946A
- Authority
- CN
- China
- Prior art keywords
- module
- image
- track fastener
- track
- fastener
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 16
- 230000007547 defect Effects 0.000 claims abstract description 25
- 238000004891 communication Methods 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims description 5
- 238000012552 review Methods 0.000 claims description 4
- 238000012423 maintenance Methods 0.000 claims description 3
- 238000012550 audit Methods 0.000 claims description 2
- 238000000034 method Methods 0.000 description 5
- 238000013500 data storage Methods 0.000 description 4
- 241001669679 Eleotris Species 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000001502 supplementing effect Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- DMBHHRLKUKUOEG-UHFFFAOYSA-N diphenylamine Chemical compound C=1C=CC=CC=1NC1=CC=CC=C1 DMBHHRLKUKUOEG-UHFFFAOYSA-N 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000003137 locomotive effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Train Traffic Observation, Control, And Security (AREA)
- Machines For Laying And Maintaining Railways (AREA)
Abstract
The invention provides a track fastener detection device based on a neural network, which comprises an image acquisition module, an image preprocessing module, an image identification module, a storage module, a communication module, a power supply module and the like. When the device is used for detecting a track fastener, the image acquisition module acquires a real-time image of the track fastener; the image preprocessing module preprocesses the image and sends the target area image to the image recognition module; the image recognition module recognizes the image based on the neural network, determines the defect type and position, and sends the detection result to the communication module and the storage module. The detection device can rapidly complete the detection work of the track fastener, inform workers in real time, improve the detection speed and precision of the track fastener and enhance the safety of a track traffic system.
Description
Technical Field
The invention relates to the field of rail detection, in particular to a rail fastener detection device based on a neural network.
Background
The rail transit is an important transportation mode, and refers to a type of transportation system in which an operating vehicle runs on a specific rail, including a general-speed railway, a passenger dedicated line, a high-speed railway, urban rail transit and the like. In most rail transit systems, the rail is fixed on the sleeper by a special fastener, and the integrity and firmness of the fastener have very important influence on the safety of the rail and the vehicle, so that the rail fastener needs to be detected and maintained daily. At present, two methods are mainly adopted by rail transit departments to detect rail fasteners: 1. a large number of patrolmen visually patrol along the line; 2. install video acquisition system on patrolling the road car, gather video image, detect the video that the staff was gathered through visualing and detect the track fastener. Both of the two modes have the defects of low efficiency and high labor intensity, and the state of the track fastener cannot be accurately measured in time. As an important means for guaranteeing the safety of a rail transit system, the automation and the intellectualization of rail fastener detection are problems to be solved urgently.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a track fastener detection device based on a neural network.
A track fastener detection device based on a neural network comprises an image acquisition module, an image preprocessing module, an image recognition module, a storage module, a communication module, a power supply module and the like. When the track fastener is detected, the image acquisition module acquires a real-time image of the track fastener and sends the real-time image to the image preprocessing module; the image recognition module recognizes the image based on the neural network, obtains characteristic parameters of the rail fastener, determines the defect type and position, and sends the detection results to the communication module and the storage module; the communication module sends the detection result to field workers and remote monitoring personnel in real time; the storage module stores the detection result for the relevant personnel to review and examine.
Furthermore, the image acquisition module comprises a high-speed digital camera, a light supplement lamp, a global positioning module and a vehicle position positioning module, the image acquisition module acquires track images, records time information when the images are acquired and position information of the detection device, and synchronously sends the information and the track images to the image preprocessing module.
Further, the image preprocessing module carries out preprocessing such as noise reduction and background removal on the obtained track image, extracts a corresponding target area according to the geometric image feature of the track fastener area, and sends the target area image and the time position information to the image identification module so as to provide data for subsequent track fastener defect detection.
Further, the image recognition module processes images of the target area by adopting a trained neural network, marks all geometric feature points of the track fastener, judges whether the current track fastener is in a safe fastening state or not according to the geometric dimension relation of the normal fastening track fastener, determines the position and the size of a defect according to the obtained geometric relation of the feature points if the current track fastener has the defect, classifies the defect, determines the defect type of the current track fastener, and finally sends the information and the time and position information of the defect to the communication module and the storage module.
Further, the communication module sends the detection result to field staff and remote monitoring personnel in real time, and provides detailed track fastener defect and position information for the track maintenance department so that the staff can maintain the track fastener in time.
Further, the storage module stores the detection result, and the detection result is stored as a track detection file for the follow-up review and audit of workers and supervisors.
Furthermore, the power supply module is connected with a train power supply or a special detection vehicle power supply to supply power for other modules.
Furthermore, the detection device of the rail fastener can be arranged at the bottom of a locomotive, the bottom of a carriage of a passenger car or the bottom of a special detection vehicle, runs on the vehicle and detects the passing rail fastener in real time.
The invention has the advantages that: the detection device can rapidly complete the detection work of the track fastener, accurately and contactlessly measure the track fastener, automatically find the defect of the track fastener, record the corresponding position, inform relevant workers in real time and completely file all information; the track fastener detection speed and precision can be improved, the labor intensity of track patrol workers is reduced, the labor efficiency is improved, the labor cost is saved, and the safety of a track traffic system is enhanced.
Drawings
Fig. 1 is a schematic diagram of a main functional structure of the track fastener detection device based on the neural network.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example (b): a track fastener detection device based on a neural network is mainly characterized in that the main functional structure and the connection relation of the track fastener detection device are shown in figure 1, and the track fastener detection device comprises an image acquisition module, an image preprocessing module, an image recognition module, a storage module, a communication module, a power supply module and the like. When the track fastener is detected, the image acquisition module acquires a real-time image of the track fastener and sends the real-time image to the image preprocessing module; the image recognition module recognizes the image based on the neural network, obtains characteristic parameters of the rail fastener, determines the defect type and position, and sends the detection results to the communication module and the storage module; the communication module sends the detection result to field workers and remote monitoring personnel in real time; the storage module stores the detection result for the relevant personnel to review and examine.
Further, the image acquisition module comprises a high-speed digital camera, a light supplement lamp, a global positioning module and a vehicle position positioning module, wherein the high-speed digital camera adopts an SUA231GC-T type industrial camera, the resolution is 1920 multiplied by 1200, the shortest exposure time is 20 microseconds, a high-definition camera lens with the focal length of 16mm is installed, and the acquired image can be ensured to exceed 5 pixels/mm at a fastener; the light supplementing lamp adopts four 42MIL white light LED light supplementing lamps, so that the brightness of the fastener is ensured to be enough to acquire clear images; the global positioning module adopts a GPS big dipper dual-mode positioning module, and can output positioning information in real time; the vehicle position positioning module obtains the accurate position of the vehicle on the track in a laser sleeper counting mode. The image acquisition module acquires the vehicle speed according to the vehicle position positioning module, calculates the image acquisition time interval, controls the high-speed digital camera to acquire the track image and ensures that the acquired image has an overlapping part and no missing part; and simultaneously recording time information when the image is acquired and position information of the detection device, and synchronously sending the information and the acquired track image to the image preprocessing module.
Further, the image preprocessing module carries out noise reduction preprocessing on the obtained track image by adopting a bilateral filtering method, background removal preprocessing is carried out by using a difference method, then a pre-trained bolt picture detector is adopted, a target area containing the track fastener is extracted according to the geometric image characteristics of the track fastener area, and the target area image and time position information are sent to the image recognition module to be used as subsequent track fastener defect detection data.
Further, the image recognition module processes images of the target area by adopting a trained convolutional neural network, marks all geometric feature points of the track fastener, judges whether the current track fastener is in a safe fastening state or not according to the geometric dimension relation of the normal fastening track fastener, determines the position and the size of a defect according to the obtained geometric relation of the feature points if the current track fastener has the defect, classifies the defect, determines the defect type of the current track fastener, and finally sends the information and the time and position information of the defect to the communication module and the storage module.
Furthermore, the communication module is connected with a network in a mobile data and wireless network mode, and the detection result is sent to field workers and remote monitoring personnel in real time through the network, so that detailed track fastener defect and position information are provided for a track maintenance department, and the workers can maintain the track fasteners in time.
Furthermore, the storage module stores the detection result into the TF data storage card, the TF data storage card is taken out after the detection is finished, the TF data storage card is inserted into a card reader on a computer, and the TF data storage card is stored as a corresponding track detection file according to the date and the track position, so that a worker and a supervisor can subsequently look up and check the track detection file.
Furthermore, the power supply module is connected with a special detection vehicle power supply to supply power to other modules.
Further, this track fastener's detection device installs in the special waterproof casing of special detection bottom of the car, and image acquisition module is towards track fastener's direction, and the vehicle-mounted operation carries out real-time detection to the track fastener that passes through.
The above description is only one preferred embodiment of the present invention and is not intended to limit the present invention. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
Claims (3)
1. The utility model provides a track fastener detection device based on neural network, including image acquisition module, image preprocessing module, storage module, communication module, power module etc, a serial communication port, the device contains the image recognition module, handle the image based on neural network, mark each geometric feature point of track fastener, and whether the geometric dimensions relation according to normal fastening track fastener judges current track fastener is in safe fastening state, if current track fastener has the defect, confirm the position and the size of defect according to the geometric relations of the characteristic point that obtains again, classify the defect simultaneously, confirm the defect type of current track fastener.
2. The track fastener detection device based on the neural network as claimed in claim 1, wherein the device communication module sends the detection result to field staff and remote monitoring personnel in real time to provide detailed track fastener defect and position information for track maintenance departments.
3. The neural network-based track fastener detection device as claimed in claim 1, wherein the storage module stores the detection result as a track detection file for subsequent review and audit by workers and supervisors.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910652711.XA CN112241946A (en) | 2019-07-19 | 2019-07-19 | Track fastener detection device based on neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910652711.XA CN112241946A (en) | 2019-07-19 | 2019-07-19 | Track fastener detection device based on neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112241946A true CN112241946A (en) | 2021-01-19 |
Family
ID=74168301
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910652711.XA Pending CN112241946A (en) | 2019-07-19 | 2019-07-19 | Track fastener detection device based on neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112241946A (en) |
-
2019
- 2019-07-19 CN CN201910652711.XA patent/CN112241946A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11620743B2 (en) | Systems and methods for determining defects in physical objects | |
CN202793389U (en) | On-line routing inspection device for high-speed rail overhead line system | |
US20070291985A1 (en) | Intelligent railyard monitoring system | |
CN105115605A (en) | Track train infrared detection system and detection method | |
CN105550654B (en) | Bullet train image capturing system, real-time license number detection system and method | |
KR102017870B1 (en) | Real-time line defect detection system | |
CN109242035B (en) | Vehicle bottom fault detection device and method | |
CN109979242A (en) | A kind of freeway tunnel lane change, overtake other vehicles prior-warning device and method | |
CN104410820A (en) | Vehicle-mounted trackside equipment box cover and cable appearance inspection system | |
WO2016107302A1 (en) | Method and system for inspecting moving vehicles | |
CN113371028A (en) | Intelligent inspection system and method for electric bus-mounted track | |
CN114067278A (en) | Railway freight goods inspection system and method thereof | |
CN110807929A (en) | Traffic police auxiliary command information analysis and research and judgment system | |
CN112330584A (en) | Track bolt detection device based on neural network | |
KR20150034860A (en) | Pole position detection system through number recognition | |
CN104269061B (en) | Yellow label car detection and management system and detection and management method based on network | |
CN109671181A (en) | A kind of automobile data recorder, vehicle insurance Claims Resolution method and vehicle insurance are settled a claim service system | |
CN110524553B (en) | Intelligent image analysis system of railway goods inspection robot | |
CN205157057U (en) | Infrared detecting system of rail train | |
CN112241946A (en) | Track fastener detection device based on neural network | |
CN112733709A (en) | Track inspection detection system and method | |
CN104477207B (en) | A kind of vehicle-mounted web of the rail connector outward appearance detecting system | |
CN104374375B (en) | A kind of vehicle-mounted transponder outward appearance detecting system | |
CN115985106A (en) | Truck violation monitoring method based on image recognition and laser point cloud technology | |
CN105844848A (en) | Safety management method and device of underground pipeline |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210119 |