CN108197610A - A kind of track foreign matter detection system based on deep learning - Google Patents
A kind of track foreign matter detection system based on deep learning Download PDFInfo
- Publication number
- CN108197610A CN108197610A CN201810107822.8A CN201810107822A CN108197610A CN 108197610 A CN108197610 A CN 108197610A CN 201810107822 A CN201810107822 A CN 201810107822A CN 108197610 A CN108197610 A CN 108197610A
- Authority
- CN
- China
- Prior art keywords
- foreign matter
- track
- image
- detection
- network model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention discloses a kind of track foreign matter detection system based on deep learning, the track foreign matter detection system includes：Installed video harvester, image transmission unit, detection device for foreign matter, machine learning device and image database, the installed video harvester be used for shoot and acquire target track and track above image, the image transmission unit is connected respectively with installed video harvester and the detection device for foreign matter, for receiving acquired image and being transferred to the detection device for foreign matter, the machine learning device builds offline neural network model, and the neural network model is trained using the image in the image database, the detection device for foreign matter is detected track foreign matter based on the image and constructed offline neural network model acquired.The system of the present invention is based on neural network model, has better real-time and accuracy relative to traditional image recognition algorithm, there is wider adaptability.
Description
Technical field
The present invention relates to track foreign bodies detection fields, and in particular to a kind of track foreign bodies detection system based on deep learning
System.
Background technology
Track foreign body intrusion refers to influence the barrier of train driving safety on railway, as bridge tunnel mountain forest trees fall
Hanger, the work mistake fallen cause to be detained the work business personnel of operation in orbit, illegal upper track people and animals' vehicle and they
Rubbish barrier left etc..Since train travel speed is fast, traditional driver's vision and conventional inspection systems are depended merely on to carry out
Foreign matter identifies, it is difficult to ensure traffic safety.Although traditional railway anti-disaster alarm, fault diagnosis are opposite with etection theory and technology
Maturation, but foreign matter alarm detection is always a difficult point.With the development of traffic new technology, detection foreign body intrusion and prediction are endangered
Danger is one of indispensable critical function of intelligent vehicle.At present, railway foreign body intrusion detection be broadly divided into active detecting and by
Two methods of dynamic detection.Active detecting is to send out certain detectable signal to the orientation that need to be detected, reflected by sensor
Signal detection barrier.This method is the detection of infringement formula, increases ambient noise, the spatial coverage of detection is limited.Passively
Detection is the intelligent image detection based on machine vision technique, is relatively had many advantages, such as with active detecting：It is examined in a manner of non-infringement
It surveys, does not increase ambient noise, space covering is wide, but its algorithm is complicated, computationally intensive.It is examined in the road foreign body intrusion of the prior art
In survey, testing principle is based on characteristics of image and cascade classifier training two kinds of algorithms of classification.When cascade classifier is trained, need
A large amount of sample is trained to obtain the cascade classifier being made of several Weak Classifiers by great amount of samples, faces the figure of processing
As information is big, operation efficiency is low etc. problems, and the foreign body intrusion that can detect is single and easy erroneous judgement.
Invention content
The present invention provides a kind of track foreign matter detection systems based on depth learning technology.
On the one hand, the present invention provides a kind of track foreign matter detection system based on deep learning, which is characterized in that the rail
Road foreign matter detection system includes：Installed video harvester, image transmission unit, detection device for foreign matter, machine learning device with
And image database, the installed video harvester be used for shoot and acquire target track and track above image, institute
It states image transmission unit respectively with installed video harvester and the detection device for foreign matter to be connected, for receiving acquired shadow
Picture is simultaneously transferred to the detection device for foreign matter, and the machine learning device builds offline neural network model, and utilizes the shadow
As the image in database is trained the neural network model, the detection device for foreign matter based on the image acquired with
And constructed offline neural network model is detected track foreign matter.
Preferably, the machine learning device is used to obtain track foreign matter image data, the image from image database
Have in data to the mark of foreign matter, be additionally operable to structure foreign bodies detection model, using acquired track foreign matter image data into
Row neural network model is trained, and obtains trained foreign bodies detection model.
Preferably, the detection device for foreign matter loads trained foreign bodies detection model and foreign bodies detection model parameter,
And foreign bodies detection is carried out using neural network model based on the orbit imagery acquired.
Preferably, the track foreign matter detection system further includes warning device, the warning device and the foreign bodies detection
Device is connected, once the detection device for foreign matter, which detects, foreign matter occurs in track video, then believes to the warning device foreign matter
Number, the warning device sends out alarm signal.
Preferably, the neural network model includes 7 convolutional layers and 5 maximum value pond layers.
On the other hand, the present invention provides a kind of track foreign matter detection system using described in claim 1 and carries out track
The method of foreign bodies detection, the method includes following step：
Step (1) obtains track foreign matter image data, has the mark to foreign matter in described image data；
Step (2), structure foreign bodies detection model, neural network model is carried out using acquired track foreign matter image data
Training, obtains trained foreign bodies detection model；
Trained foreign bodies detection model is loaded into foreign matter detection system by step (3)；
Step (4), loading foreign bodies detection model parameter；
Step (5), the video data that track is obtained using installed video harvester；
Step (6), the track video that is acquired is based on using trained foreign bodies detection model neural network carry out it is different
Analyte detection；
Step (7) then sends out alarm signal once there is foreign matter in the foreign bodies detection model inspection to track video.
Preferably, the step (2) includes：
Step (2-1), according to target alien material type and image data, normalized is done to image data input；
Step (2-2) does feature extraction to every pictures using 12 layers of neural network；
Step (2-3) carries out multiple convolution filtering process to the image of initialization；
Step (2-4), using treated image as the input of next layer of maximum value pond layer, dropped with 2 for step-length
Sampling processing；
Step (2-5), for next layer of repetition step (2-3) and (2-4), to the last one layer, obtain characteristic pattern；
Step (2-6) predicts 5 different scales and the bounding box of length-width ratio in each cell of characteristic pattern, finally
Can select in the bounding box of 5 predictions and true detection block is immediate updates network parameter as predicted value, each packet
It encloses box and predicts 4 coordinate values：Respectively abscissa center (tx), ordinate center of the bounding box on characteristic pattern cell
(ty), length scale ratio (tw), height scaling (th), for the feature as target detection；
Step (2-7), the foreign matter mark in image carry out the bounding box in characteristic pattern on the side of target detection process
Frame returns operation, obtains detection object position in the picture and confidence level, and accordingly adjust foreign bodies detection model.The step is
It is performed by machine learning device.
The operation principle of this patent depends in offline part and trains depth network based on yolo object detections frame
Foreign bodies detection and classification are carried out to track based on trained model in model and online part.To existing foreign matter video
Data are labeled, and then efficient depth network model are trained with yolo object detection frames, before completing neural network
To transmission.During train operation, camera captures image data, passes to neural network using data as input, it is examined
It surveys, specific position and classification is then identified if there is foreign matter, and alarm, then continued with if there is no foreign matter next
Frame image.
The advantages of this patent be using yolo object detection frames train come depth network model carry out in real time it is different
Analyte detection, based on yolo object detection frames train come depth network model it is more accurate compared to traditional detection algorithm
Really.
Description of the drawings
Fig. 1 shows the structure diagram of the foreign matter detection system in the embodiment of the present invention.
Fig. 2 shows the schematic diagrames of the used bounding box in the embodiment of the present invention.
Fig. 3 is the topology view of yolo neural network models.
Specific embodiment
Below in conjunction with attached drawing and embodiment, the present invention is described in detail, but not therefore by the protection model of the present invention
It encloses and is limited among the range of embodiment description.
Fig. 1 shows the structure diagram of the foreign matter detection system in the embodiment of the present invention.As shown in the figure, in the present embodiment
Foreign matter detection system include：Installed video harvester, image transmission unit, detection device for foreign matter, machine learning device with
And image database.
In terms of whole angle, system of the invention is broadly divided into track foreign matter on-line checking and track based on neural network
Two parts of foreign bodies detection model off-line training, two parts are closely coupled, and first part is vehicle-mounted, and second part is off-board.
The present invention system when in use, vehicle-mounted part mainly have installed video harvester, image transmission unit and
Detection device for foreign matter.This is partially installed in car body to detect track foreign matter in real time.Engineering in present system
Practising device and image database can be arranged in server.
Present system needs to carry out neural network model by machine learning device before formal be installed on vehicle
Structure and training.So machine learning device builds offline neural network model first, and using in image database
The offline neural network model of image training.Data in image database be all it is treated, both comprising on track have it is different
The image of object, and the image including not having foreign matter on track, also, foreign matter is marked in image.
After offline neural network model trains, it is possible to be loaded into trained offline neural network model different
In analyte detection device.In vehicle travel process, installed video harvester is for shooting and acquire target track and track
The image (image include video and/or image) of side, image transmission unit respectively with installed video harvester and described different
Analyte detection device is connected, and the image that real-time reception is acquired simultaneously is transferred to detection device for foreign matter.
Detection device for foreign matter loads trained foreign bodies detection model and foreign bodies detection model parameter, and is based on being acquired
Orbit imagery and constructed neural network model carry out foreign bodies detection.
Preferably, track foreign matter detection system of the invention further includes warning device, the warning device and the foreign matter
Detection device is connected, once the detection device for foreign matter, which detects, foreign matter occurs in track video, is then sent out to the warning device
Go out foreign matter signal, the warning device sends out alarm signal.
The vehicle-mounted detection part in detecting system of the present invention and off-line training part are described respectively separately below.
1st, off-line training neural network model：
Machine learning device is used to carry out the training of off-line training neural network model, can be to existing foreign matter video counts
According to being labeled, depth network model is trained based on yolo object detections frame, using trained network model, is automatically extracted
Foreign matter feature has excellent generalization ability, so as to fulfill the detection and classification to foreign matter.Foreign matter defined here only includes
5 kinds of common situations：People, animal, stone, trunk, cylinder are in kind.During object detection model training, first according to rule
Fixed object detection data annotation formatting is labeled the video image with foreign matter existing in database, then design grid
Network structure simultaneously carries out model training using the training algorithm that yolo frames provide, and finally tests training the model come,
According to the continuous adjusting training parameter of test result, optimize network structure so that final model meets actual demand.Detailed process
As shown in Figure 2.
In the present embodiment, applicant devises the model of a special construction.
Entire model shares 12 layers, and comprising 7 convolutional layers, 5 maximum value pond layers, applicant has found to use through overtesting
12 layers of 7 convolutional layer, 5 maximum value pond layer effects are best, are much better than other models, can omission factor be reduced at least 10%.
According to target alien material type and image data, normalized is done to data input first, picture size is adjusted to 224*
224, feature extraction is done using 12 layers of neural network to every pictures.For the figure of 3 channel of 224*224 sizes of initialization
Piece becomes 8 channels of 224*224 sizes after 8 convolution filter processing.Then the picture of 224*224*8 is made
For the input of next layer of maximum value pond layer, down-sampled processing is carried out for step-length with 2, i.e., to each nonoverlapping 2*2 regions into
Row is down-sampled, selects the maximum value in each region as output, obtains the output of 8 channel of 112*112 sizes.For under
One convolutional layer, filter passages number is double, obtains the output of a 112*112*16, then as next layer of maximum value pond
Change the input of layer, equally do the down-sampled processing of maximum value for step-length with 2, obtain the output of 56*56*16.Then as under
The input of one layer of convolutional layer obtains the output of 56*56*32, is passed to next layer of maximum value pond layer and returns to a 28*28*32
Output.Then the output of a 28*28*64 is returned after passing it through the convolutional layer processing of 64 filter passages, and is passed through
It crosses the down-sampled processing of maximum value that step-length is 2 and obtains the output of 14*14*64.Again through the convolutional layer of 128 filter passages
With step-length the output result of 7*7*128 is obtained for the 2 down-sampled processing of maximum value.Then by 256 filter passages of wheel
Convolutional layer handles to obtain the result of 7*7*256.5* (5+5)=50, according to neural network algorithm frame, used here as 50 filterings
The convolutional layer of device channel is handled, and returns to the output result of 7*7*50.The characteristic pattern finally gone out in last layer building it is every
5 bounding boxs (bounding box) of prediction in a cell (cell), each bounding box predict 5 coordinate values：tx、ty、tw、
th、to.If length of the back gauge in this cell range image upper left corner for (cx, cy) and the corresponding bounding box of the cell
Width is respectively (pw, ph), then 4 information calculation formula of the corresponding bounding box for detection object are as follows：
bx=σ (tx)+cx
by=σ (ty)+cy
According to the regression model in neural network algorithm frame, characteristics of image that last detection basis has been extracted above into
The frame of row target detection process returns operation, obtains detection object position in the picture and confidence level.And because constraint
The range of position prediction, parameter are easier to learn, and model is more stablized.
2nd, vehicle-mounted real-time detection is carried out：
Detection device for foreign matter, installed video harvester, image transmission unit are planned as a whole progress foreign matter and are detected in real time, vehicle-mounted shadow
As harvester carries out the real-time imaging acquisition of vehicle front track, image transmission unit is by the image transmission acquired to foreign matter
Detection device, detection device for foreign matter load trained neural network model, the model use that neural metwork training is generated
In being detected to real-time foreign matter.
When train is when certain circuit is run, first with the camera shooting in the installed video harvester of train head
Head captures front track data, is 224*224 for each frame Image Adjusting size；Trained model parameter is loaded simultaneously,
The forward direction that neural network algorithm frame is automatically performed neural network transmits；Then image is transmitted to neural network as input to carry out
Foreign bodies detection.If there are foreign matter in image, foreign matter position in the picture and classification are finally obtained, and start to alarm, if
There is no foreign matters, then continue with next frame image.
Although deep neural network can realize good detection result, there is the shortcomings that number of parameters is more, calculation amount, deployment
In vehicle imbedding type system, embedded type CPU computing capability is limited to, is unable to reach the effect detected in real time, we are by video
Realize that the forward direction of network model transmits operation and completed in FPGA, so as to fulfill depth in ARM in the part of acquisition and pretreatment
Network model operation accelerates, and entire detection process is made to have real-time.
ARM first reads the video data of camera acquisition, and the video resolution of camera acquisition may be different, to every frame
Image carries out scaling, and it is 224 × 224 to make every frame image resolution ratio, then is input to network model operation.The target detection net
The forward direction transmittance process of network model is mainly convolution algorithm, the convolution algorithm of the feature map and kernel of each channel
It is made of multiplication and addition, each layer of output feature map can be obtained with multi-channel parallel operation.
Applicant is tested other alternatives, but is not so good as technical scheme of the present invention in effect.
Specifically, it is different that the depth network model detection foreign matter trained with yolo object detection frames is replaced with other detections
Object model, but Detection accuracy is decreased obviously.
The foregoing is merely presently preferred embodiments of the present invention, and limitation in any form is not done to the present invention, all at this
It is any simple modification made to the above embodiment of technical spirit according to the present invention, equivalent within the spirit and principle of invention
Variation and modification, still fall within protection scope of the present invention.
Although the principle of the present invention is described in detail above in conjunction with the preferred embodiment of the present invention, this field skill
Art personnel are it should be understood that above-described embodiment is only the explanation to the exemplary implementation of the present invention, not to present invention packet
Restriction containing range.Details in embodiment is simultaneously not meant to limit the scope of the invention, in the spirit without departing substantially from the present invention and
In the case of range, any equivalent transformation based on technical solution of the present invention, simple replacement etc. obviously change, and all fall within
Within the scope of the present invention.
Claims (6)
1. a kind of track foreign matter detection system based on deep learning, which is characterized in that the track foreign matter detection system includes：
Installed video harvester, image transmission unit, detection device for foreign matter, machine learning device and image database, the vehicle
Carry image acquisition device be used for shoot and acquire target track and track above image, the image transmission unit respectively with
Installed video harvester is connected with the detection device for foreign matter, for receiving acquired image and being transferred to the foreign matter inspection
Device is surveyed, the machine learning device builds offline neural network model, and using the image in the image database to institute
It states neural network model to be trained, the detection device for foreign matter is based on the image and constructed offline nerve net acquired
Network model is detected track foreign matter.
2. the track foreign matter detection system according to claim 1 based on deep learning, which is characterized in that the engineering
It practises device to be used to obtain track foreign matter image data from image database, there is the mark to foreign matter, also in the image data
For building foreign bodies detection model, neural network model training is carried out using acquired track foreign matter image data, is instructed
The foreign bodies detection model perfected.
3. the track foreign matter detection system according to claim 2 based on deep learning, which is characterized in that the foreign matter inspection
It surveys device and loads trained foreign bodies detection model and foreign bodies detection model parameter, and utilize based on the orbit imagery acquired
Neural network model carries out foreign bodies detection.
4. the track foreign matter detection system according to claim 3 based on deep learning, which is characterized in that the track is different
Quality testing examining system further includes warning device, and the warning device is connected with the detection device for foreign matter, once the foreign bodies detection
Device, which detects, there is foreign matter in track video, then the signal for detecting foreign matter, the alarm dress are sent out to the warning device
It puts and sends out alarm signal.
5. the track foreign matter detection system according to claim 2 based on deep learning, which is characterized in that the nerve net
Network model includes 7 convolutional layers and 5 maximum value pond layers.
6. a kind of method that track foreign matter detection system using described in claim 1 carries out track foreign bodies detection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810107822.8A CN108197610A (en) | 2018-02-02 | 2018-02-02 | A kind of track foreign matter detection system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810107822.8A CN108197610A (en) | 2018-02-02 | 2018-02-02 | A kind of track foreign matter detection system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108197610A true CN108197610A (en) | 2018-06-22 |
Family
ID=62592350
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810107822.8A Pending CN108197610A (en) | 2018-02-02 | 2018-02-02 | A kind of track foreign matter detection system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108197610A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109413393A (en) * | 2018-11-26 | 2019-03-01 | 陕西瑞海工程智慧数据科技有限公司 | Live working Aerial Device safety monitoring method, device and Live working Aerial Device |
| CN109508710A (en) * | 2018-10-23 | 2019-03-22 | 东华大学 | Based on the unmanned vehicle night-environment cognitive method for improving YOLOv3 network |
| CN109697424A (en) * | 2018-12-19 | 2019-04-30 | 浙江大学 | A kind of high-speed railway impurity intrusion detection device and method based on FPGA and deep learning |
| CN109816946A (en) * | 2019-01-31 | 2019-05-28 | 上海浦江桥隧运营管理有限公司 | A kind of tunnel inspection device |
| CN110020598A (en) * | 2019-02-28 | 2019-07-16 | 中电海康集团有限公司 | A kind of method and device based on foreign matter on deep learning detection electric pole |
| CN110533640A (en) * | 2019-08-15 | 2019-12-03 | 北京交通大学 | Based on the track circuit disease discrimination method for improving YOLOv3 network model |
| CN110533023A (en) * | 2019-07-08 | 2019-12-03 | 天津商业大学 | It is a kind of for detect identification railway freight-car foreign matter method and device |
| CN110532889A (en) * | 2019-08-02 | 2019-12-03 | 南京理工大学 | Track foreign matter detecting method based on rotor unmanned aircraft and YOLOv3 |
| CN111016932A (en) * | 2019-12-02 | 2020-04-17 | 江西理工大学 | Track inspection vehicle and detection method of energy-saving air rail system |
| CN111478459A (en) * | 2019-01-23 | 2020-07-31 | 联发科技（新加坡）私人有限公司 | Related method and related device for foreign matter detection |
| CN111626203A (en) * | 2020-05-27 | 2020-09-04 | 北京伟杰东博信息科技有限公司 | Railway foreign matter identification method and system based on machine learning |
| CN112351928A (en) * | 2018-07-10 | 2021-02-09 | 铁路视像有限公司 | Railway obstacle detection method and system based on track segmentation |
| CN112508893A (en) * | 2020-11-27 | 2021-03-16 | 中国铁路南宁局集团有限公司 | Machine vision-based method and system for detecting tiny foreign matters between two railway tracks |
| CN112989931A (en) * | 2021-02-05 | 2021-06-18 | 广州华微明天软件技术有限公司 | Intelligent identification method for foreign matters in subway rail |
| CN113449617A (en) * | 2021-06-17 | 2021-09-28 | 广州忘平信息科技有限公司 | Track safety detection method, system, device and storage medium |
| CN113486726A (en) * | 2021-06-10 | 2021-10-08 | 广西大学 | Rail transit obstacle detection method based on improved convolutional neural network |
| CN112785587B (en) * | 2021-02-04 | 2024-05-31 | 上海电气集团股份有限公司 | Foreign matter detection method, system, equipment and medium in stacking production process |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1214656A (en) * | 1996-02-27 | 1999-04-21 | 航空工业有限公司 | Obstacle detection system |
| EP3027482A1 (en) * | 2013-07-31 | 2016-06-08 | Katz, Elen Josef | System and method for obstacle identification and avoidance |
-
2018
- 2018-02-02 CN CN201810107822.8A patent/CN108197610A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1214656A (en) * | 1996-02-27 | 1999-04-21 | 航空工业有限公司 | Obstacle detection system |
| EP3027482A1 (en) * | 2013-07-31 | 2016-06-08 | Katz, Elen Josef | System and method for obstacle identification and avoidance |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112351928B (en) * | 2018-07-10 | 2023-11-10 | 铁路视像有限公司 | Railway obstacle detection method and system based on track segmentation |
| CN112351928A (en) * | 2018-07-10 | 2021-02-09 | 铁路视像有限公司 | Railway obstacle detection method and system based on track segmentation |
| CN109508710A (en) * | 2018-10-23 | 2019-03-22 | 东华大学 | Based on the unmanned vehicle night-environment cognitive method for improving YOLOv3 network |
| CN109413393A (en) * | 2018-11-26 | 2019-03-01 | 陕西瑞海工程智慧数据科技有限公司 | Live working Aerial Device safety monitoring method, device and Live working Aerial Device |
| CN109697424A (en) * | 2018-12-19 | 2019-04-30 | 浙江大学 | A kind of high-speed railway impurity intrusion detection device and method based on FPGA and deep learning |
| CN111478459B (en) * | 2019-01-23 | 2023-11-24 | 联发科技（新加坡）私人有限公司 | Correlation method and correlation device for foreign matter detection |
| CN111478459A (en) * | 2019-01-23 | 2020-07-31 | 联发科技（新加坡）私人有限公司 | Related method and related device for foreign matter detection |
| CN109816946A (en) * | 2019-01-31 | 2019-05-28 | 上海浦江桥隧运营管理有限公司 | A kind of tunnel inspection device |
| CN110020598A (en) * | 2019-02-28 | 2019-07-16 | 中电海康集团有限公司 | A kind of method and device based on foreign matter on deep learning detection electric pole |
| CN110020598B (en) * | 2019-02-28 | 2022-04-15 | 中电海康集团有限公司 | Method and device for detecting foreign matters on telegraph pole based on deep learning |
| CN110533023A (en) * | 2019-07-08 | 2019-12-03 | 天津商业大学 | It is a kind of for detect identification railway freight-car foreign matter method and device |
| CN110533023B (en) * | 2019-07-08 | 2021-08-03 | 天津商业大学 | Method and device for detecting and identifying foreign matters of railway wagon |
| CN110532889A (en) * | 2019-08-02 | 2019-12-03 | 南京理工大学 | Track foreign matter detecting method based on rotor unmanned aircraft and YOLOv3 |
| CN110532889B (en) * | 2019-08-02 | 2022-08-16 | 南京理工大学 | Track foreign matter detection method based on rotor unmanned aerial vehicle and YOLOv3 |
| CN110533640B (en) * | 2019-08-15 | 2022-03-01 | 北京交通大学 | Improved YOLOv3 network model-based track line defect identification method |
| CN110533640A (en) * | 2019-08-15 | 2019-12-03 | 北京交通大学 | Based on the track circuit disease discrimination method for improving YOLOv3 network model |
| CN111016932B (en) * | 2019-12-02 | 2021-02-12 | 江西理工大学 | Track inspection vehicle and detection method of energy-saving air rail system |
| CN111016932A (en) * | 2019-12-02 | 2020-04-17 | 江西理工大学 | Track inspection vehicle and detection method of energy-saving air rail system |
| CN111626203B (en) * | 2020-05-27 | 2021-11-16 | 北京伟杰东博信息科技有限公司 | Railway foreign matter identification method and system based on machine learning |
| CN111626203A (en) * | 2020-05-27 | 2020-09-04 | 北京伟杰东博信息科技有限公司 | Railway foreign matter identification method and system based on machine learning |
| CN112508893A (en) * | 2020-11-27 | 2021-03-16 | 中国铁路南宁局集团有限公司 | Machine vision-based method and system for detecting tiny foreign matters between two railway tracks |
| CN112508893B (en) * | 2020-11-27 | 2024-04-26 | 中国铁路南宁局集团有限公司 | Method and system for detecting tiny foreign matters between double rails of railway based on machine vision |
| CN112785587B (en) * | 2021-02-04 | 2024-05-31 | 上海电气集团股份有限公司 | Foreign matter detection method, system, equipment and medium in stacking production process |
| CN112989931A (en) * | 2021-02-05 | 2021-06-18 | 广州华微明天软件技术有限公司 | Intelligent identification method for foreign matters in subway rail |
| CN112989931B (en) * | 2021-02-05 | 2022-10-18 | 广州华微明天软件技术有限公司 | Intelligent identification method for foreign matters in subway rail |
| CN113486726A (en) * | 2021-06-10 | 2021-10-08 | 广西大学 | Rail transit obstacle detection method based on improved convolutional neural network |
| CN113449617A (en) * | 2021-06-17 | 2021-09-28 | 广州忘平信息科技有限公司 | Track safety detection method, system, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108197610A (en) | A kind of track foreign matter detection system based on deep learning | |
| CN108304807A (en) | A kind of track foreign matter detecting method and system based on FPGA platform and deep learning | |
| US10480939B2 (en) | High speed stereoscopic pavement surface scanning system and method | |
| CN109460709A (en) | The method of RTG dysopia analyte detection based on the fusion of RGB and D information | |
| CN111126399B (en) | Image detection method, device and equipment and readable storage medium | |
| CN109087510B (en) | Traffic monitoring method and device | |
| Li et al. | Knowledge-based power line detection for UAV surveillance and inspection systems | |
| CN112149550B (en) | Automatic driving vehicle 3D target detection method based on multi-sensor fusion | |
| CN103593678B (en) | A kind of long-span bridge vehicle dynamic load distribution detection method | |
| CN106980113A (en) | Article detection device and object detecting method | |
| CN103279765B (en) | Steel wire rope surface damage detection method based on images match | |
| DE10326943A1 (en) | Autonomous vehicle and associated method and device for motion determination, motion control and object detection | |
| CN105678267A (en) | Scene recognition method and device | |
| DE102013208521A1 (en) | Collective learning of a highly accurate road model | |
| CN114248819B (en) | Railway intrusion foreign matter unmanned aerial vehicle detection method, device and system based on deep learning | |
| CN108596009A (en) | A kind of obstacle detection method and system for agricultural machinery automatic Pilot | |
| DE102020113848A1 (en) | ECCENTRICITY FUSION | |
| CN108645375B (en) | Rapid vehicle distance measurement optimization method for vehicle-mounted binocular system | |
| CN107273852A (en) | Escalator floor plates object and passenger behavior detection algorithm based on machine vision | |
| CN112132896A (en) | Trackside equipment state detection method and system | |
| CN107103275A (en) | The vehicle detection carried out using radar and vision based on wheel and tracking | |
| CN110516691A (en) | A kind of Vehicular exhaust detection method and device | |
| CN113361354A (en) | Track component inspection method and device, computer equipment and storage medium | |
| CN110926426B (en) | Real-time detection device and detection method for residual muck volume of subway shield segment | |
| CN107516423A (en) | A kind of vehicle heading detection method based on video |
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: 20180622 |
|
| RJ01 | Rejection of invention patent application after publication |