CN110223293A - The intelligent identification Method and identification device of train body breakage - Google Patents
The intelligent identification Method and identification device of train body breakage Download PDFInfo
- Publication number
- CN110223293A CN110223293A CN201910541627.0A CN201910541627A CN110223293A CN 110223293 A CN110223293 A CN 110223293A CN 201910541627 A CN201910541627 A CN 201910541627A CN 110223293 A CN110223293 A CN 110223293A
- Authority
- CN
- China
- Prior art keywords
- train
- identification
- model
- car body
- training
- 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
- 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]
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to technical field of vehicle detection, it discloses and the intelligent identification Method and identification device of a kind of train body breakage is provided, the intelligent identification Method of train body breakage includes: that location model for positioning the position of car body in the picture and for identification identification model of car body breakage is respectively trained using the sample image about train；Acquire the image data of train to be detected；Image data is input to location model, with location information of the car body of determination train to be detected in correspondence image；And the image data of train to be detected and location information are input to identification model, with the car body breakage of determination train to be detected.Through the above scheme, using the positioning and identification model according to the training of train sample image, automatic identification is carried out to car body breakage, realize the real time monitoring to train body breakage, reduce the operation intensity and cost of labor of car body breakage inspection and maintenance, it realizes the automatic tracing of car body damaging problem, improves working efficiency.
Description
Technical field
The present invention relates to rolling stock detection technique fields, intelligent identification Method more particularly to train body breakage and
Identification device.
Background technique
During the day-to-day operation of train, train body, EEF bogie are easy the shadow by factors such as extraneous and itself abrasion
It rings, situations such as loss of parts, damaged occurs.Especially lorry is easy on the ground such as loading, unloading goods yard along the line because being made using excavator
The operation modes such as industry, and truck body surface is caused to damage.
But artificial maintenance is relied primarily on to the inspection of the intact situation of train body at present, there are labor intensity for artificial maintenance
Greatly, it is easy to appear the problem of missing inspection.Therefore, the mode of simple artificial upkeep operation has been difficult to meet the train peace being increasingly enhanced
Full protection management requirement.
Summary of the invention
The purpose of the invention is to overcome artificial maintenance of the existing technology, there are large labor intensities, are easy to appear leakage
The problem of inspection, the intelligent identification Method and identification device of a kind of train body breakage are provided, the damaged intelligence of the train body is known
Other method can be according to the image data car body breakage of the image data automatic identification train to be detected of train to be detected.
To achieve the goals above, on the one hand the embodiment of the present invention provides the intelligent recognition side of a kind of train body breakage
Method, the intelligent identification Method of the train body breakage include: to be respectively trained using the sample image about train for positioning
The location model of the position of car body in the picture and for identification identification model of car body breakage；Acquire train to be detected
Image data；Described image data are input to the location model, with the car body of the determination train to be detected in corresponding diagram
Location information as in；And the described image data of the train to be detected and the location information are input to the identification
Model, with the car body breakage of the determination train to be detected.
Preferably, before described image data are input to the location model, the intelligence of the train body breakage
Recognition methods further include: enhancing processing is carried out to described image data.
Preferably, it is described to described image data carry out enhancing processing include: using Laplce's contrast enhancing and/or
Adaptive log transformation carries out enhancing processing to described image.
Preferably, the training is used to position the location model of the position of car body in the picture and car body is damaged for identification
The identification model of situation includes: using the convolutional neural networks CNN training location model；And/or use residual error network training
The identification model.
Preferably, the training is used to position the location model of the position of car body in the picture and car body is damaged for identification
The identification model of situation further include: the training location model and/or the identification model are accelerated using graphics processor GPU.
According to a second aspect of the embodiments of the present invention, a kind of intelligent identification device of train body breakage, the column are provided
The intelligent identification device of vehicle car body breakage includes: training module, is used for for being respectively trained using the sample image about train
The location model of positioning car body abort situation and for identification identification model of car body breakage；Acquisition module, for acquiring
The image data of train to be detected；Locating module, for described image data to be input to the location model, described in determination
Location information of the car body of train to be detected in correspondence image；And identification module, by the figure of the train to be detected
As data and the location information are input to the identification model, with the car body breakage of the determination train to be detected.
Preferably, the intelligent identification device of the train body breakage further include: preprocessing module, for by the figure
Before being input to the location model as data, enhancing processing is carried out to described image data.
Preferably, the training module includes: to position training submodule, for using convolutional neural networks CNN training institute
State location model；And/or recognition training submodule, for using identification model described in residual error network training.
Preferably, the training module further include: accelerating module, for being accelerated described in training using graphics processor GPU
Location model and/or the identification model.
In addition, the embodiment of the present invention also provides a kind of machine readable storage medium, stored on the machine readable storage medium
There is instruction, which is used for so that machine executes the intelligent identification Method of above-mentioned train body breakage.
Through the above technical solutions, using positioning and identification model according to the training of train sample image, to car body breakage
Automatic identification is carried out, realizes the real time monitoring to train body breakage, reduces the work of car body breakage inspection and maintenance
Industry intensity and cost of labor realize the automatic tracing of car body damaging problem, improve working efficiency.
The other feature and advantage of the embodiment of the present invention will the following detailed description will be given in the detailed implementation section.
Detailed description of the invention
Attached drawing is to further understand for providing to the embodiment of the present invention, and constitute part of specification, under
The specific embodiment in face is used to explain the present invention embodiment together, but does not constitute the limitation to the embodiment of the present invention.
Fig. 1 is the flow chart of the intelligent identification Method of train body breakage provided in an embodiment of the present invention；
Fig. 2 is the exemplary process of concrete application of the intelligent identification Method of train body breakage provided in an embodiment of the present invention
Figure；
Fig. 3 is the block diagram of the intelligent identification device of train body breakage provided in an embodiment of the present invention；And
Fig. 4 is the block diagram of the training module of the intelligent identification device of train body breakage provided in an embodiment of the present invention.
Description of symbols
1, training module 2, acquisition module
3, locating module 4, identification module
11, training submodule 12, recognition training submodule are positioned
Specific embodiment
It is described in detail below in conjunction with specific embodiment of the attached drawing to the embodiment of the present invention.It should be understood that this
Locate described specific embodiment and be merely to illustrate and explain the present invention embodiment, is not intended to restrict the invention embodiment.
Fig. 1 is the flow chart of the intelligent identification Method of train body breakage provided in an embodiment of the present invention, as shown in Figure 1,
May include:
S100, training location model and identification model.
Preferably, the positioning for positioning the position of car body in the picture is respectively trained using the sample image about train
Model and for identification identification model of car body breakage.
In the first preferred embodiment of the present invention, using convolutional neural networks CNN (Convolutional Neural
Networks, convolutional neural networks) train location model for positioning the position of car body in the picture.Car body is in the picture
Position refers to that compartment, car door, vehicle window etc. are carrying out the position in image acquisition process on the image that gets to train, can be with
The form of coordinate indicates.
Specifically, using convolutional network neural network CNN, in conjunction with the method for regression analysis, detection part is treated in realization
Accurate positioning, namely the positioning of the position to car body in the picture.For example, main in the specific training process of location model
Will by 75 CNN layers, 23 shortcut layers, wherein CNN layers of center size are 3 to have 38, and core size is 1 to have 37,
All activated function uses Leaky ReLU.The input of location model is sample image and body part about train in image
In coordinate ratio, indicate that body part accounts for the proportionality coefficient of sample image with x, y, w, h, output data format is class label
(car door, compartment, the vehicle window etc. in body part are indicated with different class labels), class probability (use class probability table
Show that content in the sample image is the probability of car door, compartment, vehicle window in body part etc.), coordinate proportionality coefficient x, y, w,
h.The method of network model parameter initialization uses mean value for 0, the random initializtion method of the Gaussian Profile of variance 0.01, with this
To be trained to location model.
In the second preferred embodiment of the present invention, using the identification of residual error network training car body breakage for identification
Model.
Discriminant classification is carried out using the malfunction that the method for classification treats detection part.For example, identification model
Mainly by 50 CNN layer, 16 shortcut layers in specific training process, wherein CNN layers of center size have 17 for 3, core
Size is 1 to have 33, and all activated function uses Leaky ReLU.The structure of identification model is two taxonomic structures, that is, is identified
Model goes out to break down and the non-two class situations that break down for identification, wherein car body is thought to break down there are breakage, car body
Think not break down there is no breakage.The input of identification model is that (picture size can be 256*256 for the picture of body part
Pixel), failure identification (0 or 1, wherein 0 indicate non-faulting, 1 indicate failure), probability of malfunction (i.e. output for 0 or 1 it is general
Rate value), wherein the picture of body part can use location model and obtain the coordinate of body part, then according to body part
Coordinate, from image data cut obtain the picture of body part.The method of the parameter initialization of identification model training is same
Mean value can be used for 0, the random initializtion method of the Gaussian Profile of variance 0.01.
The training process of location model and identification model requires to be iterated trained ability using more sample data
Better location model and identification model are obtained, therefore, in a preferred embodiment of the invention, training exists for positioning car body
During the location model of position in image and for identification identification model of car body breakage, it is all made of graphics processor
GPU (Graphics Processing Unit, graphics processor) accelerates training location model and/or identification model, to improve
The efficiency of model training.
The image data of S200, acquisition train to be detected.
Preferably, the image data of train to be detected is acquired by being mounted on the vehicle equipment of discrepancy section, rail side.
In a preferred embodiment of the invention, image acquisition units are installed in track two sides respectively, and when lorry passes through,
Using pick up inductor sending receiving signal control camera truck body is shot according to certain angle and other
Suitable method, to realize that the vehicle to train to be detected carries out 360 degrees omnidirection Image Acquisition.
S300, location information of the car body of train to be detected in correspondence image is determined.
Preferably, image data is input to location model, with the car body of determination train to be detected in correspondence image
Location information.
In the preferred embodiment of the present invention, the various parameter values of location model are loaded first, in the image for receiving acquisition
After data, accurate positioning of the body part of vehicle in image data is obtained using location model, intercepts train to be detected
The picture of body part, and using the picture intercepted as the input data in following step S400.
Since that there are shooting angle is different, bright and dark light is different, vehicle body pollutes journey for the picture in acquired image data
The degree difference differences such as larger, can positioning result of the car body in correspondence image to train to be detected have an impact.Therefore, this hair
In bright preferred embodiment, before image data is input to location model, enhancing processing is carried out to image data, makes image
With stronger robustness.Enhance algorithm primarily to improving the display effect contrast of image (such as change) of image,
The position of the content and body part of image in the picture is not changed.
Preferably, enhancing processing is carried out to image using the enhancing of Laplce's contrast and/or adaptive log transformation.
S400, the car body breakage for determining train to be detected.
Preferably, the image data of train to be detected and location information are input to identification model, with determination column to be detected
The car body breakage of vehicle.
In the preferred embodiment of the present invention, the car body that the image data of train to be detected is obtained according to step S300 is right
After answering the location information in image to be handled, the uninterrupted picture of car body of train to be detected is obtained, then by the figure of body part
Piece is input in identification model, with the car body breakage of determination train to be detected.Specifically, load identification model first
The picture of the body part of the train to be detected obtained by step S300 is input to identification model, obtained by various parameter values
The breakage of corresponding body part.
Illustrate the intelligent recognition side of train body breakage provided by the embodiment of the present invention with specific application example below
Method, as shown in Fig. 2, the process of car body failure evaluation is divided into two stages, the stage 1 is model training rank using in example by this
Section, the stage 2 is cognitive phase.
Training stage includes: S1, collects sample image；S2, location information mark；S3, fault message mark；S4, positioning
Model；S5, identification model.
Use position of the sample image about train being collected into after enhancing is handled, first to car body in the picture
It is labeled with fault message (i.e. car body breakage), positioning network model is respectively trained and differentiates network model.It trained
Cheng Zhong is iterated training using gradient descent method, and collected samples pictures are no less than 10000, the number of iterations 50000
Secondary, wherein 0-35000 learning rate is 0.01, and 35000-45000 time learning rate is 0.001, is for 45000-50000 times
0.0001.Training terminates obtained location model and identification model, is applied in cognitive phase.
Cognitive phase includes: S4, location model；S5, identification model；S6, acquisition image data；S7, Location vehicle；S8,
Failure evaluation；S9, recognition result is obtained.
In actual application, the various parameter values of the two models are loaded first, it is freshly harvested to be detected receiving
After the image data of train, enhancing processing is carried out to image, the test section of module to be measured is then obtained using positioning network model
The accurate positioning of part obtains the location information of body part, according to the location information of body part, intercepts the vehicle of train to be detected
The small picture of body component.In turn, the small picture of body part is input to identification model, the car body for obtaining train to be detected is damaged
Situation.
The intelligent identification Method of above-mentioned train body breakage can be flexibly deployed under windows and linux environment, right
Hardware environment is not fixed to be required, and if hardware environment supports GPU to accelerate, then can further promote the performance and effect of this method
Rate.This intelligent detecting method uses end-to-end detection mode, inputs the image data of train to be measured, using trained fixed
Handling image data for bit model and identification model, can be with the car body breakage of train to be measured in intelligent recognition, pole
The earth improves recognition efficiency and recognition accuracy to train body breakage.
Fig. 3 is the block diagram of the intelligent identification device of train body breakage provided in an embodiment of the present invention, as shown in Fig. 2, column
The intelligent identification device of vehicle car body breakage includes: training module 1, is used for for being respectively trained using the sample image about train
The location model of positioning car body abort situation and for identification identification model of car body breakage；Acquisition module 2, for acquiring
The image data of train to be detected；Locating module 3, for image data to be input to location model, with determination train to be detected
Location information of the car body in correspondence image；And identification module 4, the image data of train to be detected and location information is defeated
Enter to identification model, with the car body breakage of determination train to be detected.
Preferably, the intelligent identification device of train body breakage further include: preprocessing module (not shown) is used for
Before image data is input to location model, enhancing processing is carried out to image data.
Fig. 3 is the block diagram of the training module of the intelligent identification device of train body breakage provided in an embodiment of the present invention, such as
Shown in Fig. 3, training module 1 includes: to position training submodule 11, for using convolutional neural networks CNN training location model；
And/or recognition training submodule 12, for using residual error network training identification model.
Preferably, training module 1 further include: accelerating module (not shown), for being accelerated using graphics processor GPU
Training location model and/or identification model.
Other specific implementation details and beneficial effect of the intelligent identification device of train body breakage refer to above-mentioned train vehicle
The intelligent identification Method of body breakage, details are not described herein again.
Through the above technical solutions, using positioning and identification model according to the training of train sample image, to car body breakage
Automatic identification is carried out, realizes the real time monitoring to train body breakage, reduces the work of car body breakage inspection and maintenance
Industry intensity and cost of labor realize the automatic tracing of car body damaging problem, improve working efficiency.
The optional embodiment of the embodiment of the present invention is described in detail in conjunction with attached drawing above, still, the embodiment of the present invention is simultaneously
The detail being not limited in above embodiment can be to of the invention real in the range of the technology design of the embodiment of the present invention
The technical solution for applying example carries out a variety of simple variants, these simple variants belong to the protection scope of the embodiment of the present invention.
It is further to note that specific technical features described in the above specific embodiments, in not lance
In the case where shield, it can be combined in any appropriate way.In order to avoid unnecessary repetition, the embodiment of the present invention pair
No further explanation will be given for various combinations of possible ways.
It will be appreciated by those skilled in the art that implementing the method for the above embodiments is that can pass through
Program is completed to instruct relevant hardware, which is stored in a storage medium, including some instructions are used so that single
Piece machine, chip or processor (processor) execute all or part of the steps of each embodiment the method for the application.And it is preceding
The storage medium stated includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), magnetic or disk.
In addition, any combination can also be carried out between a variety of different embodiments of the embodiment of the present invention, as long as it is not
The thought of the embodiment of the present invention is violated, equally should be considered as disclosure of that of the embodiment of the present invention.
Claims (10)
1. a kind of intelligent identification Method of train body breakage, which is characterized in that the intelligent recognition side of the train body breakage
Method includes:
Location model for positioning the position of car body in the picture is respectively trained using the sample image about train and is used for
Identify the identification model of car body breakage；
Acquire the image data of train to be detected；
Described image data are input to the location model, with the car body of the determination train to be detected in correspondence image
Location information；And
The described image data of the train to be detected and the location information are input to the identification model, described in determination
The car body breakage of train to be detected.
2. the intelligent identification Method of train body breakage according to claim 1, which is characterized in that by described image number
According to before being input to the location model, the intelligent identification Method of the train body breakage further include:
Enhancing processing is carried out to described image data.
3. the intelligent identification Method of train body breakage according to claim 2, which is characterized in that described to described image
Data carry out enhancing processing
Enhancing processing is carried out to described image using the enhancing of Laplce's contrast and/or adaptive log transformation.
4. the intelligent identification Method of train body breakage according to claim 1, which is characterized in that the training is for fixed
The location model of the position car body position in the picture and identification model of car body breakage includes: for identification
Using the convolutional neural networks CNN training location model；And/or
Using identification model described in residual error network training.
5. the intelligent identification Method of train body breakage according to claim 4, which is characterized in that the training is for fixed
The location model of the position of position car body in the picture and for identification identification model of car body breakage further include:
The training location model and/or the identification model are accelerated using graphics processor GPU.
6. a kind of intelligent identification device of train body breakage, which is characterized in that the intelligent recognition of the train body breakage fills
It sets and includes:
Training module, for the location model for positioning car body abort situation to be respectively trained using the sample image about train
The identification model of car body breakage for identification；
Acquisition module, for acquiring the image data of train to be detected；
Locating module, for described image data to be input to the location model, with the car body of the determination train to be detected
Location information in correspondence image；And
The described image data of the train to be detected and the location information are input to the identification model by identification module,
With the car body breakage of the determination train to be detected.
7. the intelligent identification device of train body breakage according to claim 6, which is characterized in that the train body is broken
The intelligent identification device of damage further include:
Preprocessing module, for being carried out to described image data before described image data are input to the location model
Enhancing processing.
8. the intelligent identification device of train body breakage according to claim 6, which is characterized in that the training module packet
It includes:
Training submodule is positioned, for using the convolutional neural networks CNN training location model；And/or
Recognition training submodule, for using identification model described in residual error network training.
9. the intelligent identification device of train body breakage according to claim 8, which is characterized in that the training module is also
Include:
Accelerating module, for accelerating the training location model and/or the identification model using graphics processor GPU.
10. a kind of machine readable storage medium, it is stored with instruction on the machine readable storage medium, which is used for so that machine
Perform claim requires the intelligent identification Method of train body breakage described in any one of 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910541627.0A CN110223293A (en) | 2019-06-21 | 2019-06-21 | The intelligent identification Method and identification device of train body breakage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910541627.0A CN110223293A (en) | 2019-06-21 | 2019-06-21 | The intelligent identification Method and identification device of train body breakage |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110223293A true CN110223293A (en) | 2019-09-10 |
Family
ID=67814248
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910541627.0A Pending CN110223293A (en) | 2019-06-21 | 2019-06-21 | The intelligent identification Method and identification device of train body breakage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110223293A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022062242A1 (en) * | 2020-09-27 | 2022-03-31 | 广东海洋大学 | Deep learning-based underwater imaging and fishing net damage identification method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106127747A (en) * | 2016-06-17 | 2016-11-16 | 史方 | Car surface damage classifying method and device based on degree of depth study |
| CN108109137A (en) * | 2017-12-13 | 2018-06-01 | 重庆越畅汽车科技有限公司 | The Machine Vision Inspecting System and method of vehicle part |
| CN108648169A (en) * | 2018-03-20 | 2018-10-12 | 中国科学院自动化研究所 | The method and device of high voltage power transmission tower defects of insulator automatic identification |
-
2019
- 2019-06-21 CN CN201910541627.0A patent/CN110223293A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106127747A (en) * | 2016-06-17 | 2016-11-16 | 史方 | Car surface damage classifying method and device based on degree of depth study |
| CN108109137A (en) * | 2017-12-13 | 2018-06-01 | 重庆越畅汽车科技有限公司 | The Machine Vision Inspecting System and method of vehicle part |
| CN108648169A (en) * | 2018-03-20 | 2018-10-12 | 中国科学院自动化研究所 | The method and device of high voltage power transmission tower defects of insulator automatic identification |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022062242A1 (en) * | 2020-09-27 | 2022-03-31 | 广东海洋大学 | Deep learning-based underwater imaging and fishing net damage identification method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111079747B (en) | Railway wagon bogie side frame fracture fault image identification method | |
| CN111652227B (en) | Method for detecting damage fault of bottom floor of railway wagon | |
| CN110826520B (en) | Port grab bucket detection method based on improved YOLOv3-tiny algorithm | |
| CN108182383B (en) | Vehicle window detection method and device | |
| CN111080612B (en) | Truck bearing damage detection method | |
| CN111652295B (en) | Railway wagon coupler yoke key joist falling fault identification method | |
| CN112508013B (en) | Lock catch loss fault detection method, system and device | |
| CN111080600A (en) | Fault identification method for split pin on spring supporting plate of railway wagon | |
| CN107273802A (en) | A kind of detection method and device of railroad train brake shoe drill ring failure | |
| CN111091551A (en) | Method for detecting loss fault of brake beam strut opening pin of railway wagon | |
| CN108648210B (en) | Rapid multi-target detection method and device under static complex scene | |
| Santur et al. | An adaptive fault diagnosis approach using pipeline implementation for railway inspection | |
| CN112712516A (en) | High-speed rail bottom rubber strip fault detection method and system based on YOLOv5 | |
| CN110223293A (en) | The intelligent identification Method and identification device of train body breakage | |
| CN107506753A (en) | A kind of more wireless vehicle trackings towards dynamic video monitoring | |
| CN110059393A (en) | A kind of emulation test method of vehicle, apparatus and system | |
| CN108711150A (en) | A kind of end-to-end pavement crack detection recognition method based on PCA | |
| CN108446581A (en) | A kind of unmanned plane detection method in adverse circumstances | |
| Thiruganam et al. | Automatic defect detection and counting in radiographic weldment images | |
| CN116448764A (en) | Automatic crack detection method for fatigue test of aircraft structure | |
| CN115424128A (en) | Fault image detection method and system for lower link of freight car bogie | |
| CN113486951B (en) | Method for identifying and detecting impact deformation fault of windshield, storage medium and equipment | |
| CN111768357B (en) | Image detection method and device | |
| Liu et al. | An end-to-end steel strip surface defects detection framework: considering complex background interference | |
| CN112766174B (en) | Railway train carriage group bottom plate loss fault 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: 20190910 |
|
| RJ01 | Rejection of invention patent application after publication |