CN110220909A - A kind of Shield-bored tunnels Defect inspection method based on deep learning - Google Patents
A kind of Shield-bored tunnels Defect inspection method based on deep learning Download PDFInfo
- Publication number
- CN110220909A CN110220909A CN201910348834.4A CN201910348834A CN110220909A CN 110220909 A CN110220909 A CN 110220909A CN 201910348834 A CN201910348834 A CN 201910348834A CN 110220909 A CN110220909 A CN 110220909A
- Authority
- CN
- China
- Prior art keywords
- tunnel
- model
- camera
- damage
- image
- 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
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8883—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges involving the calculation of gauges, generating models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
A kind of Shield-bored tunnels Defect inspection method based on deep learning, specific implementation step are as follows: A. sets up camera and is imaged；B. the over-fitting processing of area's data is surveyed；C. the evaluation and amendment of model；D. the identification and positioning of tunnel damage；E. zone position damage automatic on-line monitoring and real-time storage are surveyed in tunnel.The invention has the advantages that 1. solve the disadvantage that the informationization that traditional artificial routing inspection efficiency is low, digitized degree is low, is unfavorable for detection, automation；2. solving the problems, such as that image transmitting distance is limited in conventional tunnel damage monitoring method, improves measurement accuracy；3. can mitigate significantly by the way that deep learning algorithm is added and identify fault as caused by environmental factor；4. camera acquired image information is transferred to computer using high speed gigabit Ethernet, real-time Transmission is fast, realizes that online data is handled immediately, monitor video can be checked in conjunction with the real-time stress state of drag-line, and the playback of event full information may be implemented.
Description
Technical field
The present invention relates to the methods that a kind of pair of shield damage carries out real-time monitoring.
Background technique
In recent years, with the continuous development of Shield-bored tunnels technology, shield method is as a kind of efficient underground construction side
Formula has become the underground of China's most cities and the main construction method of submerged tunnel engineering.But current most domestic
Be is the section of jurisdiction of concrete, is likely to occur to destroy during production, transport, construction and military service to generate damage
Wound and other damages.Current manual inspection mode low efficiency, digitized degree is low, is unfavorable for the information-based, automatic of detection
Change；And with the increase of tunnel course, the continuous rising of human cost is not able to satisfy the need of maintenance increasingly using manual inspection
It wants.
1, the damage research of general tunnel and shield tunnel
The research for tunnel damage domestic at present mainly includes tunnel defect detection technique, method, program and system；Tunnel
Road causes of disease research；Disease classification method；The research of disease Gernral Check-up method；Tunnel safety analysis and computation model；
Tunnel Repair reinforcement technique etc..It is divided in detail in " the bridge tunnel building deterioration evaluation criteria " that China came into effect since 1998
The type of tunnel defect, respectively fracturing failures of tunnel lining, liner structure leak, lining cutting deteriorate three categories diseases, and according to tight
Weight degree is divided into tetra- grades of A, B, C, D.Research for the detection of tunnel defect is concentrated both ways: first is that for detection
The research of rules and regulations, second is that the research of specific Defect inspection method and detection technique.Technology used at present mainly has following
It is several: geological radar, sound wave, CT etc..Huang Xi etc., which has studied, intersects shield tunnel dynamic response under vibration loads of train effect
And breakdown diagnosis；Li little Kun etc. has studied under the action of the sulfate attack of Southwestern China area, tunnel structure disease it is specific
Feature, and current existing Disease Characters grade separation is combined, propose quantitative division methods；In terms of shield tunnel, Lu Dai
High mountain etc. has studied the crack propagation law with tenon shield duct piece, and is directed to and does not do to influence degree of the bad construction loads to section of jurisdiction
Research has been carried out quasi- true using extended finite element method (XFEM).In terms of detection technique, current research direction is mainly concentrated
It is tested and analyzed to the construction soil body and the stress of lining cutting, deformation.Li Yuning points out shield-tunneling construction detection content
It mainly include stratum, support construction and ambient enviroment；Main project include earth's surface and deep soil sedimentation and horizontal displacement,
Launching stress, water level, building settlement, underground pipeline settlement, internal force for support and deformation etc.；Yang Xin's peace etc. devises railway bed
Vehicle is detected, which will be equipped with geological radar and static penetrometer, quickly continuous can must detect bedding subgrade defect situation.
2, the damage detection technology based on machine vision and computer vision
Since 2012, deep neural network technology had great progress in machine vision.Krizhevsky etc. is mentioned
Image Classifier is done using depth convolutional network out, this method greatly improves the accuracy of identification.Since then, depth convolutional network
Start to apply in a large amount of machine vision task.Sermane et al. has studied deep learning in image recognition, positioning and detection
The general framework of aspect.Because deep learning network needs very strong computing capability, GPU has arrived deep learning by large-scale application
Field.The appearance of Nvidia CUDA greatly accelerates calculating speed and lowers existing development cost.Later, Chetlur et al.
Propose the application threshold that further reduced deep learning for the Computational frame CuDNN of deep learning network.Meanwhile
Simonyan et al. proposes the convolutional neural networks VGG of deeper, and achieves the further promotion of accuracy.Meanwhile he
Also proposed a series of measure to guarantee the training of model.In terms of deep learning frame, the Abadi from Google is proposed
Computational frame Tensorflow be largely used in all kinds of tasks.In order to enable deep learning has mobile deployment energy
Power, Howard etc. propose a series of measures to improve speed and reduce model parameter quantity.Inferred motion structure
(Structure fromMotions, SfM) is the weight of one of task important in machine vision and all kinds of researchers concern
Point.Inferred motion structure extensive application RANSAC algorithm does the reconstruction of 3D model.Haming etc. points out, up to the present,
There are very mature inferred motion structural framing and algorithm flow.
3, damage check robot technology
Civil engineering automation aspect, more and more scholars start attention to be placed in the method for view-based access control model.
Montero points out that image method is the chief component of Tunnel testing.CCD camera is often equipped by usual such robot,
And certain distance is kept with wall in operation process.Camera will be equipped with anti-shake apparatus.Robotic arm scans auxiliary camera whole
A metope finds the position of damage by machine vision algorithm later.It is similar, robotic arm may also equip simultaneously sonar and
Camera.For shield tunnel, Yuan etc. proposes a kind of shield tunnel maintaining method of predictability.Xiongyao etc. proposes one kind
Utilize the shield tunnel detection method of GPR and 3D Laser Scanning.This method utilizes laser radar, the three-dimensional in scanning-tunnelling
Geometric shape analyzes image data using machine vision method, and judges structural health conditions accordingly.
The damage of shield tunnel is mainly lining cracking, liner structure leak, lining cutting deterioration etc., these damages can be straight
It is reversed to reflect onto outer surface, it is very suitable to using method based on computer vision.In conjunction with a large amount of city shield tunnels in China
Truth and data are runed, a kind of Shield-bored tunnels Defect inspection system based on deep learning is designed.
Summary of the invention
The present invention will overcome the shortcomings of traditional artificial method for inspecting, provide a kind of Shield-bored tunnels based on deep learning
Defect inspection method.
The equipment for implementing the method for the present invention includes hardware and software.Hardware components include the design of tool car, hardware layout,
Including camera, storage equipment, calculating treatmenting equipment etc..Software section includes the visual performance of shield tunnel damage, based on machine
The classification of the shield damage of vision and research and development and the damage reason location of recognizer.By the image data for collecting shield tunnel damage
It is marked with data set, the characteristics of further cognitive impairment, identification problem is converted to the image classification problem in machine vision.
The problem of the invention solves the following aspects:
First is that solving the overfitting problem that will appear in machine learning model.
Second is that image transmitting is apart from limitation problem in solution conventional tunnel damage monitoring method.
Third is that solving environmental factor to the interference problem of non-destructive tests.
Fourth is that solving the problems, such as to can not achieve continuous on-line monitoring in conventional tunnel damage monitoring.
A kind of Shield-bored tunnels Defect inspection method based on deep learning of the invention, specific implementation step are as follows:
A. it sets up camera and is imaged；
A1. by camera to tunnel damage measure region；
A2. adjustment adjusts lens focus, aperture size and the amplification factor etc. of camera repeatedly, so that completely damaging appearance
In the display visual field；
A3. camera time for exposure and yield value are adjusted, the clear image of tunnel tested region is obtained and it is located at
The suitable position of image finally obtains the optimized image of tunnel damage；
B. the over-fitting processing of area's data is surveyed；
B1. more shield damage datas are collected；
B2. existing damage data collection and other building materials data sets are made full use of；
B3. shift learning and data enhancing technology are made full use of；
C. the evaluation and amendment of model；
C1. training set is formed by the data distribution that camera is collected into and verifying collects；
C2. data are initialized, and is trained and identifies using deep learning model using PyTorch；
C3. model is evaluated using ROC curve and Precision Recall curve；
The identification and positioning of the tunnel D damage；
D1. the initialization pattern 3-dimensional image model and real space for transferring tunnel survey area correspond to model；
D2. multiple characteristic points in initialization pattern 3-dimensional image model are extracted, these characteristic points can sufficiently characterize tunnel
The three-dimensional geometry form in the area Dao Ce, and using these characteristic points as the relevant benchmark of three-dimensional digital image；
D3., reference templates are carried out to digital picture related in the threedimensional model of the picture construction taken by camera
Match, searches for the characteristic point of extraction in existing position in a model, and realize that multi-characteristic points track；
D4. machine learning and training are carried out to multi-characteristic points tracing process, optimizes tracking task, and adaptive algorithm is added
Interference of the variation such as light to tracking is reduced, until tracing characteristic points are met the requirements；
D5. 3-dimensional image model is transformed into practical three-dimensional space model, by practical three-dimensional space model at this time and just
Beginningization three-dimensional space model compares, and obtains the actual image in tunnel survey area；
D6. by the obtained tunnel injury region picture of camera, by optical positioning method, by the position of tool car at this time
It sets and is sent in host.
E. zone position damage automatic on-line monitoring and real-time storage are surveyed in tunnel；
E1. it requires to formulate data sampling frequency and storage strategy according to tested tunnel damage monitoring；
E2. camera is constantly taken pictures, according to the picture construction for describing to take each frame in D step model into
The tracking of row multi-characteristic points, tracking task are transformed into practical threedimensional model after completing and are identified；
E3. check whether D3 completes the acquisition strategies and acquisition store tasks of D2 proposition, if completed, tunnel damage monitoring
Task is completed.
The camera being previously mentioned in above-mentioned steps transmits acquired image using gigabit Ethernet, is stored in meter
It handles in calculation machine hard disk and immediately.
It is carrying out needing from different angles to shoot tunnel survey area with camera in image three-dimensional building, will
The image arrived carries out threedimensional model building, realizes three-dimensional stereoscopic visual function.The space angle of cut and space length of camera determine
Later, the threedimensional model that the picture construction taken obtains is one-to-one with real space model.Reality can be passed through
The object of known geometric dimension establishes the transition matrix of image and real space in the Pixel Dimensions of threedimensional model in space, and
As system calibrating.The threedimensional model to the picture construction taken is needed to chase after during threedimensional model tracing characteristic points
Track training and study, optimize tracing process.
In addition to the camera being previously mentioned and computer etc. in the present invention, additionally provides and a set of be stored in base in computer
In the tunnel damage identification and location on-line monitoring system software platform of deep learning method.
The invention has the advantages that
1. solving lacking for the informationization, automation that traditional artificial routing inspection efficiency is low, digitized degree is low, is unfavorable for detection
Point；
2. solving the problems, such as that image transmitting distance is limited in conventional tunnel damage monitoring method, improves measurement accuracy；
3. this method can be mitigated by the way that deep learning algorithm is added due to environmental factor (light variation and fog significantly
Block at edge) etc. caused by identify fault；
4. since camera acquired image information is transferred to computer, real-time Transmission using high speed gigabit Ethernet
Fastly, it realizes that online data is handled immediately, monitor video can be checked in conjunction with the real-time stress state of drag-line, event may be implemented
Full information playback；
Detailed description of the invention
Fig. 1 is the schematic device for implementing the method for the present invention.
Fig. 2 is main view of the device of implementation the method for the present invention in tunnel.
Fig. 3 is the side view in tunnel for implementing the equipment of the method for the present invention.
Flow chart Fig. 4 of the invention.
Specific embodiment
With reference to the accompanying drawing, technical solution of the present invention is further illustrated.
Referring to attached drawing, the code name in Fig. 1 is respectively indicated:
1 --- vehicle frame,
2 --- seat,
3 --- headlight,
4 --- camera and camera frame,
5 --- controller,
6 --- display and host
A kind of Shield-bored tunnels Defect inspection method based on deep learning of the invention, specific implementation step are as follows:
A. it sets up camera and is imaged；
A1. by camera to tunnel damage measure region；
A2. adjustment adjusts lens focus, aperture size and the amplification factor etc. of camera repeatedly, so that completely damaging appearance
In the display visual field；
A3. camera time for exposure and yield value are adjusted, the clear image of tunnel tested region is obtained and it is located at
The suitable position of image finally obtains the optimized image of tunnel damage；
B. the over-fitting processing of area's data is surveyed；
B1. more shield damage datas are collected；
B2. existing damage data collection and other building materials data sets are made full use of；
B3. shift learning and data enhancing technology are made full use of；
C. the evaluation and amendment of model；
C1. training set is formed by the data distribution that camera is collected into and verifying collects；
C2. data are initialized, and is trained and identifies using deep learning model using PyTorch；
C3. model is evaluated using ROC curve and Precision Recall curve；
The identification and positioning of the tunnel D damage；
D1. the initialization pattern 3-dimensional image model and real space for transferring tunnel survey area correspond to model；
D2. multiple characteristic points in initialization pattern 3-dimensional image model are extracted, these characteristic points can sufficiently characterize tunnel
The three-dimensional geometry form in the area Dao Ce, and using these characteristic points as the relevant benchmark of three-dimensional digital image；
D3., reference templates are carried out to digital picture related in the threedimensional model of the picture construction taken by camera
Match, searches for the characteristic point of extraction in existing position in a model, and realize that multi-characteristic points track；
D4. machine learning and training are carried out to multi-characteristic points tracing process, optimizes tracking task, and adaptive algorithm is added
Interference of the variation such as light to tracking is reduced, until tracing characteristic points are met the requirements；
D5. 3-dimensional image model is transformed into practical three-dimensional space model, by practical three-dimensional space model at this time and just
Beginningization three-dimensional space model compares, and obtains the actual image in tunnel survey area；
D6. by the obtained tunnel injury region picture of camera, by optical positioning method, by the position of tool car at this time
It sets and is sent in host.
E. zone position damage automatic on-line monitoring and real-time storage are surveyed in tunnel；
E1. it requires to formulate data sampling frequency and storage strategy according to tested tunnel damage monitoring；
E2. camera is constantly taken pictures, according to the picture construction for describing to take each frame in D step model into
The tracking of row multi-characteristic points, tracking task are transformed into practical threedimensional model after completing and are identified；
E3. check whether D3 completes the acquisition strategies and acquisition store tasks of D2 proposition, if completed, tunnel damage monitoring
Task is completed.
The camera being previously mentioned in above-mentioned steps transmits acquired image using gigabit Ethernet, is stored in meter
It handles in calculation machine hard disk and immediately.
It is carrying out needing from different angles to shoot tunnel survey area with camera in image three-dimensional building, will
The image arrived carries out threedimensional model building, realizes three-dimensional stereoscopic visual function.The space angle of cut and space length of camera determine
Later, the threedimensional model that the picture construction taken obtains is one-to-one with real space model.Reality can be passed through
The object of known geometric dimension establishes the transition matrix of image and real space in the Pixel Dimensions of threedimensional model in space, and
As system calibrating.The threedimensional model to the picture construction taken is needed to chase after during threedimensional model tracing characteristic points
Track training and study, optimize tracing process.
In addition to the camera being previously mentioned and computer etc. in the present invention, additionally provides and a set of be stored in base in computer
In the tunnel damage identification and location on-line monitoring system software platform of deep learning method.
Content described in this specification case study on implementation is only enumerating to the way of realization of inventive concept, guarantor of the invention
Shield range is not construed as being only limitted to the concrete form that case study on implementation is stated, protection scope of the present invention is art technology
Personnel conceive according to the present invention it is conceivable that equivalent technologies mean.
Claims (2)
1. a kind of Shield-bored tunnels Defect inspection method based on deep learning, specific implementation step are as follows:
A. it sets up camera and is imaged；
A1. by camera to tunnel damage measure region；
A2. adjustment adjusts lens focus, aperture size and the amplification factor etc. of camera repeatedly so that complete damage appear in it is aobvious
Show in the device visual field；
A3. camera time for exposure and yield value are adjusted, the clear image of tunnel tested region is obtained and it is made to be located at image
Suitable position, finally obtain tunnel damage optimized image；
B. the over-fitting processing of area's data is surveyed；
B1. more shield damage datas are collected；
B2. existing damage data collection and other building materials data sets are made full use of；
B3. shift learning and data enhancing technology are made full use of；
C. the evaluation and amendment of model；
C1. training set is formed by the data distribution that camera is collected into and verifying collects；
C2. data are initialized, and is trained and identifies using deep learning model using PyTorch；
C3. model is evaluated using ROC curve and Precision Recall curve；
The identification and positioning of the tunnel D damage；
D1. the initialization pattern 3-dimensional image model and real space for transferring tunnel survey area correspond to model；
D2. multiple characteristic points in initialization pattern 3-dimensional image model are extracted, these characteristic points can sufficiently characterize tunnel survey
The three-dimensional geometry form in area, and using these characteristic points as the relevant benchmark of three-dimensional digital image；
D3., reference templates are carried out to digital picture relevant matches in the threedimensional model of the picture construction taken by camera,
The characteristic point extracted is searched in existing position in a model, and realizes that multi-characteristic points track；
D4. machine learning and training are carried out to multi-characteristic points tracing process, optimizes tracking task, and adaptive algorithm reduction is added
Light etc. changes the interference to tracking, until tracing characteristic points are met the requirements；
D5. 3-dimensional image model is transformed into practical three-dimensional space model, by practical three-dimensional space model at this time and initialized
Three-dimensional space model compares, and obtains the actual image in tunnel survey area；
D6. the position of tool car at this time is sent out by optical positioning method by the obtained tunnel injury region picture of camera
It send into host.
E. zone position damage automatic on-line monitoring and real-time storage are surveyed in tunnel；
E1. it requires to formulate data sampling frequency and storage strategy according to tested tunnel damage monitoring；
E2. camera is constantly taken pictures, and is carried out according to the model for the picture construction for describing to take each frame in D step more
Tracing characteristic points, tracking task are transformed into practical threedimensional model after completing and are identified；
E3. check whether D3 completes the acquisition strategies and acquisition store tasks of D2 proposition, if completed, tunnel damage monitoring task
It completes.
2. a kind of Shield-bored tunnels Defect inspection method based on deep learning as described in claim 1, it is characterised in that:
Camera transmits acquired image using gigabit Ethernet, is stored in hard disc of computer and handles immediately.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910348834.4A CN110220909A (en) | 2019-04-28 | 2019-04-28 | A kind of Shield-bored tunnels Defect inspection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910348834.4A CN110220909A (en) | 2019-04-28 | 2019-04-28 | A kind of Shield-bored tunnels Defect inspection method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110220909A true CN110220909A (en) | 2019-09-10 |
Family
ID=67820043
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910348834.4A Pending CN110220909A (en) | 2019-04-28 | 2019-04-28 | A kind of Shield-bored tunnels Defect inspection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110220909A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110672632A (en) * | 2019-09-24 | 2020-01-10 | 江西交通职业技术学院 | Tunnel disease identification method |
| CN111257415A (en) * | 2020-01-17 | 2020-06-09 | 同济大学 | Tunnel damage detection management system based on mobile train vibration signal |
| CN111878174A (en) * | 2020-08-25 | 2020-11-03 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | High-speed railway tunnel lining block dropping video monitoring method and device |
| CN111967398A (en) * | 2020-08-18 | 2020-11-20 | 创新奇智(合肥)科技有限公司 | Tunnel detection method and device, electronic equipment and readable storage medium |
| CN112393796A (en) * | 2020-11-18 | 2021-02-23 | 重庆工程职业技术学院 | Suspension tunnel vibration noise detection device |
| CN112833808A (en) * | 2021-01-05 | 2021-05-25 | 厦门银江智慧城市技术股份有限公司 | Equipment deformation detection method and device based on convolutional neural network algorithm |
| CN112967221A (en) * | 2020-12-04 | 2021-06-15 | 江苏龙冠新型材料科技有限公司 | Shield constructs section of jurisdiction production and assembles information management system |
| CN112967255A (en) * | 2021-03-09 | 2021-06-15 | 暨南大学 | Shield segment defect type identification and positioning system and method based on deep learning |
| CN113325011A (en) * | 2021-05-24 | 2021-08-31 | 华能秦煤瑞金发电有限责任公司 | Concrete structure damage detection method based on deep learning |
| CN113446953A (en) * | 2021-06-22 | 2021-09-28 | 深圳市市政设计研究院有限公司 | Subway tunnel deformation monitoring system based on digital photogrammetry technology |
| CN113914933A (en) * | 2021-09-27 | 2022-01-11 | 广西大学 | Tunnel overhaul robot |
| CN115013650A (en) * | 2022-05-16 | 2022-09-06 | 华北水利水电大学 | Tracking means for intelligent image recognition of degree of deep learning |
| CN116539635A (en) * | 2023-05-17 | 2023-08-04 | 广州大学 | Unmanned aerial vehicle shield tunnel inspection method based on calculation vision and SLAM |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104990891A (en) * | 2015-06-24 | 2015-10-21 | 中国农业大学 | Method for establishing seed near infrared spectrum and spectral image qualitative analysis model |
| CN105741278A (en) * | 2016-01-26 | 2016-07-06 | 浙江大学 | On-line monitoring method of inhaul cable distribution stress on the basis of computer vision |
| CN108257114A (en) * | 2017-12-29 | 2018-07-06 | 天津市万贸科技有限公司 | A kind of transmission facility defect inspection method based on deep learning |
| CN108918539A (en) * | 2018-07-27 | 2018-11-30 | 同济大学 | A kind of apparent Defect inspection device and method of tunnel structure |
| CN109187549A (en) * | 2018-07-26 | 2019-01-11 | 上海工程技术大学 | A kind of detection method for backing layer edging defect detection station after rearview mirror |
| CN109671071A (en) * | 2018-12-19 | 2019-04-23 | 南京市测绘勘察研究院股份有限公司 | A kind of underground piping defect location and grade determination method based on deep learning |
-
2019
- 2019-04-28 CN CN201910348834.4A patent/CN110220909A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104990891A (en) * | 2015-06-24 | 2015-10-21 | 中国农业大学 | Method for establishing seed near infrared spectrum and spectral image qualitative analysis model |
| CN105741278A (en) * | 2016-01-26 | 2016-07-06 | 浙江大学 | On-line monitoring method of inhaul cable distribution stress on the basis of computer vision |
| CN108257114A (en) * | 2017-12-29 | 2018-07-06 | 天津市万贸科技有限公司 | A kind of transmission facility defect inspection method based on deep learning |
| CN109187549A (en) * | 2018-07-26 | 2019-01-11 | 上海工程技术大学 | A kind of detection method for backing layer edging defect detection station after rearview mirror |
| CN108918539A (en) * | 2018-07-27 | 2018-11-30 | 同济大学 | A kind of apparent Defect inspection device and method of tunnel structure |
| CN109671071A (en) * | 2018-12-19 | 2019-04-23 | 南京市测绘勘察研究院股份有限公司 | A kind of underground piping defect location and grade determination method based on deep learning |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110672632A (en) * | 2019-09-24 | 2020-01-10 | 江西交通职业技术学院 | Tunnel disease identification method |
| CN111257415A (en) * | 2020-01-17 | 2020-06-09 | 同济大学 | Tunnel damage detection management system based on mobile train vibration signal |
| CN111967398A (en) * | 2020-08-18 | 2020-11-20 | 创新奇智(合肥)科技有限公司 | Tunnel detection method and device, electronic equipment and readable storage medium |
| CN111967398B (en) * | 2020-08-18 | 2022-08-09 | 创新奇智(合肥)科技有限公司 | Tunnel detection method and device, electronic equipment and readable storage medium |
| CN111878174B (en) * | 2020-08-25 | 2022-03-01 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | High-speed railway tunnel lining block dropping video monitoring method and device |
| CN111878174A (en) * | 2020-08-25 | 2020-11-03 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | High-speed railway tunnel lining block dropping video monitoring method and device |
| CN112393796A (en) * | 2020-11-18 | 2021-02-23 | 重庆工程职业技术学院 | Suspension tunnel vibration noise detection device |
| CN112393796B (en) * | 2020-11-18 | 2022-11-18 | 重庆工程职业技术学院 | Suspension tunnel vibration noise detection device |
| CN112967221A (en) * | 2020-12-04 | 2021-06-15 | 江苏龙冠新型材料科技有限公司 | Shield constructs section of jurisdiction production and assembles information management system |
| CN112967221B (en) * | 2020-12-04 | 2024-05-14 | 江苏龙冠新型材料科技有限公司 | Shield segment production and assembly information management system |
| CN112833808A (en) * | 2021-01-05 | 2021-05-25 | 厦门银江智慧城市技术股份有限公司 | Equipment deformation detection method and device based on convolutional neural network algorithm |
| CN112967255A (en) * | 2021-03-09 | 2021-06-15 | 暨南大学 | Shield segment defect type identification and positioning system and method based on deep learning |
| CN113325011A (en) * | 2021-05-24 | 2021-08-31 | 华能秦煤瑞金发电有限责任公司 | Concrete structure damage detection method based on deep learning |
| CN113325011B (en) * | 2021-05-24 | 2024-03-29 | 华能秦煤瑞金发电有限责任公司 | Concrete structure damage detection method based on deep learning |
| CN113446953A (en) * | 2021-06-22 | 2021-09-28 | 深圳市市政设计研究院有限公司 | Subway tunnel deformation monitoring system based on digital photogrammetry technology |
| CN113446953B (en) * | 2021-06-22 | 2023-03-07 | 深圳市市政设计研究院有限公司 | Subway tunnel deformation monitoring system based on digital photogrammetry technology |
| CN113914933A (en) * | 2021-09-27 | 2022-01-11 | 广西大学 | Tunnel overhaul robot |
| CN115013650A (en) * | 2022-05-16 | 2022-09-06 | 华北水利水电大学 | Tracking means for intelligent image recognition of degree of deep learning |
| CN116539635A (en) * | 2023-05-17 | 2023-08-04 | 广州大学 | Unmanned aerial vehicle shield tunnel inspection method based on calculation vision and SLAM |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110220909A (en) | A kind of Shield-bored tunnels Defect inspection method based on deep learning | |
| US10818003B2 (en) | High-precision intelligent detection method for bridge diseases based on spatial position | |
| Zhao et al. | Structural health monitoring and inspection of dams based on UAV photogrammetry with image 3D reconstruction | |
| Li et al. | Recognition and evaluation of bridge cracks with modified active contour model and greedy search-based support vector machine | |
| Zou et al. | CNN-based statistics and location estimation of missing components in routine inspection of historic buildings | |
| Han et al. | Detection and location of steel structure surface cracks based on unmanned aerial vehicle images | |
| Xue et al. | Novel SfM-DLT method for metro tunnel 3D reconstruction and Visualization | |
| Wang et al. | A graphics-based digital twin framework for computer vision-based post-earthquake structural inspection and evaluation using unmanned aerial vehicles | |
| Dong et al. | Intelligent segmentation and measurement model for asphalt road cracks based on modified mask R-CNN algorithm | |
| CN114971166A (en) | Lightweight road pavement service quality inspection system | |
| Zhou et al. | Unmanned aerial vehicle-based computer vision for structural vibration measurement and condition assessment: A concise survey | |
| CN115562284A (en) | Method for realizing automatic inspection by high-speed rail box girder inspection robot | |
| CN114663840A (en) | Tunnel environment inspection equipment monitoring method and system | |
| Yiğit et al. | Automatic Crack Detection and Structural Inspection of Cultural Heritage Buildings Using UAV Photogrammetry and Digital Twin Technology | |
| CN112097743A (en) | Building earthquake-resistant characteristic analysis method based on unmanned aerial vehicle technology | |
| Peng et al. | Structural damage detection, localization, and quantification via UAV‐based 3D imaging | |
| Huang et al. | Machine inspection equipment for tunnels: a review | |
| Li et al. | Non-destructive testing research on the surface damage faced by the Shanhaiguan Great Wall based on machine learning | |
| Yan | Automated damage assessment and structural modeling of bridges with visual sensing technology | |
| Deng et al. | Identification of behavioral features of bridge structure based on static image sequences | |
| KR20210076883A (en) | State analysis and risk analysis and modelling method of sewerage pipe | |
| Ao et al. | Damage Extraction of Metro Tunnel Surface from Roughness Map Generated by Point Cloud | |
| Zhang et al. | Research on target detection of excavator in aerial photography environment based on YOLOv4 | |
| Song et al. | Pixel‐Level Crack Identification for Bridge Concrete Structures Using Unmanned Aerial Vehicle Photography and Deep Learning | |
| Gonzalez‐Libreros et al. | Influence of equipment, weather and personnel experience on the digitalization and damage detection of concrete bridges |
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 |