CN113325011B - Concrete structure damage detection method based on deep learning - Google Patents
Concrete structure damage detection method based on deep learning Download PDFInfo
- Publication number
- CN113325011B CN113325011B CN202110565229.XA CN202110565229A CN113325011B CN 113325011 B CN113325011 B CN 113325011B CN 202110565229 A CN202110565229 A CN 202110565229A CN 113325011 B CN113325011 B CN 113325011B
- Authority
- CN
- China
- Prior art keywords
- damage
- data
- model
- observation
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 56
- 238000013135 deep learning Methods 0.000 title claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000008569 process Effects 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims abstract description 9
- 230000005855 radiation Effects 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims abstract description 4
- 238000010606 normalization Methods 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 7
- 238000003384 imaging method Methods 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000009432 framing Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 2
- 238000002203 pretreatment Methods 0.000 claims 1
- 238000004422 calculation algorithm Methods 0.000 abstract description 5
- 206010063385 Intellectualisation Diseases 0.000 abstract 1
- 230000007613 environmental effect Effects 0.000 abstract 1
- 238000005286 illumination Methods 0.000 abstract 1
- 238000007689 inspection Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- 239000004566 building material Substances 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
- G01N23/04—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/02—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems using reflection of acoustic waves
- G01S15/06—Systems determining the position data of a target
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Acoustics & Sound (AREA)
- Computer Networks & Wireless Communication (AREA)
- Human Computer Interaction (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a concrete structure damage detection method based on deep learning, which comprises the steps of erecting a camera to shoot a detected member, and carrying out radiation observation on the detected member based on radiation operation to obtain observation data; preprocessing observation data and collecting historical concrete member damage data to form an observation damage comparison database; constructing a damage detection model based on a deep learning neural network, taking observed data as input parameters of the damage detection model, and taking historical concrete member damage data as comparison parameters for iterative optimization; and performing process evaluation on iterative optimization of the damage detection model by using the Lorentz curve and the confusion matrix curve respectively until the detection data meeting the confidence interval is output. The invention is non-contact detection, can avoid personnel radiation safety problem, can reduce interference problem of environmental factors such as illumination and the like to flaw detection recognition based on an algorithm of deep learning, and realizes intellectualization and digitization by calculating and processing data in real time.
Description
Technical Field
The invention relates to the technical field of neural network calculation and concrete structure damage detection, in particular to a concrete structure damage detection method based on deep learning.
Background
The concrete is used as a traditional building material and has the advantages of wide application, convenient construction, convenient material supply and the like, the high-rise building of the concrete structure is gradually raised in China, but the durability of the concrete structure is also gradually reduced along with the time, most town building communities in China are the concrete structures, the bearing capacity of the structure can be changed due to the influence of various factors on the concrete structure exposed to the atmosphere, in the processes of production, transportation, pouring and the like, the components of the concrete structure possibly cannot reach standards, the damage detection method of the general concrete structure adopts a manual inspection mode, the prior art is low in efficiency, the intelligent data degree is low, and the space positions of the structural components are unfavorable for the general manual inspection, so that the manual inspection mode is difficult to meet the detection needs of the structure along with the development and popularization of the concrete structure.
Disclosure of Invention
This section is intended to outline some aspects of embodiments of the invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section as well as in the description summary and in the title of the application, to avoid obscuring the purpose of this section, the description summary and the title of the invention, which should not be used to limit the scope of the invention.
The present invention has been made in view of the above-described problems occurring in the prior art.
Therefore, the invention provides a concrete structure damage detection method based on deep learning, which can solve the just-needed problem of the current concrete structure damage detection.
In order to solve the technical problems, the invention provides the following technical scheme: erecting a camera to shoot a tested member, and carrying out radiation observation on the tested member based on radiation operation to obtain observation data; preprocessing the observation data and collecting historical concrete member damage data to form an observation damage comparison database; constructing a damage detection model based on a deep learning neural network, taking the observed data as input parameters of the damage detection model, and taking the historical concrete member damage data as comparison parameters for iterative optimization; and performing process evaluation on the iterative optimization of the damage detection model by using the Lorentz curve and the confusion matrix curve respectively until the detection data meeting the confidence interval is output.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the step of obtaining the observation data comprises the steps of erecting the camera to aim at the tested component for shooting; the aperture and the focal length of a camera in the camera are adjusted, so that the camera can completely and clearly observe and shoot a damaged image of the tested component, and the damaged image is displayed in a display screen; and adjusting exposure time and gain value to enable the tested component to be in a reasonable photographed position, and obtaining an optimal image of the surface damage of the tested component.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the radioscopy comprises the steps of adjusting the direction and angle of the unmanned aerial vehicle, and the voltage, current, focus and shooting time parameters of the ray end through a control terminal based on the ray operation, and carrying out radioscopy on the detected component.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the preprocessing comprises the steps of marking the damage position in the optimal image by utilizing Lablemg; cleaning and framing labels by combining damage data enhancement and image normalization; and screening the processed data based on a continuous wavelet transformation strategy, and adding the historical concrete member damage data to form the observation damage comparison database.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the construction of the damage detection model includes,
O t ＝∫(V·g(U·X t +W·s t-1 ))
wherein ∈ () and g () are activation functions, U and V are argument matrices, s t For the value of the hidden layer at time t, X t The input parameter X at time t.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: weighting and summing the Q value of the damage detection model and the similarity of the observed damage comparison database by using an attention mechanism, and outputting the detection data meeting a confidence interval, wherein the steps include comparing the Q value with the similarity of the observed damage comparison database; normalizing the obtained similarity; and calculating the weight of the normalized similarity, and carrying out weighted summation on all the observed values to obtain the detection vector after the Attention processing.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the process evaluation may be performed by performing a process evaluation,
Q＝XW Q ,K＝XW K ,V＝XW V ,
wherein W is Q 、W K 、W V Q, K, V, different emmbedding matrices, d can be obtained even if the same parameter X is input k Is a normalization factor, performs normalization processing and calculates the similarity in Q, and Z is X subjected to Attention processing.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: performing damage identification positioning on the detected member by utilizing the observed data, wherein the method comprises the steps of calling an initialized pattern three-dimensional image model and an actual space corresponding model of the detected member; extracting a plurality of characteristic points in the initialized pattern three-dimensional image model to enable the characteristic points to represent the three-dimensional geometric form of the tested member, and taking the characteristic points as references related to three-dimensional digital images; carrying out semantic segmentation on the shot and scanned images by using a reference template, constructing a three-dimensional model, carrying out digital image correlation matching, searching the position of the extracted characteristic point in the current three-dimensional model, and completing multi-characteristic point tracking; performing machine learning and training on the multi-feature point tracking process, and optimizing a tracking task until feature point tracking meets the requirements; converting the three-dimensional model into the actual space model, and comparing the actual space model with the initialized three-dimensional space model to obtain an actual image of the tested member; and sending the current position of the unmanned aerial vehicle to the terminal through an ultrasonic positioning method according to the photographed picture and the imaging scanning result of the radiographic film.
As a preferable scheme of the concrete structure damage detection method based on deep learning, the invention comprises the following steps: the detected damage identification information is monitored on line and stored in real time, and the method comprises the steps of formulating a data sampling frequency and a storage strategy according to the monitoring requirement of the detected component; transmitting the pictures shot by the camera and the scanned pictures imaged by the radiographic film back to a terminal computer by using an Ethernet, and storing the pictures in the computer; and checking whether the formulated acquisition strategy and the formulated storage task are met, and circularly checking until the task is completed.
The invention has the beneficial effects that: the invention improves the cognition of damage by collecting the image of the building concrete structure, the radiographic imaging data and the data set label, converts the identification problem into the image classification problem in machine vision, combines the internal damage of the emission flaw detection concrete structure member with the surface damage of the machine vision flaw detection member, and detects the member more comprehensively.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:
fig. 1 is a schematic flow chart of a method for detecting damage to a concrete structure based on deep learning according to an embodiment of the invention;
fig. 2 is a schematic diagram of an algorithm structural layer of a deep learning neural network based on a deep learning concrete structure damage detection method according to an embodiment of the invention.
Detailed Description
So that the manner in which the above recited objects, features and advantages of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to the embodiments, some of which are illustrated in the appended drawings. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways other than those described herein, and persons skilled in the art will readily appreciate that the present invention is not limited to the specific embodiments disclosed below.
Further, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic can be included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
While the embodiments of the present invention have been illustrated and described in detail in the drawings, the cross-sectional view of the device structure is not to scale in the general sense for ease of illustration, and the drawings are merely exemplary and should not be construed as limiting the scope of the invention. In addition, the three-dimensional dimensions of length, width and depth should be included in actual fabrication.
Also in the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper, lower, inner and outer", etc. are based on the orientation or positional relationship shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The terms "mounted, connected, and coupled" should be construed broadly in this disclosure unless otherwise specifically indicated and defined, such as: can be fixed connection, detachable connection or integral connection; it may also be a mechanical connection, an electrical connection, or a direct connection, or may be indirectly connected through an intermediate medium, or may be a communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Example 1
Referring to fig. 1 and 2, for a first embodiment of the present invention, there is provided a method for detecting damage to a concrete structure based on deep learning, including:
s1: erecting a camera to shoot a tested component, and carrying out radiation observation on the tested component based on the radiation operation to obtain observation data. The obtaining observation data includes:
erecting a camera to aim at a tested component for shooting;
the aperture and focal length of the camera in the camera are adjusted, so that the camera can completely and clearly observe and shoot the damaged image of the tested component, and the damaged image is displayed in the display screen;
and adjusting the exposure time and the gain value to enable the tested member to be in a reasonable photographed position, and obtaining an optimal image of the surface damage of the tested member.
Further, the radiological observation includes:
and adjusting the direction and angle of the unmanned aerial vehicle, and the voltage, current, focus and shooting time parameters of the ray end through the control terminal based on the ray operation, and carrying out radioscopy on the detected component.
S2: preprocessing the observation data and collecting historical concrete member damage data to form an observation damage comparison database. The step is to be noted, the preprocessing includes:
labeling the damage position in the optimal image by using Lablemg;
cleaning and framing labels by combining damage data enhancement and image normalization;
and screening the processed data based on a continuous wavelet transformation strategy, and adding historical concrete member damage data to form an observation damage comparison database.
S3: and constructing a damage detection model based on the deep learning neural network, taking the observed data as input parameters of the damage detection model, and taking the historical concrete member damage data as comparison parameters for iterative optimization. Referring to fig. 2, it should be further noted that, constructing the damage detection model includes:
O t ＝∫(V·g(U·X t +W·s t-1 ))
wherein ∈ () and g () are activation functions, U and V are argument matrices, s t For the value of the hidden layer at time t, X t The input parameter X at time t.
S4: and performing process evaluation on iterative optimization of the damage detection model by using the Lorentz curve and the confusion matrix curve respectively until the detection data meeting the confidence interval is output. It should also be noted that this step is also described,
the method for detecting the damage by using the attention mechanism to carry out weighted summation on the Q value of the damage detection model and the similarity of the observed damage comparison database, and outputting detection data meeting a confidence interval comprises the following steps:
comparing the Q value with the similarity of the observed damage comparison database;
normalizing the obtained similarity;
and calculating the weight of the normalized similarity, and carrying out weighted summation on all the observed values to obtain the detection vector after the Attention processing.
Further, performing the process evaluation includes:
Q＝XW Q ,K＝XW K ,V＝XW V ,
wherein W is Q 、W K 、W V Q, K, V, different emmbedding matrices, d can be obtained even if the same parameter X is input k Is a normalization factor, performs normalization processing and calculates the similarity in Q, and Z is X subjected to Attention processing.
Preferably, in this embodiment, the damage identifying and positioning of the measured member by using the observed data includes:
invoking an initialization pattern three-dimensional image model and an actual space corresponding model of a tested component;
extracting a plurality of characteristic points in the three-dimensional image model of the initialization pattern to enable the characteristic points to represent the three-dimensional geometric form of the tested member, and taking the characteristic points as references related to the three-dimensional digital image;
carrying out semantic segmentation on the shot and scanned images by using a reference template, constructing a three-dimensional model, carrying out digital image correlation matching, searching the position of the extracted characteristic point in the current three-dimensional model, and completing multi-characteristic point tracking;
machine learning and training are carried out on the multi-feature point tracking process, and a tracking task is optimized until feature point tracking meets the requirements;
converting the three-dimensional model into an actual space model, and comparing the actual space model with the initialized three-dimensional space model to obtain an actual image of the tested member;
and sending the current position of the unmanned aerial vehicle to the terminal through an ultrasonic positioning method according to the photographed picture and the imaging scanning result of the radiographic film.
Still further, the present embodiment can perform online monitoring and real-time storage on the detected damage identification information, including:
formulating a data sampling frequency and a storage strategy according to the monitoring requirement of the tested component;
transmitting the photo shot by the camera and the scanned photo imaged by the radiographic film back to the terminal computer by using the Ethernet, and storing the photo and the scanned photo in the computer;
and checking whether the formulated acquisition strategy and the formulated storage task are met, and circularly checking until the task is completed.
Example 2
In order to better verify and explain the technical effects adopted in the method, the traditional concrete transfer learning identification detection method is selected to be compared with the method, and the test results are compared by a scientific demonstration means to verify the true effects of the method.
The traditional concrete transfer learning identification detection method has complex algorithm, needs real-time adjustment trial calculation operation, cannot leave personnel, has higher cost and longer time consumption, and has higher efficiency, stronger robustness and applicability and lower economic cost compared with the traditional method for verifying the method.
Test environment: (1) Shooting a concrete structure related image through a camera carried by the unmanned aerial vehicle, and carrying out image preprocessing by combining with a computer to obtain 100 groups of samples to be detected;
(2) Respectively inputting the operation programs of the deep learning neural network algorithm of the method and the migration learning algorithm of the traditional method into a MATLB simulation platform;
(3) And observing curve comparison information output by the display interface, and recording in a form of a table.
Table 1: the detection error and the efficiency are compared with a data table.
With reference to table 1, it can be intuitively seen that the method of the present invention has more remarkable beneficial effects than the conventional method, and under the same test sample conditions, the time efficiency and the accuracy of detection and identification are both significantly improved.
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
Claims (1)
1. A concrete structure damage detection method based on deep learning is characterized by comprising the following steps: comprising the steps of (a) a step of,
erecting a camera to shoot a tested member, and carrying out radiation observation on the tested member based on radiation operation to obtain observation data;
preprocessing the observation data and collecting historical concrete member damage data to form an observation damage comparison database;
constructing a damage detection model based on a deep learning neural network, taking the observed data as input parameters of the damage detection model, and carrying out iterative optimization by taking the historical concrete member damage data as comparison parameters;
performing process evaluation on iterative optimization of the damage detection model by using the Lorentz curve and the confusion matrix curve respectively until detection data meeting a confidence interval is output;
the obtaining of the observation data includes,
erecting the camera to aim at the tested component for shooting;
the aperture and the focal length of a camera in the camera are adjusted, so that the camera can completely and clearly observe and shoot a damaged image of the tested component, and the damaged image is displayed in a display screen;
adjusting exposure time and gain value to enable the tested component to be in a reasonable photographed position, and obtaining an optimal image of the surface damage of the tested component;
the radiological observation includes a plurality of medical procedures,
adjusting the direction and angle of the unmanned aerial vehicle, and the voltage, current, focus and shooting time parameters of the ray end through a control terminal based on the ray operation, and carrying out the radioscopy on the detected component;
the pre-treatment may comprise the steps of,
labeling the damage position in the optimal image by using Lablemg;
cleaning and framing labels by combining damage data enhancement and image normalization;
screening the processed data based on a continuous wavelet transformation strategy, and adding the historical concrete member damage data to form the observation damage comparison database;
the construction of the damage detection model includes,
O t ＝∫(V·g(U·X t +W·s t-1 ))
wherein ∈ () and g () are activation functions, U and V are argument matrices, s t For the value of the hidden layer at time t, X t The input parameter X is the time t;
the Q value of the damage detection model and the similarity of the observed damage comparison database are weighted and summed by using an attention mechanism, and the detection data meeting a confidence interval is output, wherein the detection data comprises,
comparing the Q value with the similarity of the observed damage comparison database;
normalizing the obtained similarity;
calculating the weight of the normalized similarity, and carrying out weighted summation on all the observed values to obtain a detection vector after the Attention processing;
the process evaluation may be performed by performing a process evaluation,
Q＝XW Q ,K＝XW K ,V＝XW V
wherein W is Q 、W K 、W V According to the method, even if the same parameters X are input, different EMBedding matrixes which are Q, K, V are obtained, dk is a normalization factor, normalization processing is carried out, similarity in Q is calculated, and Z is X subjected to Attention processing;
performing damage identification positioning on the detected member by using the observed data, including,
calling an initialization pattern three-dimensional image model and an actual space corresponding model of the tested component;
extracting a plurality of characteristic points in the initialized pattern three-dimensional image model to enable the characteristic points to represent the three-dimensional geometric form of the tested member, and taking the characteristic points as references related to three-dimensional digital images;
carrying out semantic segmentation on the shot and scanned images by using a reference template, constructing a three-dimensional model, carrying out digital image correlation matching, searching the position of the extracted characteristic point in the current three-dimensional model, and completing multi-characteristic point tracking;
performing machine learning and training on the multi-feature point tracking process, and optimizing a tracking task until feature point tracking meets the requirements;
converting the three-dimensional model into a model corresponding to the actual space, and comparing the model with the initialized three-dimensional model to obtain an actual image of the tested member;
according to the photographed picture and the imaging scanning result of the ray film, the current position of the unmanned aerial vehicle is sent to the terminal through an ultrasonic positioning method;
the detected damage identification information is monitored on line and stored in real time, including,
formulating a data sampling frequency and a storage strategy according to the monitoring requirement of the tested component;
transmitting the pictures shot by the camera and the scanned pictures imaged by the radiographic film back to a terminal computer by using an Ethernet, and storing the pictures in the computer;
and checking whether the formulated acquisition strategy and the formulated storage task are met, and circularly checking until the task is completed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110565229.XA CN113325011B (en) | 2021-05-24 | 2021-05-24 | Concrete structure damage detection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110565229.XA CN113325011B (en) | 2021-05-24 | 2021-05-24 | Concrete structure damage detection method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113325011A CN113325011A (en) | 2021-08-31 |
| CN113325011B true CN113325011B (en) | 2024-03-29 |
Family
ID=77416503
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110565229.XA Active CN113325011B (en) | 2021-05-24 | 2021-05-24 | Concrete structure damage detection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113325011B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110220909A (en) * | 2019-04-28 | 2019-09-10 | 浙江大学 | A kind of Shield-bored tunnels Defect inspection method based on deep learning |
| CN110569700A (en) * | 2018-09-26 | 2019-12-13 | 阿里巴巴集团控股有限公司 | method and device for optimizing damage identification result |
| CN111986164A (en) * | 2020-07-31 | 2020-11-24 | 河海大学 | Road crack detection method based on multi-source Unet + Attention network migration |
-
2021
- 2021-05-24 CN CN202110565229.XA patent/CN113325011B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569700A (en) * | 2018-09-26 | 2019-12-13 | 阿里巴巴集团控股有限公司 | method and device for optimizing damage identification result |
| CN110220909A (en) * | 2019-04-28 | 2019-09-10 | 浙江大学 | A kind of Shield-bored tunnels Defect inspection method based on deep learning |
| CN111986164A (en) * | 2020-07-31 | 2020-11-24 | 河海大学 | Road crack detection method based on multi-source Unet + Attention network migration |
Non-Patent Citations (4)
| Title |
|---|
| 哈纳提•吐尔森哈力.融合自注意力机制与深度学习的混凝土表面裂隙智能识别.铁道科学与工程学报.2021,第18卷(第14期),844-852. * |
| 基于多特征融合的文本分类算法;朱国杰;2019中国自动化大会(CAC2019)论文集;第247-250页 * |
| 基于深度神经网络的心肌缺血心电信号识别研究;张金玲;医药卫生科技辑;E062-9 * |
| 融合自注意力机制与深度学习的混凝土表面裂隙智能识别;哈纳提•吐尔森哈力;铁道科学与工程学报;第18卷(第4期);844-852 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113325011A (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9684960B2 (en) | Automated histological diagnosis of bacterial infection using image analysis | |
| CN114693661A (en) | Rapid sorting method based on deep learning | |
| CN115330876B (en) | Target template graph matching and positioning method based on twin network and central position estimation | |
| CN116524313A (en) | Defect detection method and related device based on deep learning and multi-mode image | |
| CN110427943A (en) | A kind of intelligent electric meter technique for partitioning based on R-CNN | |
| CN115170816A (en) | Multi-scale feature extraction system and method and fan blade defect detection method | |
| CN109978897B (en) | Registration method and device for heterogeneous remote sensing images of multi-scale generation countermeasure network | |
| CN114359552A (en) | Instrument image identification method based on inspection robot | |
| CN113658147A (en) | Workpiece size measuring device and method based on deep learning | |
| CN113325011B (en) | Concrete structure damage detection method based on deep learning | |
| CN116740135B (en) | Infrared dim target tracking method and device, electronic equipment and storage medium | |
| CN117133041A (en) | Three-dimensional reconstruction network face recognition method, system, equipment and medium based on deep learning | |
| CN115471782B (en) | Unmanned ship-oriented infrared ship target detection method and device | |
| CN111798516A (en) | Method for detecting running state quantity of bridge crane equipment and analyzing errors | |
| CN113887652B (en) | Remote sensing image weak and small target detection method based on morphology and multi-example learning | |
| CN116047418A (en) | Multi-mode radar active deception jamming identification method based on small sample | |
| CN111046861B (en) | Method for identifying infrared image, method for constructing identification model and application | |
| CN112508891B (en) | AI intelligent defect identification magnetic powder flaw detection system based on mobile phone and method thereof | |
| CN109993782B (en) | Heterogeneous remote sensing image registration method and device for ring-shaped generation countermeasure network | |
| Chen et al. | Video-based crack detection using deep learning and Nave Bayes data fusion | |
| CN117475091B (en) | High-precision 3D model generation method and system | |
| CN115131619B (en) | Extra-high voltage part sorting method and system based on point cloud and image fusion | |
| CN117830429A (en) | Heterogeneous sensor coordinate mapping method based on deep learning | |
| CN117437627A (en) | Digital instrument image identification reading method for petroleum refining scene | |
| CN116893162A (en) | Rare anti-nuclear antibody karyotype detection method based on YOLO and attention neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |