CN114881949A - Tunnel surface defect identification method - Google Patents
Tunnel surface defect identification method Download PDFInfo
- Publication number
- CN114881949A CN114881949A CN202210443202.8A CN202210443202A CN114881949A CN 114881949 A CN114881949 A CN 114881949A CN 202210443202 A CN202210443202 A CN 202210443202A CN 114881949 A CN114881949 A CN 114881949A
- Authority
- CN
- China
- Prior art keywords
- tunnel
- tunnel surface
- image
- point cloud
- edge
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 50
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013136 deep learning model Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 17
- 238000001514 detection method Methods 0.000 claims description 12
- 238000013519 translation Methods 0.000 claims description 12
- 238000013135 deep learning Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000005286 illumination Methods 0.000 abstract description 4
- 230000008859 change Effects 0.000 abstract description 2
- 230000009471 action Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/37—Determination of transform parameters for the alignment of images, i.e. image registration using transform domain methods
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30256—Lane; Road marking
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a tunnel surface defect identification method, which relates to the technical field of tunnel surface defect identification and specifically comprises the following steps: firstly, preprocessing an input tunnel surface image to enable the edge of an object in the tunnel surface image to be clearer; then, tunnel surface defect recognition is carried out on the preprocessed tunnel surface image through a deep learning model; finally, performing three-dimensional reconstruction on the 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and analyzing to obtain the deformation quantity of the surface of the tunnel; the identification method is not interfered by the change of external illumination; for each frame of image, the color of the size type of foreign matters is not required to be paid attention to too much, the exposure degree of the image and other external information are not required to be paid attention to, and the contrast of the target boundary is only required to be improved; and for the same tunnel, two types of data are collected at one time, defect information does not need to be considered too much, deformation of the tunnel is detected while the surface defect of the tunnel is detected, and the recognition rate is relatively high.
Description
Technical Field
The invention relates to the technical field of tunnel surface defect identification, in particular to a tunnel surface defect identification method.
Background
The prior art mostly uses a convolutional neural network scheme for tunnel surface defect identification, the scheme can train a corresponding model and identify defects on the tunnel surface, but the identification degree of the edges of the defects is not high due to diffuse reflection, so that training samples of the model are difficult to obtain, the identification rate and the false alarm rate of the model are not ideal, and the deformation of the tunnel surface cannot be detected.
Disclosure of Invention
The invention aims to: the tunnel surface defect identification method is provided for solving the problems that in the existing tunnel surface defect identification scheme, due to diffuse reflection, the identification degree of the edge of a defect is not high, a training sample of a model is difficult to obtain, the identification rate and the false alarm rate of the model are not ideal, and the deformation of the tunnel surface cannot be detected.
The technical scheme of the invention is as follows:
a tunnel surface defect identification method comprises the following steps:
step S1: preprocessing an input tunnel surface image to enable the edge of an object in the tunnel surface image to be clearer;
step S2: performing tunnel surface defect identification on the preprocessed tunnel surface image through a deep learning model;
step S3: and performing three-dimensional reconstruction on the 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and analyzing to obtain the deformation quantity of the surface of the tunnel.
Further, the detailed step of step S1 is:
step S11: marking and extracting the object edge or the edge-like object in the tunnel surface image so as to position the object edge or the edge-like object;
step S12: and carrying out sharpening operation on the object edge or the edge-like target of the tunnel surface image, and outputting the preprocessed tunnel surface image.
Further, the detailed steps in step S11 are: and calculating a Gabor filtering result matrix of each frame of image in the tunnel surface image, and mapping the result to the preprocessed tunnel surface image.
Further, the detailed step of step S12 is:
carrying out sharpening operation on each frame of image in the tunnel surface image by using a Laplace operator;
and converting the edge type target in the tunnel surface image into a foreground, and converting the rest data into a background so as to strengthen the object edge or the target similar to the edge in the tunnel surface image.
Further, the detailed step of step S2 is:
step S21: training a deep learning algorithm model according to the type of the tunnel surface defect;
step S22: using the deep learning algorithm model trained in the step S21 to perform defect detection on the preprocessed tunnel surface image;
step S23: and classifying and labeling the defect detection result, and storing the detection result image.
Further, acquiring and obtaining the 3D point cloud data of the tunnel surface by a uniform sampling method.
Further, the point cloud data fitting algorithm in step S3 is an ICP algorithm, and specifically includes:
and according to a preset constraint condition, finding a nearest neighbor point in the target point cloud P and the source point cloud Q to be matched, and calculating the minimum value of an error function, thereby iteratively calculating the distance between the adjacent points until a convergence condition is met.
Further, the detailed step of step S3 is:
step S31: performing three-dimensional reconstruction on the acquired 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and fitting to construct a 3D model of the tunnel;
step S32: performing structural geometric analysis on the three-dimensional reconstructed tunnel 3D model to obtain structural geometric data of the tunnel surface;
step S33: the deformation amount of the tunnel surface is calculated from the structural geometry data of the tunnel surface obtained in step S32 and the structural geometry data of the tunnel surface in the 3D model template of the tunnel.
Further, the ICP algorithm comprises the following specific steps:
step A: taking a point set pi from a target point cloud P, wherein the pi belongs to P;
and B: finding out a corresponding point set qi in the source point cloud Q, wherein qi belongs to Q, and min is obtained from qi to pi;
and C: calculating a rotation matrix R and a translation matrix t to minimize an error function E (R, t);
step D: performing rotation and translation transformation on pi by using the rotation matrix R and the translation matrix t obtained by calculation in the step C to obtain a new corresponding point set pi ═ { pi ═ Rpi + t, pi ∈ P };
step E: calculating the average distance d between p' and the corresponding point set qi;
step F: if d is smaller than a given threshold value or larger than a preset maximum iteration number, stopping iterative computation; otherwise, returning to the step B until the convergence condition is met.
Further, the error function E (R, t) in step C is calculated as follows:
wherein:
n-the number of nearest neighbor point pairs;
pi-a point in the target point cloud P;
qi-the closest point in the source point cloud Q corresponding to pi;
r-rotation matrix;
t-is a translation vector;
the calculation formula of the average distance d in the step E is as follows:
compared with the prior art, the invention has the beneficial effects that:
a tunnel surface defect recognition method, carry on the preconditioning to the tunnel surface picture input at first, make the edge of object in the tunnel surface picture clearer; then, tunnel surface defect recognition is carried out on the preprocessed tunnel surface image through a deep learning model; finally, performing three-dimensional reconstruction on the 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and analyzing to obtain the deformation quantity of the surface of the tunnel; the tunnel surface defect identification method is not interfered by external illumination change and is only related to the contrast ratio of the target boundary and the background in the current image; for each frame of image, because the data acquisition equipment can perform illumination compensation according to the field illumination condition, the size type color of the foreign matter does not need to be paid much attention, the exposure degree of the image and other external information do not need to be paid attention, and the contrast of the target boundary only needs to be improved; and for the same tunnel, two types of data are collected at one time, defect information does not need to be considered too much, deformation of the tunnel is detected while the surface defect of the tunnel is detected, and the recognition rate is relatively high.
Drawings
FIG. 1 is a flow chart of a method for identifying defects on a tunnel surface;
FIG. 2 is an exemplary diagram of an input tunnel surface image;
FIG. 3 is an exemplary diagram of a sharpened tunnel surface image;
FIG. 4 is an exemplary graph of Gabor filtering results;
fig. 5 is an exemplary diagram of a preprocessed tunnel surface image.
Detailed Description
It is noted that relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The features and properties of the present invention are described in further detail below with reference to examples.
Example one
The prior art mostly uses a convolutional neural network scheme for tunnel surface defect identification, the scheme can train a corresponding model and identify defects on the tunnel surface, but the identification degree of the edges of the defects is not high due to diffuse reflection, so that training samples of the model are difficult to obtain, the identification rate and the false alarm rate of the model are not ideal, and the deformation of the tunnel surface cannot be detected.
In view of the above problems, the present embodiment provides a method for identifying defects on a tunnel surface, please refer to fig. 1-5, which specifically includes the following steps:
step S1: preprocessing an input tunnel surface image to enable the edge of an object in the tunnel surface image to be clearer; thereby making the edge of the object entering step S2 clearer;
step S2: performing tunnel surface defect identification on the preprocessed tunnel surface image through a deep learning model; the defects mainly comprise: cracks, breakage, water leakage, etc.;
step S3: and performing three-dimensional reconstruction on the 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and analyzing to obtain the deformation quantity of the surface of the tunnel.
In this embodiment, specifically, the detailed step of step S1 is:
step S11: marking and extracting the object edge or the edge-like object in the tunnel surface image so as to position the object edge or the edge-like object; specifically, the detailed step of step S11 is: calculating a Gabor filtering result matrix of each frame image in the input tunnel surface image through a Gabor filter, and mapping the result to the preprocessed tunnel surface image; further, the step S11 provides a target location basis for the sharpening operation in the following step S12;
step S12: sharpening the object edge or the edge-like target of the tunnel surface image, and outputting the preprocessed tunnel surface image; specifically, the detailed step of step S12 is: carrying out sharpening operation on each frame of image in the tunnel surface image by using a Laplace operator; converting an edge target in the tunnel surface image into a foreground, and converting the rest data into a background so as to strengthen the object edge or the edge-like target in the tunnel surface image; and provides data support for step S3, that is, provides data support for foreground and background segmentation of the deep learning algorithm model when processing the image in the following step S3; the purpose of using the laplacian operator is to perform an enhancement highlighting operation on an object edge or an object shaped like an edge in an input tunnel surface image, namely, marking a boundary in the preprocessed tunnel surface image.
The mathematical expression of the two-dimensional Gabor function of the Gabor filter in the step S11 is as follows:
complex number represents:
real part:
imaginary part:
the calculation formula of x 'and y' is as follows:
x′＝xcos θ+ysin θ
y′＝-xsin θ+ycos θ
wherein:
lambda-wavelength; a wavelength parameter representing a cosine function in the Gabor kernel function; the value of λ is established in units of pixels, typically greater than or equal to 2, but not greater than 1/5 of the input image size;
theta-direction; representing the direction of the parallel strips in the Gabor filter kernel; the effective value is a real number from 0 DEG to 360 DEG;
psi-phase offset; representing a phase parameter of a cosine function in the Gabor kernel function; psi ranges from-180 deg. to 180 deg.; wherein the equations corresponding to 0 ° and 180 ° are symmetric with the origin, and the equations of-90 ° and 90 ° are centrosymmetric with respect to the origin;
y-aspect ratio; namely, the space aspect ratio determines the ellipticity of the shape of the Gabor function; when γ is 1, the shape is circular; when γ < 1, the shape elongates with the parallel stripe direction, and usually the value is γ of 0.5.
The half-response spatial frequency bandwidth b of the Gabor filter is related to the ratio of σ/λ, where σ represents the standard deviation of the gaussian factor of the Gabor function; the three have the following relations:
the value of σ cannot be set directly, it only varies with the bandwidth b; the value of the bandwidth must be positive and real, typically 1, when the standard deviation is 0.56 λ relative to the wavelength; the smaller the bandwidth, the larger the standard deviation, the larger the Gabor shape, and the greater the number of parallel stripes visible.
In the embodiment, the laplacian is adopted because the laplacian is more suitable for improving the image blur caused by the diffuse reflection of the light; the laplacian method is a commonly used edge enhancement processor, and is an isotropic second derivative.
The formula for the laplace operation is explained as follows:
for the continuous binary function there are:
wherein:
(x, y) -pixel coordinates of pixel point P in the image;
After finishing, the method comprises the following steps:
g(i，j)＝f(i+1，j)+f(i-1，j)+f(i，j+1)+f(i，j-1)-4f(i，j)
wherein:
(i, j) -pixel coordinates of the pixel point P in the image;
f (i, j) -the gray value of the pixel point P;
the coefficients preceding f (i, j) are the values of the positions that the convolution kernel should take.
The laplacian operator template can be obtained as follows:
Wherein is H 1 Is the convolution kernel for g (i, j).
In this embodiment, specifically, the detailed step of step S2 is:
step S21: training a deep learning algorithm model according to the type of the tunnel surface defect; the training method of the deep learning algorithm model is the prior art, and the training method of the deep learning algorithm model can be known by the technical personnel in the field, and is not repeated herein;
step S22: using the deep learning algorithm model trained in the step S21 to perform defect detection on the preprocessed tunnel surface image;
step S23: classifying and labeling the defect detection result, and storing a detection result image; preferably, the defect detection result is marked on each frame of the input image in the form of a marking box.
In this embodiment, the tunnel surface 3D point cloud data in step S3 is acquired by a uniform sampling method.
In this embodiment, the point cloud data fitting algorithm in step S3 specifically adopts an ICP algorithm, and specifically includes:
and according to a preset constraint condition, finding a nearest neighbor point in the target point cloud P and the source point cloud Q to be matched, and calculating the minimum value of an error function, thereby iteratively calculating the distance between the adjacent points until a convergence condition is met.
In this embodiment, specifically, the detailed step of step S3 is:
step S31: performing three-dimensional reconstruction on the acquired 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and fitting to construct a 3D model of the tunnel; preferably, 3D point cloud data of the tunnel surface is acquired while using the shot image;
step S32: performing structural geometric analysis on the three-dimensional reconstructed tunnel 3D model to obtain structural geometric data of the tunnel surface;
step S33: and calculating the deformation quantity of the tunnel surface through the structural geometric data of the tunnel surface obtained in the step S32 and the structural geometric data of the tunnel surface in the 3D model template of the tunnel.
In this embodiment, it should be noted that the specific steps of the ICP algorithm are as follows:
step A: taking a point set pi from a target point cloud P, wherein the pi belongs to P;
and B, step B: finding out a corresponding point set qi in the source point cloud Q, wherein qi belongs to Q, and min is obtained from qi to pi;
and C: calculating a rotation matrix R and a translation matrix t to minimize an error function E (R, t);
step D: performing rotation and translation transformation on pi by using the rotation matrix R and the translation matrix t obtained by calculation in the step C to obtain a new corresponding point set pi ═ { pi ═ Rpi + t, pi ∈ P };
step E: calculating the average distance d between p' and the corresponding point set qi;
step F: if d is smaller than a given threshold value or larger than a preset maximum iteration number, stopping iterative computation; otherwise, returning to the step B until the convergence condition is met.
In this embodiment, specifically, the calculation formula of the error function E (R, t) in step C is as follows:
wherein:
n-the number of nearest neighbor point pairs;
pi-a point in the target point cloud P;
qi-the closest point in the source point cloud Q corresponding to pi;
r-a rotation matrix;
t-is a translation vector;
the calculation formula of the average distance d in the step E is as follows:
example two
Referring to fig. 1-5, the deep learning algorithm model adopted in step S2 is yolov5-v3.1, which has the advantages of faster detection rate and higher detection accuracy compared with the conventional convolutional neural network; specifically, the training method of the yolov5-v3.1 model and the identification method based on the yolov5-v3.1 model are both prior art, and those skilled in the art can understand the training method and the identification method, and the details are not repeated here.
The above-mentioned embodiments only express the specific embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for those skilled in the art, without departing from the technical idea of the present application, several changes and modifications can be made, which are all within the protection scope of the present application.
Claims (10)
1. A tunnel surface defect identification method is characterized by comprising the following steps:
step S1: preprocessing an input tunnel surface image to enable the edge of an object in the tunnel surface image to be clearer;
step S2: performing tunnel surface defect identification on the preprocessed tunnel surface image through a deep learning model;
step S3: and performing three-dimensional reconstruction on the 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and analyzing to obtain the deformation quantity of the surface of the tunnel.
2. The tunnel surface defect identification method according to claim 1, wherein the detailed steps of the step S1 are as follows:
step S11: marking and extracting the object edge or the edge-like object in the tunnel surface image so as to position the object edge or the edge-like object;
step S12: and carrying out sharpening operation on the object edge or the edge-like target of the tunnel surface image, and outputting the preprocessed tunnel surface image.
3. The tunnel surface defect identification method according to claim 2, wherein the detailed steps in the step S11 are as follows: and calculating a Gabor filtering result matrix of each frame of image in the tunnel surface image, and mapping the result to the preprocessed tunnel surface image.
4. The tunnel surface defect identification method according to claim 2, wherein the detailed steps of the step S12 are as follows:
carrying out sharpening operation on each frame of image in the tunnel surface image by using a Laplace operator;
and converting the edge type target in the tunnel surface image into a foreground, and converting the rest data into a background so as to strengthen the object edge or the target similar to the edge in the tunnel surface image.
5. The tunnel surface defect identification method according to claim 1, wherein the detailed steps of the step S2 are as follows:
step S21: training a deep learning algorithm model according to the type of the tunnel surface defect;
step S22: using the deep learning algorithm model trained in the step S21 to perform defect detection on the preprocessed tunnel surface image;
step S23: and classifying and labeling the defect detection result, and storing the detection result image.
6. The method for identifying the defects on the surface of the tunnel according to claim 1, wherein the 3D point cloud data on the surface of the tunnel is acquired by a uniform sampling method.
7. The tunnel surface defect identification method according to claim 1 or 6, wherein the point cloud data fitting algorithm in the step S3 is an ICP algorithm, and specifically comprises:
and according to a preset constraint condition, finding a nearest neighbor point in the target point cloud P and the source point cloud Q to be matched, and calculating the minimum value of an error function, thereby iteratively calculating the distance between the adjacent points until a convergence condition is met.
8. The tunnel surface defect identification method according to claim 7, wherein the detailed steps of the step S3 are as follows:
step S31: performing three-dimensional reconstruction on the acquired 3D point cloud data on the surface of the tunnel by using a point cloud data fitting algorithm, and fitting to construct a 3D model of the tunnel;
step S32: performing structural geometric analysis on the three-dimensional reconstructed tunnel 3D model to obtain structural geometric data of the tunnel surface;
step S33: the deformation amount of the tunnel surface is calculated from the structural geometry data of the tunnel surface obtained in step S32 and the structural geometry data of the tunnel surface in the 3D model template of the tunnel.
9. The tunnel surface defect identification method according to claim 8, wherein the ICP algorithm comprises the following specific steps:
step A: taking a point set pi from a target point cloud P, wherein the pi belongs to P;
and B: finding out a corresponding point set qi in the source point cloud Q, wherein qi belongs to Q, and min is obtained from qi to pi;
and C: calculating a rotation matrix R and a translation matrix t to minimize an error function E (R, t);
step D: performing rotation and translation transformation on pi by using the rotation matrix R and the translation matrix t obtained by calculation in the step C to obtain a new corresponding point set pi ═ { pi ═ Rpi + t, pi ∈ P };
step E: calculating the average distance d between p' and the corresponding point set qi;
step F: if d is smaller than a given threshold value or larger than a preset maximum iteration number, stopping iterative computation; otherwise, returning to the step B until the convergence condition is met.
10. A tunnel surface defect identifying method according to claim 9, wherein the error function E (R, t) in step C is calculated as follows:
wherein:
n-the number of nearest neighbor point pairs;
pi-a point in the target point cloud P;
qi-the closest point in the source point cloud Q corresponding to pi;
r-rotation matrix;
t-is a translation vector;
the calculation formula of the average distance d in the step E is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210443202.8A CN114881949A (en) | 2022-04-26 | 2022-04-26 | Tunnel surface defect identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210443202.8A CN114881949A (en) | 2022-04-26 | 2022-04-26 | Tunnel surface defect identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114881949A true CN114881949A (en) | 2022-08-09 |
Family
ID=82672063
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210443202.8A Pending CN114881949A (en) | 2022-04-26 | 2022-04-26 | Tunnel surface defect identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114881949A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116026246A (en) * | 2022-12-15 | 2023-04-28 | 中铁隧道局集团有限公司 | Tunnel surrounding rock deformation area identification device and method based on Unet neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104502990A (en) * | 2015-01-06 | 2015-04-08 | 中国地质大学（武汉） | Geological survey method of tunnel face based on digital image |
| CN108801171A (en) * | 2018-08-23 | 2018-11-13 | 南京航空航天大学 | A kind of tunnel cross-section deformation analytical method and device |
| CN110766689A (en) * | 2019-11-06 | 2020-02-07 | 深圳微品致远信息科技有限公司 | Method and device for detecting article image defects based on convolutional neural network |
| CN111612758A (en) * | 2020-05-19 | 2020-09-01 | 南京航空航天大学 | Tunnel surface defect detection method based on convolutional neural network |
| CN112348014A (en) * | 2020-11-03 | 2021-02-09 | 招商局重庆公路工程检测中心有限公司 | Tunnel face surrounding rock grade rapid identification method based on machine vision |
-
2022
- 2022-04-26 CN CN202210443202.8A patent/CN114881949A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104502990A (en) * | 2015-01-06 | 2015-04-08 | 中国地质大学（武汉） | Geological survey method of tunnel face based on digital image |
| CN108801171A (en) * | 2018-08-23 | 2018-11-13 | 南京航空航天大学 | A kind of tunnel cross-section deformation analytical method and device |
| CN110766689A (en) * | 2019-11-06 | 2020-02-07 | 深圳微品致远信息科技有限公司 | Method and device for detecting article image defects based on convolutional neural network |
| CN111612758A (en) * | 2020-05-19 | 2020-09-01 | 南京航空航天大学 | Tunnel surface defect detection method based on convolutional neural network |
| CN112348014A (en) * | 2020-11-03 | 2021-02-09 | 招商局重庆公路工程检测中心有限公司 | Tunnel face surrounding rock grade rapid identification method based on machine vision |
Non-Patent Citations (1)
| Title |
|---|
| 程子聪;苏勇;: "基于三维点云配准的矩形隧道管片预拼装方法研究", 建筑施工, no. 04, 25 April 2017 (2017-04-25) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116026246A (en) * | 2022-12-15 | 2023-04-28 | 中铁隧道局集团有限公司 | Tunnel surrounding rock deformation area identification device and method based on Unet neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921176B (en) | Pointer instrument positioning and identifying method based on machine vision | |
| CN110349207B (en) | Visual positioning method in complex environment | |
| CN109903313B (en) | Real-time pose tracking method based on target three-dimensional model | |
| JP6660313B2 (en) | Detection of nuclear edges using image analysis | |
| CN110021024B (en) | Image segmentation method based on LBP and chain code technology | |
| CN109919960B (en) | Image continuous edge detection method based on multi-scale Gabor filter | |
| WO2019042232A1 (en) | Fast and robust multimodal remote sensing image matching method and system | |
| CN108921813B (en) | Unmanned aerial vehicle detection bridge structure crack identification method based on machine vision | |
| CN112233116B (en) | Concave-convex mark visual detection method based on neighborhood decision and gray level co-occurrence matrix description | |
| CN103440644B (en) | A kind of multi-scale image weak edge detection method based on minimum description length | |
| CN110473221B (en) | Automatic target object scanning system and method | |
| CN110288571B (en) | High-speed rail contact net insulator abnormity detection method based on image processing | |
| Pei et al. | Effects of image degradations to cnn-based image classification | |
| CN112308872B (en) | Image edge detection method based on multi-scale Gabor first derivative | |
| CN110222661B (en) | Feature extraction method for moving target identification and tracking | |
| CN111161222A (en) | Printing roller defect detection method based on visual saliency | |
| CN112396618B (en) | Grain boundary extraction and grain size measurement method based on image processing | |
| CN112651259A (en) | Two-dimensional code positioning method and mobile robot positioning method based on two-dimensional code | |
| CN117094916B (en) | Visual inspection method for municipal bridge support | |
| CN108596176B (en) | Method and device for identifying diatom types of extracted diatom areas | |
| Lou et al. | Nonlocal similarity image filtering | |
| CN113421210B (en) | Surface point Yun Chong construction method based on binocular stereoscopic vision | |
| Liu et al. | Sparsity-based image monitoring of crystal size distribution during crystallization | |
| CN114881949A (en) | Tunnel surface defect identification method | |
| CN115272306A (en) | Solar cell panel grid line enhancement method utilizing gradient operation |
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 |