CN112446852A - Tunnel imaging plane display method and intelligent defect identification system - Google Patents
Tunnel imaging plane display method and intelligent defect identification system Download PDFInfo
- Publication number
- CN112446852A CN112446852A CN201910812607.2A CN201910812607A CN112446852A CN 112446852 A CN112446852 A CN 112446852A CN 201910812607 A CN201910812607 A CN 201910812607A CN 112446852 A CN112446852 A CN 112446852A
- Authority
- CN
- China
- Prior art keywords
- tunnel
- image
- data
- dimensional
- defect
- 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.)
- Granted
Links
- 230000007547 defect Effects 0.000 title claims abstract description 94
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000003384 imaging method Methods 0.000 title claims abstract description 27
- 238000012545 processing Methods 0.000 claims abstract description 47
- 238000009826 distribution Methods 0.000 claims abstract description 11
- 238000013507 mapping Methods 0.000 claims abstract description 7
- 238000007619 statistical method Methods 0.000 claims abstract description 5
- 238000005259 measurement Methods 0.000 claims description 8
- 238000012790 confirmation Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 230000017525 heat dissipation Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000007667 floating Methods 0.000 claims description 3
- 230000003137 locomotive effect Effects 0.000 claims description 3
- 238000005096 rolling process Methods 0.000 claims description 3
- 230000001960 triggered effect Effects 0.000 claims description 3
- 238000009423 ventilation Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 11
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000000605 extraction Methods 0.000 description 10
- 238000012423 maintenance Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000013527 convolutional neural network Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000005457 optimization Methods 0.000 description 7
- 238000005070 sampling Methods 0.000 description 7
- 230000010354 integration Effects 0.000 description 6
- 230000002093 peripheral effect Effects 0.000 description 6
- 238000007781 pre-processing Methods 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 238000007635 classification algorithm Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000007639 printing Methods 0.000 description 4
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 239000013589 supplement Substances 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000002922 simulated annealing Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 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
- G06T7/0004—Industrial image inspection
-
- 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
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Computer Graphics (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Architecture (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention belongs to the technical field of tunnel imaging, and discloses a tunnel imaging plane display method and a defect intelligent identification system, wherein tunnel image data and three-dimensional contour data are acquired and preprocessed through data acquisition and processing steps, and defects contained in a tunnel image are identified through a model learning method; fusing original tunnel image data and three-dimensional contour data of the tunnel by using an image splicing and texture mapping method to form a three-dimensional image display of the tunnel; identifying defects contained in the image of the tunnel by a model learning method, and displaying the defects on the three-dimensional image of the tunnel in a form of identification; and expanding the tunnel three-dimensional image with the identified defects into a plane according to a flattening graph mode for displaying, and generating a tunnel health condition report containing time variation trend, defect distribution and defect type data on the tunnel image expanded into the plane through statistical analysis of current and historical data.
Description
Technical Field
The invention belongs to the technical field of tunnel imaging, and particularly relates to a tunnel imaging plane display method and a defect intelligent identification system.
Background
The tunnel has wide application in China and already occupies an important position in transportation and economic development in China. Tunnel quality detection is an important means for ensuring safe operation of tunnels.
The traditional tunnel quality detection method is a hole-opening or slot-opening sampling detection method, but has the great limitations of low detection efficiency, poor representativeness, great contingency, need to damage the internal structure of the tunnel and the like, and is gradually replaced by other detection methods. For example, in the existing geological radar detection method, 7 survey lines are uniformly arranged in a tunnel along the side wall, arch springing, arch waist and arch crown of the tunnel, sampled data of each survey line are analyzed to obtain 7 geological radar section images, and then the 7 geological radar section images are comprehensively analyzed to obtain the tunnel quality condition. However, the above detection method has the following disadvantages: geological radar profile images are not visual, and people cannot observe tunnel quality distribution conditions quickly.
In the prior art, as the chinese patent application publication with publication number CN105862556A, publication time of 2016, 8, 17 and titled "intelligent vehicle-mounted road information collection device and method for collecting road information", an intelligent vehicle-mounted road information collection device is disclosed, which comprises a vehicle body, the vehicle body is provided with a 360 ° panoramic camera for shooting images around the vehicle body, a laser scanning device for completing three-dimensional information data collection of road and road surroundings during vehicle driving, a thermal imager for scanning infrared thermal images around the vehicle, an auxiliary lighting device for supplementing light during collection operation of tunnel segments, an area array image sensor, a GPS positioning device, an inertial measurement device and a radar speed measurement device for accurately measuring the vehicle driving distance, which are connected with a display and a server installed inside the vehicle body as a collection information sub-device, the vehicle is internally provided with a power supply for supplying power, and the server is internally provided with a synchronous control subsystem, a time service subsystem and a data processing and analyzing subsystem, so that road disease information and other road auxiliary information can be rapidly acquired.
However, similarly, under the condition of great development of the existing imaging technology, the processing flow of the panoramic imaging detection mode is complex, the data volume is also large, and the macroscopic determination of the specific position of the problem point in the tunnel is not facilitated during the display.
Disclosure of Invention
The invention aims to provide a tunnel imaging plane display system and a display method for sampling a tunnel at a full angle, converting the tunnel into a plane and displaying the plane so as to accurately express the position of a defect point on the tunnel, aiming at the defects of the tunnel imaging technology in the prior art.
The invention discloses a tunnel imaging plane display method which is characterized by comprising a data acquisition and processing step and an image plane display step;
the data acquisition and processing step is to acquire images and three-dimensional contour data of the tunnel and preprocess the images of the tunnel, the preprocessing of the images is to detect character images and send the character images to an identification module for identification, in the image analysis, the preprocessing of the images is to perform processing before feature extraction, segmentation and matching on input images, and the preprocessing of the images mainly aims to eliminate irrelevant information in the images, recover useful real information, enhance the detectability of relevant information and simplify the data to the maximum extent, thereby improving the reliability of the feature extraction, the image segmentation, the matching and the identification; the method comprises the steps of identifying defects contained in images of tunnels by a model learning method, wherein the model learning method is an important method in the image identification technology and is a technology for establishing a model base for comparison judgment based on processing, summarizing and comparison of big data.
For example, a paper document entitled "study of image classification method based on deep learning model" published in 2018 introduces a scheme of modeling learning in the field of image recognition, wherein the problem of image classification is referred to, and the key point is how to extract more abstract feature information from an image, and the quality of feature extraction is important for a classification result. Deep learning is a multi-layer neural network learning algorithm appearing in recent years, and is widely applied to the fields of image classification, voice recognition, natural language processing and the like due to excellent feature learning capability. Firstly, on the basis of specifically analyzing the image classification and the development status of the Deep learning theory, two classic Deep learning models, namely a Deep Belief Network (DBN) and a Convolutional Neural Network (CNN), are deeply researched. Secondly, aiming at the problems of low network convergence speed and low classification precision caused by fixed learning parameters in the traditional deep belief network training process, an improved self-adaptive LBP-DBN image classification algorithm is provided. The algorithm firstly introduces a self-adaptive updating criterion into the training of a limiting Boltzmann machine; then, self-adaptive adjustment of momentum and learning rate is realized by calculating reconstruction error increment and weight updating direction before and after iteration; and finally, the verification is respectively carried out through an ORL database and a Yale database, and experimental results show that compared with the traditional LBP-DBN algorithm, the algorithm provided by the text can effectively improve the classification precision of the network and improve the convergence speed. Finally, aiming at the problem of poor network classification precision caused by random initialization weight and noise interference in the feature extraction process during the traditional convolutional neural network training, a convolutional neural network image classification algorithm based on weight integration optimization is provided. The algorithm firstly extracts hidden layer feature mapping of a sample as prior information, and an optimal solution is found by utilizing simulated annealing to serve as a full-connection layer initialization weight, so that weight updating and network convergence are accelerated; secondly, introducing a Gaussian function into the convolution layer, smoothing the image through Gaussian convolution operation, and reducing the interference of noise on feature extraction; and finally, the network after the integration optimization is applied to an MNIST database and a CIFAR-10 database, and experimental results show that the convergence rate and the classification accuracy of the network after the integration optimization are improved.
For data acquisition, due to the influence of an acquisition environment (factors such as time, illumination, a coating structure of the inner surface of a tunnel, material reflectivity and the like), optimal sampling adjustment changes, and the optimal sampling effect needs to be kept by adjusting light supplement, exposure time, wall surface data and the like in real time, so that a sampling adjustment database can be arranged, the environmental data influencing the acquisition effect are contained, an adjustment instruction of sampling equipment is correspondingly called, and autonomous sampling optimization is realized.
The image plane display step in the technical scheme of the invention comprises the following steps:
processing image data, namely reading and receiving data acquired and processed in the data acquisition and processing step, and fusing original tunnel image data and three-dimensional contour data of the tunnel by using an image splicing and texture mapping method to form a three-dimensional image display of the tunnel;
image display processing, namely identifying defects contained in the image of the tunnel by a model learning method, and displaying the defects on the three-dimensional image of the tunnel in an identification mode; and expanding the tunnel three-dimensional image with the identified defects into a plane according to a flattening graph mode for displaying, and generating a tunnel health condition report containing time variation trend, defect distribution and defect type data on the tunnel image expanded into the plane through statistical analysis of current and historical data.
And the flattened image of the three-dimensional image of the tunnel can be transversely dragged on a display medium (such as a computer screen) through a peripheral operating device (such as a mouse), and the dragging along the length direction of the tunnel can be changed. The image is checked along the length direction, if the image is dragged to the corresponding length of the tunnel, a defect mark exists if the defect mark exists, a mouse is suspended on the mark, detailed information of the corresponding defect can be displayed, further, if the defect mark exists, the image is locally amplified if a left key is clicked, operation and searching are facilitated, two-dimensional display is facilitated to view, meanwhile, the scaled image is favorable for quickly determining the position, in addition, auxiliary operation is added, a data supplement and adjustment input end can be provided for an operator, the identification result can be adjusted to a certain degree through the peripheral operation devices, and for example, the deviation of the identification position, the identification omission, the adjustment of the defect area and the like are corrected, so that the identification accuracy is ensured.
Preferably, in the data acquisition and processing step, the image acquisition of the tunnel is realized by a linear array camera array with an LED linear light source array, and further, in order to improve the service time, the linear array camera array is further provided with an air-cooled heat dissipation structure; the three-dimensional profile data acquisition of the tunnel is realized by an area array camera array with a laser array.
The three-dimensional profile data of the tunnel comprises in-tunnel profile data and profile data of tunnel accessory equipment; the tunnel inner contour data comprises contour data of a tunnel lining inner contour, a tunnel building limit, a basic building limit and a locomotive vehicle limit; the tunnel accessory equipment comprises communication equipment, ventilation equipment and the like.
And the original tunnel image data are photo images of all parts of the tunnel acquired by the linear array camera array.
The linear array camera array and the area array camera array are triggered at least once every 0.5mm, and the triggering frequency needs to reach 34 KHz; a 6-dimensional state measurement sensor is further arranged to acquire correction data of the linear array camera array and the area array camera array during data acquisition; the 6-dimensional state measurement sensor is used for detecting vibration data of the vehicle in 6 dimensions of yaw, floating, stretching, rolling, nodding and shaking, the data is used for compensating image changes caused by 6-dimensional vibration of the vehicle body, and a dynamic envelope curve of the vehicle in the process of running in a tunnel can be given.
And in the image data processing of the image plane display step, the three-dimensional image of the tunnel is cut along the length direction of the tunnel according to the structures of the left wall, the vault and the right wall of the tunnel. And the left wall, the vault, the right wall and the three-dimensional module in the length direction are cut, so that the retrieval with any length can be realized, the retrieval is carried out according to the length of the tunnel and any starting point and end point, and the model is edited and displayed.
In the image plane display processing in the image plane display step, after the defects included in the image of the tunnel identified by the model learning method are displayed on the three-dimensional image of the tunnel in the form of the identification, further confirmation needs to be performed on the defect marks by manual work, and the positions of position errors, defect mark errors and mark omission are corrected and/or supplemented, as described above, the operation can be performed by the peripheral operation devices (such as a mouse and a keyboard lamp).
Preferably, the planar tunnel image is displayed according to a tunnel defect flattening diagram mode, the horizontal direction of the flattening diagram is tunnel length information, the tunnel length information is marked by hectometer, the vertical direction of the flattening diagram is formed by a tunnel structure, and the tunnel length information sequentially comprises a left side wall, a vault and a right side wall.
The invention also discloses a tunnel imaging plane display system, which is characterized in that: the system comprises a data acquisition system, a data processing system and a data display system;
the data acquisition system is used for acquiring images in the tunnel and three-dimensional contour data of the images;
the data processing system is used for processing the data acquired by the data acquisition system, counting and analyzing historical data, identifying defect points in the tunnel image, marking the defect points, and fusing the image in the tunnel and the three-dimensional contour data thereof to form a three-dimensional display image of the tunnel;
the data display system is used for displaying a three-dimensional display image of the data processing system in a mode of installing a flat drawing, and further comprises displaying a statistical and analysis structure of historical data.
The data acquisition system comprises a linear array camera array with an LED line light source array for acquiring tunnel images and an area array camera array with a laser array for acquiring three-dimensional profile data of tunnels.
Furthermore, in order to prolong the service time, the linear array camera array is also provided with an air-cooled heat dissipation structure so as to prolong the service life and improve the stability.
Drawings
The foregoing and following detailed description of the invention will be apparent when read in conjunction with the following drawings, in which:
FIG. 1 is a logic diagram of a tunnel imaging plane display method according to the present invention;
FIG. 2 is a schematic illustration of a tunnel imaging plane display of the present invention;
FIG. 3 is a schematic illustration of the inner profile of a tunnel of the present invention;
FIG. 4 is a schematic diagram of the 6 dimensions detected by the 6-dimensional condition measuring sensor of the present invention.
Detailed Description
The technical solutions for achieving the objects of the present invention are further illustrated by the following specific examples, and it should be noted that the technical solutions claimed in the present invention include, but are not limited to, the following examples.
Example 1
As a most basic embodiment of the present invention, as shown in fig. 1, this example discloses a tunnel imaging plane display method, which includes a data acquisition and processing step and an image plane display step;
the data acquisition and processing step is to acquire images and three-dimensional contour data of the tunnel and preprocess the images of the tunnel, the preprocessing of the images is to detect character images and send the character images to an identification module for identification, in the image analysis, the preprocessing of the images is to perform processing before feature extraction, segmentation and matching on input images, and the preprocessing of the images mainly aims to eliminate irrelevant information in the images, recover useful real information, enhance the detectability of relevant information and simplify the data to the maximum extent, thereby improving the reliability of the feature extraction, the image segmentation, the matching and the identification; the method comprises the steps of identifying defects contained in images of tunnels by a model learning method, wherein the model learning method is an important method in the image identification technology and is a technology for establishing a model base for comparison judgment based on processing, summarizing and comparison of big data. (in the prior art, as a paper document entitled "study of image classification method based on Deep learning model" published in 2018, a modeling learning scheme in the field of image recognition is introduced, wherein it is mentioned that for the problem of image classification, the key point is how to extract more abstract feature information from an image, and the quality of feature extraction is important for the classification result. Deep learning is a multi-layer Neural Network learning algorithm appearing in recent years, and is widely applied to the fields of image classification, voice recognition, natural language processing and the like because of its excellent feature learning capability. firstly, two classical Deep learning models, namely Deep Belief Network (DBN) and Convolutional Neural Network (CNN), are deeply studied in the text on the basis of the specific analysis of the current situation of image classification and Deep learning theory, aiming at the problems of low network convergence speed and low classification precision caused by fixed learning parameters in the traditional deep belief network training process, an improved self-adaptive LBP-DBN image classification algorithm is provided. The algorithm firstly introduces a self-adaptive updating criterion into the training of a limiting Boltzmann machine; then, self-adaptive adjustment of momentum and learning rate is realized by calculating reconstruction error increment and weight updating direction before and after iteration; and finally, the verification is respectively carried out through an ORL database and a Yale database, and experimental results show that compared with the traditional LBP-DBN algorithm, the algorithm provided by the text can effectively improve the classification precision of the network and improve the convergence speed. Finally, aiming at the problem of poor network classification precision caused by random initialization weight and noise interference in the feature extraction process during the traditional convolutional neural network training, a convolutional neural network image classification algorithm based on weight integration optimization is provided. The algorithm firstly extracts hidden layer feature mapping of a sample as prior information, and an optimal solution is found by utilizing simulated annealing to serve as a full-connection layer initialization weight, so that weight updating and network convergence are accelerated; secondly, introducing a Gaussian function into the convolution layer, smoothing the image through Gaussian convolution operation, and reducing the interference of noise on feature extraction; and finally, the network after the integration optimization is applied to an MNIST database and a CIFAR-10 database, and experimental results show that the convergence rate and the classification accuracy of the network after the integration optimization are improved.
The image plane display step in the technical scheme of the invention comprises the following steps:
processing image data, namely reading and receiving data acquired and processed in the data acquisition and processing step, and fusing original tunnel image data and three-dimensional contour data of the tunnel by using an image splicing and texture mapping method to form a three-dimensional image display of the tunnel;
image display processing, namely identifying defects contained in the image of the tunnel by a model learning method, and displaying the defects on the three-dimensional image of the tunnel in an identification mode; and expanding the tunnel three-dimensional image with the identified defects into a plane according to a flattening graph mode for displaying, and generating a tunnel health condition report containing time variation trend, defect distribution and defect type data on the tunnel image expanded into the plane through statistical analysis of current and historical data.
Further, on the basis of the most basic embodiment, preferably, as shown in fig. 2, the flattened image of the three-dimensional image of the tunnel can be dragged transversely on a display medium (such as a computer screen) through a peripheral operating device (such as a mouse), and the dragging can be changed along the length direction of the tunnel. The image is checked along the length direction, if the image is dragged to the corresponding length of the tunnel, a defect mark exists if the defect mark exists, a mouse is suspended on the mark, detailed information of the corresponding defect can be displayed, further, if the defect mark exists, the image is locally amplified if a left key is clicked, operation and searching are facilitated, two-dimensional display is facilitated to view, meanwhile, the scaled image is favorable for quickly determining the position, in addition, auxiliary operation is added, a data supplement and adjustment input end can be provided for an operator, the identification result can be adjusted to a certain degree through the peripheral operation devices, and for example, the deviation of the identification position, the identification omission, the adjustment of the defect area and the like are corrected, so that the identification accuracy is ensured. Furthermore, in the data acquisition and processing step, the image acquisition of the tunnel is realized by a linear array camera array with an LED line light source array, and further, in order to improve the service time, the linear array camera array is also provided with an air-cooled heat dissipation structure; the three-dimensional profile data acquisition of the tunnel is realized by an area array camera array with a laser array.
As shown in fig. 3, the three-dimensional profile data of the tunnel includes in-tunnel profile data and profile data of the tunnel accessory device; the tunnel inner contour data comprises contour data of a tunnel lining inner contour, a tunnel building limit, a basic building limit and a locomotive vehicle limit; the tunnel accessory equipment comprises communication equipment, ventilation equipment and the like. And the original tunnel image data are photo images of all parts of the tunnel acquired by the linear array camera array. The linear array camera array and the area array camera array are triggered at least once every 0.5mm, and the triggering frequency needs to reach 34 KHz; a 6-dimensional state measurement sensor is further arranged to acquire correction data of the linear array camera array and the area array camera array during data acquisition; the 6-dimensional state measurement sensor is used for detecting vibration data of the vehicle in 6 dimensions of yaw, floating, stretching, rolling, nodding and shaking, the data is used for compensating image changes caused by 6-dimensional vibration of the vehicle body, and a dynamic envelope curve of the vehicle in the process of running in a tunnel can be given. And in the image data processing of the image plane display step, the three-dimensional image of the tunnel is cut along the length direction of the tunnel according to the structures of the left wall, the vault and the right wall of the tunnel. And the left side wall, the vault, the right side wall and the three-dimensional module in the length direction are cut, so that the retrieval with any length can be realized, the retrieval is carried out according to the length of the tunnel and any starting point and end point, and the edition and display of the model are carried out.
In the image plane display processing in the image plane display step, after the defects included in the image of the tunnel identified by the model learning method are displayed on the three-dimensional image of the tunnel in the form of the identification, further confirmation needs to be performed on the defect marks by manual work, and the positions of position errors, defect mark errors and mark omission are corrected and/or supplemented, as described above, the operation can be performed by the peripheral operation devices (such as a mouse and a keyboard lamp). Preferably, the planar tunnel image is displayed according to a tunnel defect flattening diagram mode, the horizontal direction of the flattening diagram is tunnel length information, the tunnel length information is marked by hectometer, the vertical direction of the flattening diagram is formed by a tunnel structure, and the tunnel length information sequentially comprises a left side wall, a vault and a right side wall.
Example 2
As a preferred embodiment of the present invention, as shown in fig. 1, this example discloses a tunnel imaging plane display method, which includes a data acquisition and processing step and an image plane display step;
the data acquisition and processing step is to acquire images and three-dimensional profile data of the tunnel, and perform compression and model learning processing;
the image plane display step comprises the following steps:
reading and receiving data acquired and processed in the data acquisition and processing step;
secondly, fusing the three-dimensional contour data of the tunnel with the original tunnel image data (image splicing- > texture mapping) to form three-dimensional display;
firstly, cutting of a 3-dimensional module (cutting of a left side wall, a vault, a right side wall and the length direction) is realized;
and (4) arbitrarily rotating the whole tunnel model.
And thirdly, realizing the retrieval of any length, retrieving according to the length of the tunnel and any starting point and end point, and editing and displaying the model.
Thirdly, attaching a defect result obtained after the data of the vehicle-mounted end is processed to the three-dimensional model in a form of identification;
when the model is browsed, the link of defect confirmation can be entered when the mark is clicked.
② the right key 'mark' can prompt the state of defect (such as suspicious, confirmed and invalid), kilometer sign or length, position (left wall, right wall or vault) and the like.
And the right key is the mark and has information display of maintenance states, such as maintenance units or people, maintenance time, maintenance completion, ongoing maintenance, non-maintenance and the like.
And fourthly, browsing the model, browsing pictures according to the tunnel composition or browsing pictures according to the camera, and identifying the defects so as to facilitate the user to enter the next defect confirmation.
Fourthly, the intelligent recognition is well known to have false recognition, especially under the condition that a defect library is not completely established, and therefore, manual further confirmation is needed;
entering a defect confirmation interface from the third step, wherein the picture in the view has the most basic picture editing function (such as dragging, moving, amplifying, zooming and the like);
secondly, the system has a defect list function, a function of screening according to defect states and shortcut keys for confirming the defect states;
in the process of reversely positioning the image, when a defect list is browsed, the situation that several defects of the same image cannot be distinguished is shown, so that the right key of the identification frame needs to be clicked to automatically position the image to the specific defect in the defect list;
and fourthly, manually marking the defect marks except the defect library, such as newly added marks of the defect library, such as marks of cracks (such as crack trend, length and the like), and also manually increasing defect description.
Fifthly, the health condition of the tunnel is described through data statistics and analysis;
statistical analysis of single detection data is used for analyzing the single detection data.
And secondly, data statistics on a time axis (year, month and week) for multiple times can be used for analyzing the change trend of the tunnel on the time axis.
Thirdly, counting according to the defect types, and analyzing the defect proportion for maintenance.
Fourthly, according to the defect distribution statistics, the defect distribution probability of the tunnel structure can be analyzed and obtained, and the method can be used for predicting the defect development trend.
And fifthly, counting according to the defect distribution of the tunnel composition, and being used for guiding a tunnel maintenance strategy.
Sixthly, analyzing the relevance of the defect, and analyzing the root cause of the defect.
And comparing historical picture data, comparing the defect state and the maintenance state of the defect, and having traceability.
And the eighthly, the evaluation function of tunnel health is provided.
Ninthly, carrying out data statistics and analysis according to regions, lines, construction technologies and the like.
And (8) performing statistics on distribution of an envelope line of the vehicle body dynamic property in the frequency band (R) for reflecting line or vehicle body problems, and considering the position relation between the vehicle body dynamic operation and the tunnel by combining a limit.
Sixthly, the data result of the tunnel detection is visually displayed to a user;
the report has the function of commander selection according to conditions (according to defect types, tunnel length, tunnel structure composition and tunnel equipment);
secondly, displaying the defect report in a tunnel defect flattening diagram mode (tunnel length information is in the horizontal direction and is marked by hectometer, and a tunnel structure is in the vertical direction and sequentially comprises a left side wall, a vault and a right side wall);
and thirdly, report printing, wherein the format is horizontal A3 printing, and the report printing needs to have line names, initial positions and printing time.
Maintaining a report form, wherein the name of the tunnel or the name of the line running, the position of the disease, the picture of the disease (before and after maintenance) and the like need to be reflected.
And fifthly, a limit defect report gives information such as defect position, defect grade, defect distribution and the like.
And sixthly, the dynamic envelope takes the tunnel length as information, and the distribution form of the dynamic envelope can be displayed in a graph form.
In addition, based on the method, the invention also discloses a tunnel imaging plane display system, which comprises a data acquisition system, a data processing system and a data display system;
the data acquisition system is used for acquiring images in the tunnel and three-dimensional contour data of the images;
the data processing system is used for processing the data acquired by the data acquisition system, counting and analyzing historical data, identifying defect points in the tunnel image, marking the defect points, and fusing the image in the tunnel and the three-dimensional contour data thereof to form a three-dimensional display image of the tunnel;
the data display system is used for displaying a three-dimensional display image of the data processing system in a mode of installing a flat drawing, and further comprises displaying a statistical and analysis structure of historical data.
The data acquisition system comprises a linear array camera array with an LED line light source array for acquiring tunnel images and an area array camera array with a laser array for acquiring three-dimensional profile data of tunnels.
Furthermore, in order to prolong the service time, the linear array camera array is also provided with an air-cooled heat dissipation structure so as to prolong the service life and improve the stability.
Claims (10)
1. A tunnel imaging plane display method is characterized by comprising a data acquisition and processing step and an image plane display step;
the data acquisition and processing step is to acquire tunnel image data and three-dimensional contour data, preprocess the tunnel image data and identify defects contained in the tunnel image by a model learning method;
the image plane display step comprises the following steps:
processing image data, namely reading and receiving data acquired and processed in the data acquisition and processing step, and fusing original tunnel image data and three-dimensional contour data of the tunnel by using an image splicing and texture mapping method to form a three-dimensional image display of the tunnel;
image display processing, namely identifying defects contained in the image of the tunnel by a model learning method, and displaying the defects on the three-dimensional image of the tunnel in an identification mode; and expanding the tunnel three-dimensional image with the identified defects into a plane according to a flattening graph mode for displaying, and generating a tunnel health condition report containing time variation trend, defect distribution and defect type data on the tunnel image expanded into the plane through statistical analysis of current and historical data.
2. The tunnel imaging plane exhibition method of claim 1, wherein: in the data acquisition and processing step, the image acquisition of the tunnel is realized by a linear array camera array with an LED line light source array, and further, in order to improve the service time, the linear array camera array is also provided with an air-cooled heat dissipation structure; the three-dimensional profile data acquisition of the tunnel is realized by an area array camera array with a laser array.
3. A tunnel imaging flat panel display method as claimed in claim 1 or 2, wherein: the three-dimensional profile data comprises in-tunnel profile data and profile data of tunnel accessory equipment;
the tunnel inner contour data comprises contour data of a tunnel lining inner contour, a tunnel building limit, a basic building limit and a locomotive vehicle limit;
the tunnel accessory equipment comprises communication equipment, ventilation equipment and the like.
4. A tunnel imaging plane exhibition method as claimed in claim 2, characterized in that: and the original tunnel image data are photo images of all parts of the tunnel acquired by the linear array camera array.
5. A tunnel imaging flat panel display method as claimed in claim 2 or 4, characterized in that: the linear array camera array and the area array camera array are triggered at least once every 0.5mm, and the triggering frequency needs to reach 34 KHz; a 6-dimensional state measurement sensor is further arranged to acquire correction data of the linear array camera array and the area array camera array during data acquisition; the 6-dimensional state measurement sensor is used for detecting vibration data of the vehicle in 6 dimensions of yaw, floating, stretching, rolling, nodding and shaking.
6. The tunnel imaging plane exhibition method of claim 1, wherein: and in the image data processing of the image plane display step, the three-dimensional image of the tunnel is cut along the length direction of the tunnel according to the structures of the left wall, the vault and the right wall of the tunnel.
7. The tunnel imaging plane exhibition method of claim 1, wherein: in the image plane display processing in the image plane display step, after the defects included in the image of the tunnel identified by the model learning method are displayed on the three-dimensional image of the tunnel in the form of identification, further confirmation of the defect marks is needed through manual work, and the positions of position errors, defect mark errors and mark omission are corrected and/or supplemented.
8. A tunnel imaging flat panel display method as claimed in claim 1 or 7, wherein: the tunnel three-dimensional image with the identified defects is displayed according to the mode of a tunnel defect flattening map, the horizontal direction of the flattening map is tunnel length information, the vertical direction of the flattening map is formed by a tunnel structure, and the tunnel three-dimensional image sequentially comprises a left side wall, a vault and a right side wall.
9. A tunnel imaging plane display system is characterized in that: the system comprises a data acquisition system, a data processing system and a data display system;
the data acquisition system is used for acquiring images in the tunnel and three-dimensional contour data of the images;
the data processing system is used for processing the data acquired by the data acquisition system, counting and analyzing historical data, identifying defect points in the tunnel image, marking the defect points, and fusing the image in the tunnel and the three-dimensional contour data thereof to form a three-dimensional display image of the tunnel;
the data display system is used for displaying a three-dimensional display image of the data processing system in a mode of installing a flat drawing, and further comprises displaying a statistical and analysis structure of historical data.
10. A tunnel imaging flat panel display system as claimed in claim 9, wherein: the data acquisition system comprises a linear array camera array with an LED line light source array for acquiring tunnel images and an area array camera array with a laser array for acquiring three-dimensional profile data of tunnels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910812607.2A CN112446852B (en) | 2019-08-30 | 2019-08-30 | Tunnel imaging plane display method and defect intelligent recognition system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910812607.2A CN112446852B (en) | 2019-08-30 | 2019-08-30 | Tunnel imaging plane display method and defect intelligent recognition system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112446852A true CN112446852A (en) | 2021-03-05 |
| CN112446852B CN112446852B (en) | 2023-12-15 |
Family
ID=74741366
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910812607.2A Active CN112446852B (en) | 2019-08-30 | 2019-08-30 | Tunnel imaging plane display method and defect intelligent recognition system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112446852B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN113252004A (en) * | 2021-06-09 | 2021-08-13 | 成都国铁电气设备有限公司 | Tunnel comprehensive detection monitoring system and method |
| CN113256714A (en) * | 2021-07-13 | 2021-08-13 | 湖南大学 | Tunnel surface image processing method and system |
| CN113447536A (en) * | 2021-06-24 | 2021-09-28 | 山东大学 | Concrete dielectric constant inversion and disease identification method and system |
| CN114943706A (en) * | 2022-05-27 | 2022-08-26 | 宁波艾腾湃智能科技有限公司 | Anti-counterfeiting authentication of plane works or products in absolute two-dimensional space state |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19513116A1 (en) * | 1995-04-07 | 1996-10-10 | Misoph Rotraud | Contactless measurement of tunnel profile or road surface e.g. motorway |
| US20060221072A1 (en) * | 2005-02-11 | 2006-10-05 | Se Shuen Y S | 3D imaging system |
| CN104048969A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | Tunnel defect recognition method |
| CN104567708A (en) * | 2015-01-06 | 2015-04-29 | 浙江工业大学 | Tunnel full-section high-speed dynamic health detection device and method based on active panoramic vision |
| CN105004732A (en) * | 2015-07-16 | 2015-10-28 | 武汉长盛工程检测技术开发有限公司 | Tunnel crack rapid detection device and method |
| CN106841216A (en) * | 2017-02-28 | 2017-06-13 | 浙江工业大学 | Tunnel defect automatic identification equipment based on panoramic picture CNN |
| CN107014352A (en) * | 2017-05-02 | 2017-08-04 | 中铁第四勘察设计院集团有限公司 | A kind of railcar and detection method for railway tunnel tunneling boring comprehensive detection |
| CN108827182A (en) * | 2018-03-30 | 2018-11-16 | 曾思远 | A kind of tunnel three-D imaging method and system |
| CN109166168A (en) * | 2018-09-04 | 2019-01-08 | 上海同岩土木工程科技股份有限公司 | A kind of three-dimensional distribution figure fast construction method of tunnel lining structure |
| WO2019033862A1 (en) * | 2017-08-14 | 2019-02-21 | 成都中信华瑞科技有限公司 | Data collection method and tunnel detection vehicle |
-
2019
- 2019-08-30 CN CN201910812607.2A patent/CN112446852B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19513116A1 (en) * | 1995-04-07 | 1996-10-10 | Misoph Rotraud | Contactless measurement of tunnel profile or road surface e.g. motorway |
| US20060221072A1 (en) * | 2005-02-11 | 2006-10-05 | Se Shuen Y S | 3D imaging system |
| CN104048969A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | Tunnel defect recognition method |
| CN104567708A (en) * | 2015-01-06 | 2015-04-29 | 浙江工业大学 | Tunnel full-section high-speed dynamic health detection device and method based on active panoramic vision |
| CN105004732A (en) * | 2015-07-16 | 2015-10-28 | 武汉长盛工程检测技术开发有限公司 | Tunnel crack rapid detection device and method |
| CN106841216A (en) * | 2017-02-28 | 2017-06-13 | 浙江工业大学 | Tunnel defect automatic identification equipment based on panoramic picture CNN |
| CN107014352A (en) * | 2017-05-02 | 2017-08-04 | 中铁第四勘察设计院集团有限公司 | A kind of railcar and detection method for railway tunnel tunneling boring comprehensive detection |
| WO2019033862A1 (en) * | 2017-08-14 | 2019-02-21 | 成都中信华瑞科技有限公司 | Data collection method and tunnel detection vehicle |
| CN108827182A (en) * | 2018-03-30 | 2018-11-16 | 曾思远 | A kind of tunnel three-D imaging method and system |
| CN109166168A (en) * | 2018-09-04 | 2019-01-08 | 上海同岩土木工程科技股份有限公司 | A kind of three-dimensional distribution figure fast construction method of tunnel lining structure |
Non-Patent Citations (7)
| Title |
|---|
| LEANNE ATTARD 等: "Tunnel inspection using photogrammetric techniques and image processing: A review", ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, pages 180 - 188 * |
| V. TABATADZE 等: "Tunnel detection and visualization using the method of auxiliary sources", pages 1 - 5, Retrieved from the Internet <URL:https://ieeexplore.ieee.org/abstract/document/8100549/> * |
| 万和平 等: "数字化隧道三维可视化技术研究", 高速铁路技术, pages 100 - 104 * |
| 李明博 等: "BIM技术在运营隧道病害检测结果三维可视化中的应用", 中外公路, pages 297 - 301 * |
| 汤一平 等: "主动式全景视觉的隧道全断面变形检测方法", 光电工程, pages 442 - 452 * |
| 汤一平 等: "基于全景图像CNN的隧道病害自动识别方法", 计算机科学, no. 2, pages 207 - 212 * |
| 雷坚强 等: "车载式公路隧道快速检测设备应用探讨", 隧道建设(中英文), no. 2, pages 47 - 53 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN113252004A (en) * | 2021-06-09 | 2021-08-13 | 成都国铁电气设备有限公司 | Tunnel comprehensive detection monitoring system and method |
| CN113252004B (en) * | 2021-06-09 | 2021-10-08 | 成都国铁电气设备有限公司 | Tunnel comprehensive detection monitoring system and method |
| CN113447536A (en) * | 2021-06-24 | 2021-09-28 | 山东大学 | Concrete dielectric constant inversion and disease identification method and system |
| CN113256714A (en) * | 2021-07-13 | 2021-08-13 | 湖南大学 | Tunnel surface image processing method and system |
| CN113256714B (en) * | 2021-07-13 | 2021-09-24 | 湖南大学 | Tunnel surface image processing method and system |
| CN114943706A (en) * | 2022-05-27 | 2022-08-26 | 宁波艾腾湃智能科技有限公司 | Anti-counterfeiting authentication of plane works or products in absolute two-dimensional space state |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112446852B (en) | 2023-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210319561A1 (en) | Image segmentation method and system for pavement disease based on deep learning | |
| CN112446852B (en) | Tunnel imaging plane display method and defect intelligent recognition system | |
| CN111855664B (en) | Adjustable three-dimensional tunnel defect detection system | |
| CN109635666B (en) | Image target rapid detection method based on deep learning | |
| Wieland et al. | Estimating building inventory for rapid seismic vulnerability assessment: Towards an integrated approach based on multi-source imaging | |
| CN111091538B (en) | Automatic identification and defect detection method and device for pipeline welding seams | |
| AU2014295972B2 (en) | System and method for detecting features in aerial images using disparity mapping and segmentation techniques | |
| CN106780284B (en) | Tunnel wall surrounding rock integrity information acquisition device and evaluation method | |
| JP4717760B2 (en) | Object recognition device and video object positioning device | |
| CN112149512A (en) | Helmet wearing identification method based on two-stage deep learning | |
| CN111339905B (en) | CIM well lid state visual detection system based on deep learning and multiple visual angles | |
| CN110837839B (en) | High-precision unmanned aerial vehicle orthographic image manufacturing and data acquisition method | |
| CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
| CN115409069A (en) | Village and town building identification method, classification method, device, electronic equipment and medium | |
| CN112581402B (en) | Road and bridge fault automatic detection method based on machine vision technology | |
| CN113139437A (en) | Helmet wearing inspection method based on YOLOv3 algorithm | |
| Lang et al. | Pavement cracking detection and classification based on 3d image using multiscale clustering model | |
| WO2017082774A1 (en) | Method and system for identifying urban objects | |
| Sezen et al. | Deep learning-based door and window detection from building façade | |
| CN117011759A (en) | Method and system for analyzing multi-element geological information of surrounding rock of tunnel face by drilling and blasting method | |
| CN111709991B (en) | Railway tool detection method, system, device and storage medium | |
| CN112541403B (en) | Indoor personnel falling detection method by utilizing infrared camera | |
| Díaz-Vilariño et al. | From LiDAR data towards HBIM for energy analysis | |
| CN112052824A (en) | Gas pipeline specific object target detection alarm method, device and system based on YOLOv3 algorithm and storage medium | |
| CN110427911A (en) | A kind of Approach for road detection, device, equipment and storage medium |
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 |