CN115511836A - Bridge crack grade evaluation method and system based on reinforcement learning algorithm - Google Patents
Bridge crack grade evaluation method and system based on reinforcement learning algorithm Download PDFInfo
- Publication number
- CN115511836A CN115511836A CN202211191901.4A CN202211191901A CN115511836A CN 115511836 A CN115511836 A CN 115511836A CN 202211191901 A CN202211191901 A CN 202211191901A CN 115511836 A CN115511836 A CN 115511836A
- Authority
- CN
- China
- Prior art keywords
- crack
- bridge
- image
- longitudinal
- grade
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 46
- 230000002787 reinforcement Effects 0.000 title claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 47
- 238000013210 evaluation model Methods 0.000 claims abstract description 45
- 238000007781 pre-processing Methods 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 230000011218 segmentation Effects 0.000 claims abstract description 8
- 238000000605 extraction Methods 0.000 claims description 29
- 230000000877 morphologic effect Effects 0.000 claims description 16
- 239000013598 vector Substances 0.000 claims description 12
- 239000011541 reaction mixture Substances 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 14
- 238000005259 measurement Methods 0.000 abstract description 4
- 230000036541 health Effects 0.000 abstract description 2
- 238000012544 monitoring process Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000012423 maintenance Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000000717 retained 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/17—Terrestrial scenes taken from planes or by drones
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/176—Urban or other man-made structures
-
- 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
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30132—Masonry; Concrete
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a bridge crack grade evaluation method and system based on a reinforcement learning algorithm, and belongs to the field of bridge structure health monitoring. The evaluation method comprises the following steps: acquiring a crack image, preprocessing the crack image, performing threshold segmentation, extracting width characteristics, length characteristics, trend characteristics, bit plane characteristics and longitudinal position information of the crack, constructing a crack basic database, and dividing the crack basic database into a training set and a test set; constructing a bridge crack grade evaluation model based on a reinforcement learning algorithm, and obtaining a stable crack grade evaluation model after training; the method comprises the steps of collecting a crack image to be evaluated, preprocessing the crack image, extracting length characteristics, width characteristics and trend characteristics of the crack, obtaining crack position and surface characteristics and longitudinal position information through space coordinates of the unmanned aircraft in cruising, inputting the crack position and surface characteristics and the longitudinal position information into a crack evaluation model, and obtaining a crack grade evaluation result. The method reduces errors caused by manual measurement, and improves the efficiency and accuracy of crack detection and evaluation.
Description
Technical Field
The invention belongs to the field of bridge structure health monitoring, and particularly relates to a bridge crack grade evaluation method and system based on a reinforcement learning algorithm.
Technical Field
The bridge evaluation work is not only an important ring of bridge management work but also one of the most basic works. In a normally operated bridge, the crack distribution and development trend of the main load-bearing structure are one of the most important indexes in the detection and evaluation process.
At present, the conventional manual mode is usually adopted for crack detection and evaluation, but the conventional manual detection and evaluation method is long in time consumption, large in workload, strong in subjectivity and unstable in evaluation result, and particularly when detection personnel face a huge number of bridge groups to be detected, corresponding problems are more prominent, so that the efficiency and the accuracy of detection and evaluation work are reduced.
Disclosure of Invention
In view of the above defects or shortcomings in the prior art, the invention aims to provide a bridge crack grade evaluation method and system based on a reinforcement learning algorithm, which is based on automatic acquisition of crack images, extracts the characteristic information of cracks through a crack identification program, digitalizes crack characteristics, reduces errors caused by manual measurement, further grades the cracks by using a crack grade evaluation model, effectively reduces the subjective influence of detection personnel, and improves the efficiency and accuracy of crack detection evaluation.
In order to achieve the above purpose, the embodiment of the invention adopts the following technical scheme:
in a first aspect, an embodiment of the present invention provides a bridge crack grade evaluation method based on a reinforcement learning algorithm, where the evaluation method includes:
collecting crack images through an unmanned aerial vehicle;
preprocessing the crack image;
performing crack image threshold segmentation on the preprocessed crack image, and extracting crack morphological characteristics; the fracture morphology characteristics comprise width characteristics, length characteristics and strike characteristics of the fracture;
acquiring a position plane of the crack image according to the flight track of the unmanned aerial vehicle, and extracting position plane characteristics;
acquiring a longitudinal coordinate of a crack central point on the axis of the bridge according to the flight track of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinate to be used as longitudinal position information of the crack;
judging the grade of the crack and generating a crack code according to the width characteristic, the length characteristic, the trend characteristic, the position face characteristic and the longitudinal position information of the crack, and constructing basic data of the crack, wherein the basic data of the crack comprises three levels of crack information which are respectively as follows: information of bridge level: bridge coding, structural systems; information of component hierarchy: structure type, member code, member longitudinal/transverse/vertical dimensions; fracture hierarchy information: crack coding, width features, length features, strike features, bit plane features, position and grade;
arranging the crack basic data corresponding to all crack images according to codes to construct a data set, and dividing the data set into a training set and a test set;
constructing a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the test set to obtain a stable crack grade evaluation model;
the method comprises the steps of collecting a crack image to be evaluated, preprocessing the crack image, extracting length characteristics, width characteristics, trend characteristics, position face characteristics and longitudinal position information of the crack, inputting the length characteristics, width characteristics, trend characteristics, position face characteristics and longitudinal position information into a model, and obtaining a crack grade evaluation result.
As a preferred embodiment of the invention, when acquiring the crack image, the method comprises the following steps:
s11, collecting a bridge image; when the bridge image is collected, all parts of the bridge are collected in an omnibearing manner;
s12, carrying out crack identification on the acquired bridge image, and marking the identified position meeting the crack condition;
and S13, carrying out local image acquisition again on the identified crack position to obtain a crack image.
As a preferred embodiment of the invention, when acquiring the crack image, the method comprises the following steps:
s11, collecting a bridge image; when the bridge image is collected, all parts of the bridge are collected in an omnibearing way;
s12, carrying out crack identification on the acquired bridge image, and marking the identified position meeting the crack condition;
and S14, segmenting the crack image at the crack position in the bridge image to obtain the crack image.
As a preferred embodiment of the present invention, the extracting fracture morphology features includes:
and S31, acquiring a starting point and an end point of the crack, calculating a central point of the crack image, dividing the crack image by using the central point, and acquiring a shortest distance edge point and a longest distance edge point which take the central point as a symmetrical center. In the step, when image segmentation is carried out, an SVM model is adopted in a core algorithm;
step S32, judging the crack trend according to the included angle between the connecting line of the starting point and the end point of the crack and the longitudinal axis of the bridge, and compiling crack trend codes as trend characteristics;
step S33, calculating width characteristics by taking the distance between the shortest distance edge points as the width of the crack;
and step S34, calculating length characteristics by taking the distance between the longest distance edge points as the crack length.
As a preferred embodiment of the present invention, the determining the fracture strike and compiling a fracture strike code as the strike characteristic in step S32 includes:
the cracks on the bottom surface or the top surface comprise a transverse crack, a longitudinal crack and an oblique crack, wherein the included angle between the transverse crack and the longitudinal axis is 75-90 degrees, the included angle between the oblique crack and the longitudinal axis is 15-75 degrees, and the included angle between the longitudinal crack and the longitudinal axis is 0-15 degrees;
the cracks on the side face comprise three types, namely vertical cracks, longitudinal cracks and oblique cracks, wherein the included angle between the vertical cracks and the longitudinal axis is 75-90 degrees, the included angle between the oblique cracks and the longitudinal axis is 15-75 degrees, and the included angle between the longitudinal cracks and the longitudinal axis is 0-15 degrees;
on the basis of judging the fracture strike, constructing a four-dimensional vector as strike coding, and coding and compiling fracture strike characteristics, wherein each strike corresponds to one vector; when the current trend is judged, the value is assigned to 1; otherwise the value is assigned to 0.
As a preferred embodiment of the present invention, the width feature is characterized by a relative width, as shown in formula (1):
in the formula (1), the reaction mixture is,
representing the relative width of the fracture; d represents the actual width of the crack; d Limit for Representing the width limit of the crack.
As a preferred embodiment of the present invention, the length feature is characterized by a relative length, as shown in formula (2):
in the formula (2), the reaction mixture is,
representing the relative length of the fracture; l represents the projection length of the crack in the projection direction, the crack length is projected in three directions of a three-dimensional space rectangular coordinate system, the longest crack length is taken as the projection length, and the coordinate axis where the longest crack is projected is taken as the projection direction; l represents the structural dimension of the component in the projection direction.
As a preferred embodiment of the invention, the bit plane of the crack comprises a bottom surface, a top surface and a side surface, and a three-dimensional vector is constructed to represent the bit plane characteristics; each dimension in the three-dimensional vector represents a bit plane, and the value of the bit plane at the current position is 1, otherwise, the value is 0.
As a preferred embodiment of the present invention, the bridge crack level evaluation model is constructed based on a reinforcement learning algorithm, an adopted algorithm network has four layers of networks, and includes an input layer, two hidden layers and an output layer, input information of the input layer includes crack position plane characteristics, crack length characteristics, crack width characteristics and crack trend characteristics in crack basic data, the hidden layer 1 has 10 unit nodes, the hidden layer 2 has 5 unit nodes, the output layer has 5 unit nodes, an activation function is a RELU function, and an output result is a crack level corresponding to a score Q of a corresponding crack evaluation result.
In a second aspect, an embodiment of the present invention further provides a bridge crack grade evaluation system based on a reinforcement learning algorithm, where the system includes: the system comprises a crack image acquisition module, an image preprocessing module, a morphological feature extraction module, a bit plane feature extraction module, a longitudinal position information extraction module, a training database, an evaluation model construction module and a result output module, wherein the crack image acquisition module, the image preprocessing module, the morphological feature extraction module, the bit plane feature extraction module, the longitudinal position information extraction module, the training database, the evaluation model construction module and the result output module are carried on an unmanned aerial vehicle; wherein:
the crack image acquisition module is used for acquiring a bridge crack image used for model training and a bridge crack image to be evaluated through an unmanned aerial vehicle;
the image preprocessing module is used for preprocessing the acquired crack image;
the morphological feature extraction module is used for performing crack image threshold segmentation on the preprocessed crack image and extracting crack morphological features; the fracture morphology characteristics comprise width characteristics, length characteristics and strike characteristics of the fracture;
the position surface feature extraction module is used for acquiring a position surface where the crack image is located according to the flight track of the unmanned aerial vehicle and extracting position surface features;
the longitudinal position information extraction module is used for acquiring longitudinal coordinates of the crack center point on the bridge axis according to the flight trajectory of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinates to be used as longitudinal position information of the crack;
the training database is used for judging the grade of the crack according to the width characteristic, the length characteristic, the trend characteristic, the position face characteristic and the longitudinal position information of the crack, generating a crack code, constructing basic data of the crack, arranging the basic data of the crack corresponding to all crack images according to the code to construct a data set, and dividing the data set into a training set and a testing set; the crack basic data comprises three levels of crack information, which are respectively as follows: information of bridge level: bridge coding, structural systems; information of component hierarchy: structure type, member code, member longitudinal/transverse/vertical dimensions; information of fracture hierarchy: crack coding, width features, length features, strike features, bit plane features, position and grade;
the evaluation model building module is used for building a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the testing set to obtain a stable crack grade evaluation model;
the result output module is used for inputting the bridge crack image to be evaluated acquired by the image acquisition module into a stable crack grade evaluation model in the evaluation model construction module and outputting the crack grade evaluation result of the bridge to be evaluated.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
according to the bridge crack grade evaluation method and system based on the reinforcement learning algorithm, provided by the embodiment of the invention, based on the automatic acquisition of crack images, the characteristic information of cracks is extracted through a crack identification program, the crack characteristics are digitalized, the errors caused by manual measurement are reduced, and further, the cracks are graded by using a crack grade evaluation model, so that the subjective influence of detection personnel is effectively reduced, and the efficiency and accuracy of crack detection evaluation are improved; meanwhile, the technical threshold of detection is reduced, so that ordinary bridge inspection maintenance personnel (technicians without profound professional literacy) can also accurately judge cracks existing in the bridge, so that a maintenance management department can make timely and reasonable maintenance decisions, and intelligent information management in the bridge management and maintenance process is realized.
Of course, it is not necessary for any product or method to achieve all of the above-described advantages at the same time for practicing the invention.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart of a bridge crack grade evaluation method based on a reinforcement learning algorithm according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating the division of the crack types on the bottom surface or the top surface in the embodiment of the present invention;
FIG. 3 is a schematic diagram of the division of the lateral fracture types in the embodiment of the present invention;
FIG. 4 is a schematic diagram of crack artwork collected in an embodiment of the present invention;
FIG. 5 is a graph showing the effect of the change of the gray level of the crack in the embodiment of the present invention;
FIG. 6 is a diagram illustrating the effect of filtering enhancement of a crack according to an embodiment of the present invention;
FIG. 7 is a schematic view of fracture splitting in an embodiment of the present invention;
FIG. 8 is a diagram of a fracture status data structure in an embodiment of the present invention;
FIG. 9 is a schematic diagram of a network structure of a fracture rating evaluation model according to an embodiment of the present invention;
FIG. 10 is a graph of the results of the crack assessment training in an embodiment of the present invention;
fig. 11 is a diagram illustrating an identification effect obtained by evaluating a crack rating in an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It is to be understood that the described embodiments are merely a few embodiments of the invention, and not all embodiments. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. It should be noted that the embodiments and features of the embodiments of the present invention may be combined with each other without conflict.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. In the description of the present invention, the terms "first," "second," "third," "fourth," etc. are used merely to distinguish one description from another, and are not intended to indicate or imply relative importance.
The embodiment of the invention provides a bridge crack grade evaluation method and system based on a reinforcement learning algorithm, which are characterized in that the characteristic information of cracks is extracted based on automatically acquired crack images, the characteristic information is digitalized, a crack grade evaluation model is constructed based on the reinforcement learning algorithm, the current crack grade is evaluated after training is finished, the crack grade evaluation efficiency is improved, the accuracy rate is up to 97%, and the accuracy of crack grade evaluation is effectively improved.
As shown in fig. 1, the bridge crack grade evaluation method based on reinforcement learning algorithm provided in the embodiment of the present invention includes the following steps:
and S1, collecting crack images through an unmanned aerial vehicle.
When crack images are acquired in the step, a preset time period or a selected bridge is taken as an object. And the unmanned plane carrying the camera device is used for image acquisition.
In one embodiment, the method comprises the following steps when image acquisition is carried out:
and S11, acquiring a bridge image. When the bridge image is collected, the girder of the bridge is collected in all directions.
And S12, carrying out crack identification on the acquired bridge image, and marking the identified position meeting the crack condition.
And S13, carrying out local image acquisition again on the identified crack position to obtain a crack image.
In another embodiment, when the crack image is acquired in this step, after the above steps S11 to S12 are completed, the following steps are performed:
and S14, segmenting the crack image at the crack position in the bridge image to obtain the crack image.
The crack image is obtained by means of re-collecting or cutting the original bridge image, and the mode is selected according to actual conditions.
And S2, preprocessing the crack image, wherein the preprocessing comprises graying, filtering enhancement and the like. By this step, the crack image in the crack image is enhanced.
And S3, performing crack image threshold segmentation on the preprocessed crack image, and extracting crack morphological characteristics. The fracture morphology features comprise a width feature, a length feature and a strike feature of the fracture. And the width characteristic, the length characteristic and the trend characteristic adopt dimensionless crack data.
The method specifically comprises the following steps:
and S31, acquiring a starting point and an end point of the crack, calculating a central point of the crack image, segmenting the crack image by using the central point, and acquiring a shortest distance edge point and a longest distance edge point which take the central point as a symmetrical center. In the step, an SVM model is adopted as a core algorithm when image segmentation is carried out.
And S32, judging the crack trend according to the included angle between the connecting line of the starting point and the end point of the crack and the longitudinal axis of the bridge, and compiling crack trend codes as trend characteristics.
In practical application, the fracture types generally include a transverse fracture, a longitudinal fracture, a vertical fracture and an oblique fracture, and correspond to four fracture trends. As shown in fig. 2, the cracks of the bottom surface (top surface) include three types, namely a transverse crack, a longitudinal crack and an oblique crack, wherein an included angle between the transverse crack and the longitudinal axis is between 75 degrees and 90 degrees, an included angle between the oblique crack and the longitudinal axis is between 15 degrees and 75 degrees, and an included angle between the longitudinal crack and the longitudinal axis is between 0 degree and 15 degrees; as shown in fig. 3, the lateral cracks include three types, namely vertical cracks, longitudinal cracks and oblique cracks, wherein an included angle between the vertical crack and the longitudinal axis is between 75 degrees and 90 degrees, an included angle between the oblique crack and the longitudinal axis is between 15 degrees and 75 degrees, and an included angle between the longitudinal crack and the longitudinal axis is between 0 degrees and 15 degrees.
On the basis of judging the fracture strike, constructing four-dimensional vectors as strike codes, and coding and compiling the fracture strike characteristics, wherein each strike corresponds to one vector; when the current trend is judged, the value is assigned to 1; otherwise the value is assigned to 0. Fracture strike codes are shown in table 1.
TABLE 1
| Type of fracture strike | Fracture strike characteristic |
| Transverse direction | (1,0,0,0) |
| Longitudinal direction | (0,1,0,0) |
| Vertical direction | (0,0,1,0) |
| In a diagonal direction | (0,0,0,1) |
And step S33, calculating width characteristics by taking the distance between the shortest distance edge points as the width of the crack.
In this step, the width characteristics are characterized by a relative width, as shown in formula (1):
in the formula (1), the reaction mixture is,
representing the relative width of the fracture; d represents the actual width of the crack (unit: mm); d Limit for Representing the width limit of the crack, and the value range is 0.05-0.2mm, preferably, in the embodiment, d Limit of Take 0.2mm.
And step S34, calculating length characteristics by taking the distance between the longest distance edge points as the crack length.
In this step, the length characteristics are characterized by using relative lengths, as shown in formula (2):
in the formula (2), the reaction mixture is,
representing the relative length of the fracture; l represents the projection length (unit: m) of the crack in the projection direction, wherein the transverse and oblique cracks positioned on the bottom surface are projected along the transverse direction of the bridge; projecting the vertical and oblique cracks on the side surfaces along the vertical direction of the bridge; longitudinal cracks on the bottom surface and the side surface grind the longitudinal projection of the bridge; l represents the structural dimension of the component in the projection direction (unit: m).
For four different fracture types, the corresponding relative length calculation formulas are respectively shown in formulas (3) to (6):
for transverse fractures:
in the formula (3), the reaction mixture is,
represents the relative length of the transverse slits; l Cross bar Represents the projection length (unit: m) of the transverse crack in the transverse bridge direction; l is Horizontal bar Represents the dimension of the member in the transverse direction (unit: m);
for longitudinal fractures:
in the formula (4), the reaction mixture is,
represents the relative length of the transverse slits; l Longitudinal direction Represents the projection length (unit: m) of the transverse crack in the transverse bridge direction; l is Longitudinal direction Represents the dimension (unit: m) of the member in the transverse direction;
for vertical fractures:
in the formula (5), the reaction mixture is,
represents the relative length of the vertical fracture; l Vertical shaft Represents the projection length (unit: m) of the vertical crack in the vertical direction; l is a radical of an alcohol Vertical Represents the dimension of the member in the vertical direction (unit: m);
for oblique fractures:
in the formula (6), the reaction mixture is,
represents the relative length of the diagonal crack; l Vertical Represents the projection length (unit: m) of the oblique crack in the vertical direction (or the transverse direction); l is a radical of an alcohol Vertical Representing the vertical (or lateral) dimension of the member (unit: m).
And S4, acquiring a position plane of the crack image according to the flight track of the unmanned aerial vehicle, and extracting position plane characteristics.
In this step, the bit plane includes three types, namely a bottom surface, a top surface and a side surface, and a three-dimensional vector is constructed to represent bit plane features. Each dimension in the three-dimensional vector represents a bit plane, and the value of the bit plane at the current position is 1, otherwise, the bit plane is 0.
The bit plane of the crack is encoded by adopting a single-hot encoding mode, and the specific encoding is shown in table 2.
TABLE 2
| Surface of crack | Bit plane characteristic |
| Bottom surface | (1,0,0) |
| Side surface | (0,1,0) |
| Top surface of the container | (0,0,1) |
In practical applications, cracks in the top surface are generally not detected. However, from the overall completion considerations, the top surface cracks still exist in individual cases in the structure, and therefore the bit plane characteristics of the top surface cracks are still retained in table 2.
And S5, acquiring a longitudinal coordinate of the central point of the crack on the axis of the bridge according to the flight track (the spatial position of the crack during image acquisition) of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinate to obtain longitudinal position information of the crack. The normalization process is shown as formula (7).
In the formula (7), the reaction mixture is,
represents the relative position of the crack center in the longitudinal direction; x represents the actual distance between the center of the crack and the starting end of the beam end, m; l represents the length of the member across which the fracture is located, m.
Step S6, judging the grade of the crack and generating a crack code according to the width characteristic, the length characteristic, the trend characteristic, the bit plane characteristic and the longitudinal position information of the crack, and constructing basic crack data, wherein the basic crack data comprise three levels of crack information, and the three levels of crack information are respectively as follows:
information of bridge level: bridge coding, structural systems;
information of component hierarchy: structure type, member coding, member longitudinal/transverse/vertical dimensions;
information of fracture hierarchy: crack coding, width features, length features, strike features, bit plane features, location, and grade.
And S7, arranging the crack basic data corresponding to all the crack images according to codes to construct a data set, and dividing the data set into a training set and a testing set.
And S8, constructing a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the test set to obtain a stable crack grade evaluation model.
In this step, the bridge crack grade evaluation model is constructed based on the reinforcement learning algorithm, the adopted algorithm network has four layers of networks, and the four layers of networks include an input layer, two hidden layers and an output layer, the input information of the input layer includes crack position face characteristics, crack length characteristics, crack width characteristics, crack trend characteristics and longitudinal position information in crack basic data, the hidden layer 1 has 10 unit nodes, the hidden layer 2 has 5 unit nodes, the output layer has 5 unit nodes, the activation function is a RELU function, and the output result is a crack grade corresponding to a score value Q of a corresponding crack evaluation result.
And 9, acquiring a crack image to be evaluated, preprocessing the crack image, extracting the length characteristic, the width characteristic, the trend characteristic, the position surface characteristic and the longitudinal position information of the crack, inputting the length characteristic, the width characteristic, the trend characteristic, the position surface characteristic and the longitudinal position information into a model, and obtaining a crack grade evaluation result.
Based on the same idea, the embodiment of the invention also provides a bridge crack grade evaluation system based on a reinforcement learning algorithm, and the system comprises: the system comprises a crack image acquisition module, an image preprocessing module, a morphological feature extraction module, a bit plane feature extraction module, a longitudinal position information extraction module, a training database, an evaluation model construction module and a result output module, wherein the crack image acquisition module, the image preprocessing module, the morphological feature extraction module, the bit plane feature extraction module, the longitudinal position information extraction module, the training database, the evaluation model construction module and the result output module are carried on an unmanned aerial vehicle; wherein:
the crack image acquisition module is used for acquiring a bridge crack image used for model training and a bridge crack image to be evaluated through an unmanned aerial vehicle;
the image preprocessing module is used for preprocessing the acquired crack image;
the morphological feature extraction module is used for carrying out crack image threshold segmentation on the preprocessed crack image and extracting crack morphological features; the fracture morphology characteristics comprise width characteristics, length characteristics and strike characteristics of the fracture;
the position surface feature extraction module is used for acquiring a position surface where the crack image is located according to the flight track of the unmanned aerial vehicle and extracting position surface features;
the longitudinal position information extraction module is used for acquiring longitudinal coordinates of the crack center point on the bridge axis according to the flight trajectory of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinates to be used as longitudinal position information of the crack;
the training database is used for judging the grade of the crack according to the width characteristic, the length characteristic, the trend characteristic, the position face characteristic and the longitudinal position information of the crack, generating a crack code, constructing basic data of the crack, arranging the basic data of the crack corresponding to all crack images according to the code to construct a data set, and dividing the data set into a training set and a testing set; the crack basic data comprises three levels of crack information, which are respectively as follows: information of bridge level: bridge coding, structural systems; information of the component hierarchy: structure type, member coding, member longitudinal/transverse/vertical dimensions; information of fracture hierarchy: crack coding, width features, length features, strike features, bit plane features, position and grade;
the evaluation model building module is used for building a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the testing set to obtain a stable crack grade evaluation model;
the result output module is used for inputting the bridge crack image to be evaluated acquired by the image acquisition module into the stable crack grade evaluation model in the evaluation model construction module and outputting the crack grade evaluation result of the bridge to be evaluated.
In the embodiment, each module is realized by a processor, and when the storage is needed, the storage is added appropriately. The Processor may be, but is not limited to, a microprocessor MPU, a Central Processing Unit (CPU), a Network Processor (NP), a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), other programmable logic devices, discrete gates, transistor logic devices, discrete hardware components, and the like. The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Alternatively, the memory may be at least one memory device located remotely from the processor.
It should be noted that, the bridge crack grade evaluation system based on the reinforcement learning algorithm in this embodiment corresponds to the bridge crack grade evaluation method based on the reinforcement learning algorithm, and the description and the limitation of the method are also applicable to the system, which is not described herein again.
The present invention is further illustrated by the following specific example.
Taking a continuous beam and a small box beam of a bridge with the code of S01 as an example, a crack image is obtained based on an image acquisition module. In this embodiment, the image acquisition module includes unmanned aerial vehicle, cloud platform, camera, infrared distancer and angle sensor. The unmanned plane and the cradle head are used as instrument carrying platforms, can fly according to a set route, and record longitudinal position information of a crack. The camera is used for collecting crack pictures. The infrared distance meter and the angle sensor provide physical conversion parameters for crack characteristic extraction in the later period.
As shown in table 3, the parameters of the components in the image capturing module used in this example are shown.
TABLE 3
| Image sensor | COMS |
| Image sensor size | 23.6*15.8mm |
| Total pixel | 1310 ten thousand |
| Effective pixel | 1230 wan |
| Color filter | RGB primary color filter |
| Maximum resolution | 4.288*2.848 |
| Size of static image | 4.288*2.848[L]，3.126*2.136[J] |
| Static image format | JPEG,RAW |
As shown in table 3, in the image acquisition module composed of the unmanned aerial vehicle, the pan-tilt, the camera, the infrared distance meter, and the angle sensor, the unmanned aerial vehicle can cruise according to a flight path planned in advance, and can perform autonomous obstacle avoidance; the holder is used for carrying a camera, an infrared distance meter and an angle sensor, and the rotating angle of the holder ranges from +85 degrees to-85 degrees; the infrared distance meter is fixed on a cradle head carried by the unmanned aerial vehicle, is used for determining the vertical distance from the image acquisition equipment to the plane where the crack is located, and is used for calibrating the crack digital characteristic information; the angle sensor is used for acquiring an included angle between a lens sight line and a structural plane where the crack is located so as to convert the actual size of the crack and the imaging size. Fig. 4 is a schematic diagram of a crack image acquired by the image acquisition module.
Graying the collected crack image, and the result is shown in fig. 5; then, the filter enhancement is performed, and the result is shown in fig. 6; maximum threshold segmentation was performed to obtain fracture morphology features, as shown in fig. 7. And extracting crack information from the existing data, wherein part of the crack information of the crack state database is shown in a table 4. As shown in table 4 and fig. 8, the constructed data structure in the fracture basic information and training database includes three levels, including a bridge level, a member level, and a fracture level.
Table 4 fracture status database example
And carrying out data normalization and dimensionless processing on the crack characteristic information in the crack state database. The collected 400 pieces of fracture data are divided into a training set and a testing set, wherein the training set comprises 300 fractures, and the testing set comprises 100 fractures. And reading the fracture data from the fracture data training set for training, wherein the structure of the fracture training model network is shown in fig. 9, and the training result is shown in fig. 10. The test set data was imported into the trained model and the prediction results are shown in table 5.
TABLE 5
As can be seen from the prediction results in Table 5, 97 samples of 100 prediction samples are correctly classified, the overall prediction accuracy reaches 97%, and the requirements are met. The 3 samples with wrong evaluation are respectively cracks numbered as S0100100100102, S0200501 and S0300501, and through analysis, the samples with wrong evaluation of crack grades are 2-type cracks and 3-type cracks, in the existing bridge detection specification, the evaluation limits of the 2-type cracks and the 3-type cracks are fuzzy, and deviation can be generated during artificial evaluation, so that the error of the concrete bridge crack grade evaluation method is controlled to be 3%. Meanwhile, the 100 samples are graded in a manual mode and a mechanical mode respectively, the time length of the crack grade evaluating method based on the reinforcement learning algorithm is 5 minutes, and the time length of the crack grade evaluating method based on the manual mode is 1 hour.
The data acquisition device acquires the crack image, extracts the morphological characteristic information of the crack, and partial extraction results are shown in fig. 11. The extracted fracture characteristic information and the longitudinal position information are subjected to dimensionless preprocessing and introduced into a fracture grade evaluation model, and the prediction results are shown in table 6.
TABLE 6
As can be seen from the prediction results of the table 6 and the graph 11, the overall prediction accuracy is 96.6%, and the crack assessment method based on the reinforcement learning algorithm improves the assessment efficiency, has the overall accuracy of 96.6% and has high feasibility.
According to the technical scheme, the bridge crack grade evaluation method and system based on the reinforcement learning algorithm are based on automatic acquisition of crack images, extract the characteristic information of cracks through a crack identification program, digitize crack characteristics, reduce errors caused by manual measurement, further utilize a crack grade evaluation model to grade the cracks, effectively reduce the subjective influence of detection personnel, and improve the efficiency and accuracy of crack detection evaluation.
The above description is only a preferred embodiment of the invention and an illustration of the applied technical principle and is not intended to limit the scope of the claimed invention but only to represent a preferred embodiment of the invention. It will be appreciated by those skilled in the art that the scope of the invention herein disclosed is not limited to the particular combination of features described above, but also encompasses other arrangements formed by any combination of the above features or their equivalents without departing from the spirit of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
Claims (10)
1. A bridge crack grade assessment method based on a reinforcement learning algorithm is characterized by comprising the following steps:
collecting a crack image through an unmanned aerial vehicle;
preprocessing the crack image;
performing crack image threshold segmentation on the preprocessed crack image, and extracting crack morphological characteristics; the fracture morphology characteristics comprise width characteristics, length characteristics and strike characteristics of the fracture;
acquiring a position plane of the crack image according to the flight track of the unmanned aerial vehicle, and extracting position plane characteristics;
acquiring a longitudinal coordinate of a crack central point on the axis of the bridge according to the flight track of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinate to be used as longitudinal position information of the crack;
judging the grade of the crack and generating a crack code according to the width characteristic, the length characteristic, the trend characteristic, the position face characteristic and the longitudinal position information of the crack, and constructing basic data of the crack, wherein the basic data of the crack comprises three levels of crack information which are respectively as follows: information of bridge level: bridge coding, structural systems; information of component hierarchy: structure type, member coding, member longitudinal/transverse/vertical dimensions; information of fracture hierarchy: crack coding, width feature, length feature, strike feature, bit plane feature, position and grade;
arranging the crack basic data corresponding to all crack images according to codes to construct a data set, and dividing the data set into a training set and a test set;
constructing a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the test set to obtain a stable crack grade evaluation model;
collecting a crack image to be evaluated, preprocessing the crack image, extracting the length characteristic, the width characteristic, the trend characteristic, the position face characteristic and the longitudinal position information of the crack, inputting the information into a model, and obtaining a crack grade evaluation result.
2. The bridge crack grade assessment method according to claim 1, wherein the crack image acquisition step comprises the following steps:
s11, collecting a bridge image; when the bridge image is collected, all parts of the bridge are collected in an omnibearing manner;
s12, carrying out crack identification on the acquired bridge image, and marking the identified position meeting the crack condition;
and S13, carrying out local image acquisition again on the identified crack position to obtain a crack image.
3. The bridge crack grade evaluation method according to claim 1, wherein the crack image acquisition comprises the following steps:
s11, collecting a bridge image; when the bridge image is collected, all parts of the bridge are collected in an omnibearing way;
s12, carrying out crack identification on the acquired bridge image, and marking the identified position meeting the crack condition;
and S14, segmenting the crack image at the crack position in the bridge image to obtain the crack image.
4. The bridge crack grade evaluation method of claim 1, wherein the extracting crack morphology features comprises:
step S31, acquiring a starting point and an end point of a crack, calculating a central point of a crack image, dividing the crack image by the central point, and acquiring a shortest distance edge point and a longest distance edge point which take the central point as a symmetric center;
step S32, judging the crack trend according to the included angle between the connecting line of the starting point and the end point of the crack and the longitudinal axis of the bridge, and compiling crack trend codes as trend characteristics;
step S33, calculating width characteristics by taking the distance between the edge points with the shortest distance as the width of the crack;
and step S34, calculating length characteristics by taking the distance between the longest distance edge points as the crack length.
5. The bridge crack grade assessment method of claim 4, wherein the step S32 of judging the crack trend and compiling a crack trend code as a trend feature comprises:
the cracks on the bottom surface or the top surface comprise a transverse crack, a longitudinal crack and an oblique crack, wherein the included angle between the transverse crack and the longitudinal axis is 75-90 degrees, the included angle between the oblique crack and the longitudinal axis is 15-75 degrees, and the included angle between the longitudinal crack and the longitudinal axis is 0-15 degrees;
the cracks on the side face comprise three types, namely vertical cracks, longitudinal cracks and oblique cracks, wherein the included angle between the vertical cracks and the longitudinal axis is 75-90 degrees, the included angle between the oblique cracks and the longitudinal axis is 15-75 degrees, and the included angle between the longitudinal cracks and the longitudinal axis is 0-15 degrees;
on the basis of judging the fracture strike, constructing four-dimensional vectors as strike codes, and coding and compiling the fracture strike characteristics, wherein each strike corresponds to one vector; when the current trend is judged, the value is assigned to 1; otherwise the value is assigned to 0.
6. The bridge crack grade assessment method of claim 4, wherein the width feature is characterized by a relative width, as shown in formula (1):
in the formula (1), the reaction mixture is,
representing the relative width of the fracture; d represents the actual width of the crack; d Limit of Representing the width limit of the crack.
7. The bridge crack grade assessment method of claim 4, wherein the length feature is characterized by a relative length, as shown in equation (2):
in the formula (2), the reaction mixture is,
represents the relative length of the fracture; l represents the projection length of the crack in the projection direction, the crack length is projected in three directions of a three-dimensional rectangular space coordinate system, the longest crack length is taken as the projection length, and the coordinate axis where the longest crack is projected is taken as the projection direction; l represents the structural dimension of the component in the projection direction.
8. The bridge crack grade assessment method of claim 1, wherein the bit plane of the crack comprises a bottom surface, a top surface and a side surface, and a three-dimensional vector is constructed to represent bit plane features; each dimension in the three-dimensional vector represents a bit plane, and the value of the bit plane at the current position is 1, otherwise, the value is 0.
9. The bridge crack level evaluation method according to claim 1, wherein a bridge crack level evaluation model is constructed based on a reinforcement learning algorithm, an adopted algorithm network has four layers of networks, and the four layers of networks comprise an input layer, two hidden layers and an output layer, input information of the input layer comprises crack position face characteristics, crack length characteristics, crack width characteristics and crack trend characteristics in crack basic data, the hidden layer 1 comprises 10 unit nodes, the hidden layer 2 comprises 5 unit nodes, the output layer comprises 5 unit nodes, an activation function is a RELU function, and an output result is a crack level corresponding to a score value Q of a corresponding crack evaluation result.
10. A bridge crack grade assessment system based on a reinforcement learning algorithm is characterized by comprising: the system comprises a crack image acquisition module, an image preprocessing module, a morphological feature extraction module, a bit plane feature extraction module, a longitudinal position information extraction module, a training database, an evaluation model construction module and a result output module, wherein the crack image acquisition module, the image preprocessing module, the morphological feature extraction module, the bit plane feature extraction module, the longitudinal position information extraction module, the training database, the evaluation model construction module and the result output module are carried on an unmanned aerial vehicle; wherein:
the crack image acquisition module is used for acquiring a bridge crack image used for model training and a bridge crack image to be evaluated through an unmanned aerial vehicle;
the image preprocessing module is used for preprocessing the acquired crack image;
the morphological feature extraction module is used for carrying out crack image threshold segmentation on the preprocessed crack image and extracting crack morphological features; the fracture morphology characteristics comprise width characteristics, length characteristics and strike characteristics of the fracture;
the position surface feature extraction module is used for acquiring a position surface where the crack image is located according to the flight track of the unmanned aerial vehicle and extracting position surface features;
the longitudinal position information extraction module is used for acquiring longitudinal coordinates of the central point of the crack on the axis of the bridge according to the flight track of the unmanned aerial vehicle and the shooting parameters, and normalizing the longitudinal coordinates to be used as longitudinal position information of the crack;
the training database is used for judging the grade of the crack according to the width characteristic, the length characteristic, the trend characteristic, the position surface characteristic and the longitudinal position information of the crack, generating crack codes, constructing basic crack data, arranging the basic crack data corresponding to all crack images according to the codes to construct a data set, and dividing the data set into a training set and a testing set; the crack basic data comprises three levels of crack information, which are respectively as follows: information of bridge level: bridge coding, structural systems; information of the component hierarchy: structure type, member code, member longitudinal/transverse/vertical dimensions; information of fracture hierarchy: crack coding, width features, length features, strike features, bit plane features, position and grade;
the evaluation model building module is used for building a bridge crack grade evaluation model based on a reinforcement learning algorithm, training the evaluation model by adopting the training set, and verifying the training model by utilizing the testing set to obtain a stable crack grade evaluation model;
the result output module is used for inputting the bridge crack image to be evaluated acquired by the image acquisition module into a stable crack grade evaluation model in the evaluation model construction module and outputting the crack grade evaluation result of the bridge to be evaluated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211191901.4A CN115511836B (en) | 2022-09-28 | 2022-09-28 | Bridge crack grade assessment method and system based on reinforcement learning algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211191901.4A CN115511836B (en) | 2022-09-28 | 2022-09-28 | Bridge crack grade assessment method and system based on reinforcement learning algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115511836A true CN115511836A (en) | 2022-12-23 |
| CN115511836B CN115511836B (en) | 2023-10-27 |
Family
ID=84506892
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211191901.4A Active CN115511836B (en) | 2022-09-28 | 2022-09-28 | Bridge crack grade assessment method and system based on reinforcement learning algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115511836B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116128380A (en) * | 2023-04-13 | 2023-05-16 | 沧州路桥工程有限责任公司 | Bridge health monitoring method and device, electronic equipment and storage medium |
| CN116434087A (en) * | 2023-04-04 | 2023-07-14 | 中建二局土木工程集团有限公司 | Concrete crack identification method and device based on GOA-SVM cooperative algorithm and unmanned aerial vehicle |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105279790A (en) * | 2014-06-13 | 2016-01-27 | 中国石油化工股份有限公司 | Fracture network 3D digital core modeling method |
| CN108051450A (en) * | 2017-11-01 | 2018-05-18 | 中铁大桥科学研究院有限公司 | A kind of bridge health monitoring system and method based on unmanned plane |
| CN108985363A (en) * | 2018-07-03 | 2018-12-11 | 长安大学 | A kind of cracks in reinforced concrete bridge classifying identification method based on RBPNN |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN110378950A (en) * | 2019-06-18 | 2019-10-25 | 上海同岩土木工程科技股份有限公司 | A kind of tunnel structure crack identification method merged based on gray scale and gradient |
| CN110866872A (en) * | 2019-10-10 | 2020-03-06 | 北京邮电大学 | Pavement crack image preprocessing intelligent selection method and device and electronic equipment |
| KR102239377B1 (en) * | 2019-12-26 | 2021-04-13 | 주식회사 딥인스펙션 | System and method for safety inspection on bridge facilities based on image and xai |
| CN112750200A (en) * | 2021-01-14 | 2021-05-04 | 中路高科交通检测检验认证有限公司 | Bridge disease implementation method and device based on BIM technology |
| CN114913162A (en) * | 2022-05-25 | 2022-08-16 | 广西大学 | Bridge concrete crack detection method and device based on lightweight transform |
| CN114937016A (en) * | 2022-05-25 | 2022-08-23 | 广西大学 | Bridge concrete crack real-time detection method and device based on edge calculation and Transformer |
-
2022
- 2022-09-28 CN CN202211191901.4A patent/CN115511836B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105279790A (en) * | 2014-06-13 | 2016-01-27 | 中国石油化工股份有限公司 | Fracture network 3D digital core modeling method |
| CN108051450A (en) * | 2017-11-01 | 2018-05-18 | 中铁大桥科学研究院有限公司 | A kind of bridge health monitoring system and method based on unmanned plane |
| CN108985363A (en) * | 2018-07-03 | 2018-12-11 | 长安大学 | A kind of cracks in reinforced concrete bridge classifying identification method based on RBPNN |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN110378950A (en) * | 2019-06-18 | 2019-10-25 | 上海同岩土木工程科技股份有限公司 | A kind of tunnel structure crack identification method merged based on gray scale and gradient |
| CN110866872A (en) * | 2019-10-10 | 2020-03-06 | 北京邮电大学 | Pavement crack image preprocessing intelligent selection method and device and electronic equipment |
| KR102239377B1 (en) * | 2019-12-26 | 2021-04-13 | 주식회사 딥인스펙션 | System and method for safety inspection on bridge facilities based on image and xai |
| CN112750200A (en) * | 2021-01-14 | 2021-05-04 | 中路高科交通检测检验认证有限公司 | Bridge disease implementation method and device based on BIM technology |
| CN114913162A (en) * | 2022-05-25 | 2022-08-16 | 广西大学 | Bridge concrete crack detection method and device based on lightweight transform |
| CN114937016A (en) * | 2022-05-25 | 2022-08-23 | 广西大学 | Bridge concrete crack real-time detection method and device based on edge calculation and Transformer |
Non-Patent Citations (3)
| Title |
|---|
| PARK, J等: "CRACK DETECTION AND REFINEMENT VIA DEEP REINFORCEMENT LEARNING", IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) * |
| 余加勇 等: "基于YOLOv5和U-Net3+的桥梁裂缝智能识别与测量", 湖南大学学报(自然科学版) * |
| 魏世银: "大跨桥梁结构健康诊断和管养决策的机器学习方法研究", 中国博士学位论文全文数据库工程科技II辑 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116434087A (en) * | 2023-04-04 | 2023-07-14 | 中建二局土木工程集团有限公司 | Concrete crack identification method and device based on GOA-SVM cooperative algorithm and unmanned aerial vehicle |
| CN116434087B (en) * | 2023-04-04 | 2024-01-23 | 中建二局土木工程集团有限公司 | Concrete crack identification method and device based on GOA-SVM cooperative algorithm and unmanned aerial vehicle |
| CN116128380A (en) * | 2023-04-13 | 2023-05-16 | 沧州路桥工程有限责任公司 | Bridge health monitoring method and device, electronic equipment and storage medium |
| CN116128380B (en) * | 2023-04-13 | 2023-06-30 | 沧州路桥工程有限责任公司 | Bridge health monitoring method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115511836B (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115511836B (en) | Bridge crack grade assessment method and system based on reinforcement learning algorithm | |
| CN110751099B (en) | Unmanned aerial vehicle aerial video track high-precision extraction method based on deep learning | |
| CN111126184B (en) | Post-earthquake building damage detection method based on unmanned aerial vehicle video | |
| CN107977656A (en) | A kind of pedestrian recognition methods and system again | |
| CN111768417B (en) | Railway wagon overrun detection method based on monocular vision 3D reconstruction technology | |
| CN109948471A (en) | Based on the traffic haze visibility detecting method for improving InceptionV4 network | |
| CN110910440B (en) | Power transmission line length determination method and system based on power image data | |
| CN114049356B (en) | Method, device and system for detecting structure apparent crack | |
| CN112960014B (en) | Rail transit operation safety online real-time monitoring and early warning management cloud platform based on artificial intelligence | |
| CN114035604B (en) | Video monitoring and unmanned aerial vehicle air-ground linkage abnormal target detection method | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN115512247A (en) | Regional building damage grade assessment method based on image multi-parameter extraction | |
| CN115526999A (en) | Watershed model generation method and generation device, electronic device and storage medium | |
| Pinto et al. | LOW-COST UAS PHOTOGRAMMETRY FOR ROAD INFRASTRUCTURES’INSPECTION | |
| CN117371949B (en) | Three-dimensional visual model-based power transmission line construction safety monitoring method and system | |
| CN111784785A (en) | Bridge dynamic displacement identification method | |
| CN116363088A (en) | Expressway concrete pavement crack detection method based on unmanned aerial vehicle, electronic equipment and storage medium | |
| CN110827264A (en) | Evaluation system for apparent defects of concrete member | |
| CN113361968B (en) | Power grid infrastructure worker safety risk assessment method based on artificial intelligence and big data | |
| CN113408916B (en) | Fire-fighting facility detection and field acceptance assessment system based on intelligent AI and mobile APP | |
| Boudet et al. | A supervised classification approach towards quality self-diagnosis of 3D building models using digital aerial imagery | |
| Fujita et al. | Rapid and Accurate Detection of Building Damage Investigation Using Automatic Method to Calculate Roof Damage Rate | |
| CN111507423B (en) | Engineering quantity measuring method for cleaning transmission line channel | |
| JIA et al. | Crack damage detection of bridge based on convolutional neural networks | |
| Avendãno et al. | Image based inspection of concrete cracks using UAV photography |
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 |