CN115112854A - Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning - Google Patents
Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning Download PDFInfo
- Publication number
- CN115112854A CN115112854A CN202210586153.3A CN202210586153A CN115112854A CN 115112854 A CN115112854 A CN 115112854A CN 202210586153 A CN202210586153 A CN 202210586153A CN 115112854 A CN115112854 A CN 115112854A
- Authority
- CN
- China
- Prior art keywords
- polarization
- scattering
- matrix
- corrosion
- steel bar
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000010287 polarization Effects 0.000 title claims abstract description 100
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 46
- 230000007797 corrosion Effects 0.000 title claims abstract description 46
- 238000005260 corrosion Methods 0.000 title claims abstract description 46
- 239000010959 steel Substances 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000010801 machine learning Methods 0.000 title claims abstract description 19
- 230000000149 penetrating effect Effects 0.000 title claims abstract description 19
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims description 36
- 238000004088 simulation Methods 0.000 claims description 19
- 239000004567 concrete Substances 0.000 claims description 13
- 230000001427 coherent effect Effects 0.000 claims description 11
- 239000011150 reinforced concrete Substances 0.000 claims description 11
- 238000012706 support-vector machine Methods 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 5
- 229910001294 Reinforcing steel Inorganic materials 0.000 claims description 3
- 230000004913 activation Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 claims description 2
- 238000012360 testing method Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 2
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 7
- 238000011161 development Methods 0.000 description 4
- 229910001220 stainless steel Inorganic materials 0.000 description 4
- 239000010935 stainless steel Substances 0.000 description 4
- 238000005259 measurement Methods 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- VEXZGXHMUGYJMC-UHFFFAOYSA-M Chloride anion Chemical compound [Cl-] VEXZGXHMUGYJMC-UHFFFAOYSA-M 0.000 description 1
- 239000011358 absorbing material Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000002161 passivation Methods 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/20—Metals
- G01N33/204—Structure thereof, e.g. crystal structure
- G01N33/2045—Defects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N17/00—Investigating resistance of materials to the weather, to corrosion, or to light
- G01N17/006—Investigating resistance of materials to the weather, to corrosion, or to light of metals
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/885—Radar or analogous systems specially adapted for specific applications for ground probing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Remote Sensing (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Electromagnetism (AREA)
- Computational Mathematics (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Medicinal Chemistry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Food Science & Technology (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Crystallography & Structural Chemistry (AREA)
- Databases & Information Systems (AREA)
- Algebra (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental & Geological Engineering (AREA)
- Environmental Sciences (AREA)
- Computing Systems (AREA)
Abstract
The utility model provides a ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning: acquiring the complete polarization data information of GPR on the underground steel bar, and acquiring the multidimensional polarization attribute by adopting series polarization decomposition to provide more information for target identification; meanwhile, the obtained data is used for training and testing a machine learning algorithm based on the multilayer perceptron, and the method is used for classifying whether the steel bars are corroded or not. The method fully excavates and utilizes the polarization characteristics of the steel bar, and can obtain a more stable and efficient identification effect.
Description
Technical Field
The invention relates to the technical field of radars, in particular to a ground penetrating radar technology for identifying underground rusted steel bars.
Background
A Ground Penetrating Radar (GPR) is a nondestructive testing tool for road detection, underground exploration and the like, and the principle is mainly to emit short high-frequency electromagnetic pulses and record reflection signals of an underground target body. With the development of instruments and high-frequency antennas, ground penetrating radars are also used for evaluating the state of materials, such as characterization of moisture or chloride ion content in concrete and detection of corrosion of steel bars. The obvious dielectric property difference exists between the steel bar corrosion products (iron rust and cracks) and the reinforced concrete components, which is the theoretical basis that the ground penetrating radar can be used for steel bar corrosion detection. However, the conventional GPR signal based on single polarization contains limited underground information, and most of the conventional GPR signals directly use maximum amplitude, peak phase, or radar profile, so that uncertainty exists in the identification result.
Disclosure of Invention
In view of this, the present disclosure provides a ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning, which can obtain higher corrosion identification accuracy and improve the problem of uncertainty in the conventional ground penetrating radar detection method.
The ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning comprises the following steps:
acquiring full-polarization data of reinforced concrete, and establishing a full-polarization scattering matrix of each point;
performing a series of polarization decompositions on the fully polarized scattering matrix to obtain a polarization attribute data set of each point, including: polarization entropy, polarization scattering anisotropy, polarization scattering angle, surface scattering power, bulk scattering power, secondary scattering power, and polarization-like properties of surface scattering, secondary scattering, and bulk scattering;
and adopting a machine learning device, and identifying the corrosion reinforcing steel bars after learning training and parameter adjustment optimization are carried out on the polarization attribute data set.
Further, the method for acquiring fully polarized data comprises the following steps: the method comprises the steps of directly obtaining the signals in a full polarization radar detection mode, respectively collecting and synthesizing the signals through receiving and transmitting antennas in four polarization modes, and generating the signals through simulation software.
Further, the polarization decomposition of the fully polarized scattering matrix specifically includes the following steps:
obtaining a three-dimensional polarization coherent matrix and a three-dimensional polarization covariance matrix from the fully polarized scattering matrix;
performing H/A/alpha decomposition on the polarization coherent matrix to extract a three-dimensional polarization attribute, wherein the three-dimensional polarization attribute comprises the following steps: polarization entropy, polarization scattering anisotropy degree and polarization scattering angle;
performing Freeman-Durden decomposition on the polarization covariance matrix to extract a three-dimensional polarization attribute, which comprises the following steps: surface scattering power, volume scattering power and secondary scattering power;
based on the polarization coherent matrix, extracting a polarization similarity attribute rs of the target body, wherein rs is:
wherein k is P Is the Pauly scattering vector, T, of a typical scatterer 3 Trace (-) is the trace operation of the matrix for the said polarized coherent matrix obtained from the measured data.
Further, the machine learning device adopts a support vector machine or a multilayer perceptron.
Furthermore, the multilayer perceptron is a fully-connected neural network comprising two hidden layers, the number of nodes of an output layer is set to be 1, and softmax is used as an activation function.
Furthermore, the method for identifying the rusted steel bars further comprises the step of constructing a rusted reinforced concrete simulation model for generating full polarization data by using simulation software; the simulation model comprises four elements of concrete, steel bars, corrosion products and cracks, and is structurally characterized by comprising the steps of reducing the diameter of the steel bars in the corrosion process, forming a corrosion layer, opening the concrete cracks and filling the corrosion products in the cracks.
The method acquires the full polarization data information of the GPR on the steel bars in the concrete, adopts a series of polarization decomposition methods to obtain the multidimensional polarization attribute, and provides more information for target identification; meanwhile, binary classification is realized on the obtained data by using a machine learning algorithm which mainly uses a multilayer perceptron, a method combining machine learning and polarization decomposition is developed, the identification accuracy is further improved, the problems that the characteristic information contained in a single-polarization GPR signal is limited and the identification result is uncertain are solved, and the precise identification of the corrosion of the steel bar is realized.
Compared with the prior art, this disclosed beneficial effect lies in: (1) compared with the prior art that H/A/alpha decomposition is used, and only H and alpha are analyzed, the method extracts more dimensional polarization characteristics of the reinforced concrete and obtains more identification information; (2) machine learning based on a neural network is introduced, uncertainty caused by single polarization radar signal identification is overcome, and a more stable and efficient identification effect can be obtained; (3) the multi-layer Perceptron (MLP) built by the method is proposed and verified to be a better choice than a commonly-used Support Vector Machine (SVM) in the aspect of identifying the rusted steel bars; (4) a more detailed and actual corrosion reinforced concrete simulation model is established, and the method is used for learning training and parameter optimization of machine learning, and can further improve the recognition efficiency and accuracy.
Drawings
FIG. 1 is a flowchart of a method for identifying reinforcement corrosion by a ground penetrating radar according to an exemplary embodiment;
FIG. 2 is a schematic diagram of the rust stage of reinforced concrete;
FIG. 3 is a fully polarized data storage form of an exemplary embodiment;
FIG. 4 is an exemplary multi-tier perceptron;
FIG. 5 is an exemplary concrete rebar corrosion simulation model;
FIG. 6 is a result of 500 random training sessions in a simulation scenario;
FIG. 7 shows the corrosion steel bars (a) and the stainless steel bars (b) measured;
fig. 8 shows the results of 500 random training sessions of the measured scene. .
Detailed Description
The present disclosure is described in detail below by way of examples with reference to the accompanying drawings.
The purpose of the disclosure is to overcome the defects of the traditional ground penetrating radar corrosion steel bar detection method and provide a ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning. The method is suitable for detecting the initial stage of the steel bar corrosion (surface passivation film damage and rust generation) and also can be used for detecting the development period of the steel bar corrosion (cracks appear in concrete due to continuous rust generation).
Fig. 1 shows a flow chart of an exemplary embodiment of the present disclosure. The method mainly comprises the following steps:
step 1: acquiring full-polarization data of reinforced concrete, and establishing a full-polarization scattering matrix model;
the full polarization data can be directly acquired through a full polarization radar detection mode, and can also be acquired and synthesized through receiving and transmitting antennas in four polarization modes respectively. And full polarization data of a simulation scene can be obtained by adopting ground penetrating radar simulation software such as gprMax and the like based on a simulation model.
The data acquisition adopts two-dimensional scanning measurement, a measurement matrix is arranged on each observation point, the measurement matrix is composed of echo data in four polarization modes, the echo data is subjected to noise and clutter suppression, and the full polarization data of each pixel point of a target can be represented by a 2 multiplied by 2 polarization scattering matrix as follows:
the measured full polarization data storage format is shown in FIG. 3, in which HH, HV, VH, VV represent polarization channel types, S HH 、S HV 、S VH 、S VV The complex scattering coefficients of the four polarization channels are respectively, and i and j are pixel point indexes of scanning data.
Step 2: and decomposing the complete polarization scattering matrix by adopting a multi-polarization decomposition algorithm to obtain a multi-dimensional polarization attribute data set. The method specifically comprises the following steps:
assuming that GPR satisfies the single-station backscatter regime, then reciprocity-constrained scatter matrix S is a symmetric matrix, i.e., has S HV =S VH From which a three-dimensional polarized coherence matrix T can be constructed 3 And three-dimensional polarization covariance matrix C 3 :
Based on the polarization coherent matrix, extracting three-dimensional polarization attributes through H/A/alpha decomposition: polarization entropy H, polarization scattering anisotropy A and polarization scattering angle alpha, wherein the polarization entropy H and the polarization scattering anisotropy A are measures of randomness of the scattering mechanism, and the polarization scattering angle alpha characterizes the scattering mechanism.
Extracting three-dimensional polarization attributes through Freeman-Durden decomposition based on the polarization covariance matrix: surface scattered power P S Volume scattering power P V Secondary scattering power P D 。
Based on the polarization coherent matrix, the polarization similarity rs of the target body can be extracted. rs is defined as:
wherein k is P Is the Pauly scattering vector, T, of a typical scatterer 3 Trace (-) is the trace operation of matrix for the polarization coherent matrix obtained by actually measuring data; rs represents the similarity between a typical scatterer and an actually measured scatterer;
consider three typical scatterers, namely surface scatterers, secondary scatterers, and bulk scatterers, whose respective pauli scattering vectors are:
since the three Pagli scattering vectors are orthogonal to each other, the corresponding similarity attribute rs s ,rs d ,rs v Satisfies the following conditions:
rs s +rs d +rs v =1
therefore, each pixel point can obtain a data set containing nine polarization characteristics, which can be expressed as:
D=[H,A,alpha,P S ,P D ,P V ,rs s ,rs d ,rs v ]
theoretically, the more feature information is obtained, the more the identification of the target is facilitated.
And step 3: and (3) utilizing a machine learning device to perform learning training and parameter tuning by using the polarization attribute data set, and then identifying the corroded steel bar. The method comprises the following steps: enough pixel points are selected for the rust steel bars and the stainless steel bars respectively, nine polarization attributes of each pixel point are used as input characteristics, labels are rust (1) and stainless corrosion (0), and supervised learning training and parameter adjustment optimization are carried out on a machine learning device.
For the two classification problems of whether the steel bar is corroded or not, a Support Vector Machine (SVM) can be used to obtain good results. However, simulation and practice show that better results can be obtained by applying the multi-layer sensor provided by the disclosure through optimization. As shown in fig. 4, a fully-connected neural network, i.e. a multi-layer perceptron, having two hidden layers is preferably constructed for supervised learning to distinguish rusted steel bars from rustless steel bars. Since it is a two-class problem, it is preferable that the number of output layer nodes is set to 1, and its activation function uses softmax.
As optimization, the ground penetrating radar corrosion steel bar identification method further comprises the steps of constructing a simulation model of corrosion steel bars in concrete, generating full polarization simulation data, and using the full polarization simulation data for machine learning training and parameter adjustment optimization.
FIG. 2 is a schematic diagram of different corrosion stages of reinforced concrete, the development period of corrosion of steel bars is mainly studied in the exemplary embodiment, and the proposed three-dimensional model comprises four elements of concrete, steel bars, corrosion products and cracks; typical relative permittivity, conductivity and relative permeability of the medium used in the model are similar to those of real materials; the characteristic elements considered structurally include: the diameter of the steel bar is reduced in the corrosion process, a corrosion layer is formed, the opening degree of the concrete crack is opened, and a corrosion product in the crack is filled.
Simulation experiment:
as shown in fig. 5, the development period of concrete reinforcement corrosion was modeled. And (3) constructing a simulation model of the corrosion of the steel bars in the concrete, wherein the parameters are shown in the table 1 and the table 2. The stainless steel bars to which they were compared had a radius of 1.2cm, no cracks and rust products and the other parameters were the same as in the table.
Table 1 simulation parameter settings (geometric parameters)
Table 2 simulation parameter settings (dielectric properties of the medium)
And the ground penetrating radar simulation software gprMax is adopted for simulation, full polarization data of a simulation scene are obtained, and polarization characteristics are obtained by polarization decomposition of the full polarization data. The corrosion reinforcing steel bar and the stainless steel bar are respectively subjected to 9-dimensional polarization characteristics of 682 pixel points, 500 times of random training is carried out by utilizing the neural network of fig. 4, and the sample proportion of the training set and the test set is 2: 1, the relationship between the training times and the accuracy is shown in fig. 6, the average accuracy is 85.67%, and the MSE is less than 0.1. Under the same conditions, the kernel function of the SVM which is compared with the kernel function selects a polynomial function, the average precision is 83.06 percent, and the error recognition rate is 0.1626. The built multilayer perceptron is superior to the SVM.
Actually measured test:
the rusted steel bars and the rustless steel bars shown in the figure 7 are adopted for experiments, the radar system adopts the LTD-2600 ground penetrating radar, the radar is fixed on the sliding guide rail, and wave absorbing materials are paved around the scene.
The 9-dimensional polarization characteristics of 682 pixel points are respectively taken for actually measured corrosion steel bars and stainless corrosion steel bars, 500 times of random training is carried out by utilizing the neural network of figure 4, and the sample proportion of the training set and the test set is 2: 1, the relationship between the training times and the accuracy is shown in fig. 8, the average accuracy is 84.47%, and the MSE is less than 0.1. Under the same conditions, the kernel function of the SVM to be compared selects a polynomial function, the average precision is 83.09%, and the error recognition rate is 0.1674. The built neural network is superior to the SVM.
The result shows that the method can fully excavate reinforced concrete corrosion information, can obtain higher corrosion identification precision and certainty compared with the traditional ground penetrating radar detection method, and belongs to a steady and efficient corrosion steel bar identification method.
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. A method for identifying a corrosion steel bar of a ground penetrating radar based on polarization characteristics and machine learning is characterized by comprising the following steps:
acquiring full-polarization data of reinforced concrete, and establishing a full-polarization scattering matrix of each point;
performing a series of polarization decompositions on the fully polarized scattering matrix to obtain a polarization attribute data set of each point, including: polarization entropy, polarization scattering anisotropy, polarization scattering angle, surface scattering power, bulk scattering power, secondary scattering power, and polarization-like properties of surface scattering, secondary scattering, and bulk scattering;
and adopting a machine learning device, and identifying the corrosion reinforcing steel bars after learning training and parameter adjustment optimization are carried out on the polarization attribute data set.
2. The method for identifying the rusted steel bar according to claim 1, wherein the method for acquiring the full polarization data comprises the following steps: the method comprises the steps of directly obtaining the signals in a full polarization radar detection mode, respectively collecting and synthesizing the signals through receiving and transmitting antennas in four polarization modes, and generating the signals through simulation software.
3. The method for identifying the rusted steel bar according to claim 1, wherein the polarization decomposition of the fully polarized scattering matrix specifically comprises the following steps:
obtaining a three-dimensional polarization coherent matrix and a three-dimensional polarization covariance matrix from the fully polarized scattering matrix;
performing H/A/alpha decomposition on the polarization coherent matrix to extract a three-dimensional polarization attribute, wherein the three-dimensional polarization attribute comprises the following steps: polarization entropy, polarization scattering anisotropy degree and polarization scattering angle;
performing Freeman-Durden decomposition on the polarization covariance matrix to extract a three-dimensional polarization attribute, which comprises the following steps: surface scattering power, volume scattering power and secondary scattering power;
based on the polarization coherent matrix, extracting a polarization similarity attribute rs of the target body, wherein rs is:
wherein k is P Is the Pauly scattering vector, T, of a typical scatterer 3 Trace (-) is the trace operation of the matrix for the said polarized coherent matrix obtained from the measured data.
4. The method of claim 1, wherein the machine learning device employs a support vector machine or a multi-layer perceptron.
5. The method for identifying the rusty steel bar according to claim 4, wherein the multilayer perceptron is a fully-connected neural network comprising two hidden layers, the number of nodes of an output layer is set to 1, and softmax is used as an activation function.
6. The method for identifying the rusted steel bars according to any one of claims 1 to 5, further comprising the step of constructing a rusted reinforced concrete simulation model for generating the complete polarization data by using simulation software; the simulation model comprises four elements of concrete, steel bars, corrosion products and cracks, and is structurally characterized by comprising the steps of reducing the diameter of the steel bars in the corrosion process, forming a corrosion layer, opening the concrete cracks and filling the corrosion products in the cracks.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210586153.3A CN115112854A (en) | 2022-05-26 | 2022-05-26 | Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210586153.3A CN115112854A (en) | 2022-05-26 | 2022-05-26 | Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115112854A true CN115112854A (en) | 2022-09-27 |
Family
ID=83326594
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210586153.3A Pending CN115112854A (en) | 2022-05-26 | 2022-05-26 | Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115112854A (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109145424A (en) * | 2018-08-10 | 2019-01-04 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | It is a kind of for the bridge data recognition methods of Coherent Noise in GPR Record and system |
CN109685011A (en) * | 2018-12-25 | 2019-04-26 | 北京华航无线电测量研究所 | A kind of underground utilities detection recognition method based on deep learning |
CN110826643A (en) * | 2019-11-20 | 2020-02-21 | 上海无线电设备研究所 | Offshore target identification method based on polarized Euler feature fusion deep learning |
CN111323764A (en) * | 2020-01-21 | 2020-06-23 | 山东大学 | Underground engineering target body intelligent identification method and system based on ground penetrating radar |
CN111444629A (en) * | 2020-04-15 | 2020-07-24 | 中国二冶集团有限公司 | Reinforcing steel bar corrosion parameter prediction method based on support vector machine |
CN111751392A (en) * | 2020-07-30 | 2020-10-09 | 广州大学 | Steel bar corrosion detection method based on dual-polarization ground penetrating radar |
CN112882019A (en) * | 2021-01-14 | 2021-06-01 | 长春工程学院 | Full-polarization target identification and classification method based on rotary single-polarization ground penetrating radar |
CN113344219A (en) * | 2021-06-11 | 2021-09-03 | 广东电网有限责任公司 | Concrete reinforcement corrosion state evaluation method, system, terminal and storage medium |
-
2022
- 2022-05-26 CN CN202210586153.3A patent/CN115112854A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109145424A (en) * | 2018-08-10 | 2019-01-04 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | It is a kind of for the bridge data recognition methods of Coherent Noise in GPR Record and system |
CN109685011A (en) * | 2018-12-25 | 2019-04-26 | 北京华航无线电测量研究所 | A kind of underground utilities detection recognition method based on deep learning |
CN110826643A (en) * | 2019-11-20 | 2020-02-21 | 上海无线电设备研究所 | Offshore target identification method based on polarized Euler feature fusion deep learning |
CN111323764A (en) * | 2020-01-21 | 2020-06-23 | 山东大学 | Underground engineering target body intelligent identification method and system based on ground penetrating radar |
CN111444629A (en) * | 2020-04-15 | 2020-07-24 | 中国二冶集团有限公司 | Reinforcing steel bar corrosion parameter prediction method based on support vector machine |
CN111751392A (en) * | 2020-07-30 | 2020-10-09 | 广州大学 | Steel bar corrosion detection method based on dual-polarization ground penetrating radar |
CN112882019A (en) * | 2021-01-14 | 2021-06-01 | 长春工程学院 | Full-polarization target identification and classification method based on rotary single-polarization ground penetrating radar |
CN113344219A (en) * | 2021-06-11 | 2021-09-03 | 广东电网有限责任公司 | Concrete reinforcement corrosion state evaluation method, system, terminal and storage medium |
Non-Patent Citations (12)
Title |
---|
匡纲要 等: "《极化合成孔径雷达基础理论及其应用》", 30 June 2011, 《国防科技大学出版社》, pages: 58 - 59 * |
李平湘 等: "《雷达干涉测量原理与应用》", 31 December 2016, 《国防大学出版社》, pages: 139 * |
肖可可: "基于极化SAR数据的地物要素提取", 《现代测绘》 * |
肖可可: "基于极化SAR数据的地物要素提取", 《现代测绘》, vol. 35, no. 3, 31 May 2012 (2012-05-31), pages 8 * |
郑树剑等: "主成分分析和支持向量机的方法在混凝土结构钢筋腐蚀检测中的应用", 《电子测量技术》 * |
郑树剑等: "主成分分析和支持向量机的方法在混凝土结构钢筋腐蚀检测中的应用", 《电子测量技术》, no. 09, 15 September 2007 (2007-09-15), pages 16 - 18 * |
郑树剑等: "基于支持向量机的混凝土结构中钢筋腐蚀的判别", 《电子器件》 * |
郑树剑等: "基于支持向量机的混凝土结构中钢筋腐蚀的判别", 《电子器件》, no. 05, 15 October 2007 (2007-10-15), pages 1935 - 1938 * |
钟景阳: "基于极化探地雷达的钢筋早期锈蚀检测方法研究初探", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
钟景阳: "基于极化探地雷达的钢筋早期锈蚀检测方法研究初探", 《中国优秀硕士学位论文全文数据库 工程科技II辑》, 15 May 2022 (2022-05-15), pages 1 * |
韩萍: "基于典型散射差异指数的PolSAR图像Lee滤波", 《系统工程与电子技术》 * |
韩萍: "基于典型散射差异指数的PolSAR图像Lee滤波", 《系统工程与电子技术》, vol. 40, no. 2, 28 February 2018 (2018-02-28), pages 288 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liu et al. | GPRInvNet: Deep learning-based ground-penetrating radar data inversion for tunnel linings | |
Gurbuz et al. | A compressive sensing data acquisition and imaging method for stepped frequency GPRs | |
US8717223B2 (en) | Classification of subsurface objects using singular values derived from signal frames | |
CN105137498A (en) | Underground target detection and recognition system and method based on feature fusion | |
CN107402386B (en) | A kind of estimation method of underground metalliferous round tube radius and buried depth based on BP neural network | |
CN107390213B (en) | A kind of time lag curve extracting method of the Ground Penetrating Radar record section based on sliding window | |
Todkar et al. | One-class SVM based outlier detection strategy to detect thin interlayer debondings within pavement structures using Ground Penetrating Radar data | |
Persico et al. | Microwave tomography for processing of GPR data at Ballachulish | |
Ozkaya et al. | Deep dictionary learning application in GPR B-scan images | |
Sun et al. | Estimating parameters of the tree root in heterogeneous soil environments via mask-guided multi-polarimetric integration neural network | |
CN114638830B (en) | Training method of tunnel reinforcing steel bar recognition model and tunnel reinforcing steel bar recognition method | |
Kuchipudi et al. | Automated assessment of reinforced concrete elements using ground penetrating radar | |
Gonzalez-Huici | A strategy for landmine detection and recognition using simulated GPR responses | |
Hanabusa et al. | Deep-learning-based calibration in contrast source inversion based microwave subsurface imaging | |
CN115112854A (en) | Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning | |
Li et al. | 3D ground penetrating radar road underground target identification algorithm using time-frequency statistical features of data | |
Hu et al. | Defects identification and location of underground space for ground penetrating radar based on deep learning | |
Juan et al. | A new Wavelet Prediction method for GPR clutter elimination Based on LSTM network | |
Guo et al. | Research on tunnel lining image target recognition method based on YOLOv3 | |
Zhang et al. | A Deep Learning Method for Rough Surface Clutter Reduction in GPR Image | |
Tarai et al. | A Practical GPR Imaging Scheme for Buried Objects | |
CN112859007B (en) | Method for detecting and identifying weak and small targets under sea clutter background based on polarization decomposition | |
Caorsi et al. | Electromagnetic infrastructure monitoring: The exploitation of GPR data and neural networks for multi-layered geometries | |
Li et al. | Training data augmentations for improving hyperbola recognition in ground penetrating radar B-scan image for tree roots detection | |
Ge et al. | Wavelet-GAN: A GPR Noise and Clutter Removal Method Based on Small Real Datasets |
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 |