CN117406190A

CN117406190A - Non-excavation pull rod corrosion detection method, device and equipment based on radar signals

Info

Publication number: CN117406190A
Application number: CN202311358362.3A
Authority: CN
Inventors: 周春晓; 夏立伟; 胡洪炜; 尹洪; 吴启进; 张楚谦; 付子峰; 吴嘉琪; 周炜
Original assignee: Super High Voltage Co Of State Grid Hubei Electric Power Co ltd
Current assignee: Super High Voltage Co Of State Grid Hubei Electric Power Co ltd
Priority date: 2023-10-19
Filing date: 2023-10-19
Publication date: 2024-01-16

Abstract

The invention provides a method, a device and equipment for detecting the corrosion of a trenchless pull rod based on radar signals, which comprises the following steps: collecting a grounding grid target echo signal; extracting the echo waveform characteristics of the ground grid scattering in the echo signals of the ground grid targets to form a ground grid echo signal fingerprint library; constructing a grounding grid A-SCAN characteristic waveform library under different soil environments, materials and burial depths according to the grounding grid echo signal fingerprint library; performing two-dimensional imaging of the grounding grid according to the A-SCAN characteristic waveform library of the grounding grid by adopting an offset imaging method to obtain a focused B-SCAN image; constructing a C-SCAN three-dimensional image of the grounding grid according to the focused B-SCAN image; and (3) intercepting a horizontal section of the B-SCAN image, and identifying fault points of the pull rods in the grounding grid according to the structure of the horizontal section. The detection method can be used for rapidly, accurately and independently limited by on-site operation conditions, and monitoring of the state of the grounding network is realized.

Description

Non-excavation pull rod corrosion detection method, device and equipment based on radar signals

Technical Field

The invention belongs to the technical field of power failure detection, and particularly relates to a method, a device and equipment for detecting trenchless pull rod corrosion based on radar signals.

Background

The design, standard installation and maintenance detection of the grounding network of the ultra-high voltage transmission line and the transformer substation are the basis for ensuring the safe operation of the line and the system, and the grounding performance is valued by the power grid design and the production operation departments for a long time. Along with the construction of an extra-high voltage transmission line and the large-scale implementation of urban power grid transformation, the requirements of an electric power system on the aspects of laying, operation and maintenance of a grounding device are more and more strict. In the operation process of the grounding grid, the contradiction between the design of the grounding grid and the actual safe operation is the key point of the safety problem faced by the grounding grid, and the effective detection mode of the electric power grounding grid is searched as a necessary means for realizing the safe operation of the electric network due to the existence of various threats of the safety problem of the grounding grid.

In the construction of the grounding grid, the design of the pull rod is important. And because of the reasons of imprecision in the construction process, inaccuracy in the test process and the like, great hidden danger is caused to the safe and stable operation of the power transmission line and the transformer substation. The phenomenon that the grounding grid is not in accordance with the design often occurs in the construction of the grounding grid, such as the small size of the grounding body pull rod, the inaccurate welding spots, poor welding, missing welding and the like, which lead to the deterioration of the electrical connection performance of the grounding grid. In addition, the pull rod usually runs in soil, and the running environment is abominable, easily contacts moist, harmful gas and soil acidity etc. environment, causes the earth screen to corrode, causes ground conductor or lead wire to rust, warp, even fracture to destroy original design structure, lead to the performance variation, make personal and equipment's safety receive serious threat. Therefore, the accurate detection of the actual physical state and physical details of the pull rod in the power transmission line and the transformer substation grounding network has important significance, and the regular detection and maintenance are key for ensuring the normal operation of the system.

Because the pull rod in the grounding grid is a concealed device which is different from other equipment and is independently buried underground, the detection and maintenance are inconvenient, and the detection of the grounding grid of the transformer substation and the transmission line is limited by the concealment and the obstruction of the ground surface building. Meanwhile, the grounding grid has the characteristics of large scale, complex structure, large laying depth and the like. In actual engineering, compared with original measuring means of a pull rod in a grounding grid, the traditional measuring method is direct excavation measurement or indirect resistance value measurement, has blindness, and has the defects of large workload, low speed, low measuring precision and the like. Meanwhile, the extra-high voltage transmission line and the transformer substation bear the heavy duty of ensuring the normal running of national production and people life, and the power failure maintenance can inevitably bring about huge economic loss, so that the detection of the grounding grid in the actual engineering is difficult.

The conventional evaluation methods for the pull rod nondestructive test (NDT) in the grounding grid mainly comprise an electric network analysis method, an electromagnetic field analysis method, an electrochemical method and the like, but the methods are obvious in the problems of large workload, low efficiency, lower measurement precision and the like due to the fact that the grounding grid is dependent on a design construction drawing, low positioning precision or power failure of a transformer substation is required, the operation of a power system is influenced.

Disclosure of Invention

The invention provides a method, a device and equipment for detecting the non-excavation pull rod corrosion based on radar signals, and aims to solve the problems of large workload, low efficiency, low measurement precision and the like of a pull rod nondestructive detection method in a grounding network of a power system.

In order to achieve the above object, the present invention provides the following technical solutions:

a radar signal-based trenchless pull rod corrosion detection method comprises the following steps:

collecting a grounding grid target echo signal;

extracting the waveform characteristics of the ground grid scattering echo in the ground grid target echo signal by using a matching tracking MP and an orthogonal matching tracking method to form a ground grid echo signal fingerprint library;

constructing a grounding grid A-SCAN characteristic waveform library under different soil environments, materials and burial depths according to the grounding grid echo signal fingerprint library;

performing two-dimensional imaging of the grounding grid according to the A-SCAN characteristic waveform library of the grounding grid by adopting an offset imaging method to obtain a focused B-SCAN image;

constructing a C-SCAN three-dimensional image of the grounding grid according to the focused B-SCAN image;

and (3) intercepting a horizontal section of the B-SCAN image, and identifying fault points of the pull rods in the grounding grid according to the structure of the horizontal section.

Preferably, the ground network target echo signals are collected by a UWB pulse radar hardware system, wherein the UWB pulse radar hardware system comprises a group of receiving and transmitting antennas, a radar module and an I/0 module.

Preferably, when the UWB pulse radar hardware system is adopted to collect the target echo signals of the grounding grid, the relation between the detection time and the electromagnetic wave speed is as follows:

wherein X is the distance between a transmitting antenna and a receiving antenna, Z is the target burial depth, t is the double-way travel time of the ground penetrating radar GPR, and v is the propagation speed of electromagnetic waves in soil.

Preferably, removing background noise in the ground plane target echo signal by a gain joint-RPCA method is further included before extracting the ground plane scattered echo waveform characteristic in the ground plane target echo signal.

Preferably, after removing the background noise in the ground network target echo signal, the method further comprises performing data enhancement on the ground network target echo signal from which the noise is removed by adopting a resolution enhancement method.

Preferably, the method for extracting the waveform characteristics of the ground plane scattering echo in the ground plane target echo signal by using the matching pursuit MP and the orthogonal matching pursuit method comprises the following steps: and extracting characteristic waveforms based on the signal subspace, extracting the characteristic of the echo waveform of the ground network scattering by utilizing a matching pursuit MP and orthogonal matching pursuit method, and integrating the ground network echo signal fingerprint library.

Preferably, the method for performing two-dimensional imaging of a ground grid according to the ground grid A-SCAN characteristic waveform library to obtain a focused B-SCAN image specifically comprises the following steps:

Let e (x, z, t) be the echo signal received by the antenna of ground penetrating radar GPR, described using the scalar helmholtz wave equation:

wherein x represents the horizontal moving distance of the antenna, z represents the underground depth, t represents the time when the antenna receives an echo signal, and v is the propagation speed of electromagnetic waves in soil;

the offset imaging process of the ground penetrating radar extends to an offset section e (x, z=0, t) and particularly relates to the geometric repositioning of a return signal, and offset refocusing imaging is completed;

i.e. fourier transform E (k) of the recording profile E (x, z=0, t) _x Z=0, ω) to determine the fourier transform a (k) of the offset profile e (x, z, t=0) _x ,k _z ) Inverse transformation results in an offset profile e (x, z, t=0);

the propagation time of the electromagnetic wave of the ground penetrating radar in the underground medium is double-pass time, and the propagation speed of the electromagnetic wave adopts half of the actual propagation speed v; thus, k _x ，k _z The ω relationship is expressed as:

wherein v is constant in a homogeneous medium;

in the frequency wavenumber domain, the wavefield extrapolation at depth z is expressed as:

the offset profile e (x, z, t=0) is:

and obtaining an offset refocusing section of the target space from the recording section of the space, and completing offset imaging of the echo image signals.

Preferably, the entropy of the image after the echo image signal completes offset imaging is defined as:

Wherein S represents the number of sampling points, and T represents the number of sampling channels;

and when the electromagnetic wave propagation speed of the preset soil medium is estimated, obtaining a sectional view of the wire drawing rod by using an F-K offset imaging algorithm.

The invention also provides a device for detecting the trenchless pulling rod corrosion based on the radar signal, which comprises:

the signal acquisition module is used for acquiring a grounding grid target echo signal;

the signal extraction module is used for extracting the waveform characteristics of the ground grid scattering echo in the ground grid target echo signals by utilizing a matching tracking MP and an orthogonal matching tracking method to form a ground grid echo signal fingerprint library;

the waveform construction module is used for constructing a grounding grid A-SCAN characteristic waveform library under different soil environments, materials and burial depths according to the grounding grid echo signal fingerprint library

The B-SCAN image construction module is used for carrying out two-dimensional imaging on the grounding grid according to the grounding grid A-SCAN characteristic waveform library by adopting an offset imaging method to obtain a focused B-SCAN image;

the C-SCAN three-dimensional image construction module is used for constructing a grounding grid C-SCAN three-dimensional image according to the focused B-SCAN image;

and the fault identification module is used for intercepting the horizontal section of the B-SCAN image and identifying the fault point of the pull rod in the grounding grid according to the structure of the horizontal section.

The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory, the processor executing the computer program to implement the steps of any one of the method for detecting the trenchless pull rod corrosion based on radar signals.

According to the radar signal-based trenchless pull rod corrosion detection method, the ground network echo signal fingerprint library is constructed, the ground network A-SCAN characteristic waveform library under different soil environments, materials and burial depths is constructed according to the ground network echo signal fingerprint library, the design construction drawing is not needed, the characteristic waveforms are sequentially converted into the A-SCAN characteristic waveform library, the B-SCAN image and the ground network C-SCAN three-dimensional image through a series of operations, the horizontal section of the B-SCAN image is finally intercepted, the fault points of the pull rods in the ground network are identified according to the structure of the horizontal section, power failure processing is not needed, and the monitoring of the ground network state can be realized quickly, accurately and without being limited by on-site operation conditions.

Drawings

In order to more clearly illustrate the embodiments of the present invention and the design thereof, the drawings required for the embodiments will be briefly described below. The drawings in the following description are only some of the embodiments of the present invention and other drawings may be made by those skilled in the art without the exercise of inventive faculty.

FIG. 1 is a flow chart of a radar signal-based trenchless pull rod corrosion detection method according to embodiment 1 of the present invention;

fig. 2 is a Vivaldi antenna; wherein (a) is a traditional Vivaldi antenna model, and (b) is a Vivaldi radiation direction;

FIG. 3 is a radar transceiver pulse signal; wherein (a) a spectrogram, (b) a pulse signal;

FIG. 4 is a schematic view of a gprMax model space;

FIG. 5 is a first Fresnel zone of the transmit and receive antennas;

FIG. 6 is a gprMax simulation model;

FIG. 7 is a hand-held GPR radar transceiver system and test sand pool; wherein (a) hand-held ultra-wideband GPR radar, (b) and (C) sand pool experiments;

FIG. 8 is an exponential gain joint-RPCA clutter suppression method; wherein (a) raw data, (b) target echo data, (c) clutter and noise data;

FIG. 9 is an exponential gain joint-RPCA clutter suppression method; wherein (a) raw data, (b) target echo data, (c) clutter and noise data;

FIG. 10 is an A-scan analysis of the effect of RPCA constraint parameters on target signal extraction; wherein (a) a one-dimensional signal of trace=44, (b) a one-dimensional signal of data trace=42;

FIG. 11 is an experimental view of a grounding grid- -round steel 1; wherein (a) simulation model, (b) radar signal, (c) velocity-entropy curve, (d) F-K offset, (e) F-K offset (true velocity);

FIG. 12 is an experimental view of the grounding grid- -round steel 2; wherein (a) simulation model, (b) radar signal, (c) velocity-entropy curve, (d) F-K offset, (e) F-K offset (true velocity);

fig. 13 is an experimental view of a ground screen-flat steel 1; wherein (a) simulation model, (b) radar signal, (c) velocity-entropy curve, (d) F-K offset, (e) F-K offset (true velocity);

fig. 14 is an experimental view of ground screen-steel flat 2; wherein (a) simulation model, (b) radar signal, (c) velocity-entropy curve, (d) F-K offset, (e) F-K offset (true velocity);

FIG. 15 is a drawing of a vertical insertion experiment of a pull rod; wherein (a) experimental scene, (b) original radar chart, (c) filtering clutter image, (d) horizontal placement speed-entropy curve, (e) F-K offset;

FIG. 16 is a drawing of a pull rod oblique insertion experiment; wherein (a) experimental scene, (b) original radar chart, (c) filtering clutter image, (d) horizontal placement speed-entropy curve, (e) F-K offset;

FIG. 17 is a diagram of a dual objective pull rod experiment; wherein (a) the experimental scene, (b) the original radar chart, (c) the clutter image is filtered, (d) the speed-entropy curve is horizontally placed, and (e) the F-K offset is carried out.

Detailed Description

The present invention will be described in detail below with reference to the drawings and the embodiments, so that those skilled in the art can better understand the technical scheme of the present invention and can implement the same. The following examples are only for more clearly illustrating the technical aspects of the present invention, and are not intended to limit the scope of the present invention.

In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the technical solutions of the present invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention.

Furthermore, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. In the description of the present invention, it should be noted that, unless explicitly specified or limited otherwise, the terms "connected," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; may be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances. In the description of the present invention, unless otherwise indicated, the meaning of "a plurality" is two or more, and will not be described in detail herein.

Example 1

The grounding device usually runs in soil, has a severe running environment, is easy to contact with humid, harmful gas, soil acidity and other environments, causes corrosion and carbonization of a grounding grid, and researches show that the annual corrosion rate of metal in corrosive soil is 2.0-8.0 mm, and the grounding conductor or the lead is easy to rust, deform and even break, thereby damaging the original design structure. In recent years, domestic and foreign investigation shows that the frequent occurrence of electric power system accidents caused by ground network faults brings about huge economic loss, meanwhile, the ground network is a concealed device which is independently buried underground, the buried depth is usually in the range of 0.3-0.5m of the shallow earth surface, the deepest can reach 1.5m, and the traditional direct excavation sampling or indirect resistance value measuring method has limitations and inefficiency and cannot form visual and objective detection. How to detect the electric connection faults of the grounding network caused by factors such as construction technology, soil medium, manual fault current electrodynamic action, soil corrosion and the like under the condition of not excavating, has become a focus and a difficult problem of attention of various countries in the world.

The invention mainly aims at researching the application feasibility of the ground penetrating radar for detecting the construction integrity of a high-voltage grounding network. And in the construction stage of the ground grid without finishing the construction of the transformer substation, the material property of the ground grid is estimated by analyzing the frequency/phase/amplitude characteristics between the target scattering signal detected by the GPR and the reference signal, so as to find out fault points such as ground grid fracture, welding leakage, outer edge disconnection and the like in the construction process and corrosion points generated by the influence of the environment on the ground grid.

Based on the above, the invention provides a radar signal-based trenchless pull rod corrosion detection method, which specifically is shown in fig. 1 and comprises the following steps:

and step 1, collecting a target echo signal of the grounding grid.

Specifically, the invention uses UWB pulse radar hardware system to collect ground network target echo signals. The UWB pulse radar hardware system is a GPR radar and mainly comprises the following parts: a group of receiving and transmitting antennas, a radar module and an I/0 module. The underlying software includes a drive and I/O module database, which is provided by the I/O module manufacturer and is used for specific I/O module programming. The middle layer is called Rardalib3 for intercommunication between the I/O modules and the user applications. And at the top layer, the user application program receives and processes the data in the radar module and finally displays the data through the terminal.

(1) And a radar module.

The radar signal generator port transmits a first-order Gaussian pulse signal with the transmission bandwidth of 0.45-3.55 GHz (-10 dB), the pulse repetition frequency (Pulse Repetition Frequency, PRF) is 48MHz, and parallel sampling with the maximum frame depth of 512 can be realized.

(2) I/O module

The I/O module is a USB to SPI adapter, which is itself a connection cable with various built-in electronic components, for communication between the radar module and the computer. The main characteristics are as follows: standard USB to SPI I/O adapter, do not require firmware development; the development is simple and convenient and has strong functions through computer support; for connectors that interface with Novelda radar modules, the USB2.0 high-speed interface (480 Mb/s) provides faster rates for high frame rate applications; the USB power supply mode provides power for the radar module;

(3) Antenna module

The ultra-wideband pulse GPR of the invention adopts a traditional linear polarization antenna and a Vivaldi antenna. The antenna is composed of three parts, namely a feeder line, a circular open cavity and a gradual change index slot line, as shown in fig. 2 (a), wherein the index gradual change curve of the antenna is an important parameter for determining the bandwidth of the antenna, the circular open cavity and the gradual change index slot line are arranged on the same side of a dielectric plate, and the feeder line is arranged on the other side of the dielectric plate. Theoretically, the vivaldi antenna bandwidth is not affected by the infinite bandwidth of frequency, but the actual main bandwidth is limited by the distance between the starting point and the ending point of the antenna radiation port curve, and the distance between the starting point and the ending point determines the highest frequency point of the antenna, and then corresponds to the lowest frequency of the antenna bandwidth. The metallic patch layer on the surface of the antenna can bind the radiation of the electric field, so that the binding capability of the metallic patch to the surface electric field at the opening direction of the exponential taper slot line of the antenna is minimum, corresponding to the maximum radiation direction of the antenna, and the far field directional radiation mode is shown in fig. 2 (b). The bandwidth range of the vivaldi antenna adopted by the invention is 0.9 GHz-5 GHz. The lobe width of the main lobe is E-plane and H-plane, and the antenna size. And combining the frequency band range of the radar signal generation module, and transmitting a first-order Gaussian pulse signal with the center frequency of 2.235GHz and the frequency band of 0.9-3.55 GHz by the ultra-wideband pulse ground penetrating radar. The radar transceiver pulse signal is shown in fig. 3.

(4) Display interface

The display of ultra-wideband pulse GPR is one NET radar application RadarScope. Vivaldi is SMA through the double-end, and the resistance value is 50 ohm's connecting wire and the transmitting end and the receiving end of antenna link to each other, and I/O module is connected through a 20PIN connector between the radar module.

When the hardware is well connected, the hardware and the computer are correctly connected through an adapter for converting the I/O module into the SPI after the radar application software package and the driver are correctly installed. The method comprises the steps of carrying out real-time updating on data captured by the radar in all windows during the operation of the radar through software, and storing sampling B-scan data. When shallow target measurement is carried out in sand pool experiments, radar parameter files are important to the subsequent estimation of the depth of the target in the vertical direction, and in addition, the sampling time of a radar receiving end needs to be zeroed before each measurement. Frame offset (Frame offset) refers to the time from the start of the transmission of a radar pulse to the sampling circuit to the sampling/listening of a received signal. The time required for a radar pulse to travel this distance is largely dependent on the medium through which the signal passes. In estimating depth information of echo data in a-scan/B-scan, it is important to ensure that the sampling time zero point of a-scan is consistent with the time of transmitting a pulse signal by a transmitting antenna, so that the propagation time of the pulse signal in a sampling circuit is measured, most CMOS circuits are influenced by environments such as temperature and time critical circuits (such as delay elements), and therefore, the propagation time of the signal in a hardware circuit is measured before each measurement. The radar sampling time "zeroing" is achieved by adjusting the parameters "SampleDelayToReference" (distance between the internal circuit and the reference point, in m), "sampledelayfromtreference" (time between the reference point and the start of the frame, in s) and "offsetdistancefromtreference" (distance between the reference point and the start of the sampling, in m).

When the UWB pulse radar hardware system is adopted to collect the target echo signals of the grounding grid, the relation between the detection time and the electromagnetic wave speed is as follows:

During the microwave nondestructive detection, the radar belongs to a very important technology, and the physical detection work of the grounding grid can be finished under the condition of no power failure by adopting the radar with high penetrability. The ultra-wideband transmitting signal can enter the grounding conductor, then the physical condition of the grounding conductor is mastered by adopting multi-dimensional measuring information in the echo signal, and the high output resolution is mastered by adopting the antenna technology, so that a very realistic visual image can be provided, and good data support is brought.

And 2, removing background noise in the ground network target echo signal by a gain joint-RPCA method, and performing data enhancement on the noise-removed ground network target echo signal by a resolution enhancement method.

Step 3, extracting the ground grid scattering echo waveform characteristics in the ground grid target echo signals after data enhancement by using a matching pursuit MP and an orthogonal matching pursuit method to form a ground grid echo signal fingerprint library, wherein the method specifically comprises the following steps: and extracting characteristic waveforms based on the signal subspace, extracting the characteristic of the echo waveform of the ground network scattering by utilizing a matching pursuit MP and orthogonal matching pursuit method, and integrating the ground network echo signal fingerprint library.

And 4, constructing a grounding grid A-SCAN characteristic waveform library under different soil environments, materials and burial depths according to the grounding grid echo signal fingerprint library.

And 5, performing two-dimensional imaging of the grounding grid according to the A-SCAN characteristic waveform library of the grounding grid by adopting an offset imaging method to obtain a focused B-SCAN image, wherein the method specifically comprises the following steps of: .

wherein v is constant in a homogeneous medium;

the offset profile e (x, z, t=0) is:

The entropy of the image after the echo image signal completes offset imaging is defined as:

And 6, constructing a C-SCAN three-dimensional image of the grounding grid according to the focused B-SCAN image.

And 7, intercepting a horizontal section of the B-SCAN image, and identifying fault points of the pull rods in the grounding grid according to the structure of the horizontal section.

In the following, the detection method provided by the invention is verified by constructing the ground penetrating radar experiment software and the hardware platform.

And building a GprMax forward simulation platform, wherein a GprMax simulation model is shown in figure 6.

The gprMax is open source software which is mainly written by Python and simulates electromagnetic wave propagation. Maxwell's equations describe The relationship between basic electromagnetic fields and their dependence on excitation sources, with which all electromagnetic phenomena can be described, gprMax uses The Finite-differential Time-Domain-Time-Domain (FDTD) method to discretize spatial and temporal continuity into Finite units, as in fig. 4, to solve iteratively on each unit for The maxwell's equation value of geometry and initial conditions (excitation signals emitted by GPR antennas) in a satisfied 3D simulation model, in order to simulate GPR responses from a specific target or set of targets.

(1) Description of the method

Theoretically, the smaller the discretized three-dimensional space x, y, z and the time t, the more truly the FDTD model can represent the electromagnetic wave propagation characteristics in the model, but is limited by the limited memory capacity and the processing speed of a computer, kane Yee determines the set criteria of the minimum grid building block (unit) size of the FTDT space discretization, namely the size of Deltax, deltay and Deltaz, and names grid units as Yee cells. Fig. 4 illustrates the coordinate system of gprMax, with the electric field component tangential to the matrix interface, the magnetic field component perpendicular to the matrix interface, and no field component in the center of the matrix for each cell matrix. The space discretization is required according to the precision, source pulse frequency and target geometry required by GPR numerical simulation, in addition, the electromagnetic wave in the real scene propagates around at the same speed, is not influenced by the direction and frequency (assuming no dispersion medium and far field conditions), but the situation is not satisfied in the simulation, the error caused by the simulation can be reduced by following a rule of thumb, and the discrete step size (deltax, deltay, deltaz) is ensured to be less than one tenth of the minimum wavelength of the propagating electromagnetic field.

Another most challenging problem that GPR forward simulation needs to solve is: the simulation of a real scene where a field propagates in space to infinity is zero. The space defined when the simulation model is built in gprMax is of finite size, i.e., z maximum is not infinite, so Maxwell's equation computation domain is truncated at a distance from the target finite distance. The gprMax uses an approximation, the perfect matching layer (Perfectly Matched Layer, PML) absorbs the boundary conditions (Absorbing Boundary Condition, ABC) and applies to the distance from the field source to the cut-off boundary, thus limiting the computation space. ABC functions to absorb any waves striking the cut-off boundary, simulating infinite space, thus avoiding the placement of excitation sources in these layers. In addition, due to the imperfection of ABCs, there is a reflection wave at the cut-off boundary of the computational domain (i.e., model) that is much smaller than the amplitude of the reflection wave in the model, in order not to introduce significant artificial reflection to interfere with the echo information of the target, when the size of the model domain is designed, it is necessary to satisfy that all targets are at least 15 Yee cells from the cut-off boundary, and that the air medium of at least 15-20 Yee cells in the GPR model is located in the area above the source (antenna).

(2) GPR model construction

FDTD simulations first define the dielectric parameters of the material. The simplest GPR detects at least three materials of different electromagnetic properties in the shallow mine and UXO models, air, dielectric half space (soil) and metal/nonmetal UXO, respectively. The dielectric parameters of air (free space) have been built into gplmax, which can be accessed and used using the identifier free_space, and the dielectric parameters of the metallic material have also been modeled in gplmax, which is accessed and used using the identifier pec. The definition of the dielectric half-space is the focus of the gprMax simulation, the soil is usually a non-magnetic material, μ _r ＝1，σ _* =0, definition and naming of the dielectric parameters, material identifiers by the term #material.

And then determining the waveform type and the center frequency of the antenna excitation source, and determining the center frequency and the pulse type of the radar signal according to the maximum reachable depth and the transverse/longitudinal resolution required by actual detection. In sand basin experiments, the GPR system uses a first order gaussian pulse with a center frequency of 2.235GH, so in corresponding GPR numerical simulations, a first order gaussian pulse radar signal waveform with an amplitude of 1 and a center frequency of 2.0GHz is created with a #wave form command, a Z-direction polarized hertz dipole source perpendicular to the measurement direction is commanded with a #hetzian_dipole, and an initial position of the transmitting antenna.

The computational spatial resolution (grid) and computational domain size are then defined. The mesh size is defined based on the shortest wavelength at which the excitation is effectively resolved and the detailed characteristics of the geometry in the model. To determine the minimum wavelength, the highest frequency and lowest speed present in the model are required. The highest frequency is not the center frequency of the first order gaussian pulse and it is necessary to examine the pulse waveform spectrum confirmation, which corresponds to the wavelength in the dielectric half space (with low wave velocity) as equation (9). The size of the computational domain should be sufficient to accommodate objects such as subsurface targets, interferents, etc., and at least 10 Yee cells required to account for the PML absorption boundary conditions and at least 10 Yee cells spaced between the PML and objects in the dielectric half-space.

Finally, the simulated total time window must be set long enough to ensure that the electromagnetic wave propagates through the dielectric half-space to the target object and is reflected back to the receiving antenna after being transmitted from the antenna. Dielectric half space range and mine therein and UXO geometric models can be constructed using the gprMax physical object creation command.

B-scan image resolution enhancement and subsurface target feature waveform extraction based on sparse representation.

Sparse representation of a signal refers to mapping the signal to a sparse domain (sparse basis) and representing useful information of the signal with as few non-zero transform domain elements as possible. For the observed echo data containing clutter interference, removing clutter components based on a sparse representation method of original data, and obtaining target signals with clutter components as few as possible, namely predicted data, wherein the target signals cannot be far away from the observed data. Consider a set of functions d= { g _k K=1, 2,.. _k Is a unit vector of the number of units,the function set D is called dictionary, g _k Called atoms (or groups). The theory of highly linear estimation focuses on solving the following problem, given set g _k And an observation matrix X, how to construct the signal S by self-adaptive approximation using M bases in the dictionary D, i.e. approximation of XResults X _M The expression is:

wherein c= { c ₁ ,c ₂ ,...c _K Sparsity of sparse representation, I _M Is an index set, corresponding to non-zero elements in c, the number of non-zero elements in the set is M, and the corresponding approximation error is expressed as:

since M is much smaller than N, this signal representation method is called sparse representation. For redundant dictionaries (K > N), the unit vectors therein must be non-linearly independent, so that equation (11) has many different sets of representations. The focus of the sparse decomposition study is: when the error isApproaching a constant value, how to select the set with the highest sparsity from all possible combinations, i.e. how to get the minimum problem for M in equation (11).

Finding the best sparse representation of observed data, i.e. solving for i ₀ Norm minimization problem:

wherein c ₀ Representing the number of non-zero entries in c.

In practical applications, equation (12) is typically converted to a sparse approximation problem to approximate a solution, in the form:

min||c|| ₀ s.t.||X-Dc||≤ε (13)

But due to l ₀ The Non-convex function of norms, the computation of the best M-approximation from any redundant dictionary, is an NP-hard problem (Non-deterministic Polynomial hard problem, NP-hard). Therefore, it is necessary to rely on research calculationsThe method obtains good but non-optimal approximate values according to the signal characteristics of the actual application scene.

RPCA sparse representation denoising analysis.

The development of robust principal component analysis can be traced back to principal component analysis (Principal Component Analysis, PCA), which cannot effectively extract the target signal from highly corrupted measurement signals, i.e. observed data containing outliers and sparsity noise of too high amplitude. The RPCA effectively connects the effective bridge of two problems of sparse representation and matrix recovery, overcomes the defects of a PCA method by introducing sparse representation, enhances the robustness of an algorithm and recovers a low-rank matrix. In the application scene of land mine detection, RPCA decomposes GPR acquired data into a low-rank clutter matrix and a sparse matrix containing target response, and PCA is used as one of the performance analysis comparison methods of clutter denoising.

(1) Basic principle of

The echo of the target in the ground penetrating radar detection soil can be regarded as the sum of the responses of a plurality of strong scattering centers to be approximated, and is a sparse signal in a time domain, but a covariance matrix in PCA is very sensitive to data containing abnormal points or sparse signals with large amplitude, and can not effectively de-correlate an interference signal with coupling from the target signal, so that clutter components in the reconstructed target signal B-scan data can be restrained. Cande et al propose a robust principal component analysis method incorporating sparse representation to overcome the shortcoming of PCA, and for the radar signal clutter suppression field, RPCA decomposes m×n-sized raw data X into two parts, namely a low-rank matrix A (background noise/clutter interference) and a sparse matrix E (target echo), corresponding to a mathematical model representation as shown in formula (14), wherein A is a low-rank matrix with a small number of eigenvalues other than zero, and E represents a sparse matrix with only a few non-zero terms. The RPCA model can be defined as the mathematical optimization problem of (15)

X＝A+E (14)

Where rank () represents the rank of the matrix, I.I ₀ Is l ₀ And the norm is the number of non-zero elements in the matrix, and lambda is a positive weighting parameter used for balancing the proportion between the low-rank matrix A and the sparse matrix E. But the optimized function of equation (15) -Low function sum l ₀ The norm is non-convex, no effective algorithm can solve the optimal solution, and the global optimal solution is guaranteed. Wright et al propose to sum the kernel norms and l ¹ The norms are respectively convex relaxed as a value function and l ₀ The principal component tracing method of the norm effectively solves the optimization problem of solving the optimal solution in the formula (15). As long as matrix E is sufficiently sparse compared to a, equation (15) is converted to solve the convex optimization problem below, decomposing the observation matrix X:

wherein I II _* Is the kernel norm (sum of singular values), I.I ₁ Represents the maximum value of the sum of the absolute values of the matrix row (column) elements, λ being a positive weighting parameter.

(2) Optimization model solution

The above optimization can be regarded as solving for the minimization l ¹ The problem of combining norms and nuclear norms, also known as convex optimization. The interior point method relies on the second order information of the objective function, whose iterative computation step direction complexity is O (m ⁶ ) The method is not suitable for a matrix with large dimension; iterative thresholding (Iterative Thresholding, IT), slow convergence computation takes too long; the APG and gradient rising method combined with the rapid extension technology has higher calculation speed than the IT method, but the convergence is not in line with the actual application requirement; the invention adopts the matrix recovery technology of the non-precise augmented Lagrangian multiplier (InexactAugmented Lagrange Multipliers, IALM) proposed by Lin et al to solve the problem of convex optimization, and the IALM is an improved method of the precise augmented Lagrangian multiplier method (Exact Augmented Lagrange Multipliers, EALM) to obtain the precise solution, and has the characteristics of high precision and high convergence speed.

Applying EALM to RPCA to solve the convex optimization problem of (16), then:

X＝(A,E)，f(X)＝||A|| _* +λ||E|| ₁ ，h(X)＝X-A-E (17)

the corresponding Lagrangian function form is:

however, the solution of (18) involves the sub-problem formula (19), { mu } _k The total number of SVD calculations is affected by the process, since this process consumes most of the calculation time, when the constraint factor μ _k If the setting is too low, the convergence rate of the iterative thresholding method for solving the equation becomes slow, and the literature indicates that the IALM can effectively solve this problem. The detailed procedure of the IALM-based RPCA method is described in algorithm 1.

/>

Wherein J (X) = max (| Y ₂ ,λ ^-1 ||Y|| _∞ )，||·|| _∞ Representing the maximum value of the sum of absolute values of each row, the loop termination condition of algorithm 1 is:

||X-A _k -E _k || _F /||X|| _F ＜ε ₁ ，μ _k-1 ||E _k -E _k-1 || _F /||X|| _F ＜ε ₂ (20)

μ _k the update criteria of (2) are:

wherein λ=1/sqrt (max [ m, n)])，μ ₀ ＝0.5/||sgn(X)|| ₂ ，ρ＝1.6，ε ₁ ＝10 ^-7 ，ε ₂ ＝10 ^-5 。

In the invention, when GPR sampling data of shallow target detection are processed, parameters rho, mu and lambda in a signal decomposition model of RPCA are adjusted according to the distribution characteristics between clutter and target signals, so as to obtain an optimal result.

Gain processing in ground penetrating radar echoes.

Pulse radars are attenuated during the propagation of the soil medium due to absorption, dispersion, spherical divergence, and the like. Absorption causes energy losses associated with thermal conversion, associated with amplitude decay; dispersion refers to pulse deformation (typically expressed as a time extension) related to the frequency dependence of the wave parameters; spherical divergence refers to the scattering of energy from a wavefront, a pure geometric factor. The high frequency pulse components are more significantly affected by attenuation than the low frequency components, and radar image resolution suffers. In low and medium loss dielectrics, the signal varies significantly over only its amplitude spectrum, the phase spectrum remains almost unchanged, and the one-dimensional wave equation solution is in the general form:

A(x,t)＝A ₀ exp(-αx)exp[iω(t-x/ν)] (22)

Wherein A (x, t) represents the electromagnetic field value at the horizontal sampling point x, A ₀ ＝A(0,0)，Alpha is the attenuation coefficient, v is the phase velocity, ω=2pi f, for pulsed radar f=f _c . The attenuation coefficient is typically frequency dependent with the phase velocity. Attenuation effects can be measured by a quality factor Q, expressed as (23) representing the ratio between stored and scattered energy during the cycle, GPR performance being better at high values of Q. Substituting equation (23) into equation (22) the signal amplitude can be expressed as (24).

In summary, the electromagnetic wave amplitude is exponentially decaying over time, which effect is particularly pronounced for high frequency components. From a practical point of view, well known time gain and spectral balancing tools can effectively correct for attenuation effects, are suitable for non-uniform media and do not require precise knowledge of attenuation parameters. According to the invention, an exponential gain is adopted to apply a gain function to a tracked continuous area (a time window with a specified length) in time, and an exponential fitting of a relational expression as shown in a formula (25) is carried out on echo data in a specified time window range of a specified horizontal sampling point, so that parameters A and B are determined, and m is the number of longitudinal sampling points. The gain function is also of the form (25), where B takes the negative of its exponentially fitting corresponding value, and a=1.

g(x)＝A*exp(B*m) (25)

And (3) performing offset imaging and grounding grid detection on the entropy constraint speed field.

Aiming at the embedded posture of the tension rod, a rapid migration imaging method is researched, the characteristic extraction of the underground target is carried out, the embedded posture and the geometric information of the underground target are estimated, and the method is very important for the detection and maintenance of a power transmission line; and (3) researching an RPCA nonexplosive drug characteristic extraction method based on F-K offset imaging aiming at an underground target made of low metal or plastic materials.

Based on the coupling relation between the echo characteristics and the geometric shape and the gesture of the echo characteristics, the high-precision detection of the geometric shape and the gesture of the tension rod is realized, but the horizontal and transverse resolution of GPR radar detection is not high. According to the fluctuation theory, the main reflection energy of the target comes from the first Fresnel zone, and the size of the reflection echo of the underground target is larger than the actual size due to the occurrence of the Fresnel zone, so that the boundary of the underground target is blurred, and the resolution in the horizontal direction and the imaging precision of the underground target are reduced. The first fresnel zone radius approximately characterizes the horizontal resolution, the larger the first fresnel zone radius, the lower the horizontal resolution of identifying adjacent targets. High resolution imaging is a key basis for subsurface target pose estimation. In order to improve the detection accuracy of the underground target and realize the extraction of geometric shapes, the first Fresnel radius must be reduced, and the resolution in the horizontal direction is improved.

The first fresnel zone is horizontal to GPR resolution. The region of the ground penetrating radar pulse traveling downwards is a cone as shown in fig. 5, and the limited size of this region affects both the vertical resolution and the horizontal resolution. Wherein the vertical resolution is approximately c/2 Deltaf (epsilon) _r ) ^1/2 =λ/2, where c is the radar wave velocity in air, ε _r Is the dielectric constant of the formation, lambda is the wavelength.

The radius of the first Fresnel zone is r= [ v/2 ]][t/f] ^0.5 Is frequency; f is the bi-directional propagation time; v is the wave velocity in the medium. Accurate speed estimation is seen as a key issue.

The invention is based on the ground penetrating radar wave speed estimation principle of the minimum principle of the offset image entropy, the invention researches the F-K offset algorithm of the minimum speed estimation constraint of the entropy, and compares and analyzes the F-K offset with the kirchhoff offset and split-step Fourier (SSF) offset method; the method has the advantages that the local GPR pulse propagation speed of the shallow surface target area under various buried postures is rapidly and effectively estimated, underground target pickup and high-resolution imaging under the background noise of strong coupling in the direction of the survey line are realized, the boundary of the horizontal direction of the non-explosive is clear, the resolution of the geometric form detection of the underground target is improved, the interpretation difficulty is reduced, the buried postures, the size and the buried depth of the shallow surface target are determined, and the detection efficiency of the underground grounding grid and the tension rod is further improved.

The invention also provides an electromagnetic wave speed estimation method based on the entropy constraint of the F-K offset image information, and local constant speed field estimation is realized. When the detected target is shallow (0-25 cm), the local subsurface can be assumed to be a uniform medium, and the electromagnetic wave velocity can be regarded as approximately a local constant velocity. The electromagnetic wave velocity in the underground medium determines the F-K offset imaging quality, and the F-K offset algorithm is quite sensitive to small changes of velocity. Therefore, the invention adopts a method of scanning the wave velocities one by one on the sequence frequency-wave number domain offset image, and selects the wave velocity with the optimal imaging quality as the optimal wave velocity.

The focus quality of offset imaging is judged by using an image entropy value, the larger the entropy value is, the larger the offset calculation error is, the smaller the entropy value is, the clearer the offset image is, the higher the resolution in the horizontal direction is, and the smaller the error is. The main purpose of GPR offset imaging is to generate an image with significant peaks at the target location and no significant structure in non-target areas, which has a small entropy and can accurately identify the detected target. The method uses the maximum norm of the variance as the approximation of the entropy, has the advantages of low calculation complexity and good stability, and can avoid the complex situation of calculating zero logarithm. The entropy of the B-SCAN radar image is defined as:

Where S represents the number of sampling points and T represents the number of sampling channels.

When the entropy in the sequence offset (focusing) image is minimum, the estimated speed of the F-K offset image is the optimal offset speed, and the estimated speed is closest to the real speed of the electromagnetic wave propagating in the underground medium, so that the energy focusing performance of the generated offset image is the best. When the underground target detection is carried out in the field, the propagation speed of electromagnetic waves in an underground soil medium needs to be estimated rapidly and effectively, and an underground section view of the non-explosive agent along the direction of the survey line is formed by using a proper offset imaging method based on the speed, so that the shape, the size, the placement posture, the burial depth and other information of the non-explosive agent are estimated. The algorithm speed is considered, compared with offset methods such as SSF, kirchhoff and the like, the F-K offset method is the fastest in speed, and the offset precision and accuracy can meet the requirements of non-explosive detection. F-K local constant velocity offset imaging algorithm based on offset image entropy constraint is provided.

The main algorithm flow is that a group of speeds with corresponding sizes are set according to the approximate estimated dielectric constant in a medium where an underground target is located, a certain speed interval is taken, then F-K offset imaging is carried out on B-scan images filtered by clutter, sorting is carried out from large to small according to the entropy of offset images corresponding to each speed, speed output corresponding to the minimum entropy is taken, F-K offset is carried out, an offset section is output, and further judgment and analysis of the shape, the size and the like of the underground target are carried out.

As shown in fig. 8 to 17, the present invention also performed a number of experiments to verify the proposed method.

And (5) extracting experiments and analysis of the gain combination-RPCA underground target characteristic waveform.

As shown in fig. 8 to 10, in the case that the target signal is completely submerged in the surface strong reflection signal, after the values of the parameters ρ and λ in the RPCA are adjusted, the decoupled target echo data E and the clutter and noise data a are observed, so that the surface strong echo data with the depth ranging from 3.07 to 7.11 can be obviously and effectively inhibited, the hyperbolic characteristic of the target signal in the sparse matrix E can be effectively reserved, the lower surface echo data (the lower surface target echo with the depth of 14 cm) of the geometric model of the tension rod can also be effectively reserved, the signal amplitude is increased by exponential gain, and the geometric shape B-scan of the target is restored by the offset processing. When serious aliasing is not generated, RPCA can effectively inhibit the surface standing wave and other clutter components, but the extraction effect of lower surface echo signals (the signal depth is 17 cm) at deeper positions is poor, and the repeated reflection wave residues of an upper interface cause interference to the identification of lower interface echo signals of a target.

F-K offset algorithm experiment and analysis based on offset image entropy constraint.

(1) Ground net simulation model under uniform medium

And carrying out a round steel ground grid B-SCAN radar imaging simulation experiment by using a matGPR ground penetrating radar simulation platform, and verifying the effectiveness and feasibility of the algorithm.

And respectively taking round steel with phi of 20mm and flat steel with 60mm x 8mm as grounding grid materials to establish a simulation model. The buried depth of the grounding grid is set to be 0.75m, three horizontal supporting pieces are transversely arranged, the interval is 1m, the dielectric constant of the soil environment is 8, the resistivity is 1000 Ω & m, the emission frequency of the ground penetrating radar is 2130MHz, and the size of the region model is 4m multiplied by 2m. The movement of the antenna is stepped by 0.052m and the sampling time interval 5.217e-2ns, as in fig. 8 (a), 9 (a), 10 (a), 11 (a).

The simulation model is built on a matGPR platform based on matlab to obtain a longitudinal plane view and a target echo view of the ground net, as shown in fig. 8 (b), 9 (b), 10 (b) and 11 (b), respectively. The blue marked circular or rectangular targets in the simulation model are longitudinal planing surfaces of the horizontal/vertical support members, and are transversely distributed in soil at certain intervals. In the target echo diagram, as the medium of the simulation model is a single soil component, the echo only has the reflected signal from the target. F-K offset algorithm based on entropy constraint is respectively carried out on the simulation models, and an imaging result after offset is obtained.

Sand pool experiments and data analysis.

In order to verify the effectiveness of finding the position and shape of the target by using the offset imaging method under the condition that the target signal and the surface direct wave are seriously aliased. The transmitted signal is a first-order Gaussian pulse with the center frequency of 2.23GHz, and the GPR antenna is kept at a certain height on the surface of a sand pit and scans along a track from left to right. The medium carrying the target is sandy soil, and the surface relief height varies within 2 cm.

The experimental models in fig. 15 (a), fig. 16 (a) and fig. 17 (a) are subjected to sand pool experiments by using an autonomously built handheld radar device and a test sand pool (as shown in fig. 7), the radar device scans a detection area along the front side surface of the sand pool to obtain radar echo data as shown in fig. 15 (b), fig. 16 (b) and fig. 17 (b), and the serious interference of ground direct waves and soil non-target signals can be found. The results of the filtering processing of the radar echo data are shown in fig. 15 (c), 16 (c) and 17 (c), the wave velocity is scanned one by one within the interval of 0-0.30m/ns, the scanning interval is taken to be 0.01m/ns, the entropy curves of the F-K offset imaging results are shown in fig. 15 (d), 16 (d) and 17 (d), the local constant velocity in each case is estimated, and the final offset imaging results are shown in fig. 15 (e), 16 (e) and 17 (e) by using the local constant velocity.

Test experiment of the rod sand pool:

as shown in fig. 12, the pull rod is inserted vertically:

for the case of double targets, the offset targets are located at coordinates (0.5067,0.1638) and (0.4999,0.3222), the zero point coordinates at this time are the initial positions of antenna scanning, the actual embedded depths of the two targets should be 0.16m and 0.32m, and the heights of the antenna from the ground are 0.005m, so that the errors of the offset depths are-0.12 cm and-0.28 cm respectively.

The effect of shifting the target nearer to the ground is remarkable, while the effect of shifting the target farther from the ground is weaker because the soil is not a lossless medium, and the signal strength of the electromagnetic wave signal when it propagates underground decreases with the increase of the propagation depth.

The invention provides a GPR radar system integration for ground network construction detection, which integrates earlier-stage researches on signal propagation characteristics and rules of radar signals under different soil mediums, parameter estimation of a ground network topological structure and geometric forms and a high-resolution imaging algorithm, combines detection methods such as an instantaneous electromagnetic method, resistance measurement and the like, researches a set of ground network diagnosis system based on signal receiving-target extraction-two-dimensional imaging-breakpoint identification of a ground penetrating radar, and realizes multi-dimensional method evaluation of the ground network; combining the numerical simulation of a matGPR, GPRMax simulation platform, a laboratory semi-physical model and ground network detection real data, a set of analysis software for detecting the integrity of the ground network and identifying fault points is researched, and a ground network detection system combining hardware detection and data management software is realized.

Aiming at the non-excavation detection of the ultra-wideband pulse GPR of the grounding grid, the invention provides offset imaging and attitude estimation based on an entropy constraint speed field, so that the propagation speed of pulse electromagnetic waves of the GPR of the shallow earth surface is rapidly and effectively estimated, high-resolution refocusing imaging of an underground target under the ground strong-coupling direct wave background is realized, information such as the buried position, the shape and the size of an underground grounding grid assembly is determined, the estimation error of depth information is displayed within 2cm as a result, the boundary of the target in the horizontal direction is clear, and the horizontal resolution is improved. The method can identify echo signals of various postures of the underground target, has universality, can rapidly obtain high-precision offset images, and further provides technical support for detecting the integrity of the grounding grid for detection personnel, and has important significance for nondestructive detection of the integrity of the grounding grid.

The invention also provides a radar signal-based trenchless bar corrosion detection device, which comprises a signal acquisition module, a signal extraction module, a waveform construction module, a B-SCAN image construction module, a C-SCAN three-dimensional image construction module and a fault identification module.

Specifically, the signal acquisition module is used for acquiring a target echo signal of the grounding grid; the signal extraction module is used for extracting the waveform characteristics of the ground grid scattering echo in the ground grid target echo signals by utilizing a matching tracking MP and an orthogonal matching tracking method to form a ground grid echo signal fingerprint library; the waveform construction module is used for constructing a B-SCAN image construction module of a grounding grid A-SCAN characteristic waveform library under different soil environments, materials and burial depths according to the grounding grid echo signal fingerprint library, and is used for carrying out two-dimensional imaging of the grounding grid according to the grounding grid A-SCAN characteristic waveform library by adopting an offset imaging method to obtain a focused B-SCAN image; the C-SCAN three-dimensional image construction module is used for constructing a grounding grid C-SCAN three-dimensional image according to the focused B-SCAN image; the fault identification module is used for intercepting the horizontal section of the B-SCAN image and identifying fault points of the pull rod in the grounding grid according to the structure of the horizontal section.

The modules in the trenchless bar corrosion detection device based on radar signals can be all or partially realized by software, hardware and the combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory, the processor executing the computer program to implement the steps in the embodiments of the method for detecting the trenchless pulling rod corrosion based on radar signals. The specific implementation method may refer to a method embodiment, and will not be described herein.

It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The method provided by the invention is mainly used in the construction stages of the incomplete construction of the transformer substation and the grounding grid, and the quality attribute of the target scattering signal detected by the ground penetrating radar GPR is estimated through the analysis of the frequency/phase/amplitude characteristic between the target scattering signal and the reference signal, so that fault points such as grounding grid fracture, welding leakage, outer edge disconnection and the like in the construction process and corrosion points generated by the influence of the environment on the grounding grid are found.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It should be noted that the above-described embodiments will enable those skilled in the art to more fully understand the invention, but do not limit it in any way. Thus, although the present invention has been described in detail with reference to the present specification and examples, it should be understood by those skilled in the art that the present invention may be modified or equivalents; all technical schemes and improvements which do not depart from the spirit and scope of the invention are covered by the protection scope of the invention. Any reference sign in a claim should not be construed as limiting the claim concerned.

The above embodiments are merely preferred embodiments of the present invention, the protection scope of the present invention is not limited thereto, and any simple changes or equivalent substitutions of technical solutions that can be obviously obtained by those skilled in the art within the technical scope of the present invention disclosed in the present invention belong to the protection scope of the present invention.

Claims

1. The trenchless pull rod corrosion detection method based on radar signals is characterized by comprising the following steps of:

collecting a grounding grid target echo signal;

2. The radar signal-based trenchless radars corrosion detection method of claim 1, wherein a UWB pulse radar hardware system is employed to collect ground network target echo signals, the UWB pulse radar hardware system comprising a set of transmit receive antennas, radar modules, and I/0 modules.

3. The radar signal-based trenchless bar corrosion detection method according to claim 2, wherein when the UWB pulse radar hardware system is used to collect the target echo signal of the ground net, the relation between the detection time and the electromagnetic wave speed is as follows:

4. The radar signal-based trenchless dragline corrosion detection method of claim 1 further comprising removing background noise in the ground plane target echo signal by a gain joint-RPCA method prior to extracting the ground plane scattered echo waveform characteristics in the ground plane target echo signal.

5. The radar signal-based trenchless dragline corrosion detection method of claim 3 further comprising data enhancing the noise-removed ground net target echo signal using a resolution enhancement method after removing background noise from the ground net target echo signal.

6. The radar signal-based trenchless bar corrosion detection method according to claim 1, wherein the extracting the characteristics of the ground grid scattering echo waveform in the ground grid target echo signal by using the matching pursuit MP and the orthogonal matching pursuit method, and forming the ground grid echo signal fingerprint library specifically comprises: and extracting characteristic waveforms based on the signal subspace, extracting the characteristic of the echo waveform of the ground network scattering by utilizing a matching pursuit MP and orthogonal matching pursuit method, and integrating the ground network echo signal fingerprint library.

7. The radar signal-based trenchless bar corrosion detection method of claim 3, wherein the adopting of the offset imaging method performs two-dimensional imaging of the ground grid according to the ground grid a-SCAN signature library to obtain a focused B-SCAN image, specifically comprising the steps of:

Wherein v is constant in a homogeneous medium;

the offset profile e (x, z, t=0) is:

8. The radar signal-based trenchless rod erosion detection method of claim 7 wherein the entropy of the image after the echo image signal completes the offset imaging is defined as:

9. The utility model provides a no excavation pull rod corrosion detection device based on radar signal which characterized in that includes:

10. A computer device comprising a memory, a processor and a computer program stored on the memory, characterized in that the processor executes the computer program to carry out the steps of the method according to any one of claims 1 to 8.