A method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar

By using ground-penetrating radar detection method, a database of electromagnetic parameters of the medium is established, the dielectric constant of the medium is corrected in real time, and the weighted least squares method is used to fit the vertex of the reflected wave, realizing rapid, non-destructive, and high-precision detection of the burial depth of cement foundations for transportation facilities, thus solving the detection problems in existing technologies.

CN123111151APending Publication Date: 2026-11-17HUAIAN HIGHWAY DEV CENT +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611237226.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-14
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing ground-penetrating radar (GPR) detection methods for detecting the burial depth of cement foundations in transportation facilities suffer from problems such as insufficient parameter adaptability, deviation in medium characteristics, insufficient positioning accuracy of reflected waves, discontinuous detection results, and significant impact from systematic errors, failing to meet the requirements for rapid, non-destructive, and high-precision detection.

Method used

The detection method based on ground penetrating radar is adopted. By establishing a database of electromagnetic parameters of the medium, matching the antenna element with the center frequency, configuring the scanning parameters, dynamically correcting the dielectric constant of the medium in real time, fitting the vertex of the reflected wave using the weighted least squares method, supporting dual-channel control unit and flexible scanning mode, performing system calibration and result reliability verification, removing interference signals, and generating a burial depth distribution map.

Benefits of technology

It improves the quality of echo signals, enhances the accuracy of electromagnetic wave propagation speed calculation, ensures the continuity and integrity of detection results, reduces the error in burial depth calculation, and provides accurate information on the location and depth of cement foundations, thus providing reliable data for the construction, acceptance, and maintenance of traffic facilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111151A_ABST
    Figure CN123111151A_ABST
Patent Text Reader

Abstract

This invention discloses a method for detecting the burial depth of cement foundations for traffic facilities based on ground-penetrating radar (GPR). It relates to the field of GPR detection technology and employs a general-purpose pulsed GPR system to collect shallow medium foundation data of the detection area and match antenna elements with corresponding center frequencies. After configuring scanning parameters and combining an odometer, continuous non-contact scanning is performed in a mobile mode. The raw echo data is preprocessed, and the electromagnetic wave propagation speed is corrected based on the difference in dielectric constant between cement and the medium. The actual burial depth of the cement foundation is calculated by extracting the vertex features of the reflected wave using a hyperbolic fitting algorithm. This invention achieves non-destructive and rapid detection without interfering with normal traffic operation. It has strong parameter adaptability, can dynamically correct the dielectric constant, improve the accuracy of burial depth calculation, and provides continuous and reliable detection results. It is applicable to the detection of cement foundations for various types of traffic facilities.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of ground-penetrating radar (GPR) detection technology, and in particular to a method for detecting the burial depth of cement foundations for transportation facilities based on GPR. Background Technology

[0002] As the core supporting structure of various transportation facilities, the burial depth of cement foundations directly affects the overall structural stability and long-term service performance of these facilities. Accurately obtaining the burial depth information of cement foundations is a necessary prerequisite for subsequent work in the construction quality acceptance, periodic maintenance inspection, and safety risk assessment of transportation facilities. Traditional cement foundation burial depth detection mainly relies on the excavation method, which involves manually or mechanically excavating the surface cover layer of the inspection area to directly measure the burial depth of the cement foundation. This method can cause irreversible damage to the road surface, roadbed, and other transportation facility structures, requiring additional repair work after the inspection. Furthermore, the inspection process is time-consuming and often necessitates the closure of part or all of the traffic sections, significantly disrupting normal traffic flow in the area. In addition, the excavation method is a single-point sampling inspection, only obtaining burial depth data at discrete points, failing to comprehensively cover the entire inspection area and thus unable to fully reflect the continuous distribution of cement foundations, potentially leading to missed safety hazards.

[0003] Ground-penetrating radar (GPR), a non-destructive testing technology based on the principle of electromagnetic wave reflection, is increasingly being applied in the field of underground structure inspection due to its non-contact and rapid scanning characteristics. This technology transmits broadband electromagnetic pulses underground and receives reflected echoes generated at the interfaces of different media. After data processing, it can visually present the distribution of underground structures. However, existing GPR detection methods still have many technical limitations when applied to the detection of the burial depth of cement foundations for transportation facilities. Most existing detection methods use standardized antenna elements and fixed scanning parameter configurations, failing to fully consider the differences in the shallow media types and water content of the detection area. The electromagnetic properties of the medium directly affect the propagation attenuation and reflection effect of electromagnetic waves. Insufficient parameter adaptability leads to a decrease in the signal-to-noise ratio of the echo signal, making it impossible to clearly identify the reflected signals at the interface between the cement foundation and the surrounding medium.

[0004] In the burial depth calculation stage, existing methods generally rely on fixed dielectric constant values ​​obtained from offline laboratory measurements. However, the moisture content of the soil in the field changes dynamically due to factors such as climate and geological conditions. This discrepancy between the fixed dielectric constant and the actual characteristics of the soil leads to inaccurate calculations of electromagnetic wave propagation speed, ultimately resulting in significant errors in the burial depth calculation. Although soil physics has established models such as the Topp equation (a cubic polynomial) and the CRIM mixed media model to describe the relationship between soil moisture content and dielectric constant, these general models are not specifically calibrated for compacted sandy, loamy, and clayey media commonly found in roadbeds for transportation facilities. Furthermore, they cannot achieve online real-time correction based on radar echo signals, making it difficult to meet the needs of rapid and high-precision on-site detection.

[0005] Existing ground-penetrating radar (GPR) detection methods typically apply the same weight to all echo points during reflected wave feature extraction and fitting, failing to consider the importance of data near the reflected wave vertex for positioning accuracy, resulting in insufficient positioning accuracy at the reflected wave vertex. Some detection methods lack a complete on-site system calibration process, failing to eliminate the impact of inherent equipment errors on the detection results. For complex detection scenarios with obstacles, existing methods struggle to seamlessly stitch together segmented scan data, affecting the continuity and completeness of the detection results. These technical issues limit the application effectiveness of GPR technology in detecting the burial depth of concrete foundations for transportation facilities, failing to meet the practical needs of the transportation infrastructure sector for rapid, non-destructive, and high-precision detection work. Summary of the Invention

[0006] This invention proposes a method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar, in order to solve the problems mentioned in the prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar, comprising: A pulse ground-penetrating radar system was used as the detection equipment. This system is an active radar device that uses the reflection of radio waves to measure distance. The ground-penetrating radar system includes a single-channel or dual-channel control unit, an integrated transceiver antenna unit adapted to multiple center frequencies, a wheeled odometer, a power supply unit, and a portable computer with data acquisition and processing software installed. The control unit is connected to the portable computer via an Ethernet interface. The control unit has a built-in voltage decoder that can provide 15V to 22V DC power to the portable computer. The antenna unit transmits pulses with a duration of 1ns to 10ns and is connected to the control unit via a dedicated AB cable. Fiber optic antenna units require the use of an optical converter. The detection method includes the following steps: collecting basic data of the surface and shallow media in the detection area; determining the type of overburden medium, the volumetric water content of the medium, and the depth of the groundwater level in the detection area through on-site sampling and laboratory analysis; recording the surface flatness and surrounding electromagnetic environment information of the detection area; establishing a database of electromagnetic parameters of the medium in the detection area; and matching antenna units with corresponding center frequencies according to the database. Configure the ground-penetrating radar scanning parameters through data acquisition and processing software, connect the odometer and control unit through a dedicated cable, enable the mobile scanning mode in the software, calibrate the wheel diameter parameters and pulse coefficient of the odometer, perform continuous non-contact scanning, use an antenna bracket with shock absorption device to cope with uneven ground, and record the scanning trajectory and corresponding radar echo data at the same time. The raw radar echo data was preprocessed to remove random noise, direct wave interference and background reflection signals in sequence, and the effective reflected wave signal at the interface between the cement foundation and the surrounding medium was extracted. Based on the electromagnetic wave reflection characteristics caused by the difference in dielectric constant between cement and the covering layer medium, a preset dielectric property table is used to determine the initial propagation speed of electromagnetic waves in the covering layer medium of the detection area. The dielectric property table includes the relative dielectric constant, attenuation coefficient and propagation speed parameters of different types of media at different moisture contents. The dielectric constant of the medium is dynamically corrected in real time. The real-time volumetric water content of the medium is calculated based on the amplitude attenuation characteristics of the radar echo signal collected during the detection process. The real-time relative dielectric constant is obtained through the correlation model between the medium humidity and the dielectric constant, and then the electromagnetic wave propagation speed is corrected. The hyperbola fitting algorithm is used to extract the vertex features of the effective reflected wave, determine the two-way propagation time corresponding to the vertex of the reflected wave, and calculate the actual burial depth of the cement foundation by combining the corrected electromagnetic wave propagation speed, and generate an inspection report containing the horizontal position and burial depth information of the cement foundation.

[0008] Furthermore, this includes conducting on-site system calibration before the formal scanning. At least three cement standard piles with known burial depths are selected as calibration points in the testing area. The calibration points are evenly distributed at locations with different media characteristics in the testing area. Each calibration point is scanned 2 to 5 times using the same scanning parameters as the formal testing. The average value of the multiple test results is taken as the test burial depth of the calibration point. The deviation between the test burial depth of the calibration point and the actual burial depth is calculated. The system correction coefficient is obtained based on the linear correction method. This correction coefficient is used to correct the system error of all subsequent test data.

[0009] Furthermore, the correlation model between the medium humidity and the dielectric constant is a quadratic polynomial: ;in The real-time relative permittivity of the medium; The water content by volume is expressed as a decimal. , which is the dimensionless fitting coefficient corresponding to the type of overburden medium.

[0010] Furthermore, in the antenna unit matching step, the intensity of environmental electromagnetic interference is detected by the signal monitoring function of the data acquisition and processing software. When the background noise amplitude exceeds 20% to 40% of the effective signal amplitude, it is determined that there is strong electromagnetic interference. At this time, an enhanced antenna unit with an electromagnetic shielding shell is selected, and the stack number is set to 32 to 64 to improve the signal-to-noise ratio. The scanning speed is reduced to 0.3m / s to 0.8m / s, and wireless communication equipment and high-voltage electrical equipment in the detection area are turned off.

[0011] Furthermore, in the scanning parameter configuration step, when a dual-channel control unit is used, two antenna units with different center frequencies are connected simultaneously to collect shallow and deep radar echo data respectively. The shallow antenna unit is used to detect cement foundations with a burial depth of less than 2 meters, and the deep antenna unit is used to detect cement foundations with a burial depth of more than 2 meters. The two sets of data are fused based on the location information recorded by the odometer using a time synchronization matching method.

[0012] Furthermore, in the mobile scanning mode step, when there are obstacles in the detection route that prevent continuous scanning, the mode is switched to step scanning mode. A scanning start point and end point are set on both sides of the obstacle, and a radar trace is collected every 20mm to 100mm. After scanning both sides of the obstacle is completed, the overlapping area matching method is used to select no less than 10 overlapping traces in the overlapping area of ​​the two segments of scanning data. Seamless stitching is achieved by adjusting the time reference and position reference of the two segments of data.

[0013] Furthermore, in the process of extracting the vertex features of the reflected wave using the hyperbola fitting algorithm, a weighted least squares method is used to fit the hyperbola of the reflected wave. Higher weight values ​​are assigned to echo points closer to the vertex to improve vertex positioning accuracy. "Closer to the vertex" refers to a range extending no more than five radar tracks to either side of the track where the reflected wave vertex is located. Echo points within this range have a weight coefficient of 2, while the weight coefficients for the remaining echo points are set to 1. The objective function of the weighted least squares method is: ;in The objective function value; This represents the total number of echo points participating in the fitting process. For the first Weighting coefficients for each echo point; For the first The two-way propagation time corresponding to each echo point; The two-way propagation time is the time corresponding to the vertex of the reflected wave. For the first The horizontal distance of each echo point relative to the vertex; This represents the corrected propagation speed of electromagnetic waves in the medium.

[0014] Furthermore, in the radar echo data preprocessing step, firstly, a mean filtering algorithm with a 3×3 to 5×5 window is used to remove random noise. Then, 5 to 15 sampling points in front of the radar profile are selected as direct wave templates, and a direct wave cancellation algorithm is used to remove direct wave interference. Next, a Wiener deconvolution algorithm with an operator length of 16 to 64 points is used to compress the radar pulse width. Finally, a sliding window background removal algorithm with a window length of 30 to 100 traces is used to eliminate background reflection signals generated by the underground homogeneous medium.

[0015] Furthermore, in the cement foundation burial depth calculation step, when multiple cement foundation reflected waves are detected, a hyperbola is fitted to each reflected wave to obtain the two-way propagation time of the vertex corresponding to each reflected wave. Combined with the corrected electromagnetic wave propagation speed, the burial depth of each cement foundation is calculated. At the same time, the horizontal position information of each cement foundation is recorded, and a burial depth distribution profile is generated with the horizontal distance as the abscissa and the burial depth as the ordinate. Different colored blocks are used to represent different burial depth ranges, and the cement foundation number and parameters are marked.

[0016] Furthermore, the detection method also includes verifying the reliability of the detection results. At least 10% of the detection points are randomly selected for excavation verification. The excavation depth is 1.1 to 1.5 times greater than the radar detection depth to fully expose the bottom of the cement foundation. The actual burial depth is measured with a steel tape measure, and the relative error between the actual burial depth of the excavation verification and the radar detection depth is calculated. When the relative error is greater than 5%, the scanning parameters are readjusted and a second scan is performed on the area. If the relative error of the second scan is still greater than 5%, an antenna unit with a higher center frequency is replaced and the detection is repeated.

[0017] Compared with existing technologies, the beneficial effects of this invention are: This invention collects basic data of the shallow medium in the detection area and establishes a corresponding electromagnetic parameter database. It matches the antenna unit with the corresponding center frequency according to the actual characteristics of the medium and configures the scanning parameters accordingly. This can effectively improve the quality of the echo signal, adapt to the detection needs of different medium conditions and different burial depths, and ensure that the reflected signal of the cement foundation can be effectively obtained in various complex field environments.

[0018] This invention dynamically corrects the dielectric constant of the medium in real time. Based on the characteristics of the echo signal obtained during on-site testing, the dielectric constant value is adjusted to better reflect the actual electromagnetic properties of the medium, improving the accuracy of electromagnetic wave propagation speed calculations and thus enhancing the precision of cement foundation burial depth calculations. By using the weighted least squares method to fit the hyperbola of the reflected wave and differentiating the weight values ​​of the echo points, the position of the reflected wave peak can be located more accurately, further reducing the error in burial depth calculations.

[0019] This invention supports the simultaneous connection of two antenna units with different center frequencies to a dual-channel control unit. By fusing shallow and deep radar echo data through time synchronization matching technology, it can meet both detection depth and resolution requirements, achieving full-depth coverage detection. When obstacles exist along the detection route, preventing continuous scanning, it can flexibly switch scanning modes and seamlessly stitch segmented data through overlapping area matching technology, ensuring the continuity and integrity of the detection results.

[0020] This invention incorporates an on-site system calibration step before formal scanning, effectively eliminating inherent equipment errors and improving the consistency of results across different testing batches. A comprehensive reliability verification process for test results allows for timely identification and correction of problems during the testing process, ensuring the reliability of the results. A complete data preprocessing workflow effectively removes various interference signals such as random noise and direct waves, extracting clear and effective reflected wave signals from the cement foundation. The generated test report visually presents the horizontal position and burial depth distribution of the cement foundation, providing accurate data for the construction acceptance, daily maintenance, and safety assessment of transportation facilities. Attached Figure Description

[0021] Figure 1 This is a schematic block diagram of the overall process for detecting the burial depth of cement foundations based on ground penetrating radar proposed in this invention. Figure 2 This is a schematic block diagram of the radar echo data preprocessing flowchart proposed in this invention; Figure 3 This is a flowchart of the antenna matching and scanning parameter configuration proposed in this invention; Figure 4 This is a flowchart of the obstacle scanning and seamless data stitching process proposed in this invention; Figure 5 This is a flowchart of the system calibration and result reliability verification proposed in this invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Reference Figures 1 to 5 A method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar, comprising: The OKO-3 general-purpose pulse ground-penetrating radar system was used as the detection equipment. This system is an active radar device that uses radio wave reflection for distance measurement. The OKO-3 ground-penetrating radar system includes a single-channel or dual-channel control unit, an integrated transceiver antenna unit adapted to multiple center frequencies, a DP-32U wheeled odometer, a BP-9 / 12 nickel-metal hydride power supply unit, and a portable computer equipped with GeoScan32 data processing software. The control unit connects to the portable computer via an Ethernet interface, supporting both wired and wireless data transmission. The control unit has a built-in voltage decoder that can provide 15V to 22V DC power to the portable computer, with a power output of up to 60W. The antenna unit transmits pulses with a duration of 1 ns to 10 ns, corresponding to a center frequency of 100 MHz to 1700 MHz. It connects to the control unit via a dedicated AB cable. Fiber optic antenna units require the use of an optical converter. The detection method includes the following steps: collecting basic data of the surface and shallow media in the detection area; determining the type of overburden medium, the volumetric water content of the medium, and the depth of the groundwater level in the detection area through on-site sampling and laboratory analysis; recording the surface flatness and surrounding electromagnetic environment information of the detection area; and establishing a database of electromagnetic parameters of the medium in the detection area, including the relative permittivity of the medium and the electromagnetic wave attenuation coefficient. According to the electromagnetic parameter database of the medium in the detection area, the antenna unit with the corresponding center frequency is matched. When the volume moisture content of the medium in the detection area is less than 10% and the designed burial depth of the cement foundation is less than 2 meters, the AB-700M3 antenna unit is selected. When the volume moisture content of the medium in the detection area is between 10% and 20% and the designed burial depth of the cement foundation is between 2 meters and 6 meters, the AB-400M3 antenna unit is selected. When the volume moisture content of the medium in the detection area is higher than 20% and the designed burial depth of the cement foundation is greater than 6 meters, the AB-250M3 antenna unit is selected. The GeoScan32 software was used to configure the ground-penetrating radar scanning parameters. The sample size was set from 512 to 2048 points, the stack number from 1 to 64, the time scale from 10 ns to 200 ns, the trace interval from 50 mm to 200 mm, and the gain factor from 1 to 100 with a linear gain profile. The software has multiple built-in color palettes to enhance the contrast of the radar image and can adjust the parameters in real time during the scanning process to optimize the echo signal display effect. Connect the DP-32U odometer and control unit via a dedicated cable, enable the moving scan mode in the GeoScan32 software, calibrate the wheel diameter parameters and pulse coefficient of the odometer, control the antenna unit to move along the preset detection route at a uniform speed of 0.5m / s to 1.5m / s to perform continuous non-contact scanning, keep the antenna unit parallel to the ground and the vertical distance less than 5cm during the scanning process, use an antenna bracket with shock absorption device to deal with uneven ground, and record the scanning trajectory and corresponding radar echo data at the same time. The raw radar echo data was preprocessed to remove random noise, direct wave interference and background reflection signals in sequence, and the effective reflected wave signal at the interface between the cement foundation and the surrounding medium was extracted. The preprocessed data was stored in GeoScan32 standard format for subsequent analysis. Based on the electromagnetic wave reflection characteristics caused by the difference in dielectric constant between cement and the covering layer medium, a preset dielectric property table is used to determine the initial propagation speed of electromagnetic waves in the covering layer medium of the detection area. The dielectric property table includes the relative dielectric constant, attenuation coefficient and propagation speed parameters of different types of media at different moisture contents. The dielectric constant of the medium is dynamically corrected in real time. The real-time volumetric water content of the medium is calculated based on the amplitude attenuation characteristics of the radar echo signal collected during the detection process. The real-time relative dielectric constant is obtained through the correlation model between the medium humidity and the dielectric constant, and then the electromagnetic wave propagation speed is corrected. The hyperbola fitting algorithm is used to extract the vertex features of the effective reflected wave, determine the two-way propagation time corresponding to the vertex of the reflected wave, and calculate the actual burial depth of the cement foundation by combining the corrected electromagnetic wave propagation speed, and generate an inspection report containing the horizontal position and burial depth information of the cement foundation.

[0024] This invention also includes on-site system calibration before formal scanning. At least three cement standard piles with known burial depths are selected as calibration points in the detection area. The calibration points are evenly distributed at locations with different media characteristics in the detection area. Each calibration point is scanned three times using the same scanning parameters as the formal detection. The average value of the three detection results is taken as the detection burial depth of the calibration point. The deviation between the detection burial depth of the calibration point and the actual burial depth is calculated. The system correction coefficient is obtained based on the linear correction method. The correction coefficient is used to correct the system error of all subsequent detection data.

[0025] This invention also includes real-time dynamic correction of the dielectric constant of the medium. The real-time volumetric water content of the medium is calculated based on the amplitude attenuation characteristics of the radar echo signal collected during the detection process. The real-time relative dielectric constant is obtained through a quadratic polynomial correlation model between the medium's humidity and its dielectric constant. The correlation model is as follows: ;in is the real-time relative permittivity of the medium, which is dimensionless; The water content of the medium is dimensionless and expressed as a decimal (i.e., a percentage value divided by 100). The coefficients are dimensionless fitting coefficients corresponding to the types of overburden media. The fitting coefficients for sandy media are 120, 28, and 3.2, respectively; for loamy media, they are 250, 45, and 3.2, respectively; and for clayey media, they are 380, 62, and 2.4, respectively.

[0026] In this invention, during the antenna element matching step, the intensity of environmental electromagnetic interference is detected using the signal monitoring function of GeoScan32 software. When the background noise amplitude exceeds 30% of the effective signal amplitude, it is determined that there is strong electromagnetic interference. At this time, an enhanced antenna element with an electromagnetic shielding shell is selected, and the stack number is set to 32 to 64 to improve the signal-to-noise ratio. The scanning speed is reduced to 0.3m / s to 0.8m / s, and wireless communication equipment and high-voltage electrical equipment in the detection area are turned off to reduce external interference.

[0027] In this invention, during the scanning parameter configuration step, when a dual-channel control unit is used, two antenna units with different center frequencies are connected simultaneously to collect radar echo data from shallow and deep layers respectively. The shallow antenna unit is the AB-1000P3 type, used to detect cement foundations with a burial depth of less than 2 meters, and the deep antenna unit is the AB-250M3 type, used to detect cement foundations with a burial depth of more than 2 meters. The time synchronization matching method is used to fuse the two sets of data based on the location information recorded by the odometer, covering the detection requirements of the entire depth range.

[0028] In this invention, during the mobile scanning mode step, when there are obstacles in the detection route that prevent continuous scanning, the system switches to a step-by-step scanning mode. A scanning start point and an end point are set on both sides of the obstacle, and a radar trace is collected every 50mm. After scanning both sides of the obstacle, the overlapping area matching method is used to select no less than 10 overlapping traces in the overlapping area of ​​the two segments of scanning data. Seamless stitching is achieved by adjusting the time reference and position reference of the two segments of data.

[0029] In this invention, during the process of extracting the vertex features of the reflected wave using the hyperbola fitting algorithm, a weighted least squares method is used to fit the hyperbola of the reflected wave. Higher weight values ​​are assigned to echo points closer to the vertex to improve vertex localization accuracy. The objective function of the weighted least squares method is: ;in The objective function value, in units of ; The total number of echo points involved in the fitting, dimensionless; For the first The weighting coefficients for each echo point are dimensionless. The weighting coefficient for echo points near the vertex is set to 2, and the weighting coefficient for echo points far from the vertex is set to 1. For the first The two-way propagation time corresponding to each echo point is expressed in nanoseconds. The two-way propagation time corresponding to the vertex of the reflected wave is expressed in nanoseconds. For the first The horizontal distance of each echo point relative to the vertex is expressed in meters. The corrected propagation speed of electromagnetic waves in a medium is expressed in m / ns.

[0030] In this invention, the radar echo data preprocessing step first uses a 3×3 window mean filtering algorithm to remove random noise, then selects the first 10 sampling points of the radar profile as the direct wave template, uses a direct wave cancellation algorithm to remove direct wave interference, then uses a Wiener deconvolution algorithm with an operator length of 32 points to compress the radar pulse width to improve depth resolution, and finally uses a sliding window background removal algorithm with a window length of 50 traces to eliminate background reflection signals generated by the underground homogeneous medium.

[0031] In this invention, during the cement foundation burial depth calculation step, when multiple cement foundation reflected waves are detected, a hyperbola fit is performed on each reflected wave to obtain the two-way propagation time of the vertex corresponding to each reflected wave. Combined with the corrected electromagnetic wave propagation speed, the burial depth of each cement foundation is calculated. At the same time, the horizontal position information of each cement foundation is recorded, and a burial depth distribution profile is generated with horizontal distance as the abscissa and burial depth as the ordinate. Different colored blocks are used to represent different burial depth ranges, and the cement foundation number and parameters are labeled.

[0032] In this invention, the detection method further includes verifying the reliability of the detection results. At least 10% of the detection points are randomly selected for excavation verification. The excavation depth is greater than 1.2 times the radar detection depth to fully expose the bottom of the cement foundation. The actual burial depth is measured using a steel tape measure with an accuracy of 1 mm. The relative error between the actual burial depth of the excavation verification and the radar detection depth is calculated. When the relative error is greater than 5%, the scanning parameters are readjusted and a second scan is performed on the area. If the relative error of the second scan is still greater than 5%, an antenna unit with a higher center frequency is replaced and the detection is repeated.

[0033] The dielectric property table of this invention includes the electromagnetic parameters of three common media for roadbed construction in transportation facilities—sand, loam, and clay—at different volumetric moisture contents. Representative values ​​are shown in the table below: Table 1: Reference Table of Typical Dielectric Properties

[0034] Note: The above parameters are calibration results for compacted roadbeds used in traffic facilities. Adjustments can be made based on on-site sampling results during actual application. 2. Calibration method for the correlation model of dielectric humidity and dielectric constant The quadratic polynomial correlation model used in this invention is calibrated using the following method: Standard medium samples with different volumetric moisture contents were prepared, ranging from 0% to 30%. The relative permittivity of the medium at various moisture contents was determined using a vector network analyzer. The least squares method was used to perform a quadratic polynomial fitting on the measured data to obtain the fitting coefficients a, b, and c for the corresponding medium types. The fitting coefficients were corrected through on-site excavation verification to ensure that the model accuracy met the testing requirements.

[0035] The following two examples further illustrate the specific implementation of this system: This invention employs a general-purpose pulse ground-penetrating radar system to detect the burial depth of cement foundations for transportation facilities. This system mainly consists of a control unit, an integrated transceiver antenna unit, a wheeled odometer, a nickel-metal hydride power supply unit, and a portable computer with dedicated data processing software. The control unit establishes a communication connection with the portable computer via an Ethernet interface, supporting both wired and wireless data transmission modes. A built-in voltage conversion module provides a stable DC power supply to the portable computer. The antenna unit can emit broadband electromagnetic pulses with different center frequencies, with pulse durations ranging from 1 ns to 10 ns. It connects to the control unit via a dedicated signal cable. Fiber optic antenna units require the use of a photoelectric conversion device. The following two specific embodiments provide a detailed description of the technical solution of this invention. These two embodiments address different medium conditions and different burial depth ranges for detecting cement foundations for transportation facilities, fully covering all the technical content protected by the claims of this invention. Example 1

[0036] This embodiment focuses on the burial depth detection of the cement foundation of a corrugated beam guardrail on a secondary highway. The designed burial depth of the cement foundation in this section is 1.2 meters to 2.5 meters. The detection area is 2.0 kilometers long. The surface cover is mainly sandy medium, the terrain is relatively flat, and there are a few low-voltage transmission lines in the vicinity, so the electromagnetic environment is relatively simple.

[0037] Before commencing the testing, basic data on the surface and shallow subsurface media of the testing area were first collected. Sampling points were set every 50 meters along the testing route, and surface media samples within a depth of 0 to 1 meter were collected using the ring sampler method. Three parallel samples were collected from each sampling point. The collected samples were sealed and sent to the laboratory, where the volumetric moisture content of the media was determined using the drying method. The drying temperature was set to 105 degrees Celsius, and the drying time was no less than 24 hours, until the sample mass stabilized. The determination showed that the volumetric moisture content of the sandy media in the testing area ranged from 6% to 9%, and the groundwater level was greater than 5 meters deep, which had no significant impact on this testing. Based on the measured media type and moisture content, a pre-set dielectric property table was used to obtain the relative permittivity, electromagnetic wave attenuation coefficient, and propagation velocity parameters under the corresponding conditions, establishing a database of electromagnetic parameters for the media in the testing area.

[0038] Based on the electromagnetic parameter database of the medium, an antenna unit with the corresponding center frequency was matched. Since the volumetric moisture content of the medium in the detection area was less than 10% and the concrete foundation was buried at a depth of less than 2 meters, an integrated transceiver antenna unit with a center frequency of 700MHz was selected. The antenna unit was fixed on a hand-push detection bracket with shock absorption. The horizontality and height of the antenna unit were adjusted to ensure it was parallel to the ground, with the vertical distance controlled between 3 and 5 centimeters. The nickel-metal hydride power supply unit was placed in the storage box of the detection bracket and connected to the control unit and antenna unit via power cables to provide power to the entire system.

[0039] Before the formal scanning, on-site system calibration was performed. Three cement standard piles with known burial depths were selected as calibration points in the testing area, evenly distributed at the start, middle, and end points of the testing section. Each calibration point was scanned three times using the exact same scanning parameters as the formal test, and the average of the three test results was taken as the detected burial depth of the calibration point. The deviation between the detected burial depth and the actual burial depth of the calibration point was calculated, and a system correction coefficient was obtained based on the linear correction method. This correction coefficient was then used to correct the system error of all data from this test.

[0040] Turn on the power switches of the portable computer and control unit, and start the dedicated data processing software. The software automatically identifies the connected control unit and antenna unit. Enter the scan parameter configuration interface, set the sample size to 1024 points, the stack number to 16, the time scale to 50ns, the trace interval to 100mm, the gain factor to 40, and select the linear gain profile. The software has multiple built-in color palettes; in this test, a grayscale color palette was used to enhance the radar image contrast. During the scan, the gain factor can be finely adjusted in real time according to the display effect of the echo signal to ensure that the reflected signal from the concrete foundation is clearly presented.

[0041] Connect the wheel-mounted odometer to the odometer interface of the control unit via a dedicated cable and enable the moving scan mode in the software. Perform on-site calibration of the odometer by selecting a 10-meter-long flat road section in the test area, marking both ends of the road surface, and pushing the test bracket across this road section at a constant speed. Compare the distance traveled by the software with the actual marked distance. If there is a discrepancy, adjust the wheel diameter parameters and pulse coefficient in the odometer settings interface of the software, and repeat the calibration process until the distance measurement error is less than 1%.

[0042] After calibration, the detection bracket is moved along the preset detection route at a constant speed of 0.8 m / s to 1.2 m / s to perform continuous non-contact scanning. During the scanning process, the relative height of the antenna unit to the ground is kept stable. The moving speed is appropriately reduced when encountering uneven road surfaces to avoid collisions between the antenna unit and the ground. The software automatically records the scanning trajectory and corresponding radar echo data, which is stored in real-time on the portable computer's hard drive.

[0043] After scanning, the acquired raw radar echo data is imported into the data processing module for preprocessing. First, a 3×3 window mean filtering algorithm is used to remove random noise from the raw data, preserving the signal's edge features during the filtering process. Then, the first 10 sampling points of the radar profile are selected as the direct wave template, and a direct wave cancellation algorithm is used to remove direct wave interference, eliminating the obstruction of shallow reflection signals by the antenna's direct coupling signal. Next, a Wiener deconvolution algorithm with a 32-point operator length is used to compress the radar pulse width, improving the detection depth resolution and enabling clear differentiation between adjacent reflection interfaces. Finally, a sliding window background removal algorithm with a window length of 50 traces is used to eliminate background reflection signals generated by the homogeneous underground medium, highlighting the effective reflection waves from the concrete foundation. After preprocessing, the effective reflection wave signal at the interface between the concrete foundation and the surrounding medium is extracted, and the processed data is saved in the software's standard format.

[0044] Based on the electromagnetic wave reflection characteristics arising from the difference in dielectric constants between cement and sandy media, the dielectric constant of the media is dynamically corrected in real time. According to the amplitude attenuation characteristics of radar echo signals collected during the detection process, the real-time volumetric moisture content of the media at each detection location is calculated. The real-time relative dielectric constant is obtained through a correlation model between media humidity and dielectric constant, thereby correcting the propagation speed of electromagnetic waves in the media. A weighted least squares method is used to perform hyperbolic fitting on the extracted effective reflected waves. During the fitting process, higher weight values ​​are assigned to echo points closer to the vertex to accurately locate the position of the reflected wave vertex and determine the corresponding two-way propagation time. Combined with the corrected electromagnetic wave propagation speed, the actual burial depth of each cement foundation is calculated.

[0045] During the result generation phase, the software automatically identifies the reflected waves from all cement foundations, calculates their burial depth and horizontal position, and generates a burial depth distribution profile with horizontal distance as the x-axis and burial depth as the y-axis. Different grayscale blocks represent different burial depth ranges in the profile, and each cement foundation is labeled with its number and corresponding burial depth value. Finally, 10% of the detection points are randomly selected for excavation verification. The excavation depth is greater than 1.2 times the radar detection depth, fully exposing the bottom of the cement foundation. A steel tape measure with 1mm accuracy is used to measure the actual burial depth, and the difference between the radar detection results and the actual burial depth is compared to verify the reliability of the detection results.

[0046] Table 2. Echo signal-to-noise ratio under different scanning parameter combinations

[0047] Table 1 shows the data from the field measurements taken at the secondary highway testing site in Example 1, illustrating the signal-to-noise ratio (SNR) test results for different combinations of stack count and scanning speed. The data indicates that as the stack count increases and the scanning speed decreases, the SNR of the echo signal gradually improves, and the signal quality is enhanced. However, excessively high stack counts and excessively low scanning speeds significantly increase testing time and reduce testing efficiency. This example uses a parameter combination of 16 stack counts and a scanning speed of 1.0 m / s, which ensures sufficient SNR while maintaining testing efficiency, making the reflected signal from the cement foundation clearly discernible. This parameter selection balances testing efficiency and signal quality, is suitable for most testing scenarios with good medium conditions and weak electromagnetic interference, and can meet the routine requirements for detecting the burial depth of cement foundations in transportation facilities. Example 2

[0048] This embodiment focuses on the burial depth detection of the cement foundation of a signal on a conventional railway. The designed burial depth of the cement foundation of the signal on this railway section is 2.8 meters to 5.5 meters. The detection area is 1.2 kilometers long. The surface cover is mainly loamy medium, with clay interlayers in some areas. The medium has a high water content. There are railway communication signal equipment and high-voltage contact networks in the surrounding area, making the electromagnetic environment relatively complex.

[0049] Before the testing began, basic data on the surface and shallow subsurface media of the testing area were collected. Sampling points were set up every 30 meters along the testing route, collecting media samples at depths ranging from 0 to 2 meters. Each sampling point was sampled in two layers: a 0-1 meter layer and a 1-2 meter layer. The collected samples were sent to the laboratory to determine the volumetric moisture content. The results showed that the volumetric moisture content of the loamy media in the testing area ranged from 12% to 18%, with some local clay interlayers reaching up to 22%. The groundwater level was approximately 6 meters deep. Based on the measured media type and moisture content distribution, a database of electromagnetic parameters for the media in the testing area was established. Different depths and moisture content areas corresponded to different dielectric constants and attenuation coefficients.

[0050] Based on the electromagnetic parameter database of the medium, an antenna unit with a corresponding center frequency was matched. Since the water content of the medium in the detection area is between 10% and 20%, and the designed burial depth of the cement foundation is between 3 and 6 meters, an integrated transceiver antenna unit with a center frequency of 400MHz was selected. Simultaneously, to ensure detection resolution for the shallow cement foundation, a dual-channel control unit was used, connecting both a shallow antenna unit with a center frequency of 1000MHz and a deep antenna unit with a center frequency of 400MHz. The two antenna units were fixed side-by-side on the detection bracket, and their horizontal spacing and height were adjusted to ensure that the scanning trajectories of the two antenna units overlapped.

[0051] Launch the dedicated data processing software and enter the scanning parameter configuration interface to set the scanning parameters for the two antenna elements. For the shallow antenna element, set the sample size to 512 points, the stack size to 8, the time scale to 30 ns, and the trace spacing to 80 mm. For the deep antenna element, set the sample size to 2048 points, the stack size to 32, the time scale to 150 ns, and the trace spacing to 100 mm. Both antenna elements use a linear gain profile, with gain factors set to 35 and 55 respectively. The software's signal monitoring function was used to detect the ambient electromagnetic interference intensity. The detection showed that the background noise amplitude was approximately 25% of the effective signal amplitude, which did not reach the strong electromagnetic interference threshold; therefore, the above parameter settings were kept unchanged.

[0052] Connect the wheeled odometer to the control unit, activate the mobile scanning mode, and calibrate the odometer parameters. Since the detection area is located beside the railway embankment, the road surface is relatively narrow and contains some gravel. Move the detection support along the detection route at a constant speed of 0.5 m / s to 0.8 m / s. During the scan, the dual-channel control unit simultaneously acquires radar echo data from both antenna units. Based on the location information recorded by the odometer, the software uses a time-synchronization matching method to fuse the shallow and deep echo data, generating a radar profile covering the entire depth range.

[0053] When the detection reached a certain location, a utility pole obstacle was found in the detection route, preventing continuous scanning. At this point, the system switched to a step-by-step scanning mode, setting a scan start and end point on both sides of the obstacle, and acquiring a radar track every 50mm. After scanning both sides of the obstacle, 15 overlapping tracks were selected in the overlapping area of ​​the two scan segments. By adjusting the time and position references of the two data segments, seamless stitching of the segmented scan data was achieved, ensuring the continuity of the detection results.

[0054] After scanning, the fused radar echo data is preprocessed. Mean filtering, direct wave cancellation, Wiener deconvolution, and background removal are performed sequentially. To address the issue of significant signal attenuation at deeper depths, gain compensation in the deeper regions is appropriately increased to enhance the reflected signal from deep concrete foundations. The effective reflected wave signal at the interface between the concrete foundation and the surrounding medium is extracted, and the dielectric constant of the medium is dynamically corrected in real time. The propagation speed of electromagnetic waves is adjusted layer by layer based on the differences in water content at different depths. A weighted least squares method is used to fit the reflection hyperbola of each concrete foundation, accurately extracting vertex features and calculating the actual burial depth of the concrete foundation.

[0055] Before the formal scanning, four cement standard piles with known burial depths were selected for on-site system calibration. The calibration points covered areas with different moisture contents and burial depths. Each calibration point was scanned three times, and the system correction coefficient was calculated to correct for system errors in all detection data. During the result generation phase, the software automatically generated a burial depth distribution profile, marking the location and burial depth information of each signal's cement foundation. 12% of the detection points were randomly selected for excavation verification, comparing the differences between the radar detection results and the actual burial depth. For detection points with relatively large errors, the scanning parameters were readjusted for a second scan to ensure the accuracy of the detection results.

[0056] Table 3. Relative Error of Burial Depth Detection under Different Dielectric Constant Correction Methods

[0057] Table 2 shows the data from the excavation verification results of the railway signal foundation inspection in Example 2. The true values ​​are the actual burial depth of the cement foundation measured with a steel tape measure. It compares the relative errors of the burial depth detection obtained using the fixed dielectric constant method and the real-time dynamic correction method of this invention. The data shows that the relative error of the fixed dielectric constant method is generally high, and the error varies significantly with fluctuations in the medium's moisture content. In contrast, the real-time dynamic correction method adjusts the dielectric constant value according to the actual characteristics of the medium on site, significantly reducing the relative error of the burial depth detection and resulting in more accurate results. This result demonstrates that this invention, by dynamically correcting the dielectric constant of the medium in real time, can effectively eliminate the influence of changes in medium moisture content on the detection results. It is suitable for complex detection scenarios with uneven medium characteristics and large fluctuations in moisture content, and can provide reliable data support for the quality acceptance and maintenance inspection of cement foundations for transportation facilities.

[0058] Reference Figure 2 This paper details the step-by-step preprocessing of radar echo data. Raw data is subject to various interferences during underground propagation, necessitating purification to ensure the accuracy of subsequent depth calculations. First, the system employs a mean filtering algorithm to process the radar profile, effectively removing random noise in space. Next, the system selects the sampling points at the front of the radar profile as the direct wave template and uses a cancellation algorithm to completely eliminate shallow interference caused by the direct wave. After removing the direct wave, to improve the radar signal's resolution in the depth direction, the system uses a deconvolution algorithm to compress the radar pulse width. Finally, a sliding window algorithm is used to eliminate background reflection signals generated by the homogeneous underground medium. After these four consecutive processing steps, the interface reflection signal between the concrete foundation and the surrounding medium is clearly extracted and stored in a standard format for subsequent analysis.

[0059] Reference Figure 3This section demonstrates the equipment selection and parameter configuration logic in the pre-detection phase. The system comprehensively assesses the on-site environment to select the optimal hardware combination. First, based on the combination of medium moisture content and target burial depth, different models of shallow, mid-, or deep antenna units are selected. During this process, the system also monitors the intensity of electromagnetic interference in the environment in real time. If strong interference is detected, it directly switches to an enhanced shielded antenna and simultaneously reduces the scanning speed and increases the stack count to ensure the signal-to-noise ratio. In the parameter setting phase, operators configure the sample size, time scale, and gain factor. If there is a need for detection at extremely deep depths, the system also supports a dual-channel mode, simultaneously connecting both high- and low-frequency antennas. Through time synchronization matching, shallow and deep echo data are fused to achieve accurate detection across the entire depth range.

[0060] Reference Figure 4 This paper describes the scanning response and data stitching strategy when encountering obstacles in complex traffic conditions. Under normal circumstances, the radar uses continuous moving scans. However, when an insurmountable obstacle exists on the preset route, the system immediately switches to a step-by-step scanning mode. Operators redefine the scan start and end points on both sides of the obstacle and acquire radar tracks on both sides of the obstacle using a fixed-distance acquisition method. After acquiring two data segments, to ensure the continuity of the underground profile, the system introduces an overlapping area matching method. By extracting a sufficient number of overlapping tracks within the overlapping areas of the two data segments, the system performs fine alignment and adjustment of their time and position references, ultimately eliminating spatial discontinuities and achieving seamless stitching of the two interrupted data segments, ensuring the integrity of the detection profile.

[0061] Reference Figure 5 This demonstrates the error control and result reliability verification mechanism throughout the entire process, from pre- and post-detection. Before the formal large-area scanning, the system first enters the calibration phase. Multiple standard piles with known accurate burial depths are selected on-site as benchmarks, and the same parameters are used for repeated scanning, with the average value taken. The radar measured values ​​are compared with the physical true values, and the deviation is calculated to generate a system correction coefficient to eliminate systematic errors in the equipment. After all detection tasks are completed, the system enters the verification phase. A fixed proportion of the detected points are randomly selected for on-site excavation, and the actual burial depth is directly measured using physical measuring tools. If the relative error between the measured value and the radar calculated value exceeds the allowable range, the system will force parameter adjustments for a second regional re-measurement, or even replace the antenna with a higher-specification one to re-execute the task, thus forming a strict closed-loop quality control.

[0062] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar, characterized in that, include: A pulse ground-penetrating radar system was used as the detection equipment. This system is an active radar device that uses the reflection of radio waves to measure distance. The ground-penetrating radar system includes a single-channel or dual-channel control unit, an integrated transceiver antenna unit adapted to multiple center frequencies, a wheeled odometer, a power supply unit, and a portable computer with data acquisition and processing software installed. The control unit is connected to the portable computer via an Ethernet interface. The control unit has a built-in voltage decoder that can provide 15V to 22V DC power to the portable computer. The antenna unit transmits pulses with a duration of 1ns to 10ns and is connected to the control unit via a dedicated AB cable. Fiber optic antenna units require the use of an optical converter. The detection method includes the following steps: collecting basic data of the surface and shallow media in the detection area; determining the type of overburden medium, the volumetric water content of the medium, and the depth of the groundwater level in the detection area through on-site sampling and laboratory analysis; recording the surface flatness and surrounding electromagnetic environment information of the detection area; establishing a database of electromagnetic parameters of the medium in the detection area; and matching antenna units with corresponding center frequencies according to the database. Configure the ground-penetrating radar scanning parameters through data acquisition and processing software, connect the odometer and control unit through a dedicated cable, enable the mobile scanning mode in the software, calibrate the wheel diameter parameters and pulse coefficient of the odometer, perform continuous non-contact scanning, use an antenna bracket with shock absorption device to cope with uneven ground, and record the scanning trajectory and corresponding radar echo data at the same time. The raw radar echo data was preprocessed to remove random noise, direct wave interference and background reflection signals in sequence, and the effective reflected wave signal at the interface between the cement foundation and the surrounding medium was extracted. Based on the electromagnetic wave reflection characteristics caused by the difference in dielectric constant between cement and the covering layer medium, a preset dielectric property table is used to determine the initial propagation speed of electromagnetic waves in the covering layer medium of the detection area. The dielectric property table includes the relative dielectric constant, attenuation coefficient and propagation speed parameters of different types of media at different moisture contents. The dielectric constant of the medium is dynamically corrected in real time. The real-time volumetric water content of the medium is calculated based on the amplitude attenuation characteristics of the radar echo signal collected during the detection process. The real-time relative dielectric constant is obtained through the correlation model between the medium humidity and the dielectric constant, and then the electromagnetic wave propagation speed is corrected. The hyperbola fitting algorithm is used to extract the vertex features of the effective reflected wave, determine the two-way propagation time corresponding to the vertex of the reflected wave, and calculate the actual burial depth of the cement foundation by combining the corrected electromagnetic wave propagation speed, and generate an inspection report containing the horizontal position and burial depth information of the cement foundation.

2. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, This also includes conducting on-site system calibration before the formal scanning. At least three cement standard piles with known burial depths are selected as calibration points in the detection area. The calibration points are evenly distributed in different media characteristics locations in the detection area. Each calibration point is scanned 2 to 5 times using the same scanning parameters as the formal detection. The average value of the multiple detection results is taken as the detection burial depth of the calibration point. The deviation between the detection burial depth of the calibration point and the actual burial depth is calculated. The system correction coefficient is obtained based on the linear correction method. This correction coefficient is used to correct the system error of all subsequent detection data.

3. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, The correlation model between the medium humidity and the dielectric constant is a quadratic polynomial: ;in The real-time relative permittivity of the medium; The water content by volume is expressed as a decimal. , which is the dimensionless fitting coefficient corresponding to the type of overburden medium.

4. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the antenna unit matching step, the intensity of environmental electromagnetic interference is detected by the signal monitoring function of the data acquisition and processing software. When the background noise amplitude exceeds 20% to 40% of the effective signal amplitude, it is determined that there is strong electromagnetic interference. At this time, an enhanced antenna unit with an electromagnetic shielding shell is selected, and the stack number is set to 32 to 64 to improve the signal-to-noise ratio. The scanning speed is reduced to 0.3m / s to 0.8m / s, and wireless communication equipment and high-voltage electrical equipment in the detection area are turned off.

5. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the scanning parameter configuration step, when a dual-channel control unit is used, two antenna units with different center frequencies are connected simultaneously to collect shallow and deep radar echo data respectively. The shallow antenna unit is used to detect cement foundations with a burial depth of less than 2 meters, and the deep antenna unit is used to detect cement foundations with a burial depth of more than 2 meters. The two sets of data are fused based on the location information recorded by the odometer using a time synchronization matching method.

6. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the mobile scanning mode step, when there are obstacles in the detection route that prevent continuous scanning, switch to the step-by-step scanning mode. Set the scanning start point and end point on both sides of the obstacle, and collect a radar trace every 20mm to 100mm. After completing the scanning of both sides of the obstacle, use the overlapping area matching method to select no less than 10 overlapping traces in the overlapping area of ​​the two segments of scanning data. Seamless stitching is achieved by adjusting the time reference and position reference of the two segments of data.

7. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the process of extracting the vertex features of the reflected wave using the hyperbola fitting algorithm, a weighted least squares method is used to fit the hyperbola of the reflected wave. Higher weight values ​​are assigned to echo points closer to the vertex to improve vertex localization accuracy. "Closer to the vertex" refers to a range extending no more than five radar tracks to either side of the track where the reflected wave vertex is located. Echo points within this range have a weight coefficient of 2, while the weight coefficients for other echo points are set to 1. The objective function of the weighted least squares method is: ;in The objective function value; This represents the total number of echo points involved in the fitting process. For the first Weighting coefficients for each echo point; For the first The two-way propagation time corresponding to each echo point; The two-way propagation time is the time corresponding to the vertex of the reflected wave. For the first The horizontal distance of each echo point relative to the vertex; This represents the corrected propagation speed of electromagnetic waves in the medium.

8. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the radar echo data preprocessing step, firstly, a mean filtering algorithm with a 3×3 to 5×5 window is used to remove random noise. Then, 5 to 15 sampling points in front of the radar profile are selected as direct wave templates, and a direct wave cancellation algorithm is used to remove direct wave interference. Next, a Wiener deconvolution algorithm with an operator length of 16 to 64 points is used to compress the radar pulse width. Finally, a sliding window background removal algorithm with a window length of 30 to 100 traces is used to eliminate background reflection signals generated by the underground homogeneous medium.

9. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, In the cement foundation burial depth calculation step, when multiple cement foundation reflected waves are detected, a hyperbola is fitted to each reflected wave to obtain the two-way propagation time of the vertex corresponding to each reflected wave. Combined with the corrected electromagnetic wave propagation speed, the burial depth of each cement foundation is calculated. At the same time, the horizontal position information of each cement foundation is recorded to generate a burial depth distribution profile with horizontal distance as the abscissa and burial depth as the ordinate. Different colored blocks are used to represent different burial depth ranges and the cement foundation number and parameters are marked.

10. The method for detecting the burial depth of cement foundations for transportation facilities based on ground-penetrating radar according to claim 1, characterized in that, The detection method also includes verifying the reliability of the detection results. At least 10% of the detection points are randomly selected for excavation verification. The excavation depth is 1.1 to 1.5 times greater than the radar detection depth to fully expose the bottom of the cement foundation. The actual burial depth is measured with a steel tape measure, and the relative error between the actual burial depth of the excavation verification and the radar detection depth is calculated. When the relative error is greater than 5%, the scanning parameters are readjusted and the area is scanned a second time. If the relative error of the second detection is still greater than 5%, the antenna unit with a higher center frequency is replaced and the detection is repeated.