Roadbed grouting material quality evaluation method based on ground penetrating radar image recognition method

Abstract

This invention provides a method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition, belonging to the field of grouting material evaluation technology. The method includes the following steps: selecting grouting materials for road grouting; using a trolley to pull a ground-penetrating radar to scan the road for rapid data acquisition; preprocessing the acquired data; performing image comparison and analysis and setting evaluation indicators; then performing normal distribution curve analysis, expected value analysis of the normal distribution curve, and standard deviation analysis of the normal distribution curve; and evaluating the grouting quality based on the standard deviation. This invention, through comparative analysis of 80 sets of planar amplitude grayscale images before and after grouting, clearly shows that the grayscale image after grouting has higher consistency than the grayscale image before grouting. Intuitive quantitative observation shows that the grouting process effectively reinforces hidden road defects.

Description

Technical Field

[0001] This invention relates to the field of grouting material evaluation technology, and in particular to a method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition. Background Technology

[0002] With the rapid development of China's transportation infrastructure, the focus of highway transportation has shifted from large-scale construction to sustainable maintenance. Simultaneously, China's rapid economic growth has increased the demand for highway transportation, posing significant challenges to maintenance practices. Therefore, timely and effective handling of highway problems is crucial. Traditional excavation-based maintenance methods are time-consuming, costly, and cause significant disruption to traffic and the environment. In contrast, grouting technology offers advantages such as short construction cycles, low cost, and minimal impact on traffic and the environment, making it a commonly used and effective method for solving highway problems.

[0003] Grouting technology has been used in highway maintenance for many years, yielding rich results in theoretical and materials research. However, research on grouting quality assessment is limited. Current Chinese standards still employ traditional methods for grouting quality assessment, such as core sampling or deflection testing. Due to the uneven distribution of grout underground, core sampling and deflection testing often lack representativeness. Ground penetrating radar (GPR), a non-destructive testing technology based on electromagnetic field theory, has been widely used in the highway sector. While GPR is frequently used for tunnel grouting assessment, research on using GPR to assess grouting quality in the highway sector is scarce. This may be because hidden defects beneath the road surface are often not obvious. Furthermore, the multi-resolution characteristics of GPR images also present challenges for quantitatively describing grouting quality using GPR.

[0004] The S511 highway from Guigang City to Laibin City in Guangxi Zhuang Autonomous Region is a cement concrete pavement. The structural layers, from top to bottom, are: a 26 cm thick C30 cement concrete surface layer; a 1.5 cm thick asphalt macadam seal layer; a 20 cm thick cement-stabilized crushed stone layer; a 15 cm thick graded crushed stone layer; and a 15 cm thick crushed stone subbase. Since its completion and opening to traffic in 2017, the road has experienced a significant increase in traffic load. Consequently, multiple instances of slab bottom voids and mud leakage have been observed along the entire route. When slab bottom voids develop to a certain extent, they can lead to cracking or even fracture of the slab, posing a significant risk to the road structure quality and overall traffic safety. Based on the need to improve road surface quality and extend road service life, grouting is used to treat the slab bottom void areas. Therefore, it is necessary to design a quality evaluation method for roadbed grouting materials based on ground-penetrating radar image recognition to ensure the quality of the grouting. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition, solving the technical problem that existing grouting evaluation methods in highway maintenance cannot cover all grouting areas and are inaccurate. To assess grouting quality, three-dimensional ground-penetrating radar technology is used to perform comparative scans of the road before and after grouting.

[0006] Three-dimensional ground-penetrating radar (GPR) technology was used to scan the road base layer before and after grouting. The main objective was to obtain a planar amplitude grayscale image that effectively displays the changes in grouting depth. Image recognition technology was used to obtain the grayscale values ​​of each point in the planar amplitude grayscale image. Subsequently, statistical analysis was performed on the distribution, expected value, and standard deviation of the grayscale values ​​in the planar amplitude grayscale images before and after grouting. Finally, a roadbed grouting quality evaluation index based on the GPR image recognition method was proposed.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition, the method comprising the following steps: Step 1: Select grouting materials and grout the highway; Step 2: Use a trolley to pull the ground-penetrating radar for rapid data collection; Step 3: Preprocess the collected data; Step 4: Perform image comparison analysis and set evaluation indicators; Step 5: Normal distribution curve analysis, expected value analysis of the normal distribution curve, and standard deviation analysis of the normal distribution curve; Step 6: Evaluate the grouting quality based on the standard deviation.

[0008] Further, the specific process of step 1 is as follows: In order to solve the problem of local defects in ordinary cement grout, cement-based solid waste grouting material is selected as the highway grouting material. The grouting material has good fluidity, self-compacting and micro-expansion properties, effectively filling the void area at the bottom of the slab. The water-cement ratio of the grouting material is 0.4. The specific grouting process is as follows: Hole layout: Each concrete slab has an area of ​​4.5 meters × 5 meters, with 5 holes per slab. Drilling: To effectively address the issue of mud seepage from the bottom of the slab, the holes penetrate the entire structural layer, with a drilling depth of 80 centimeters and a hole diameter of 38 millimeters. Grouting: Low-pressure, slow-injection method is used, with the maximum grouting pressure not exceeding 1.5 MPa. Grouting can be stopped when the grouting pressure stabilizes for more than one minute. Please note that if grout seeps from adjacent holes during the grouting process, the leaking holes must be sealed and grouting must continue. The final setting time of the grouting material should not exceed 12 hours. To ensure grouting quality, the project requires 24-hour curing to allow the grout to fully solidify.

[0009] Furthermore, the specific process of step 2 is as follows: the ground-penetrating radar is fixed to the back of the vehicle, and the ground-penetrating radar is equipped with two wheels. The vehicle is driven at a set speed on the road, and the ground-penetrating radar follows the vehicle and collects road data in real time, thereby improving the collection efficiency. When electromagnetic waves propagate through underground media, they are reflected when they encounter non-homogeneous materials. The reflection coefficient is a key characteristic of these reflected waves and is closely related to the dielectric constant of the medium. When the dielectric constants of two media differ significantly, the intensity of the reflected signal increases. The intensity of the reflected signal depends on the reflection coefficient, and its relationship with the dielectric constant of the medium is as follows: in e 1 and e 2 are the dielectric constants of medium 1 and medium 2, respectively. According to the above formula, when an electromagnetic wave propagates from medium 1 to medium 2, if... e 1> e 2, R If it is positive, e 1< e 2, R If the value is negative, in the amplitude grayscale image, a region will be black or white, indicating that the R value of the corresponding region is greater than the set value, reflecting the inhomogeneity of the medium in the corresponding region. Usually, black and white appear at the same time.

[0010] Furthermore, the specific process of step 3 is as follows: the scanned data is processed using the professional software CONDOR before and after grouting. In the preprocessing stage, DC removal and DC drift removal filtering algorithms are used to eliminate DC offset during the data recording process. In the postprocessing stage, normalization, amplitude correction, antenna residual vibration removal and interpolation are applied to process the data in sequence. Two-dimensional image grayscale analysis can accurately determine the changes in the road base layer before and after grouting by simultaneously analyzing two parallel images, thus providing a comprehensive understanding of the grouting effect. To ensure the consistency of ground-penetrating radar (GPR) image quality, the image size acquired from GPR images is standardized to a resolution of 450×500. To evaluate the changes in GPR images before and after grouting, the grayscale value of each pixel in the image is used as a key evaluation indicator. Statistical analysis of the grayscale value changes in two parallel images before and after grouting can provide a reference for the post-grouting effect. Descriptive statistical analysis methods are used to quantify the changes between the two parallel images captured before and after the grouting process, as shown in the following formula. in X Represents grayscale value, N Represents a normal distribution. m Represents the expected value. s Represents standard deviation.

[0011] Further, the specific process of step 4 is as follows: In the grouting area, 9 plates are selected as the analysis objects. After performing three-dimensional ground radar scanning, the amplitude grayscale map before and after grouting is obtained using CONDOR. Planar amplitude grayscale maps of 10 depths for each plate are extracted for further analysis. The term "depth" indicates the distance from the ground horizontal plane. Based on the analysis of 90 sets of planar amplitude grayscale maps before and after grouting, the consistency of the planar amplitude grayscale map after grouting is improved compared with the respective maps before grouting. Grouting eliminates or reduces the inhomogeneity of the medium in the corresponding area. To assess grouting quality based on ground-penetrating radar amplitude grayscale images, image recognition methods were used to extract grayscale values ​​from the images for analysis. The distribution, expected value, and standard deviation of grayscale values ​​in the amplitude grayscale images before and after grouting were compared and analyzed.

[0012] Furthermore, the specific process of normal distribution curve analysis in step 5 is as follows: based on the normal distribution curve of the amplitude grayscale images before and after grouting, which contain 10 sets of grayscale images representing different depths for each plate, the grayscale values ​​in the image range from 0 to 255, with an average value of 127.5. The further the grayscale value of a region deviates from 127.5, the more severe the disease in the corresponding region. Histogram analysis revealed that the grayscale values ​​before grouting were mainly distributed at both ends, while after grouting they were concentrated in the middle, indicating a more homogeneous formation and a reduction in defects after grouting. Comparing the normal distribution curves before and after grouting, it was noted that the curve after grouting was smoother and had a lower peak value than the curve before grouting. The mathematical expression for the normal distribution curve is: N ( m , s ²), where N represents a normal distribution. m Indicates the expected value. s It represents the standard deviation.

[0013] Furthermore, the specific process of the expected value analysis of the state distribution curve in step 5: comparison of the expected values ​​of the amplitude grayscale images before and after grouting, and the synchronous increase or decrease of the expected value with the grouting depth; A comparison of the standard deviations of the amplitude grayscale images before and after grouting reveals that the standard deviation of the amplitude grayscale images after grouting is significantly lower than that before grouting. From a mathematical perspective, the standard deviation reflects the dispersion of individuals within a group. The larger the standard deviation, the greater the variation among individuals and the greater the deviation from the mean. In the context of amplitude grayscale images, a larger standard deviation means that the image is more uneven and the grayscale values ​​deviate more from the mean. After grouting, the amplitude grayscale images become more uniform, resulting in a decrease in the accuracy of the amplitude grayscale images compared to before grouting. Therefore, the standard deviation is considered a reliable indicator for evaluating the quality of grouting.

[0014] Furthermore, the specific process of step 6 is as follows: based on standard deviation d The grouting quality is defined by the following formula: in s b and s a These are the standard deviations of the amplitude grayscale images before and after grouting. Further analysis of the data in the images determines the grouting quality of a total of 80 depth profiles across 9 slabs.

[0015] The present invention, by adopting the above-described technical solution, has the following beneficial effects: This invention compares and analyzes 80 sets of planar amplitude grayscale images before and after grouting, clearly showing that the grayscale images after grouting have higher consistency than those before grouting. Visual observation indicates that the grouting process effectively reinforces the road base layer. Using image recognition technology to extract grayscale values ​​from the amplitude grayscale images and plotting normal distribution curves, it is noted that the curve after grouting is smoother and the peak value is lower than that before grouting. There is no significant correlation between the expected value and grouting quality, while the standard deviation shows a significant correlation with grouting quality. It is suggested that the grouting quality be qualitatively described by examining the normal distribution curve of the amplitude grayscale values, and quantitatively evaluated using the standard deviation. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a photograph of the actual grouting process of this invention; Figure 3 This is the ground-penetrating radar amplitude diagram of the present invention; Figure 4 This is a field image of the 3D ground-penetrating radar vehicle of this invention; Figure 5 This is a diagram of the CONDOR software interface and processing program of the present invention; Figure 6 is a grayscale image of the planar amplitude of the nine plates at different depths in this invention; Figure 7 This is a normal distribution curve of the amplitude gray value of the present invention; Figure 8 This is a comparison chart of the expected values ​​of the amplitude grayscale before and after grouting according to the present invention; Figure 9 This is a comparison chart of the standard deviation of the grayscale amplitude before and after grouting according to the present invention; Figure 10 This is a graph showing the grouting quality evaluation results based on standard deviation of this invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and preferred embodiments. However, it should be noted that many details listed in the specification are merely to provide the reader with a thorough understanding of one or more aspects of the present invention, and these aspects of the invention can be implemented even without these specific details.

[0018] like Figure 1 As shown, a method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition is described, and the method includes the following steps: Step 1: The selection of grouting material plays a decisive role in the effectiveness of grouting treatment. To address the localized defects of ordinary cement grout, a novel cement-based solid waste grouting material was chosen as an alternative. This grouting material possesses excellent fluidity, self-compacting properties, and micro-expansion characteristics. It can effectively fill voids under the slab, thereby improving the pavement's load-bearing capacity and durability. Based on field practice, a water-cement ratio of 0.4 was determined.

[0019] The grouting process is as follows: (a) Hole Layout: Each concrete slab has an area of ​​4.5 meters (width) x 5 meters (length). Five holes are arranged on each slab, such as... Figure 2 As shown in (a).

[0020] (b) Drilling: To effectively solve the problem of mud seepage from the bottom of the slab, we aim to drill through the entire structural layer, such as... Figure 2 As shown in (b). Therefore, the drilling depth is set to 80 cm and the hole diameter to 38 mm.

[0021] (c) Grouting: Grouting shall be carried out using a low-pressure, slow-injection method, with a maximum grouting pressure not exceeding 1.5 MPa. Grouting may be stopped when the grouting pressure has been stable for more than one minute. Please note that if grout seeps out from adjacent holes during the grouting process, the leaking holes must be sealed and grouting must continue.

[0022] (d) Curing: The final setting time of the grouting material shall not exceed 12 hours. To ensure the quality of grouting, the project requires 24-hour curing to allow the grout to fully solidify within this time.

[0023] Step 2: Ground-penetrating radar (GPR) is an electromagnetic device that emits ultra-high-frequency electromagnetic waves. These pulsed waves are reflected and scattered when they encounter changes in the properties of the medium. When used for road detection, the emitted electromagnetic waves are captured by the receiving antenna when they encounter changes in the medium's properties (such as interfaces, loose areas, voids, cracks, or fissures). GPR reveals the presence and location of underground objects by generating images that include the amplitude of the reflected waves and their two-way travel times.

[0024] During electromagnetic wave propagation, the wave velocity v and the relative permittivity ε of the medium are related. r The relationship is as follows: (1) in c It is the speed of electromagnetic waves in a vacuum (≈3×10). 8 m / s).

[0025] When electromagnetic waves propagate through underground media, they are reflected when they encounter non-homogeneous materials. The reflection coefficient is a key characteristic of these reflected waves and is closely related to the dielectric constant of the medium. When the difference in dielectric constant between two media is significant, the intensity of the reflected signal increases. Therefore, the intensity of the reflected signal depends on the reflection coefficient, and its relationship with the dielectric constant of the medium is as follows: (2) in e 1 and e 2 are the dielectric constants of medium 1 and medium 2, respectively. According to formula (2), when an electromagnetic wave propagates from medium 1 to medium 2, if e 1> e 2, R If e1 <ε 2 , R The value is negative. In an amplitude grayscale image, a region that is black or white indicates a large R-value, reflecting the inhomogeneity of the medium in that region. Typically, black and white appear simultaneously, such as... Figure 3 As shown.

[0026] The Raptor-45 3D ground-penetrating radar, manufactured by Deep Radar, was used to survey the area before and after grouting. Within a given space, the 3D ground-penetrating radar is equipped with multiple transmitting antennas, simultaneously emitting multiple beams of electromagnetic waves towards the object being measured, forming an electromagnetic wave wall that effectively illuminates the object and records its echo reflections. Simultaneously, the 3D ground-penetrating radar performs a comprehensive scan of a large area of ​​the object being measured through relative motion, ultimately generating a continuous scan and full-space 3D image of the target. Figure 4 The exhibition showcased the on-site 3D ground-penetrating radar vehicle, along with a magnified view of the 3D ground-penetrating radar.

[0027] A comprehensive 3D ground-penetrating radar was used to scan the area before and after grouting to ensure complete coverage. The scan data was processed using the specialized software CONDOR before and after grouting. In the preprocessing stage, DC removal and dewow filtering algorithms were employed to eliminate DC offset during data recording. Post-processing sequentially applied normalization (Muting), amplitude correction, antenna ringdown removal, and interpolation. The functions of these filtering algorithms are as follows: (a) Normalization: By normalizing the amplitude, the algorithm can eliminate extreme amplitudes and ensure that the amplitude is evenly distributed within a specific range, thereby improving the visibility of anomalous signals in ground-penetrating radar images.

[0028] (b) Amplitude correction: Electromagnetic waves weaken gradually with distance as they propagate underground. This algorithm is used to amplify the weakened weak signal.

[0029] (c) Antenna ringdown removal: During data acquisition, interference signals called ringing are generated. This algorithm can eliminate this ringing interference.

[0030] (d) Interpolation: This algorithm can perform interpolation on areas not covered by the scan.

[0031] Step 3: Typically, assessing grouting quality often requires destructive methods such as core drilling or excavation. These destructive assessment methods are not only limited in scope but may also lead to additional traffic congestion. Image analysis has proven to be an effective tool for characterizing particle size and shape distribution. This technique helps capture subtle material features, which is crucial for a more rigorous and effective assessment of grouting quality. This study employs grayscale analysis of two-dimensional images, a method that can accurately determine changes in the road base layer before and after grouting by simultaneously analyzing two parallel images, thereby providing a comprehensive understanding of the grouting effect. To ensure the consistency of the quality of the ground-penetrating radar images, the image size obtained from the ground-penetrating radar images is standardized to a resolution of 450×500 (length×width). To assess the changes in the ground-penetrating radar images before and after grouting, we use the grayscale value of each pixel in the image as a key evaluation indicator. Statistical analysis of the grayscale value changes of two parallel images before and after grouting can provide a reference for the post-grouting effect, as shown in Figure 6. Descriptive statistical analysis methods are used to quantify the changes between the two parallel images captured before and after the grouting process, as shown in Formula (3).

[0032] in X Represents grayscale value, N Represents a normal distribution. m Represents the expected value. s Represents standard deviation.

[0033] Step 4: Image analysis. Nine slabs were selected as analysis objects in the grouting area. After 3D ground-penetrating radar scanning, amplitude grayscale images before and after grouting were obtained using CONDOR. Planar amplitude grayscale images at 10 depths were extracted from each slab for further analysis, such as... Figure 7As shown. The term "depth" refers to the distance from the horizontal plane of the ground. Based on 90 sets of planar amplitude grayscale images before and after grouting, as shown... Figure 8 As shown, it is evident that the grayscale image of the planar amplitude after grouting shows significantly improved consistency compared to the images before grouting. This indicates that grouting eliminates or reduces the inhomogeneity of the medium in this area.

[0034] The evaluation index is based on the ground-penetrating radar amplitude grayscale image to assess grouting quality. Image recognition methods are used to extract grayscale values ​​from the amplitude grayscale image for analysis. The distribution, expected value, and standard deviation of grayscale values ​​in the amplitude grayscale images before and after grouting are compared and analyzed.

[0035] Step 5: Normal distribution curve Figure 7 Each plate contains 10 sets of grayscale images representing different depths of grouting before and after the grouting process. Figure 8 The normal distribution curves of the grayscale amplitude images of slab #8 before and after grouting are shown. For detailed analysis, let's focus on... Figure 9 (b) And examine the displayed image data. The grayscale value ranges from 0 to 255, with an average value of 127.5. According to formula (2), the further the grayscale value of a region deviates from 127.5, the more severe the disease in that region.

[0036] Histogram analysis revealed that the grayscale values ​​before grouting were mainly distributed at both ends, while after grouting they were concentrated in the middle. This indicates a more homogeneous formation and a reduction in defects after grouting. Furthermore, comparing the normal distribution curves before and after grouting showed that the curve after grouting was smoother and had a lower peak value than the curve before grouting. In short, the mathematical expression for the normal distribution curve is: N ( m , s ²), where N Indicates a normal distribution. m Indicates the expected value. s It represents the standard deviation.

[0037] These observations are not limited to slab #8, but are consistent in the bar charts and normal distribution curves before and after grouting at different depths for all eight slabs. Figure 7 (a) and (c)-(j) provide a visual representation of these findings. Furthermore, the 80 sets of data corresponding to the remaining 8 slabs also follow the same pattern. Extensive data supports these conclusions, demonstrating the feasibility of using image recognition methods to assess grouting quality.

[0038] In the past, the interpretation of ground-penetrating radar images was subjective. The introduction of image recognition technology has provided an objective and scientific method for assessing grouting quality. However, while the normal distribution curve can confirm that the formation has been improved through grouting, it cannot quantify the degree of improvement. To assess the degree of improvement, it is necessary to further analyze two parameters in the normal distribution curve: the expected value and the standard deviation.

[0039] The expected value of the normal distribution curve. Figure 9 The comparison of expected values ​​of amplitude grayscale images before and after grouting is presented. The expected values ​​increase and decrease synchronously with grouting depth. However, the observed synchronous change is not significant, possibly due to two factors: first, the expected value may not be a suitable indicator for evaluating grouting quality; second, the complex engineering conditions at the grouting site may lead to poor data quality. To further investigate the impact of grouting on the expected values ​​of amplitude grayscale images, detailed indoor experiments are necessary in the future.

[0040] The standard deviation of the normal distribution curve Figure 10 This paper compares the standard deviations of amplitude grayscale images before and after grouting. Notably, the standard deviation of the amplitude grayscale image after grouting is significantly lower than that before grouting. Mathematically, standard deviation reflects the dispersion of individuals within a group. A larger standard deviation indicates greater inter-individual variation and a greater deviation from the mean. Against the backdrop of amplitude grayscale images, a larger standard deviation means a more non-uniform image and a greater deviation of grayscale values ​​from the mean. After grouting, the amplitude grayscale image becomes more uniform, resulting in a decrease in accuracy compared to before grouting. Therefore, standard deviation is considered a reliable indicator for evaluating grouting quality.

[0041] Grouting quality evaluation based on standard deviation d The grouting quality is defined as shown in formula (4): in s b and s a These are the standard deviations of the grayscale amplitude images before and after grouting.

[0042] Using formula (4) as a basis, the data in Figure 6 can be further analyzed to determine the grouting quality of the 80 depth profiles across the 9 slabs. This information is as follows: Figure 10 As shown, the quality of the vast majority of grouted areas improved to varying degrees, with only five data points displaying negative values. The negative values ​​are attributed to complex on-site engineering conditions or improper grouting techniques.

[0043] This invention presents a case study of road base grouting engineering, utilizing 3D ground-penetrating radar technology to obtain planar amplitude grayscale images of the road base at different depths before and after grouting. By applying image recognition technology to extract grayscale values ​​from the amplitude grayscale images and performing statistical analysis, the following conclusions were drawn: Comparative analysis of 80 sets of planar amplitude grayscale images before and after grouting clearly shows that the grayscale images after grouting have higher consistency than those before grouting. This intuitive observation indicates that the grouting process effectively reinforces the road base. Using image recognition technology to extract grayscale values ​​from the amplitude grayscale images and plotting normal distribution curves, it was noted that the curve after grouting is smoother and the peak value is lower than that before grouting. There was no significant correlation between the expected value and grouting quality, while the standard deviation showed a significant correlation with grouting quality. It is recommended to qualitatively describe the grouting quality by examining the normal distribution curve of the amplitude grayscale values ​​and to quantitatively evaluate it using the standard deviation.

[0044] Statistical analysis was performed using grayscale values ​​extracted through image recognition technology. The study concluded the following: First, comparing 80 sets of grayscale images revealed higher uniformity in the grayscale images after grouting, indicating that the grouting process effectively reinforced the road base. Second, the normal distribution curve plotted using the extracted grayscale values ​​showed that the curve after grouting was smoother and had lower peak values. Finally, there was no significant correlation between the expected value and grouting quality, while the standard deviation showed a significant correlation. Therefore, it is recommended to conduct a qualitative assessment of grouting quality by examining the normal distribution curve and a quantitative assessment using the standard deviation. These findings provide valuable insights for evaluating grouting quality in similar projects and contribute to improving road construction and maintenance practices.

[0045] Matters not covered in this invention are common knowledge.

[0046] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition, characterized in that: The method includes the following steps: Step 1: Select grouting materials and grout the highway; Step 2: Use a trolley to pull the ground-penetrating radar for rapid data collection; Step 3: Preprocess the collected data; Step 4: Perform image comparison analysis and set evaluation indicators; Step 5: Normal distribution curve analysis, expected value analysis of the normal distribution curve, and standard deviation analysis of the normal distribution curve; Step 6: Evaluate the grouting quality based on the standard deviation.

2. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of step 1 is as follows: In order to solve the problem of local defects in ordinary cement grout, cement-based solid waste grouting material is selected as the grouting material for highways. The grouting material has good fluidity, self-compacting and micro-expansion properties, and fills the void area at the bottom of the slab. The water-cement ratio of the grouting material is 0.

4. The specific grouting process is as follows: Hole layout: Each concrete slab has an area of ​​4.5 meters × 5 meters, with 5 holes per slab. Drilling: To effectively solve the problem of mud seeping from the bottom of the slab, the drilling penetrates the entire structural layer. The drilling depth is set at 80 centimeters, and the hole diameter is 38 millimeters. Grouting: Low-pressure slow grouting is used. The maximum grouting pressure does not exceed 1.5 MPa. Grouting can be stopped when the grouting pressure stabilizes for more than one minute. Please note that if grout seeps from adjacent holes during the grouting process, the leaking holes should be sealed and grouting should continue. The final setting time of the grouting material should not exceed 12 hours. To ensure grouting quality, the project requires 24-hour curing to allow the grout to fully solidify.

3. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of step 2 is as follows: the ground-penetrating radar is fixed to the back of the vehicle. The ground-penetrating radar is equipped with two wheels. The vehicle is driven at a set speed on the road. The ground-penetrating radar follows the vehicle and collects data on the road it passes in real time, thereby improving the collection efficiency. When electromagnetic waves propagate through underground media, they are reflected when they encounter non-homogeneous materials. The reflection coefficient is a key characteristic of the reflected wave and is closely related to the dielectric constant of the medium. When the difference in dielectric constant between two media is greater than a set value, the intensity of the reflected signal will increase. The intensity of the reflected signal depends on the reflection coefficient, and its relationship with the dielectric constant of the medium is as follows: in ε 1 and ε 2 are the dielectric constants of medium 1 and medium 2, respectively. According to the above formula, when an electromagnetic wave propagates from medium 1 to medium 2, if... ε 1> ε 2, R If it is positive, ε 1< ε 2, R If the value is negative, in the amplitude grayscale image, a region will be black or white, indicating that the R value of the corresponding region is greater than the set value, reflecting the inhomogeneity of the medium in the corresponding region. Black and white will appear simultaneously.

4. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of step 3 is as follows: The scanned data is processed using the professional software CONDOR before and after grouting. In the preprocessing stage, DC removal and DC drift removal filtering algorithms are used to eliminate DC offset during the data recording process. In the postprocessing stage, normalization, amplitude correction, antenna residual vibration removal and interpolation are applied to process the data in sequence. Two-dimensional image grayscale analysis was employed to determine the changes in the roadbed before and after grouting by simultaneously analyzing two parallel images, thus comprehensively understanding the grouting effect. To ensure the consistency of ground-penetrating radar (GPR) image quality, the image size acquired from the GPR images was standardized to a resolution of 450×500. To evaluate the changes in GPR images before and after grouting, the grayscale value of each pixel in the image was used as a key evaluation indicator. Statistical analysis of the grayscale value changes in the two parallel images before and after grouting can provide a reference for the post-grouting effect. Descriptive statistical analysis methods were used to quantify the changes between the two parallel images captured before and after the grouting process, as shown in the following formula. in X Represents grayscale value, N Represents a normal distribution. μ Represents the expected value. σ Represents standard deviation.

5. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of step 4 is as follows: In the grouting area, 9 plates are selected as the analysis objects. After three-dimensional ground radar scanning, the amplitude grayscale map before and after grouting is obtained using CONDOR. Planar amplitude grayscale maps of 10 depths of each plate are extracted for further analysis. The term "depth" indicates the distance from the horizontal plane of the ground. Based on the analysis of 90 sets of planar amplitude grayscale maps before and after grouting, the consistency of the planar amplitude grayscale map after grouting is improved compared with the respective maps before grouting. Grouting eliminates or reduces the inhomogeneity of the medium in the corresponding area. To assess grouting quality based on ground-penetrating radar amplitude grayscale images, image recognition methods were used to extract grayscale values ​​from the amplitude grayscale images for analysis. The distribution, expected value, and standard deviation of grayscale values ​​in the amplitude grayscale images before and after grouting were compared and analyzed.

6. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of normal distribution curve analysis in step 5 is as follows: Based on the normal distribution curve of the grayscale images of amplitude before and after grouting, which contain 10 sets of grayscale images of amplitude before and after grouting representing different depths for each plate, the grayscale values ​​in the image range from 0 to 255, with an average value of 127.

5. The further the grayscale value of a region deviates from 127.5, the more severe the disease in the corresponding region. Histogram analysis revealed that the grayscale values ​​before grouting were mainly distributed at both ends, while after grouting they were concentrated in the middle, indicating a more homogeneous formation and a reduction in defects after grouting. Comparing the normal distribution curves before and after grouting, it was noted that the curve after grouting was smoother and had a lower peak value than the curve before grouting. The mathematical expression for the normal distribution curve is: N ( μ , σ ²), where N represents a normal distribution. μ Indicates the expected value. σ It represents the standard deviation.

7. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of the expected value analysis of the state distribution curve in step 5: compare the expected values ​​of the amplitude grayscale images before and after grouting, and the expected value increases or decreases synchronously with the grouting depth; A comparison of the standard deviations of amplitude grayscale images before and after grouting shows that the standard deviation of the amplitude grayscale image after grouting is lower than that before grouting. The standard deviation reflects the dispersion of individuals within a group. The larger the standard deviation, the greater the variation between individuals and the greater the deviation from the mean. Against the background of the amplitude grayscale image, a standard deviation greater than the set value indicates that the image is more uneven and the grayscale value deviates more from the mean. After grouting, the amplitude grayscale image becomes more uniform, but the accuracy of the amplitude grayscale image is lower than before grouting. The standard deviation is a reliable indicator for evaluating the quality of grouting.

8. The method for evaluating the quality of roadbed grouting materials based on ground-penetrating radar image recognition as described in claim 1, characterized in that: The specific process of step 6 is as follows: based on standard deviation δ The grouting quality is defined by the following formula: in σ b and σ a These are the standard deviations of the amplitude grayscale images before and after grouting. Further analysis of the data in the images will determine the grouting quality of the depth profile.

Images