A Method and System for Detecting and Evaluating Settlement of Buried Pipelines Based on Ground Penetrating Radar

Abstract

This invention discloses a method and system for detecting and evaluating the settlement of buried pipelines based on ground-penetrating radar (GPR). The detection method acquires B-Scan mask images of shallow-buried pipelines, extracts the overall and local characteristics of target hyperbolic reflection signal feature points in the images, removes singular points, performs curve fitting on all valid target hyperbolic reflection signal feature points, and calculates and scores the depth and wave velocity values ​​corresponding to the target feature points to obtain the pipeline burial depth. The evaluation method, based on the pipeline settlement detection results, identifies the location of maximum settlement deformation by using the angle formed by the height change between two points on the pipeline, and predicts the location where the risk of fracture in the detected section may occur. The detection system implements this detection method, and the evaluation system implements this evaluation method. This application solves the problem in engineering measurements where hyperbolic reflection signals with characteristic damage caused by electromagnetic wave attenuation and coupling cannot be effectively calculated, significantly improving the data's processability and analyzability.

Description

Technical Field

[0001] This invention belongs to the field of non-destructive surveying and settlement evaluation technology for underground pipelines, and particularly relates to a method and system for detecting and evaluating the settlement of buried pipelines based on ground penetrating radar. Background Technology

[0002] Foundation settlement, especially uneven settlement, can have a significant impact on buried pipelines. Due to the subsidence and movement of the foundation, the pipeline may not only experience pipe wall rupture due to additional stress, but also experience sealing failure due to excessive joint angles, thereby compromising the structural integrity of the pipeline.

[0003] This degradation of pipelines caused by foundation settlement not only shortens their service life but can also lead to serious consequences such as leaks, environmental pollution, and safety accidents (e.g., road collapses). Therefore, how to monitor the condition of buried pipelines under the influence of foundation settlement and conduct quantitative early warning of pipeline settlement disasters has become a pressing technical challenge in the field of pipeline engineering. Summary of the Invention

[0004] One of the objectives of this invention is to provide a method for detecting the settlement of buried pipelines based on ground penetrating radar, which solves the problem that hyperbolic reflection signals with characteristic damage caused by electromagnetic wave attenuation and coupling issues cannot be effectively calculated in engineering measurements.

[0005] The technical solution for realizing the present invention is:

[0006] A method for detecting settlement of buried pipelines based on ground-penetrating radar includes:

[0007] Step 1: Use ground penetrating radar to acquire B-Scan mask images of shallow buried pipelines;

[0008] Step 2: Extract the global and local characteristics of the target hyperbolic reflection signal feature points in the B-Scan mask image, remove the singular points in the feature points, and output the remaining target hyperbolic reflection signal feature points after removing the singular points;

[0009] Step 3: Map the position information of the target hyperbolic reflection signal feature points from the image space to the physical space, establish a pixel transformation matrix, and assign a specific physical meaning to each feature point. All feature point data are saved in the form of digital text.

[0010] Step 4: Based on the physical coordinate information of the feature points of the target hyperbolic reflection signal, an adaptive curve fitting method with a penalty term is used to fit and invert the hyperbolic reflection signal. The optimal solution that meets the conditions is iteratively searched, the fitted hyperbolic reflection signal is output, and it is stored in a structured data format.

[0011] Step 5: Based on the fitted hyperbolic reflection signal obtained in Step 4, select a certain number of target feature points on the fitted hyperbolic reflection signal, calculate the depth value and wave velocity value corresponding to each target feature point, and obtain the depth set and wave velocity set of the target feature points;

[0012] Step 6: Filter and score each element in the depth set and wave velocity set of the target feature points point by point, search for the optimal solution in the solution space, and finally output the optimal radar wave velocity value and the corresponding target depth value.

[0013] In some embodiments, step 5 calculates the depth and wave velocity values ​​corresponding to each target feature point as follows:

[0014] A model coordinate system is established using the ground-penetrating radar time window range as the y-axis range and the sampling distance as the x-axis range. In step 4, a certain number of target feature points are uniformly selected on the fitted hyperbolic reflection signal to obtain the positional coordinates of the target feature points and the vertices of the fitted hyperbolic reflection signal. ;

[0015] Based on the electromagnetic wave imaging principle of ground-penetrating radar and the assumption of medium homogeneity, a functional expression is established relating the target burial depth to the electromagnetic wave two-way propagation time, the pipe radius, the distance to the target feature point, and the distance to the hyperbolic reflection signal vertex:

[0016]

[0017] The burial depth and wave velocity information represented by this feature point are calculated using the following formula:

[0018]

[0019]

[0020] In the formula, the vertex position of the fitted hyperbolic reflection signal is obtained. The corresponding electromagnetic wave two-way propagation time is Fit the i-th target feature point of the hyperbolic reflection signal. The two-way propagation time of electromagnetic waves is The speed of electromagnetic wave propagation in underground media is The pipe radius is The distance between the target feature point and the vertex of the fitted hyperbolic reflection signal is... ;

[0021] Each target feature point can be calculated to have a depth value and a wave velocity value. The calculated maximum and minimum wave velocities among all target feature points constitute the wave velocity range, and the maximum and minimum depths constitute the burial depth range.

[0022] In some embodiments, step 2 extracts the global and local characteristics of the target hyperbolic reflection signal feature points in the B-Scan mask image, and removes singular points from the feature points as follows:

[0023] The center point with the strongest hyperbolic reflection signal of the target is extracted as the feature point of the hyperbolic reflection signal of the target;

[0024] Isolated noise points in the top region of the target hyperbolic reflection signal feature points are removed, and the first row without isolated noise points is taken as the vertex row. The number of target hyperbolic reflection signal feature points in the vertex row is determined. If the number is odd, the center point is selected as the vertex of the target hyperbolic reflection signal feature point. If the number is even, the average value of the two middle feature points is selected as the vertex of the target hyperbolic reflection signal feature point.

[0025] The system checks whether the hyperbolic reflection signal features conform to hyperbolic features by proceeding from the vertex row to the tail region. It also checks whether the ordinate of the feature points near the tail is greater than the coordinate of the feature points near the vertex. If they do not conform, it is considered that there is a warping at the tail or a protrusion in the line body, and these feature points are deleted.

[0026] Based on the range of the target hyperbolic reflection signal feature points from the vertex of the target hyperbolic reflection signal feature points, feature points within 90% of the vertex are retained, while those exceeding the range are deleted.

[0027] In some embodiments, step 3, which maps the position information of the target reflection signal feature points from the image space to the physical space, involves using a feature point mapping algorithm to map the obtained target hyperbolic reflection signal feature points to the inherent coordinate system of the current B-Scan mask image, using the time window range as the y-axis range of the coordinates and the sampling distance as the x-axis range of the coordinates, and establishing a transformation matrix from image pixel coordinates to actual physical coordinates.

[0028] In some embodiments, step 6, which involves point-by-point filtering and scoring of each element in the depth set and wave velocity set of the target feature points, is as follows:

[0029] The error function is constructed based on the detection principle, as follows:

[0030]

[0031] In the formula, Let j be the wave velocity of the j-th error value to be determined in the wave velocity set. Let the distance from the i-th target feature point to the vertex of the fitted hyperbolic reflection signal be the distance. To fit the two-way propagation time of the electromagnetic wave at the vertex position of the hyperbolic reflection signal, Let R be the two-way propagation time of the electromagnetic wave at the i-th target feature point, R be the pipe radius, and N be the number of target feature points.

[0032] When calculating the error of the first target feature point, the wave velocity of the first target feature point is... Substitute the information into the above formula, and then input the location information of each target feature point. Substitute them in separately, at this time The error for each target feature point, obtained using the wave velocity of the first target feature point, was calculated; after summing, the error was... The overall error;

[0033] Repeat the above process to calculate... The error value for each wave velocity is calculated and compared; the wave velocity corresponding to the target feature point with the smallest error value is selected as the optimal wave velocity, and the target depth information D is calculated based on this.

[0034]

[0035] In the formula, To select the optimal wave velocity obtained from the calculation, To fit the two-way propagation time of the electromagnetic wave at the vertex position of the hyperbolic reflection signal.

[0036] In some embodiments, step 4 includes:

[0037] Define the parameterized mathematical model of the target hyperbolic reflection signal, specify the geometric shape characteristics and basic parameter information of the target hyperbolic reflection signal, select the x-coordinate corresponding to the vertex of the feature point of the target hyperbolic reflection signal as the x-coordinate of the center of symmetry, and initialize the current penalty term weight λ and error tolerance ε.

[0038] An objective function is constructed that includes a fitting error term and a constraint penalty term. The penalty term is established to compensate for the difference between the vertex of the fitted hyperbolic reflection signal and the vertex of the feature point of the target hyperbolic reflection signal. The penalty term weight is introduced to control the degree of attention to the vertex of the feature point of the target hyperbolic reflection signal. The weight before the penalty term is automatically iterated, with faster growth when the difference is large and slower growth when the difference is small, so as to meet the requirements of speed and accuracy.

[0039] The algorithm minimizes the objective function and fits the feature points of the target hyperbolic reflection signal within the parameter space, searching for the optimal solution for the fitted hyperbolic reflection signal parameters. Simultaneously, it calculates constraint violations and the computational error between the fitted hyperbolic reflection signal vertex and the target hyperbolic reflection signal feature point vertex under the current parameters. It then determines whether the termination condition is met; if it is, the algorithm terminates; otherwise, it continuously updates and iterates the weights. Once the termination condition is met, it calculates the final goodness of fit and the goodness of fit obtained from the aforementioned weight calculations, determining whether the average value has been reached.

[0040] The second objective of this invention is to provide a method for evaluating the settlement of buried pipelines based on the settlement detection results of ground penetrating radar. This method identifies the location of maximum settlement deformation by using the angle formed by the height change between two points on the pipeline, thereby predicting the location where the risk of fracture in the detected section may occur.

[0041] The technical solution for realizing the present invention is:

[0042] A method for evaluating the settlement of buried pipelines based on ground-penetrating radar includes:

[0043] S1. Using the starting point as a reference point, measure and obtain the ground elevation data at each location along the pipeline at regular intervals.

[0044] S2. Using the starting point as a reference point, at regular intervals along the pipeline, the depth from the ground to the top of the pipe, i.e. the pipeline elevation, is calculated using the aforementioned ground-penetrating radar-based buried pipeline settlement detection method.

[0045] S3. Calculate the actual inclination angle of the pipeline. The calculation method is as follows:

[0046]

[0047] In the formula, This refers to the difference in ground elevation. For pipeline elevation difference; The distance between the nth measuring point and the i-th measuring point; where:

[0048]

[0049] In the formula, Let n be the ground elevation of the nth measuring point. Let i be the ground elevation of the i-th measuring point;

[0050]

[0051] In the formula, Let n be the pipeline elevation at the nth measuring point. Let be the pipeline elevation at the i-th measuring point;

[0052] S4. Evaluate the settlement risk. If the pipeline deflection angle exceeds the design allowable angle, it is considered a high-risk area for settlement.

[0053] In some embodiments, during detection in S1, the ground elevation at each location is measured using a level and a leveling rod at 1m intervals along the line.

[0054] In some embodiments, ground-penetrating radar is used in S2. During detection, ground-penetrating radar is used to measure the depth from the ground to the top of the pipe at each 1m interval along the line, which is the pipeline elevation.

[0055] The third objective of this invention is to provide a ground-penetrating radar-based buried pipeline settlement detection system, including a ground-penetrating radar control unit, a high-precision odometer wheel, a transmitting antenna, and a receiving antenna. A data transmission line connects the ground-penetrating radar control unit and a high-performance data processing unit. The depth signal of the buried pipeline is displayed in real time on the display unit. When the high-performance data processing unit executes, it implements the ground-penetrating radar-based buried pipeline settlement detection method.

[0056] The fourth objective of this invention is to provide a ground-penetrating radar-based buried pipeline settlement evaluation system, including a ground-penetrating radar control unit, a high-precision odometer wheel, a transmitting antenna, and a receiving antenna. A data transmission line connects the ground-penetrating radar control unit and a high-performance data processing unit. The depth signal of the buried pipeline is displayed in real time on the display unit. When the high-performance data processing unit executes, it implements the ground-penetrating radar-based buried pipeline settlement evaluation method.

[0057] Compared with existing technologies, the method and system for detecting and evaluating the settlement of buried pipelines based on ground penetrating radar provided in this application have the following advantages:

[0058] (1) This invention addresses the problems of underutilization of ground-penetrating radar B-Scan mask images and the ineffective selection of feature points and large positioning deviations commonly encountered in underground surveying. It extracts the overall and local characteristics of target hyperbolic reflection signal feature points from the B-Scan mask image, removes singular points, and ensures effective utilization of overall features while considering local features. An adaptive hyperbolic fitting method with a penalty term is used to fit curves to all effective target hyperbolic reflection signal feature points. Compared to traditional fitting methods, the reconstruction accuracy of hyperbolic reflection signals is higher, ensuring the effectiveness and robustness of this type of hyperbolic reflection signal. The relative positions of target feature points on the fitted hyperbolic reflection signal are used for inversion, solving the problem of large inversion errors in traditional methods that directly utilize parameters. Simultaneously, the pipeline radius is introduced as a key physical parameter, overcoming the limitation of simplifying the reflector to a point target in traditional methods, significantly improving the accuracy of buried pipeline depth measurement. Field experiments have verified that the wave velocity calculation accuracy of this invention is better than ±7.5%, and the depth estimation error is controlled within ±7.5%. This invention is also applicable to ordinary hyperbolic reflections. In numerical simulations, the accuracy of wave velocity calculation is better than ±3%, and the depth estimation error is controlled within ±3%.

[0059] (2) This invention addresses the technical challenge of accurately extracting feature points of hyperbolic reflection signals from targets in B-Scan mask image processing for ground-penetrating radar. During the feature point extraction process, this invention fully considers the influence of environmental noise and interference factors. Through a dual feature point denoising mechanism (denoising the tail and vertices of the hyperbola), it can automatically remove singular feature points, avoiding the limitations of traditional methods that rely on manual intervention. A unified selection standard is established, improving the accuracy and robustness of feature point extraction from hyperbolic reflection signals. This achieves accurate extraction of feature points from hyperbolic reflection signals in images and efficiently converts image information into digital text information, significantly improving the data's processability and analyzability.

[0060] (3) The adaptive hyperbolic fitting method with a penalty term proposed in this invention addresses the problem of poor inversion accuracy of hyperbolic reflection signals damaged in actual engineering projects. This fitting method not only considers the overall characteristics of the echo signal but also introduces a penalty term to prioritize the local vertex regions with stronger reflections, thereby maximizing the fitting and inversion accuracy of the damaged hyperbolic echo signal. This method uses penalty term weights to control the degree of attention paid to the vertex features of the target hyperbolic reflection signal, automatically optimizing the iterative penalty term coefficients instead of manually preset thresholds, so that each fitted hyperbolic reflection signal can automatically obtain the optimal penalty term weight configuration based on its own characteristics. This adaptive mechanism ensures that local features are prioritized and the time for obtaining accurate feature points is achieved, while also ensuring that the overall fitting goodness is optimal. By introducing this mechanism, this method can automatically identify and compensate for curve distortion caused by scattering, attenuation, and other problems during electromagnetic wave propagation, controlling the fitting error within a preset threshold range and finding the hyperbolic reflection signal that best approximates reality.

[0061] (4) Based on the principle of ground-penetrating radar electromagnetic wave imaging and the assumption of medium homogeneity, this invention establishes a calculation model related to the target burial depth and the electromagnetic wave two-way propagation time, the pipe radius, the distance of the target feature point to the hyperbolic reflection signal vertex, and so on, to obtain the depth set and wave velocity set of the target feature points. This model, by comprehensively considering the geometric characteristics of the pipe and the electromagnetic wave reflection characteristics, achieves accurate inversion of the spatial position and structural parameters of the pipe. At the same time, this invention adopts a dynamic sampling strategy, using parameters such as sampling interval, number of sampling channels and sampling time to replace the traditional fixed sampling distance and sampling time, thereby theoretically realizing efficient calculation of B-Scan images of arbitrary size, getting rid of the limitation of fixed-size B-Scan image calculation in traditional methods; this invention constructs an objective function based on the principle of electromagnetic wave measurement to systematically screen, score and sort the velocity set of feature point data, thereby accurately selecting the optimal velocity solution and depth solution. Compared with the traditional average value idea, weighted calculation idea and parameter solution idea, this invention completely abandons the limitations of these methods and significantly improves the accuracy and reliability of the calculation results. Traditional methods often fail completely when dealing with incomplete or distorted hyperbolas, leading to significant errors in the calculation results. However, the screening calculation technique of this invention, by introducing an objective function optimization mechanism, can effectively handle complex geological conditions and data anomalies, ensuring the accuracy of the calculation results even in scenarios with incomplete or distorted hyperbolas. For the feature point set in the measured hyperbolic echo signal, a screening function is used to select the optimal wave velocity solution from the point set. Compared with traditional methods, this method does not introduce calculation errors from other points, thus avoiding increased wave velocity errors.

[0062] (5) This invention uses the starting point as a reference point and measures the ground elevation data at regular intervals along the pipeline. The aforementioned ground-penetrating radar-based buried pipeline settlement detection method is used to calculate the depth from the ground to the top of the pipe, i.e., the pipeline elevation. Through quantitative evaluation of the actual inclination angle of the pipeline, it can accurately determine whether the pipeline's deflection angle meets safety requirements, providing early warning of potential pipeline settlement failure risks. This can effectively avoid large-scale repair and environmental remediation costs caused by pipeline damage. This invention achieves precise spatial positioning and multi-dimensional detection and real-time monitoring of the structural status of shallow underground pipelines, significantly improving the efficiency and scientific nature of handling underground pipeline emergencies in complex environments. Simultaneously, it promotes the development of underground pipeline operation and maintenance management towards intelligence and smart technology. Attached Figure Description

[0063] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the technical description will be briefly introduced below.

[0064] Figure 1 A flowchart of the buried pipeline settlement detection method based on ground penetrating radar provided in this application;

[0065] Figure 2 This is a schematic diagram of the integrated assembly and measurement of the ground-penetrating radar module provided in this application;

[0066] Figure 3 A schematic diagram illustrating the multi-parameter ground-penetrating radar target calculation model provided in this application;

[0067] Figure 4 A schematic diagram of the centerline, horizontal line and elevation profile of the buried pipeline provided in this application.

[0068] Explanation of reference numerals in the attached figures:

[0069] 1. Ground penetrating radar control unit; 2. High-precision odometer wheel; 3. Transmitting antenna; 4. Receiving antenna; 5. Data processing unit; 6. Display unit; 7. Distance between transmitting and receiving antennas; 8. Data transmission line; 9. Pipe burial depth x0; 10. Pipe radius R; 11. Ground penetrating radar travel distance S; 12. Electromagnetic wave reflection path d; 13. Hyperbolic imaging distance d h 14. Target hyperbolic signal; 15. Sampling distance S t 16. Radar system sampling window range t t 17. Shallow underground pipelines; 18. Echo delay t of characteristic points on the hyperbolic reflection signal of a target. h 19. Target hyperbolic reflection signal vertex echo delay t o . Detailed Implementation

[0070] The following detailed description provides further details on specific implementation methods.

[0071] Figures 1 to 3 A schematic diagram of a ground-penetrating radar-based method for detecting settlement of buried pipelines is presented. The method includes the following steps:

[0072] Step 1: Measurement and Acquisition of B-Scan Mask Images for Shallow-Buried Pipelines Figure 2 The ground-penetrating radar control unit 1, high-precision odometer wheel 2, transmitting antenna 3, and receiving antenna 4 are integrated and assembled. During assembly, the antenna configuration is optimized to minimize the baseline distance 7 between the transmitting and receiving antennas. A data transmission line 8 connects the ground-penetrating radar control unit 1 and the high-performance data processing unit 5, and the radar wave reflection signal is displayed in real time on the display unit 6. The data processing unit 5 obtains the mask image of the B-Scan image.

[0073] In recent years, with the development of machine learning technology, B-Scan mask images have emerged as a new image format. A B-Scan mask image is a binary image obtained after semantic segmentation via a neural network and extraction of hyperbolic target features. This invention uses B-Scan mask images as the basis for subsequent data processing, effectively solving the problems of underutilization of ground-penetrating radar B-Scan mask images and the ineffective selection of feature points and large positioning errors commonly encountered in underground surveying.

[0074] The second step is to use adaptive feature point extraction and denoising technology to process the B-Scan mask image, extract the overall and local characteristics of the target feature points, monitor and remove singular points in the target hyperbolic reflection signal feature points, and output the denoised target hyperbolic reflection signal feature points.

[0075] First, the center point with the strongest hyperbolic reflection signal of the target is extracted as the feature point of the hyperbolic reflection signal of the target, and is used as the overall characteristic for extraction.

[0076] Secondly, denoising is performed on the feature points of the target hyperbolic reflection signal, mainly targeting the top and tail regions. First, isolated noise points in the top region are removed, and the first row without isolated noise points is selected as the vertex row. The number of vertex rows is then determined; if it is odd, the center point is selected as the vertex; if it is even, the average of the two middle feature points is selected as the vertex. The above extraction of feature points from the target hyperbolic reflection signal is considered a local characteristic.

[0077] Subsequently, the system checks column by column from the vertex row to the tail region to see if it conforms to the hyperbolic feature. Pixel by pixel, it checks whether the ordinate of the feature point near the tail is greater than the coordinate of the feature point near the vertex. If they do not conform, it is considered that there is a warping at the tail or a protrusion in the line body, and these target hyperbolic reflection signal feature points are deleted.

[0078] Finally, the characteristic reflection at the tail of the hyperbola is weak, resulting in a larger fitting error. Based on the number of feature points in the target reflection signal, feature points within 90% of the vertex are retained, while remote feature points are discarded. This completes the adaptive feature point extraction and noise removal.

[0079] The third step involves mapping the obtained target hyperbolic reflection signal feature points to the inherent coordinate system of the current B-Scan mask image using a feature point mapping algorithm. A transformation matrix from image pixel coordinates to actual physical coordinates is established, using the time window range as the y-axis range and the sampling distance as the x-axis range. This maps the positional information of the target reflection signal feature points from image space to physical space, achieving pixel-level high-precision conversion between image pixel coordinates and actual physical coordinates, and assigning a specific physical meaning to each feature point. All feature point data is stored in digital text format, facilitating rapid loading and efficient fitting calculation of subsequent target feature points.

[0080] The fourth step employs an adaptive curve fitting technique with a penalty term. Based on the physical coordinate information of the feature points of the target hyperbolic reflection signal, hyperbolic reflection signal is inverted through hyperbolic fitting. The penalty term weight optimization is automatically iterated, ensuring that each fitted hyperbolic reflection signal automatically obtains the optimal penalty term weight configuration according to its own characteristics. This automatically identifies and compensates for curve distortion caused by scattering and attenuation during electromagnetic wave propagation, ensuring that the local error between the fitted hyperbolic reflection signal and the measured data is controlled within a preset threshold range. The adaptive mechanism guarantees both the accuracy of local features and the overall goodness of fit. The fitting results output key geometric parameters such as the vertex coordinates, aperture parameters, and curvature characteristics of the fitted hyperbolic reflection signal, stored in a structured data format on the processing hard drive, providing accurate geometric feature descriptions for subsequent velocity analysis and depth calculations.

[0081] The adaptive curve fitting process with a penalty term is as follows:

[0082] First, the parameterized mathematical model of the hyperbolic echo signal is defined as follows:

[0083]

[0084] In the formula, x0 is the abscissa of the center point of the fitted hyperbolic reflection signal, a is the length of the real semi-axis, b is the length of the imaginary semi-axis, and y0 is the ordinate of the center point of the fitted hyperbolic reflection signal.

[0085] The x-coordinates corresponding to the vertices of the aforementioned target hyperbolic reflection signal feature points are specified as the x-coordinates of the fitted hyperbolic reflection signal's center of symmetry, strictly ensuring the accuracy of the extraction time. Furthermore, the current penalty term weight λ and error tolerance are initialized. Key factors such as the selection of the vertices of the target hyperbolic reflection signal feature points. The hyperbola is transformed into a hyperbola with known x0, reducing the number of fitting parameters by one and lowering the computational complexity.

[0086] Secondly, for the extracted target hyperbolic reflection signal feature point set data after removing singular points... A target function is constructed that includes a fitting error term and a constraint penalty term. Mean squared error is used to measure the fitting effect, and weighting coefficients are introduced to dynamically adjust the strength of the constraints. The hyperbolic fitting function value is expressed as... , Let be the objective function. The objective function expression can be represented as:

[0087]

[0088]

[0089]

[0090] in, For data residuals; To constrain residuals;

[0091] Subsequently, an adaptive method for updating penalty term weights is defined, which automatically optimizes the iterative penalty term coefficients instead of manually preset thresholds, and dynamically adjusts the penalty term weights according to the degree of constraint. The increase in value. Weight adjustment amount. and deviation Inversely proportional; when the deviation is large, the step size... When growth is rapid and deviation is small, the step size is... The growth is slow. A sensitivity coefficient is introduced. To control response sensitivity, The larger the value, the slower the function increases, and the less likely it is to rapidly increase the weight; conversely, the smaller the value, the slower the function increases, and the less likely it is to increase the weight rapidly. The smaller the value, the faster the weight increases. This is a very small constant, which can be set to 10⁻⁶ to prevent the denominator from being zero. The weight adjustment amount obtained from the k-th calculation. for:

[0092]

[0093] In the formula, The error between the hyperbola vertex calculated using the current hyperbola parameters and the specified feature point vertex.

[0094] Finally, an iterative optimization strategy is adopted, with five iterations. By gradually increasing the weight of the penalty term, the original constrained optimization problem is transformed into a series of unconstrained optimization problems. The objective function is minimized to search within the parameter space. The optimal solution is obtained, while imposing boundary constraints on the parameters: a and b are non-negative real numbers, and y0 is any real number. The parameter values ​​obtained from the k-th iteration. :

[0095]

[0096] Calculate the constraint violation amount, and calculate the vertices of the feature points of the target hyperbolic reflection signal relative to the fitted hyperbolic reflection signal under the current parameters. The calculation error, of which Let y be the longitudinal deviation between the vertex of the fitted hyperbolic reflection signal and the vertex of the feature point of the target hyperbolic reflection signal calculated in the k-th iteration. t The ordinate of the vertex of the feature point of the target hyperbolic reflection signal.

[0097]

[0098] Determine if the termination conditions are met; if they are met, terminate the process. The maximum iteration value is specified. This is the weight value for the kth iteration.

[0099] or

[0100] Based on the above weight update principle, after updating the weights, the iteration continues the calculation.

[0101]

[0102] Once the objective function converges, the difference between the final goodness of fit and the aforementioned goodness of fit is calculated. It is then determined whether the average goodness of fit has been reached. If it has, the fitting result is directly output; otherwise, the fitting has failed, indicating a severe loss of hyperbolic features. This suggests potential issues near the pipe, such as excessively high dielectric conductivity, antenna frequency mismatch, or severe clutter interference. (Goodness of fit) The calculation formula is as follows, where These are the fitted values. The mean, The ordinate of the feature point of the target hyperbolic reflection signal is given.

[0103]

[0104] To supplement the goodness-of-fit metric, a penalty for the number of independent variables is considered to avoid overfitting. To consider the number of independent variables k and the sample size N, the goodness of fit is as follows:

[0105]

[0106] Step 5: Construct a multi-dimensional parameter ground-penetrating radar target calculation model ( Figure 3 By setting key physical parameters in the model, including the pipe's geometric radius R10 and the radar system's time window range t t 16 and the sampling distance S determined by the number of sampling channels and the sampling interval. t 15. Establish a complete computational framework boundary; substitute the target feature points on the fitted hyperbolic signal obtained by the adaptive hyperbolic fitting method with penalty term into the computational model, and calculate the depth value and wave velocity value corresponding to each target feature point through a multi-point computation strategy to obtain the depth set and wave velocity set of the target feature points.

[0107] The model first defines the coordinate system boundaries, with the time window range tt16 defined as the y-axis range and the sampling distance St15 defined as the x-axis range, assuming a uniform medium distribution within the detection area. The sampling distance St is obtained by multiplying the sampling interval and the number of channels. Target feature points (random selection is also possible, but 30 points are used as an example) are uniformly selected on the fitted hyperbolic reflection signal, and the vertices of the fitted hyperbolic reflection signal are obtained to acquire the positional and temporal information of these points. Based on the electromagnetic wave imaging principle of ground-penetrating radar and the assumption of medium homogeneity, the following equations can be established. (Reference) Figure 2 Fitting the vertex position of the hyperbolic reflection signal The burial depth is The corresponding electromagnetic wave two-way propagation time is Fitting hyperbolic reflection signal target feature points The two-way propagation time of electromagnetic waves is The distance to the center of the pipe is The image imaging distance is The speed of electromagnetic wave propagation underground is The radius of the pipe is The distance between the two sampling points is .

[0108]

[0109]

[0110] The above formula uses a combination of measurement and imaging principles for calculation. Therefore, the propagation speed of electromagnetic waves can be expressed as:

[0111]

[0112] Solving the above three equations simultaneously, we can establish a function relating the pipe burial depth to the electromagnetic wave two-way propagation time, the pipe radius, and the distance of the sampling point from the top of the pipe:

[0113]

[0114] This is used to calculate the burial depth and wave velocity information represented by the target feature points:

[0115]

[0116]

[0117] The location and time coordinate information of 30 target feature points Substituting into the above formula, each target feature point and the vertex of the fitted hyperbolic reflection signal can be used to calculate a depth value x0 and a wave velocity value v, resulting in a total of 30 wave velocity and depth values. The maximum and minimum wave velocities among these 30 points constitute the wave velocity range, and the maximum and minimum depths constitute the depth range.

[0118] Step 6: A filtering calculation technique is applied to all target information elements in the obtained depth and wave velocity sets. Each element in the depth and wave velocity sets is filtered and scored individually. By sorting the scores of each element, the optimal solution is searched in the solution space, and finally, the optimal radar wave velocity value and the corresponding target depth value are output. The wave velocity calculation accuracy is better than ±7.5%, and the depth estimation error is controlled within ±7.5%.

[0119] After obtaining the wave velocity set consisting of 30 wave velocity point sets, the wave velocity set is substituted into the defined error function, the expression of which is as follows, where v j S is the wave velocity value of the j-th error value to be determined in the wave velocity set, where j = 1, 2, 3, 4, ... 30. i Let ti be the distance from the i-th target feature point to the vertex of the fitted hyperbolic reflection signal, i = 1, 2, 3, 4, ... 30, and t0 be the two-way propagation time of the electromagnetic wave at the vertex position of the fitted hyperbolic reflection signal. i Let R be the two-way propagation time of the electromagnetic wave at the i-th target feature point, and let R be the radius of the pipe, i=1,2,3,4,...30.

[0120]

[0121] To verify the error of the wave velocity value of the first target feature point, substitute the wave velocity v1 of the first target feature point into the above formula, and input the position information of each target feature point {S1,t1},{S2,t2}...{S 30 ,t 30 Substituting the values ​​into the above formula in double-precision form, a total of 30 DF error values ​​can be obtained. The sum of these 30 error values ​​is taken as the error of the first target feature point; the error values ​​of the remaining 29 points are then calculated. Subsequently, the magnitudes of the 30 error values ​​are compared, and the wave velocity corresponding to the point with the smallest error value is selected as the optimal wave velocity. Based on this, the target depth information D is calculated.

[0122]

[0123] In the formula, v n To filter the optimal wave velocity obtained from the calculation, t0 is the two-way propagation time of the electromagnetic wave at the vertex position of the fitted hyperbolic reflection signal.

[0124] Using the above method on-site, the acquired B-Scan masked image radar data can be processed quickly to obtain precise positioning information of shallow-buried pipelines in real time, enabling accurate location of shallow-buried underground pipelines. Furthermore, historical data can be easily queried and compared with measurement data in a timely manner, achieving the acquisition of pipeline location information, scientific quantitative assessment of pipeline structure, and identification and location of pipeline settlement.

[0125] This invention provides a method for evaluating the settlement of buried pipelines based on ground penetrating radar. This method uses the aforementioned method for detecting the settlement of buried pipelines to calculate the estimated depth of the pipeline.

[0126] The method for evaluating the settlement of buried pipelines based on ground-penetrating radar includes the following steps:

[0127] The first step is to use the starting point as the reference point (zero elevation) during the inspection. At 1-meter intervals along the line, use a level and leveling rod to measure the ground elevation at each location. Figure 4 (For illustrative purposes only).

[0128] The second step involves using ground-penetrating radar (GPR) to measure the depth from the ground to the top of the pipe. During the detection, GPR is used to measure the depth from the ground to the top of the pipe at 1-meter intervals along the pipeline, which is the pipeline elevation. The pipeline elevation is given by the steps of the above-mentioned GPR-based buried pipeline settlement detection method.

[0129] Step 3: Calculate the actual inclination angle of the pipe using formula (1):

[0130] Equation (1)

[0131] Equation (2)

[0132] In the formula, Let n be the ground elevation of the nth measuring point. Let i be the ground elevation of the i-th measuring point; The distance between the nth measuring point and the i-th measuring point;

[0133]

[0134] In the formula, Let n be the pipeline elevation at the nth measuring point. Let be the pipeline elevation at the i-th measuring point.

[0135] Step 4: Settlement Risk Assessment. Fiberglass and ductile iron pipes are allowed a certain angle of inclination during installation. If the angle exceeds the design allowable angle, it is considered a high-risk area for settlement detection. The design allowable angle of inclination can be referenced in the "Code for Construction and Acceptance of Water Supply and Drainage Pipeline Engineering" (GB50268-2008). Simulation can also be used to evaluate the impact of different pipe materials, dimensions, and settlement levels on pipe safety.

[0136] This invention also provides a ground-penetrating radar (GPR)-based system for detecting the settlement of buried pipelines. Based on the GPR module, the GPR-based method for detecting the settlement of buried pipelines is programmed into a computer program and loaded into a high-performance data processing unit 5. When executed, the high-performance data processing unit 5 implements the GPR-based method for detecting the settlement of buried pipelines.

[0137] This invention also provides a ground-penetrating radar (GPR)-based system for evaluating the settlement of buried pipelines. Based on the GPR module, the GPR-based method for evaluating the settlement of buried pipelines is programmed into a computer program and loaded into a high-performance data processing unit 5. When executed, the high-performance data processing unit 5 implements the GPR-based method for evaluating the settlement of buried pipelines.

[0138] The above description is only a specific embodiment of this application, but the protection scope of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application.

Claims

1. A method for detecting settlement of buried pipelines based on ground-penetrating radar, characterized in that, include: Step 1: Use ground penetrating radar to acquire B-Scan mask images of shallow buried pipelines; Step 2: Extract the global and local characteristics of the target hyperbolic reflection signal feature points in the B-Scan mask image, remove the singular points in the feature points, and output the remaining target hyperbolic reflection signal feature points after removing the singular points; Step 3: Map the position information of the target hyperbolic reflection signal feature points from the image space to the physical space, establish a pixel transformation matrix, and assign a specific physical meaning to each feature point. All feature point data are saved in the form of digital text. Step 4: Based on the physical coordinate information of the feature points of the target hyperbolic reflection signal, an adaptive curve fitting method with a penalty term is used to fit and invert the hyperbolic reflection signal. The optimal solution that meets the conditions is iteratively searched, and the fitted hyperbolic reflection signal is output and stored in a structured data format. Step 4 includes: Define the parameterized mathematical model of the target hyperbolic reflection signal, specify the geometric shape characteristics and basic parameter information of the target hyperbolic reflection signal, select the x-coordinate corresponding to the vertex of the feature point of the target hyperbolic reflection signal as the x-coordinate of the center of symmetry, and initialize the current penalty term weight λ and error tolerance ε. An objective function is constructed that includes a fitting error term and a constraint penalty term. The penalty term is established to compensate for the difference between the vertex of the fitted hyperbolic reflection signal and the vertex of the feature point of the target hyperbolic reflection signal. The penalty term weight is introduced to control the degree of attention to the vertex of the feature point of the target hyperbolic reflection signal. The weight before the penalty term is automatically iterated, with faster growth when the difference is large and slower growth when the difference is small, so as to meet the requirements of speed and accuracy. The algorithm minimizes the objective function and fits the feature points of the target hyperbolic reflection signal within the parameter space, searching for the optimal solution for the fitted hyperbolic reflection signal parameters. Simultaneously, it calculates constraint violation, determines the computational error between the fitted hyperbolic reflection signal vertex and the target hyperbolic reflection signal feature point vertex under the current parameters, and checks if the termination condition is met. If the condition is met, the algorithm terminates; otherwise, it continuously updates and iterates the weights. Once the termination condition is met, the final goodness of fit is calculated and compared with the goodness of fit calculated for each weight, determining whether the average value has been reached. Step 5: Based on the fitted hyperbolic reflection signal obtained in Step 4, select a certain number of target feature points on the fitted hyperbolic reflection signal, calculate the depth value and wave velocity value corresponding to each target feature point, and obtain the depth set and wave velocity set of the target feature points; Step 6: Filter and score each element in the depth set and wave velocity set of the target feature points point by point, search for the optimal solution in the solution space, and finally output the optimal radar wave velocity value and the corresponding target depth value.

2. The method for detecting settlement of buried pipelines based on ground-penetrating radar as described in claim 1, characterized in that, Step 5 calculates the depth and wave velocity values ​​corresponding to each target feature point as follows: A model coordinate system is established using the ground-penetrating radar time window range as the y-axis range and the sampling distance as the x-axis range. In step 4, a certain number of target feature points are uniformly selected on the fitted hyperbolic reflection signal to obtain the position and time coordinate information of these target feature points. And the position coordinates of the vertices of the fitted hyperbolic reflection signal; Based on the electromagnetic wave imaging principle of ground-penetrating radar and the assumption of medium homogeneity, a functional expression is established relating the target burial depth to the electromagnetic wave two-way propagation time, the pipe radius, the distance to the target feature point, and the distance to the hyperbolic reflection signal vertex: The burial depth and wave velocity information represented by this feature point are calculated using the following formula: In the formula, the two-way propagation time of the electromagnetic wave corresponding to the vertex position of the fitted hyperbolic reflection signal is obtained as follows: The two-way propagation time of the electromagnetic wave corresponding to the position of the i-th target feature point in the fitted hyperbolic reflection signal is: The speed of electromagnetic wave propagation in underground media is The pipe radius is The distance between the target feature point and the vertex of the fitted hyperbolic reflection signal is... ; Each target feature point is calculated to have a depth value and a wave velocity value. The maximum and minimum wave velocities calculated for all target feature points constitute the wave velocity range, and the maximum and minimum depths constitute the burial depth range.

3. The method for detecting settlement of buried pipelines based on ground-penetrating radar as described in claim 1 or 2, characterized in that, Step 2 extracts the global and local characteristics of the target hyperbolic reflection signal feature points in the B-Scan mask image, and removes singular points from the feature points as follows: The center point with the strongest hyperbolic reflection signal of the target is extracted as the feature point of the hyperbolic reflection signal of the target; Isolated noise points in the top region of the target hyperbolic reflection signal feature points are removed, and the first row without isolated noise points is taken as the vertex row. The number of target hyperbolic reflection signal feature points in the vertex row is determined. If the number is odd, the center point is selected as the vertex of the target hyperbolic reflection signal feature point. If the number is even, the average value of the two middle feature points is selected as the vertex of the target hyperbolic reflection signal feature point. The system checks whether the hyperbolic reflection signal features conform to hyperbolic features by proceeding from the vertex row to the tail region. It also checks whether the ordinate of the feature points near the tail is greater than the coordinate of the feature points near the vertex. If they do not conform, it is considered that there is a warping at the tail or a protrusion in the line body, and these feature points are deleted. Based on the range of the target hyperbolic reflection signal feature points from the vertex of the target hyperbolic reflection signal feature points, feature points within 90% of the vertex are retained, while those exceeding the range are deleted.

4. The method for detecting settlement of buried pipelines based on ground-penetrating radar as described in claim 1 or 2, characterized in that, Step 3 involves mapping the location information of the target reflection signal feature points from the image space to the physical space. The obtained target hyperbolic reflection signal feature points are mapped to the inherent coordinate system of the current B-Scan mask image using a feature point mapping algorithm. The time window range is used as the y-axis range of the coordinates, and the sampling distance is used as the x-axis range of the coordinates. A transformation matrix from image pixel coordinates to actual physical coordinates is then established.

5. The method for detecting settlement of buried pipelines based on ground-penetrating radar as described in claim 1 or 2, characterized in that, Step 6 involves point-by-point filtering and scoring of each element in the depth set and wave velocity set of the target feature points, as follows: The error function is constructed based on the detection principle, as follows: In the formula, Let j be the wave velocity of the j-th error value to be determined in the wave velocity set. Let the distance from the i-th target feature point to the vertex of the fitted hyperbolic reflection signal be the distance. To fit the two-way propagation time of the electromagnetic wave at the vertex position of the hyperbolic reflection signal, Let R be the two-way propagation time of the electromagnetic wave at the i-th target feature point, R be the pipe radius, and N be the number of target feature points. When calculating the error of the first target feature point, the wave velocity of the first target feature point is... Substitute the information into the above formula, and then input the location information of each target feature point. Substitute them in separately, at this time The error of each target feature point, obtained using the wave velocity of the first target feature point, was calculated; wave velocity The error value is the sum of the errors of each target feature point obtained by using the wave velocity of the first target feature point; Repeat the above process to calculate... The error value for each wave velocity is calculated and compared; the wave velocity corresponding to the target feature point with the smallest error value is selected as the optimal wave velocity, and the target depth information D is calculated based on this. In the formula, To select the optimal wave velocity obtained from the calculation, To fit the two-way propagation time of the electromagnetic wave at the vertex position of the hyperbolic reflection signal.

6. A method for evaluating the settlement of buried pipelines based on ground-penetrating radar, characterized in that, include: S1. Using the starting point as a reference point, measure and obtain the ground elevation data at each location along the pipeline at regular intervals. S2. Using the starting point as a reference point, at certain intervals along the pipeline, calculate the depth from the ground to the top of the pipe using the ground-penetrating radar-based buried pipeline settlement detection method as described in any one of claims 1 to 5. S3. Calculate the actual inclination angle of the pipeline. The calculation method is as follows: In the formula, This refers to the difference in ground elevation. For pipeline elevation difference; Let be the distance between the nth measuring point and the i-th measuring point; where: In the formula, Let n be the ground elevation of the nth measuring point. Let i be the ground elevation of the i-th measuring point; In the formula, Let n be the pipeline elevation at the nth measuring point. Let i be the pipeline elevation at the i-th measuring point; S4. Evaluate the settlement risk. If the pipeline deflection angle exceeds the design allowable angle, it is considered a high-risk area for settlement.

7. The method for evaluating the settlement of buried pipelines based on ground-penetrating radar as described in claim 6, characterized in that, During the detection in S1, the ground elevation at each 1m interval along the line is measured using a level and a leveling rod. In S2, the ground penetrating radar method is used, and the depth from the ground to the top of the pipe at each 1m interval along the line is measured using ground penetrating radar.

Images