Methods, apparatus, equipment and storage media for inversion of buried pipelines

Abstract

This application provides a method, apparatus, equipment, and storage medium for inverting buried pipelines, relating to the field of pipeline data inversion technology, capable of predicting the burial depth and orientation of buried pipelines. The method includes: acquiring magnetic field strength data of a region to be measured, the magnetic field strength data describing the magnetic field distribution of the region, which includes the buried pipeline; preprocessing the magnetic field strength data to obtain a magnetic anomaly image; inputting the magnetic anomaly image into an inversion network for feature extraction and mapping processing to obtain a pipeline simulation model of the buried pipeline, the inversion network using wavelet transform to downsample the magnetic anomaly image to retain high-frequency feature information of the magnetic field gradient; and extracting parameters from the pipeline simulation model to obtain the location information of the buried pipeline.

Description

Technical Field

[0001] This application relates to the field of pipeline data inversion technology, and in particular to a method, apparatus, equipment and storage medium for inverting buried pipelines. Background Technology

[0002] With the acceleration of urbanization and the increasing service life of underground oil and gas pipelines, pipeline damage accidents caused by lost ground markings, missing records, and illegal construction by third parties are frequent. Therefore, accurately obtaining information on the spatial location, depth, and direction of buried pipelines is crucial for achieving pipeline integrity management and accident prevention.

[0003] Among related technologies, magnetic detection has become the most widely used detection method due to its high sensitivity to ferromagnetic pipelines. Inverting underground magnetic structures by measuring surface magnetic anomaly signals is an effective way to obtain three-dimensional information about pipelines. In conventional convolutional neural networks (such as the standard U-Net network), feature maps are downsampled through pooling layers.

[0004] However, the above technical solutions will result in the loss of high-frequency edge features of magnetic anomaly signals, leading to blurred pipeline imaging boundaries and low positioning accuracy for buried pipelines. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, equipment and storage medium for inverting buried pipelines, which can predict the burial depth and direction of buried pipelines and improve the contour restoration and positioning accuracy of buried pipelines.

[0006] To achieve the above objectives, this application adopts the following technical solution: Firstly, a method for locating buried pipelines is provided, the method comprising: Magnetic field strength data of the area to be tested, including buried pipelines, is acquired to describe the magnetic field distribution. The magnetic field strength data is preprocessed to obtain magnetic anomaly images. These images are then input into an inversion network for feature extraction and mapping to obtain a pipeline simulation model. The inversion network uses wavelet transform to downsample the magnetic anomaly images to preserve high-frequency features of the magnetic field gradient.

[0007] The technical solution provided in this application involves inputting a magnetic anomaly image into an inversion network to obtain a pipeline simulation model of the buried pipeline. Parameters are extracted from the pipeline simulation model to obtain the spatial location, burial depth, and orientation information of the buried pipeline, thereby enabling the prediction of the burial depth and orientation. Furthermore, compared to the traditional pooling layer-based downsampling in the U-NET network, this application utilizes wavelet transform for downsampling. Leveraging the multi-resolution analysis characteristics of wavelets, it effectively reduces feature dimensions while fully preserving high-frequency components, thus solving the problem of lost pipeline boundary information due to downsampling and improving the contour restoration accuracy of the buried pipeline.

[0008] One possible implementation of the buried pipeline inversion method provided in this application further includes: updating the parameters of the inversion network based on the pipeline simulation model by constructing a physical constraint loss function. By constructing the physical constraint loss function, the forward modeling operator is embedded into the network training closed loop, forcing the inversion results to not only geometrically approximate the true label, but also conform to the potential field theory described by Maxwell's equations in terms of physical response, fundamentally eliminating non-physical spurious solutions.

[0009] Another possible implementation, based on a pipeline simulation model, involves updating the parameters of the inversion network by constructing a physical constraint loss function. Specifically, this can be achieved as follows: Calculate the geometric loss between the pipeline simulation model and the labeled pipeline model to obtain the data loss. Based on the pipeline simulation model, map the pipeline simulation model to theoretical magnetic field data using a forward modeling operator. Calculate the residual norm between the theoretical magnetic field data and the observed magnetic field data to obtain the physical loss. Based on the data loss and physical loss, calculate the total loss using the physical constraint loss function. Based on the total loss, update the parameters of the inversion network. By verifying the consistency between the theoretical magnetic field data generated by the pipeline simulation model and the observed magnetic field data, the inversion network is forced to output a solution consistent with the potential field theory, effectively avoiding "illusion" errors or non-physical anomalies caused by complex geological conditions, and significantly improving the reliability of the inversion results.

[0010] Another possible implementation, the physical constraint loss function can be expressed as:

[0011] in, For the total loss, For data loss, For physical loss, This represents the weight of physical loss.

[0012] Another possible implementation involves inputting the magnetic anomaly image into an inversion network to obtain a pipeline simulation model of the buried pipeline. Specifically, this can be achieved by: decomposing the magnetic anomaly image into low-frequency and high-frequency components using wavelet transform; upsampling the low-frequency components; and fusing the information obtained from the upsampling with the high-frequency components to obtain the pipeline simulation model. This approach reduces the feature dimensionality while fully preserving the high-frequency components, effectively solving the problem of lost pipeline boundary information caused by downsampling and significantly improving the contour reconstruction accuracy of the buried pipeline.

[0013] Another possible implementation involves upsampling the low-frequency components. Specifically, this can be achieved by convolving the low-frequency components to obtain a deep feature map. Based on this deep feature map, a spatial attention mechanism is used to weight it, resulting in a spatial weight map. This spatial weight map enhances the features of the buried pipeline area and suppresses background interference. The spatial weight map serves as an indicator of the feature map obtained by weighting the deep feature map. Upsampling is then applied to the spatial weight map to restore spatial resolution. This approach adaptively focuses on the effective target area in the magnetic anomaly image, suppressing background field and geological noise interference.

[0014] Another possible implementation involves preprocessing the magnetic field strength data to obtain a magnetic anomaly image. The preprocessing includes: polarization, network interpolation, normalization, and dataset construction.

[0015] Secondly, a buried pipeline inversion device is provided, the device comprising: The acquisition module is used to acquire magnetic field strength data of the area to be measured. The magnetic field strength data is used to describe the magnetic field distribution of the area to be measured, which includes buried pipelines.

[0016] The processing module is used to preprocess the magnetic field strength data to obtain magnetic anomaly images.

[0017] The inversion module is used to input the magnetic anomaly image into the inversion network for feature extraction and mapping processing to obtain the pipeline simulation model of the buried pipeline. The inversion network uses wavelet transform to downsample the magnetic anomaly image to retain the high-frequency feature information of the magnetic field gradient.

[0018] The extraction module is used to extract parameters from the pipeline simulation model to obtain the location information of the buried pipeline.

[0019] One possible implementation of the buried pipeline inversion device provided in this application includes an update module, which is used to update the parameters of the inversion network based on the pipeline simulation model by constructing a physical constraint loss function.

[0020] Another possible implementation involves the update module further used to: calculate the geometric loss between the pipeline simulation model and the labeled pipeline model to obtain the data loss; map the pipeline simulation model to theoretical magnetic field data using a forward modeling operator based on the pipeline simulation model; calculate the residual norm between the theoretical magnetic field data and the observed magnetic field data to obtain the physical loss; calculate the total loss using a physical constraint loss function based on the data loss and the physical loss; and update the parameters of the inversion network based on the total loss.

[0021] Another possible implementation, the physical constraint loss function can be expressed as:

[0022] in, For the total loss, For data loss, For physical loss, This represents the weight of physical loss.

[0023] Another possible implementation, the inversion module, is also used to: decompose the magnetic anomaly image into low-frequency and high-frequency components using wavelet transform. The low-frequency components are then upsampled. The information obtained from the upsampled low-frequency components is then fused with the high-frequency components to obtain a pipeline simulation model for the buried pipeline.

[0024] Another possible implementation, the inversion module, is also used to: convolve the low-frequency components to obtain a deep feature map. Based on the deep feature map, a spatial attention mechanism is used to weight the deep feature map to obtain a spatial weight map, which enhances the features of the buried pipeline area and suppresses background interference. The spatial weight map is used to indicate the feature map obtained by weighting the deep feature map. The spatial weight map is then upsampled to restore the spatial resolution.

[0025] Another possible implementation is that the processing module is also used to: preprocess the magnetic field strength data to obtain a magnetic anomaly image. The data preprocessing includes: polarization, network interpolation, normalization, and dataset construction.

[0026] The technical effects of any implementation method in the second aspect can be found in the technical effects of any implementation method in the first aspect mentioned above, and will not be repeated here.

[0027] Thirdly, a computer device is provided, comprising: a processor and a memory, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the buried pipeline inversion method described above.

[0028] Fourthly, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and the at least one computer program is loaded and executed by a processor to implement the buried pipeline inversion method described above.

[0029] The solutions provided in the third and fourth aspects above are used to implement the method provided in the first aspect above, and their specific implementations will not be elaborated further. The technical effects corresponding to any implementation method in the solutions provided in the third and fourth aspects above can be found in the technical effects corresponding to any implementation method in the first aspect above, and will not be elaborated further here.

[0030] It should be noted that any of the possible implementations of any of the above aspects can be combined, provided that the solutions do not contradict each other. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 A schematic diagram of the system architecture for a buried pipeline inversion method provided in this application embodiment; Figure 2 A schematic flowchart illustrating a buried pipeline inversion method provided in this application embodiment; Figure 3 A schematic flowchart of another buried pipeline inversion method provided in this application embodiment; Figure 4 A schematic flowchart of another buried pipeline inversion method provided in this application embodiment; Figure 5 A schematic diagram of the structure of a buried pipeline inversion system provided in this application embodiment; Figure 6 This is a schematic diagram of the structure of an inversion network provided in an embodiment of this application; Figure 7 A schematic diagram of the structure of a buried pipeline inversion device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0034] In the description of this application, it should be understood that the terms "upper," "lower," "left," "right," "front," "rear," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or relative positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and for simplification, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Unless otherwise specified, the above-mentioned orientational descriptions can be flexibly set in practical applications, provided that the relative positional relationships shown in the accompanying drawings are satisfied.

[0035] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0036] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "communication" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection. They can refer to a direct connection or an indirect connection through an intermediate medium, or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0037] In embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, article, or apparatus that includes that element.

[0038] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0039] In the embodiments of this application, at least one can also be described as one or more, and multiple can be two, three, four or more, and this application does not impose any restrictions.

[0040] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0041] To facilitate understanding, the terms used in the embodiments of this application will be explained first.

[0042] Forward modeling is a physics term primarily used in geophysical exploration research. It involves constructing mathematical or physical models to calculate or observe theoretical values ​​or numerical values ​​of geophysical effects on geological bodies.

[0043] Magnetic bodies: refer to underground geological bodies that possess magnetic properties (such as iron pipes, basalt, and ore bodies).

[0044] Voxel: The smallest unit in three-dimensional space, used to represent three-dimensional objects or data, equivalent to a pixel in a two-dimensional image.

[0045] Magnetic susceptibility is a physical quantity that describes the degree to which a material is magnetized under the influence of an external magnetic field. It represents the ratio of the magnetization intensity to the external magnetic field intensity.

[0046] Magnetic anomaly images are contour maps that show the difference between actual measured values ​​and theoretical values ​​of the Earth's magnetic field, used to reveal the distribution of magnetic bodies in the Earth's crust and geological structural characteristics.

[0047] Pooling is a technique that reduces data dimensionality and extracts key features by aggregating local information. It has wide applications in deep learning and computer resource management.

[0048] Upsampling: In deep learning, upsampling is the operation of scaling up a low-resolution feature map to a high-resolution one.

[0049] Downsampling is a data compression operation in deep learning that expands the receptive field, extracts abstract features, and reduces computation by reducing the spatial resolution of feature maps (i.e., shrinking their size).

[0050] Receptive field: In a convolutional neural network, the receptive field refers to the region of the input image that a point on the feature map can see. In other words, the point on the feature map is calculated from the size of the receptive field region in the input image.

[0051] It should be noted that the information (including but not limited to device information, personal information of the subject, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the subject or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0052] Underground oil and gas pipeline networks serve as the "main arteries" of national energy transportation, and their safe operation is directly related to public safety and energy strategy security. However, with accelerated urbanization and the increasing service life of pipelines, pipeline damage accidents caused by lost ground markings, missing records, and illegal construction by third parties are frequent. Therefore, accurately obtaining information on the spatial location, depth, and direction of buried pipelines is crucial for achieving pipeline integrity management and accident prevention.

[0053] Magnetic detection has become the most widely used detection method due to its high sensitivity to ferromagnetic pipelines. Inverting underground magnetic structures by measuring surface magnetic anomaly signals is an effective way to obtain three-dimensional information about pipelines.

[0054] For example, the industry-standard magnetic inversion method often uses convolutional neural networks to process magnetic anomaly signals, which will be briefly explained below.

[0055] Magnetic anomaly signals are obtained through magnetic detection. Based on these signals, convolutional neural networks (such as the standard U-Net network) are used to process them to obtain predicted images of buried pipelines.

[0056] However, existing conventional convolutional neural networks have the following technical bottlenecks in magnetic inversion: First, traditional pooling operations cause the loss of high-frequency edge features of magnetic anomaly signals, resulting in blurred imaging boundaries and low positioning accuracy of deep small-diameter pipes; Second, pure data-driven models lack geophysical field theory constraints, have weak generalization ability, and are prone to producing false predictions that violate physical laws under complex geological conditions, leading to high risks in engineering excavation verification.

[0057] Based on this, this application provides a buried pipeline inversion method. By inputting a magnetic anomaly image into an inversion network, a pipeline simulation model is obtained. Parameters are extracted from the pipeline simulation model to obtain the spatial location, burial depth, and orientation information of the buried pipeline, enabling the prediction of the burial depth and orientation. Furthermore, compared to the traditional pooling layer-based downsampling in the U-NET network, this application utilizes wavelet transform for downsampling. Leveraging the multi-resolution analysis characteristics of wavelets, high-frequency components are fully preserved while reducing feature dimensions, effectively solving the problem of lost pipeline boundary information due to downsampling and improving the contour restoration accuracy of the buried pipeline.

[0058] The solutions provided by the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0059] Figure 1 A schematic diagram of the system architecture of a buried pipeline inversion method provided in this application embodiment is shown below. Figure 1 As shown, the system includes: a computer device 101, which can be implemented as a terminal or a server.

[0060] When the computer device 101 is implemented as a server, the computer device 101 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0061] When the computer device 101 is implemented as a terminal, the computer device 101 can be a smartphone, tablet computer, laptop computer, desktop computer, etc.

[0062] Optionally, the system described above includes one or more computer devices 101. This application embodiment does not limit the number of computer devices 101.

[0063] Specifically, in Figure 1 In the illustrated system architecture, computer device 101 acquires magnetic field strength data of the area to be measured, performs data preprocessing on the magnetic field strength data to obtain a magnetic anomaly image. Then, the magnetic anomaly image is input into an inversion network for forward propagation processing to obtain a pipeline simulation model of the buried pipeline. Finally, parameters are extracted from the pipeline simulation model to obtain the location information of the buried pipeline. The specific implementation of computer device 101 is described in the following method embodiment.

[0064] Figure 2This is a flowchart illustrating a buried pipeline inversion method provided in an embodiment of this application. The method can be executed by a computer device. The computer device can be, for example,... Figure 1 The computer device 101 shown is shown.

[0065] like Figure 2 As shown in the embodiments of this application, the buried pipeline inversion method may include: Step S201: The computer device acquires the magnetic field strength data of the area to be measured.

[0066] The area to be measured refers to the target geographical area where the spatial location, burial depth, and direction of the buried pipeline need to be determined. The area to be measured includes the buried pipeline.

[0067] Magnetic field strength data is the total scalar sum of the Earth's magnetic field vector and the induced magnetization magnetic field vector of underground magnetic objects. In the embodiments of this application, the magnetic field strength data reflects the total magnetic field strength generated after the Earth's magnetic field is affected by underground magnetic objects (such as iron pipes).

[0068] Specifically, data can be collected in the area to be measured using a magnetometer or a drone to obtain magnetic field strength data. The coordinates of the area to be measured are also recorded.

[0069] The coordinates of the area to be measured refer to the geographical location information of the area to be measured. For example, the coordinates of the area to be measured may include the coordinates of the four corner points of the area to be measured and the spatial location information of each measurement point.

[0070] In step S202, the computer equipment performs data preprocessing on the magnetic field strength data to obtain a magnetic anomaly image.

[0071] Among them, magnetic anomaly images are used to describe the spatial variation of the surface magnetic field intensity caused by buried pipelines relative to the background field.

[0072] Data preprocessing includes, but is not limited to: polarization, networked interpolation, normalization, and dataset construction.

[0073] Among them, pole conversion refers to the data processing method that transforms magnetic anomalies under inclined magnetization into vertical component magnetic anomalies under perpendicular magnetization. Magnetic anomaly pole conversion is equivalent to the magnetic anomalies caused by magnetization conditions when a magnetic body is located in a magnetic pole region, simplifying the morphology of the magnetic anomaly and eliminating the asymmetry of magnetic anomalies caused by the tilt angle and deflection angle of the magnetization field.

[0074] Due to variations in geomagnetic tilt, the magnetization direction of geological bodies changes with latitude. The same geological body will produce different magnetic anomalies under magnetization in different directions, increasing the complexity of interpretation. Except near the geomagnetic poles, most of the Earth's surface is located in mid- to low-latitude regions with significant tilted magnetization. Due to the influence of tilted magnetization, the morphology of magnetic anomalies changes; the center of the anomaly does not correspond directly above the geological body, but rather shifts and moves to varying degrees in the opposite direction to the horizontal projection of the tilted magnetization intensity vector. This presents difficulties and inconveniences in determining the spatial location, morphology, and distribution range of magnetic geological bodies. Therefore, magnetic anomaly polarity is crucial for the analysis and interpretation of magnetic anomalies.

[0075] Specifically, Fourier transform or potential field transformation algorithms can be used to polarize the magnetic field strength data.

[0076] Grid interpolation refers to the process of converting irregularly distributed magnetic field strength data into data on regular network nodes.

[0077] Field survey points for magnetic surveys are often limited by terrain, obstacles, or construction conditions, resulting in an irregular distribution (i.e., data points are not located on neat grid vertices). Subsequent neural network processing (such as U-Net) and visualization imaging require input data to be regular two-dimensional (or three-dimensional) matrices (i.e., pixel / voxel grids like a "chessboard").

[0078] Specifically, gridded interpolation of magnetic field strength data can be performed using methods such as Kriging interpolation, inverse distance weighting, and minimum curvature method.

[0079] Normalization refers to scaling the numerical range of magnetic field strength data to a fixed standard range (such as 0-1) through mathematical transformation.

[0080] Specifically, Z-score standardization and Min-Max normalization can be used to normalize the magnetic field strength data.

[0081] The standard input tensor is generated by performing polarization, gridding interpolation, and normalization on the magnetic field strength data.

[0082] The standard input tensor is a multidimensional array that meets the input requirements of the inversion network.

[0083] Specifically, the dataset can be constructed using the following method: Based on different pipeline geometric parameters (burial depth, diameter, and orientation) and magnetization parameters, a paired "magnetic anomaly map-underground 3D model" sample library is generated using forward modeling, and the data in the sample library is divided into training and validation sets according to a ratio. For example, the data in the sample library can be divided into training and validation sets in an 8:2 ratio.

[0084] The training set refers to the set of data used to train the inversion network.

[0085] The validation set is a set of data used to evaluate the performance of the inversion network.

[0086] In step S203, the computer equipment inputs the magnetic anomaly image into the inversion network for feature extraction and mapping processing to obtain the pipeline simulation model of the buried pipeline.

[0087] The inversion network utilizes wavelet transform to downsample the magnetic anomaly image, preserving high-frequency features of the magnetic field gradient. The inversion network is a deep learning network that inverts a magnetic anomaly image into a pipeline model. For example, the inversion network could be a U-NET neural network.

[0088] The inversion network mainly consists of an encoding module and a decoding module. The input magnetic anomaly image enters the encoding module, and the magnetic anomaly image is decomposed into low-frequency approximate components and high-frequency detail components through wavelet transform.

[0089] The encoder is used for downsampling and feature extraction (through convolution and wavelet transform).

[0090] The decoder is used for upsampling and feature reconstruction.

[0091] A pipeline simulation model refers to a three-dimensional pipeline model of a buried pipeline obtained through an inversion network.

[0092] In some embodiments, the pipeline simulation model may also be referred to as a three-dimensional magnetic susceptibility distribution or probability map, a three-dimensional probability density volume, a three-dimensional physical property distribution model, an underground three-dimensional prediction model, etc. This application does not limit the name of the pipeline simulation model.

[0093] The 3D probability density volume is a 3D matrix output by the inversion network after inverting and imaging a magnetic anomaly image. Each voxel (3D pixel) in the matrix has a probability value between 0 and 1, representing the probability that the location of that voxel belongs to a buried pipeline, as predicted by the inversion network.

[0094] In one possible implementation, the computer device decomposes the magnetic anomaly image into low-frequency and high-frequency components using wavelet transform. The low-frequency components are then upsampled, and the information obtained from the upsampled low-frequency components is fused with the high-frequency components to obtain a pipeline simulation model of the buried pipeline.

[0095] After obtaining the low-frequency and high-frequency components, the low-frequency components are used for the next level of deep feature extraction, while the high-frequency components are directly transmitted to the decoding end through skip connections to ensure that pipeline boundary information is not lost.

[0096] Skip connections refer to directly passing (i.e. skipping) the feature map obtained from a certain layer in the encoder to the corresponding resolution layer in the decoder without going through intermediate operations such as pooling, convolution, and downsampling, and then fusing it with the feature map obtained by the decoder through upsampling.

[0097] Specifically, low-frequency components are convolved to obtain a deep feature map. Based on the deep feature map, a spatial attention mechanism is used to weight the deep feature map to obtain a spatial weight map, which enhances the features of the buried pipeline area and suppresses background interference. The spatial weight map is then upsampled to restore spatial resolution.

[0098] The spatial weight map is used to indicate the feature map obtained by weighting the deep feature map. The spatial weight map represents the importance of each position (pixel) in the deep feature map. High values ​​in the spatial weight map correspond to the buried pipeline target, while low values ​​correspond to background noise. It can suppress environmental background noise and enhance the feature response of the pipeline target area.

[0099] First, spatial attention mechanisms can be used to analyze the spatial dependencies of features, and then spatial weight graphs can be generated based on these spatial dependencies.

[0100] Spatial dependency is used to describe the relationship between a pixel and its surrounding environment.

[0101] By introducing a spatial attention mechanism, the spatial weight matrix can be adaptively learned based on the distribution characteristics of the magnetic anomaly signal. Furthermore, it can adaptively focus on the effective target area within the magnetic anomaly signal while suppressing interference from the environmental background field and random geological noise, ensuring accurate positioning even in low signal-to-noise ratio environments.

[0102] To address the issue of texture loss caused by traditional max pooling, this application embodiment can employ a Haar wavelet basis to perform a two-dimensional discrete wavelet transform (2D-DWT) on the feature map.

[0103] For example, wavelet transform coding is implemented as follows: First, the deep feature map is decomposed into four non-overlapping frequency band components in the frequency domain: low frequency component, horizontal high frequency component, vertical high frequency component, and diagonal high frequency component.

[0104] Low-frequency components are used to characterize the main energy and main contour of deep feature maps.

[0105] The horizontal high-frequency component is used to characterize the high-frequency changes of deep feature maps in the horizontal direction.

[0106] Vertical high-frequency components are used to characterize the high-frequency variations of deep feature maps in the vertical direction.

[0107] Diagonal high-frequency components are used to characterize the high-frequency variations of deep feature maps in the diagonal direction.

[0108] Then, the low-frequency components are selected as the input to the next convolutional layer to achieve dimensionality reduction by halving the resolution while preserving the main energy of the image.

[0109] The three high-frequency components—horizontal, vertical, and diagonal—are concatenated along the channel dimension and then passed to the corresponding layer of the decoder via skip connections. This process is used to restore the sharp edges of the pipeline during the image reconstruction stage.

[0110] By using skip connections to directly transmit the high-frequency details (horizontal, vertical, and diagonal components) retained by the encoder to the decoder, the problem of blurred boundaries and lost textures in the inversion of deep small-diameter pipes is effectively solved, and the reconstruction accuracy of the pipe's geometric contour is significantly improved.

[0111] Specifically, the forward propagation stage of the network, i.e., after the magnetic anomaly map is input into the inversion network, includes three stages: encoding, bottleneck, and decoding. In the encoding stage, the "convolution-wavelet decomposition-gating" operation is performed; in the bottleneck stage, the "feature compression-attention weighting" operation is performed; and in the decoding stage, the "upsampling-high-frequency feature fusion" operation is performed, ultimately outputting the predicted direction and depth information of the pipeline.

[0112] In step S204, the computer equipment extracts parameters from the pipeline simulation model to obtain the location information of the buried pipeline.

[0113] Among them, the location information of buried pipelines is used to describe the spatial location, burial depth and direction of buried pipelines.

[0114] In some embodiments, a threshold segmentation is performed on the pipeline simulation model (i.e., a three-dimensional probability density volume). For example, the threshold is set to 0.5, and regions in the three-dimensional probability density volume with probability values ​​greater than 0.5 are classified as high-probability regions; regions with probability values ​​less than 0.5 are classified as low-density regions.

[0115] Then, the centroid coordinates of the high-probability region are extracted as the center position of the buried pipeline, and the burial depth and direction of the pipeline are calculated based on the voxel distribution.

[0116] In summary, the technical solution provided in this application obtains a pipeline simulation model of the buried pipeline by inputting a magnetic anomaly image into an inversion network. Parameters are extracted from the pipeline simulation model to obtain the spatial location, burial depth, and orientation information of the buried pipeline, thus enabling the prediction of the burial depth and orientation. Furthermore, compared to the traditional pooling layer-based downsampling in the U-NET network, this application utilizes wavelet transform for downsampling. Leveraging the multi-resolution analysis characteristics of wavelets, it effectively reduces feature dimensions while fully preserving high-frequency components, thus solving the problem of lost pipeline boundary information due to downsampling and improving the contour restoration accuracy of the buried pipeline.

[0117] After the computer device obtains the location information of the buried pipeline, the computer device can also update the parameters of the inversion network, that is, step S205 is included after step S204.

[0118] Figure 3 This is a schematic flowchart of another buried pipeline inversion method provided in this application embodiment. The method can be executed by a computer device.

[0119] Step S205: The computer equipment updates the parameters of the inversion network based on the pipeline simulation model by constructing a physical constraint loss function.

[0120] In one possible implementation, the computer calculates the geometric loss between the pipeline simulation model and the labeled pipeline model to obtain the data loss. Based on the pipeline simulation model, a forward modeling operator is used to map the pipeline simulation model to theoretical magnetic field data. The residual norm between the theoretical magnetic field data and the observed magnetic field data is calculated to obtain the physical loss. Based on the data loss and the physical loss, the total loss is calculated using a physical constraint loss function. Based on the total loss, the parameters of the inversion network are updated.

[0121] Among them, the labeled pipeline model refers to the artificially set, known, and "standard answer" three-dimensional digital model of underground pipelines when generating a sample library through forward modeling during the model training phase.

[0122] In some embodiments, the labeled pipe model may also be referred to as a real underground structure. This application does not limit the name of the labeled pipe model.

[0123] Specifically, the mean squared error (MSE) can be used to calculate the error between the pipeline simulation model and the labeled pipeline model to obtain the data loss.

[0124] Then, the pipeline simulation model is input into the forward modeling operator, which maps the pipeline simulation model to theoretical magnetic field data through forward modeling. The forward modeling operator is embedded into the inversion network training loop, which forces the inversion results to not only geometrically approximate the true label, but also conform to the potential field theory described by Maxwell's equations in terms of physical response, thus fundamentally eliminating non-physical spurious solutions.

[0125] A forward modeling operator is a physical mapping function used to simulate geophysical processes. Specifically, a forward modeling operator takes a three-dimensional subsurface physical property model (such as a magnetic susceptibility distribution model of a pipe in space) as input, and then calculates the theoretical magnetic field data that this subsurface model should produce at the location of the observation point on the surface, based on the fundamental physical laws described by Maxwell's equations (or simplified potential field theory, such as the Poisson equation).

[0126] The specific implementation of the forward modeling operator is as follows: based on a convolution algorithm in the frequency domain or spatial domain, a formula for the magnetic field response generated by the magnetic body is defined. For the predicted magnetic susceptibility value of each voxel output by the network, the magnetic induction intensity component generated on the observation surface is calculated by integration, and the responses of all voxels are superimposed to obtain the theoretical magnetic anomaly map of the entire region.

[0127] After obtaining the theoretical magnetic field data, the residual norm between the theoretical and observed magnetic field data is calculated, forcing the inversion network to learn the generation mechanism of the magnetic field and eliminating non-physical spurious solutions. By verifying the consistency between the theoretical magnetic field data generated by the pipeline simulation model and the observed magnetic field data obtained from actual observations, the output of the inversion network is forced to conform to the solution of the potential field theory, effectively avoiding non-physical anomalies caused under complex geological conditions and significantly improving the reliability of the inversion results.

[0128] Non-physical pseudo-solutions refer to false solutions that violate basic physical laws. For example, underground models generated by inversion networks during training or prediction may mathematically fit the observed data, but are physically impossible or untenable.

[0129] Finally, based on the data loss and physical loss, the total loss is calculated according to the physical constraint loss function.

[0130] The physical constraint loss function can be expressed as: .

[0131] in, For the total loss, For data loss, For physical loss, This represents the weight of physical loss.

[0132] After calculating the total loss, the backpropagation algorithm is used to update the inversion network weights based on the total loss until the loss function converges, and the trained model parameters are saved.

[0133] For example, the loss function is considered to have converged when the total loss decreases by less than a preset threshold over multiple consecutive training epochs. The preset threshold can be set according to actual circumstances, and this application does not impose any limitations on it.

[0134] Figure 4 This is a schematic flowchart of another buried pipeline inversion method provided in this application embodiment. The method can be executed by a computer device.

[0135] Step S401: The computer device acquires magnetic field strength data and performs data preprocessing on the magnetic field strength data to obtain a magnetic anomaly image.

[0136] For a description of this step, please refer to steps S201 and S202; further details will not be provided here.

[0137] In step S402, the computer device performs wavelet transform encoding.

[0138] For a description of this step, please refer to step S203; it will not be elaborated further here.

[0139] In step S403, the computer device performs spatial attention focusing.

[0140] For a description of this step, please refer to step S203; it will not be elaborated further here.

[0141] In step S404, the computer device performs feature reconstruction and decoding to obtain the pipeline simulation model.

[0142] For a description of this step, please refer to step S203; it will not be elaborated further here.

[0143] In step S405, the computer equipment inputs the pipeline simulation model into the forward simulation layer to obtain theoretical magnetic field data.

[0144] The forward modeling layer is used to map the pipeline simulation model to theoretical magnetic field data through forward modeling.

[0145] Specifically, an untrainable forward simulation layer is constructed, and based on the SimPEG framework, the pipeline simulation model is mapped to theoretical magnetic field data.

[0146] Step S406: The computer equipment calculates the geometric loss between the pipeline simulation model and the labeled pipeline model to obtain the data loss.

[0147] For a description of this step, please refer to step S205; it will not be elaborated upon here.

[0148] In step S407, the computer equipment calculates the residual norm between the theoretical magnetic field data and the observed magnetic field data to obtain the physical loss.

[0149] For a description of this step, please refer to step S205; it will not be elaborated upon here.

[0150] In step S408, the computer device calculates the total loss based on data loss and physical loss using a physical constraint loss function.

[0151] For a description of this step, please refer to step S205; it will not be elaborated upon here.

[0152] Step S409: The computer device updates the parameters of the inversion network based on the total loss.

[0153] For a description of this step, please refer to step S205; it will not be elaborated upon here.

[0154] In step S410, the computer equipment locates and extracts parameters of the buried pipeline based on the pipeline simulation model.

[0155] For a description of this step, please refer to step S204; it will not be elaborated further here.

[0156] Figure 5 This is a schematic diagram of the structure of a buried pipeline inversion system provided in an embodiment of this application.

[0157] like Figure 5 As shown, the buried pipeline inversion system of this application includes the following modules: data acquisition and preprocessing module 501, wavelet transform coding module 502, spatial attention bottleneck module 503, feature reconstruction decoding module 504, and physical constraint training module 505.

[0158] The data acquisition and preprocessing module 501 is used to acquire magnetic field strength data of the area to be measured and to preprocess the magnetic field strength data to obtain a magnetic anomaly image.

[0159] The wavelet transform coding module 502 is used to extract multi-scale features from the magnetic anomaly image of the input inversion network, and uses discrete wavelet transform to replace the traditional pooling layer for downsampling.

[0160] Multi-scale feature extraction can be understood as decomposing a magnetic anomaly image into different frequency bands through wavelet transform.

[0161] The spatial attention bottleneck module 503 is located between the encoder and decoder in the inversion network. It is used to generate a spatial weight map and automatically focus on the sparse target area where the buried pipeline is located.

[0162] The sparse target area can be understood as the discrete local space occupied by the buried pipeline itself in the area to be measured by magnetic detection, which is very limited relative to the vast background area.

[0163] The feature reconstruction decoding module 504 is used to map low-dimensional features back to the original size and fuse high-frequency details through skip connections to output a pipeline simulation model.

[0164] Low-dimensional features refer to the feature maps that reach the end of the encoder after a magnetic anomaly image has undergone multiple layers of encoding.

[0165] The physical constraint training module 505 includes a forward modeling operator, which is used to calculate the theoretical magnetic field response corresponding to the pipeline simulation model and to feed back the difference between the theoretical magnetic field data and the observed magnetic field data in order to optimize the parameters of the inversion network.

[0166] Figure 6 This is a schematic diagram of the structure of an inversion network provided in an embodiment of this application.

[0167] like Figure 6 As shown, after the magnetic anomaly image is input into the inversion network, it is first subjected to two-dimensional convolution and then enters the encoder. At the encoder end, it goes through four encoding stages: encoding stage 1, encoding stage 2, encoding stage 3, and encoding stage 4.

[0168] In each encoding stage, the magnetic anomaly image is decomposed into low-frequency and high-frequency components using wavelet transform. The high-frequency components are then connected to the corresponding layers of the decoder (e.g., encoding stage 1 corresponds to decoding stage 4, encoding stage 2 corresponds to decoding stage 3, encoding stage 3 corresponds to decoding stage 2, and encoding stage 4 corresponds to decoding stage 1). The low-frequency components are then input into the next convolutional layer. This process is repeated for the low-frequency components input into the next convolutional layer (connecting the high-frequency components to the corresponding layers of the decoder and inputting the low-frequency components into the next convolutional layer), finally yielding the encoder output.

[0169] Then, feature extraction is performed on the encoder output. After feature extraction, the output enters the decoder (which goes through four decoding stages at the encoder: decoding stage 1, decoding stage 2, decoding stage 3, and decoding stage 4). After the upsampling modules of decoding stages 1, 2, 3, and 4 process the features sequentially and pass them up, the decoder output is obtained.

[0170] Spatial attention is used to process the decoder output to suppress environmental background noise and enhance the feature response of the pipeline target area. Finally, the predicted burial depth information of the buried pipeline is output through ReLU+0.5, and the prediction confidence is output through Sigmoid. Both ReLU and Sigmoid functions are activation functions.

[0171] In summary, the technical solution provided in this application obtains a pipeline simulation model of the buried pipeline by inputting a magnetic anomaly image into an inversion network. Parameters are extracted from the pipeline simulation model to obtain the spatial location, burial depth, and orientation information of the buried pipeline, thus enabling the prediction of the burial depth and orientation. Furthermore, compared to the traditional pooling layer-based downsampling in the U-NET network, this application utilizes wavelet transform for downsampling. Leveraging the multi-resolution analysis characteristics of wavelets, it effectively reduces feature dimensions while fully preserving high-frequency components, thus solving the problem of lost pipeline boundary information due to downsampling and improving the contour restoration accuracy of the buried pipeline.

[0172] Furthermore, it eliminates the reliance on empirical parameters such as structural indices found in traditional methods (such as Eulerian deconvolution), enabling direct mapping from magnetic survey data to underground three-dimensional structures. It can also simultaneously output the horizontal position, burial depth, and geometric orientation of buried pipelines, providing efficient and intuitive decision-making support for the digital management and excavation maintenance of underground pipe networks.

[0173] Through actual testing and verification, the method provided in this application embodiment has achieved sub-meter level positioning accuracy under complex working conditions and can output three-dimensional imaging results with confidence. It effectively solves the problems of "inaccurate depth determination and unclear direction" of traditional methods, and provides a highly reliable technical means for non-destructive detection and digital management of urban underground pipe networks.

[0174] The foregoing mainly describes the solution provided in this application. Accordingly, this application also provides a buried pipeline inversion device, which is used to implement the above-described method embodiments.

[0175] like Figure 7 The schematic diagram shown illustrates the structure of a buried pipeline inversion device. The device may include an acquisition module 701, a processing module 702, an inversion module 703, and an extraction module 704. The acquisition module 701 is used for execution... Figure 2 and Figure 3 The illustrated method includes step S201; the processing module 702 is used to execute... Figure 2 and Figure 3 The illustrated method includes step S202; the inversion module 703 is used to execute... Figure 2 and Figure 3 The illustrated method includes step S203; the extraction module 704 is used to perform this operation. Figure 2 and Figure 3 The illustrated method includes step S204.

[0176] In some embodiments, the buried pipeline inversion device includes hardware structures and / or software modules corresponding to the execution of each function in order to achieve the above-described functions. Those skilled in the art will readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0177] This application embodiment can divide the buried pipeline inversion device into functional modules according to the above method embodiment. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one buried pipeline inversion module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0178] like Figure 8 As shown, the computer device provided in this application embodiment may include a processor 801, a bus 802, a communication interface 803, and a memory 804. The processor 801, memory 804, and communication interface 803 communicate with each other via the bus 802. It should be understood that this application does not limit the number of processors and memories in the network device.

[0179] The 802 bus can be a PCI bus, an Extended Industry Standard Architecture (EISA) bus, or a UB bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 8 The bus 802 may be represented by a single line, but this does not mean that there is only one bus or one type of bus. The bus 802 may include a path for transmitting information between various components of the network device (e.g., memory 804, processor 801, communication interface 803).

[0180] Processor 801 may include any one or more processors such as CPU, graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0181] The memory 804 may include volatile memory, such as random access memory (RAM). The processor 801 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0182] The communication interface 803 uses transceiver modules, such as, but not limited to, network interface cards and transceivers, to enable communication between network devices and other devices or communication networks.

[0183] The memory 804 stores executable program code, and the processor 801 executes the executable program code to implement the functions of the aforementioned method embodiments. That is, the memory 804 stores instructions for executing the above-described buried pipeline inversion method.

[0184] In another aspect, a computer-readable storage medium is provided, which stores at least one computer program, which is loaded and executed by a processor to implement the buried pipeline inversion method provided in the above-described method embodiments.

[0185] In another aspect, a computer program product is provided, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the buried pipeline inversion method provided in the above-described method embodiments is implemented.

[0186] Through the above description of the implementation methods, those skilled in the art will clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the module can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, modules, and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0187] Since the buried pipeline inversion module, computer-readable storage medium, and computer program product in the embodiments of the present invention can be applied to the above method, the technical effects that can be obtained can also be referred to the above method embodiments. The embodiments of the present invention will not be repeated here.

[0188] The method steps in this embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in a network device. Of course, the processor and storage medium can also exist as discrete components in the network device.

[0189] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer programs or instructions. When a computer program or instruction is loaded and executed on a computer, the processes or functions of the embodiments of this application are performed, in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, or other programmable module. The computer program or instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, a computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video disc (DVD); or it can be a semiconductor medium, such as a solid-state drive (SSD).

[0190] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations 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 scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for inverting buried pipelines, characterized in that, The method includes: Acquire magnetic field strength data of the area to be tested, the magnetic field strength data being used to describe the magnetic field distribution of the area to be tested, the area to be tested including buried pipelines; The magnetic field strength data is preprocessed to obtain a magnetic anomaly image; The magnetic anomaly image is input into an inversion network for feature extraction and mapping to obtain a pipeline simulation model of the buried pipeline. The inversion network uses wavelet transform to downsample the magnetic anomaly image to retain high-frequency feature information of the magnetic field gradient. The parameters of the pipeline simulation model are extracted to obtain the location information of the buried pipeline.

2. The method according to claim 1, characterized in that, The method further includes: Based on the pipeline simulation model, the parameters of the inversion network are updated by constructing a physical constraint loss function.

3. The method according to claim 2, characterized in that, The process of updating the parameters of the inversion network based on the pipeline simulation model by constructing a physical constraint loss function includes: Calculate the geometric loss between the pipeline simulation model and the labeled pipeline model to obtain the data loss; Based on the pipeline simulation model, the pipeline simulation model is mapped to theoretical magnetic field data through forward modeling operators; Calculate the residual norm between the theoretical magnetic field data and the observed magnetic field data to obtain the physical loss; Based on the data loss and the physical loss, the total loss is calculated using the physical constraint loss function; Based on the total loss, the parameters of the inversion network are updated.

4. The method according to claim 3, characterized in that, The physical constraint loss function is expressed as follows: in, For the total loss, For the data loss, For the physical loss, This represents the weight of physical loss.

5. The method according to claim 1, characterized in that, The step of inputting the magnetic anomaly image into the inversion network to obtain a pipeline simulation model of the buried pipeline includes: Based on the magnetic anomaly image, the magnetic anomaly image is decomposed into low-frequency components and high-frequency components by wavelet transform. The low-frequency components are upsampled; The information obtained by upsampling the low-frequency components is fused with the high-frequency components to obtain the pipeline simulation model of the buried pipeline.

6. The method according to claim 5, characterized in that, The upsampling of the low-frequency component includes: Convolution is performed on the low-frequency components to obtain a deep feature map; Based on the deep feature map, a spatial attention mechanism is used to weight the deep feature map to obtain a spatial weight map, which enhances the features of the buried pipeline area and suppresses background interference. The spatial weight map is used to indicate the feature map after weighting the deep feature map. The spatial weight map is upsampled to restore the spatial resolution.

7. The method according to claim 1, characterized in that, The magnetic field strength data is preprocessed to obtain a magnetic anomaly image. The data preprocessing includes: polarization, network interpolation, normalization, and dataset construction.

8. A buried pipeline inversion device, characterized in that, The device includes: The acquisition module is used to acquire magnetic field strength data of the area to be measured. The magnetic field strength data is used to describe the magnetic field distribution of the area to be measured, which includes buried pipelines. The processing module is used to preprocess the magnetic field strength data to obtain a magnetic anomaly image; The inversion module is used to input the magnetic anomaly image into the inversion network for feature extraction and mapping processing to obtain the pipeline simulation model of the buried pipeline. The inversion network uses wavelet transform to downsample the magnetic anomaly image to retain the high-frequency feature information of the magnetic field gradient. The extraction module is used to extract parameters from the pipeline simulation model to obtain the location information of the buried pipeline.

9. A computer device, characterized in that, The computer device includes a processor and a memory, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the buried pipeline inversion method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to implement the buried pipeline inversion method as described in any one of claims 1 to 7.

Images