Tunnel detection method and device based on multi-beam ground penetrating radar and computer equipment

Abstract

The application relates to a tunnel detection method and device based on a multi-beam ground penetrating radar and a computer device. The method comprises the following steps: acquiring multiple frames of initial radar data; for each frame of initial radar data, performing channel interpolation between two adjacent initial detection channels in the initial radar data through wavelet transform to obtain interpolated radar data; performing grid conversion on the interpolated radar data based on a preset resolution to determine a grid range corresponding to the interpolated radar data; for each grid point in the grid range, performing bilinear interpolation on the detection data of two target detection channels in the interpolated radar data to obtain detection data corresponding to the grid point; obtaining target radar data corresponding to the initial radar data based on the detection data of all grid points in the grid range; and performing tunnel detection analysis based on the target radar data corresponding to the multiple frames of initial radar data to obtain a tunnel detection result. The method can improve detection accuracy.

Description

Technical Field

[0001] This application relates to the field of tunnel non-destructive testing and visualization technology, and in particular to a tunnel detection method, device and computer equipment based on multibeam ground penetrating radar. Background Technology

[0002] With the development of infrastructure, the total number of highway and railway tunnels in my country has exceeded 20,000. Projects such as the Sichuan-Tibet Railway, urban subways, and mountain expressways are continuously progressing, and tunnels are developing towards longer, larger, deeper, and more geologically complex structures. Adverse geological conditions such as karst, fault fracture zones, water-rich layers, and weak surrounding rock are common. Water inrush, mudslides, and collapses are the most fatal risks in tunnel construction. Traditional point drilling is insufficient for full coverage and early warning. Ground penetrating radar (GPR), as a non-destructive testing technology, has become an indispensable core technology for the safety monitoring of tunnel projects throughout their entire lifecycle.

[0003] In related technologies, two-dimensional detection is mainly used to ensure tunnel detection efficiency. However, two-dimensional detection has low resolution, which leads to low detection accuracy. Summary of the Invention

[0004] Therefore, it is necessary to provide a tunnel detection method, device, and computer equipment based on multi-beam ground-penetrating radar that can improve detection accuracy in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a tunnel detection method based on multi-beam ground-penetrating radar, comprising:

[0006] Acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points.

[0007] For each frame of the initial radar data, inter-channel interpolation is performed between two adjacent initial detection channels in the initial radar data by wavelet transform to obtain interpolated radar data.

[0008] The interpolated radar data is rasterized based on a preset resolution to determine the raster range corresponding to the interpolated radar data.

[0009] For each grid point within the grid range, bilinear interpolation is performed based on the detection data from the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point; the target detection channel is determined based on the position of the grid point.

[0010] Based on the detection data of all grid points within the grid range, the target radar data corresponding to the initial radar data is obtained;

[0011] Based on the target radar data corresponding to each of the multiple frames of initial radar data, tunnel detection analysis is performed to obtain tunnel detection results.

[0012] Secondly, this application also provides a tunnel detection device based on multi-beam ground-penetrating radar, comprising:

[0013] The acquisition module is used to acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of the initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points.

[0014] The channel interpolation module is used to perform inter-channel interpolation between two adjacent initial detection channels in each frame of the initial radar data by wavelet transform to obtain interpolated radar data.

[0015] The conversion module is used to perform raster conversion on the interpolated radar data based on a preset resolution to determine the raster range corresponding to the interpolated radar data.

[0016] An interpolation module is used to perform bilinear interpolation on each grid point within the grid range, based on the detection data from two target detection channels in the interpolated radar data, to obtain the detection data corresponding to the grid point; the target detection channels are determined based on the position of the grid point.

[0017] The output module is used to obtain the target radar data corresponding to the initial radar data based on the detection data of all grid points within the grid range;

[0018] The analysis module is used to perform tunnel detection analysis based on the target radar data corresponding to each of the multiple frames of initial radar data, and to obtain tunnel detection results.

[0019] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0020] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0021] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0022] The aforementioned tunnel detection method, device, and computer equipment based on multi-beam ground-penetrating radar acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line. First, wavelet transform is used to interpolate adjacent initial detection channels, compensating for data gaps caused by the intervals between multi-beam channels. Then, a grid conversion is performed using a preset resolution to define a regular grid range. Based on the grid point positions, bilinear interpolation is used to calculate the detection values ​​of each grid point corresponding to two sets of target detection channels. Finally, uniform and regular target radar data is generated and combined with multiple frames of data to complete tunnel analysis. In this process, wavelet channel interpolation improves the continuity of multi-beam channel dimensional data, and gridded bilinear interpolation achieves high-precision uniform resampling in the spatial dimension. This fully preserves the details of echo intensity from different detection directions, eliminates data discontinuities caused by channel discrepancies and uneven spatial sampling, and unifies the spatial resolution of the data, giving the multi-frame radar data a consistent and regular data format. This effectively reduces misjudgments caused by spatial distortion and clutter interference in multi-beam radar data, significantly improving the spatial positioning accuracy and identification reliability of tunnel lining, voids, cracks, and other defects. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is an application environment diagram of a tunnel detection method based on multibeam ground-penetrating radar in one embodiment;

[0025] Figure 2 This is a flowchart illustrating a tunnel detection method based on multibeam ground-penetrating radar in one embodiment.

[0026] Figure 3 This is a schematic diagram of the inter-channel interpolation process in one embodiment;

[0027] Figure 4 This is a schematic diagram illustrating the application process of a tunnel detection method in one embodiment;

[0028] Figure 5 This is a schematic diagram of inter-channel interpolation in one embodiment;

[0029] Figure 6 This is a schematic diagram illustrating the effect of 3D imaging versus the original B-scan image in one embodiment;

[0030] Figure 7This is a structural block diagram of a tunnel detection device based on multibeam ground-penetrating radar in one embodiment;

[0031] Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0033] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0034] The tunnel detection method based on multi-beam ground-penetrating radar provided in this application embodiment can be applied to, for example... Figure 1 The application environment shown is illustrated. Terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be integrated onto server 102, or it can be located in the cloud or on another network server.

[0035] The inspection personnel initiate a tunnel detection task through terminal 101. Server 102 responds to the task by acquiring multiple frames of initial radar data collected at multiple locations along the tunnel detection line. Each frame of initial radar data includes detection data corresponding to multiple initial detection channels. Different initial detection channels are used to detect different directions. The detection data includes the echo intensity of multiple detection points. For each frame of initial radar data, wavelet transform is used to perform inter-channel interpolation between two adjacent initial detection channels to obtain interpolated radar data. The interpolated radar data is then rasterized based on a preset resolution. The process involves several steps: first, determining the grid range corresponding to the interpolated radar data; second, for each grid point within the grid range, performing bilinear interpolation based on the detection data from the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point; third, determining the target detection channels based on the grid point's position; fourth, obtaining the target radar data corresponding to the initial radar data based on the detection data from all grid points within the grid range; fifth, performing tunnel detection analysis based on the target radar data corresponding to each of the multiple frames of initial radar data to obtain the tunnel detection results; and finally, returning the obtained tunnel detection results to the terminal 101 for inspection personnel to review.

[0036] Terminal 101 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 102 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0037] In one exemplary embodiment, such as Figure 2 As shown, a tunnel detection method based on multi-beam ground-penetrating radar is provided, which is then applied to... Figure 1 Taking the server in the example, the explanation includes the following steps 201 to 206. Wherein:

[0038] Step 201: Acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points.

[0039] The initial radar data is radar data collected by the multibeam ground-penetrating radar. Furthermore, at each location of the tunnel detection line, the multibeam ground-penetrating radar synchronously emits multiple electromagnetic waves at different angles and receives a corresponding independent detection beam through multiple initial detection channels. All beams radiate outward in a fan shape with the radar probe as the center, thus obtaining a frame of initial radar data.

[0040] Among them, the tunnel detection line refers to the planned trajectory line along the tunnel by the detection equipment equipped with multi-beam ground penetrating radar, which is used to collect radar data at equal intervals or fixed distances along the line to obtain multiple frames of initial radar data.

[0041] In some embodiments, a detection point refers to a discrete depth sampling point that is equally spaced along the underground depth direction of a single detection beam corresponding to a single initial detection channel; each point corresponds to a set of sampling signals reflected back after the electromagnetic wave propagates to the depth interface, and the complete echo curve of a single channel along the depth can be formed by connecting all detection points in series, which is the detection data corresponding to the initial detection channel.

[0042] Echo intensity refers to the amplitude of the reflected echo signal captured by the radar receiving antenna after an electromagnetic wave is incident. It can be understood that the greater the difference in impedance between different media, the more obvious the difference in echo intensity values. It is a core quantitative characteristic parameter for distinguishing tunnel internal defects, lining structures, and surrounding rock media.

[0043] Step 202: For each frame of initial radar data, inter-channel interpolation is performed between two adjacent initial detection channels in the initial radar data using wavelet transform to obtain interpolated radar data.

[0044] Interpolated radar data refers to the complete dataset generated after inter-channel interpolation is completed. It includes both the detection data corresponding to each of the original acquisition multiple initial detection channels, which can be described as the original detection data, and the detection data corresponding to each of all interpolated channels obtained by interpolation, which can be described as the interpolated data. By combining the original detection data and the interpolated data, a multi-channel radar signal that is horizontally continuous and without obvious discontinuities can be obtained, which is interpolated radar data.

[0045] It is understandable that wavelet transform can be used to decompose multi-channel radar echo signals into layers, thereby extracting low-frequency profile features and high-frequency detail abrupt changes in echo intensity between adjacent detection channels. By reconstructing the signals between high and low frequency components, the signal gaps between channels can be smoothly filled, thus achieving a continuous and distortion-free signal transition between adjacent detection channels.

[0046] It should be noted that, due to limitations in the number of radar beams and the spacing of the fan-shaped beams, multi-beam ground-penetrating radar has low resolution in the lateral detection direction of tunnels, and there are data gaps between channels. If channel interpolation is skipped and data raster interpolation is performed directly, the fault features of discrete channels will be forcibly smoothed, and the echo features of small defects such as cavities and cracks will be easily submerged, resulting in data distortion. However, by interpolating between channels, the number of lateral detection channels can be increased, the data gaps between channels can be reduced, and a data foundation can be provided for subsequent raster interpolation.

[0047] For example, assuming the initial radar data includes N initial detection channels, and wavelet transform is used to complete a set of signal interpolation between adjacent channels, that is, one interpolation channel is added between every two initial detection channels, then the total number of channels can be 2N-1 detection channels, including the original N initial detection channels and the N-1 interpolation channels added between each pair of channels, thereby doubling the channel density of the fan-shaped detection dimension.

[0048] Of course, in other embodiments, wavelet transform can be used again to perform inter-channel interpolation processing on the interpolated radar data, and new interpolation channels can be generated in the gaps between adjacent channels based on the previously encrypted channels, thereby further improving the overall density of the detection channels.

[0049] Step 203: Perform raster conversion on the interpolated radar data based on the preset resolution to determine the raster range corresponding to the interpolated radar data.

[0050] The preset resolution refers to the spatial resolution of the rasterized coordinates, which is used to define the size of a single raster to facilitate raster transformation.

[0051] In some embodiments, the raster transformation converts the polar coordinates of all detection points in the interpolated radar data into rectangular coordinates, and then statistically obtains the minimum coordinates (x, y) of the raster matrix. min y min ), and the maximum coordinate (x) max y max This allows us to determine the grid range corresponding to the interpolated radar data.

[0052] It is understandable that all grid points contained within the grid range can correspond to a fixed two-dimensional spatial position in the actual tunnel space. By solving for the echo intensity value corresponding to each grid point, the mapping of the fan-shaped distributed discrete radar sampling points to a regular uniform grid space can be realized.

[0053] Step 204: For each grid point in the grid range, bilinear interpolation is performed based on the detection data of the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point; the target detection channel is determined based on the position of the grid point.

[0054] For each grid point within the grid range, the target detection channel refers to the detection channel adjacent to the actual spatial location corresponding to the grid point. For example, for the first grid point, its corresponding spatial location is the first position, which is located between the first detection channel and the second detection channel. Therefore, the first detection channel and the second detection channel are the two target detection channels for the first grid point. It can be understood that the two target detection channels must be adjacent.

[0055] In some embodiments, for each grid point in the grid range, coordinate transformation can be performed based on the rectangular coordinates of the grid point to obtain the corresponding polar coordinates, and the two target detection channels adjacent to it can be determined based on the polar coordinates.

[0056] In some embodiments, for each grid point within the grid range, the target interpolation channel can be determined by the azimuth difference; specifically, since two target detection channels are necessarily adjacent, adjacent detection channels in the interpolated radar data are combined to obtain multiple adjacent detection channel combinations. For each adjacent detection channel combination, using the right beam as the reference beam, the grid point is converted to polar coordinates (r, ), calculate the azimuth difference of the grid point relative to each adjacent combination of detection channels, i.e., azimuth difference ratio = ( - ) / ( - ),in, The angle of the left detection channel in an adjacent detection channel combination. The angle of the right detection channel in an adjacent detection channel combination is given. If the azimuth difference is less than 1 and greater than or equal to 0, it means that the grid point is located between the two detection channels of the adjacent detection channel combination, that is, the two detection channels of the adjacent detection channel combination are the target interpolation channels of the grid point; otherwise, it means that the azimuth of the grid point is not between the angles of the two channels in this group. Continue to traverse the next group of adjacent detection channel combinations and repeat the polar coordinate transformation and azimuth difference calculation until an adjacent channel combination that meets the azimuth difference value condition is matched, thereby determining the two target detection channels of the grid point.

[0057] Bilinear interpolation is a two-dimensional spatial weighted interpolation algorithm. In this embodiment, linear interpolation is performed separately along the two dimensions of fan-shaped beam angle and detection depth, and then fused. This allows for simultaneous consideration of the spatial correlation between the grid point angle and underground depth. By combining the echo amplitude and spatial weight of adjacent dual channels, continuous and smooth echo numerical solutions are achieved, avoiding the problems of transverse beam faults and abrupt changes in echoes in the depth direction caused by single linear interpolation.

[0058] In some embodiments, bilinear interpolation is performed on the grid points, which involves linear interpolation using echo intensity and linear interpolation using a saliency factor. The saliency factor reflects the contribution weight and importance of the detection data from the two target detection channels to the location corresponding to the grid point. Specifically, for the grid points, two reference points are determined to be at the same depth as the spatial location of the two target detection channels and the grid points. Echo intensity interpolation and saliency factor interpolation are performed on the two reference points respectively. Then, the interpolation results of the two reference points are fused to determine the detection data corresponding to the grid points.

[0059] Step 205: Based on the detection data of all grid points within the grid range, obtain the target radar data corresponding to the initial radar data.

[0060] Specifically, all grid points within the grid range are traversed, and the detection data corresponding to each grid point are arranged in a regular manner to construct a complete two-dimensional grid dataset, which is the target radar data.

[0061] Step 206: Based on the target radar data corresponding to each of the multiple initial radar data frames, perform tunnel detection analysis to obtain tunnel detection results.

[0062] In some embodiments, multiple frames of target radar data can be registered and stitched together according to the spatial acquisition location coordinates to form a complete two-dimensional radar cross-section map covering the entire tunnel detection area. Then, data features of each grid point in the map, such as echo intensity, amplitude change interval, and abnormal area size, are uniformly extracted. Combined with radar echo feature thresholds corresponding to typical defects such as tunnel lining, surrounding rock, voids, cracks, and cavitation, the lateral position, burial depth, and extension range of various defects inside the tunnel are located. Finally, the tunnel detection results containing quantitative information on defect location, defect type, and defect size are output.

[0063] In the aforementioned tunnel detection method based on multi-beam ground-penetrating radar, multiple frames of initial radar data are acquired from multiple locations along the tunnel detection line. First, wavelet transform is used to interpolate adjacent initial detection channels to compensate for data gaps caused by the interval between multi-beam channels. Then, grid conversion is performed in combination with a preset resolution to define a regular grid range. Based on the grid point position matching, bilinear interpolation is performed to calculate the detection value of each grid point. Finally, uniform and regular target radar data is generated and combined with multiple frames of data to complete tunnel analysis. In the above process, wavelet channel interpolation is used to improve the continuity of multi-beam channel dimensional data. Then, gridded bilinear interpolation is used to achieve high-precision uniform resampling in spatial dimension. This not only fully preserves the details of echo intensity in different detection directions and eliminates data discontinuities caused by channel discreteness and uneven spatial sampling, but also unifies the spatial resolution of the data, so that the multi-frame radar data has a consistent and regular data format. This can effectively reduce misjudgments caused by spatial distortion and clutter interference of multi-beam radar data, and significantly improve the spatial positioning accuracy and identification reliability of tunnel lining, voids, cracks and other defects.

[0064] In one exemplary embodiment, such as Figure 3 As shown, two adjacent initial detection channels include a first initial detection channel and a second initial detection channel; inter-channel interpolation includes steps 301 to 304. Wherein:

[0065] Step 301: Based on the detection directions corresponding to the first and second initial detection channels in the initial radar data, determine the interpolation direction corresponding to the interpolation channel.

[0066] In some embodiments, the middle direction of the detection directions corresponding to the first initial detection channel and the second initial detection channel can be used as the interpolation direction corresponding to the interpolation channel.

[0067] For example, if the first initial detection channel corresponds to a 20° direction and the first initial detection channel corresponds to a 40° direction, then the interpolation direction corresponding to the interpolation channel can be a 30° direction.

[0068] In other embodiments, the detection directions corresponding to the first initial detection channel and the second initial detection channel may be weighted according to other rules to obtain the interpolation direction corresponding to the interpolation channel.

[0069] Step 302: Perform wavelet transform on the detection data of the first initial detection channel and the second initial detection channel respectively to obtain the first transformation result of the first initial detection channel and the second transformation result of the second initial detection channel.

[0070] Among them, wavelet transform can perform multi-scale hierarchical decomposition of the detection data of a single detection channel, thereby separating the low-frequency sub-band coefficients that characterize the overall medium distribution (such as stratigraphic contours, large-layer interfaces, etc.) and the high-frequency sub-band coefficients that reflect the structural information of underground targets (such as cracks, cavities, etc.).

[0071] The first transformation result includes low-frequency subband coefficients and multi-layer high-frequency subband coefficients obtained by decomposing the detection data of the first initial detection channel; similarly, the second transformation result includes low-frequency subband coefficients and multi-layer high-frequency subband coefficients obtained by decomposing the detection data of the second initial detection channel.

[0072] Step 303: Based on the first transformation result and the second transformation result, fuse them to obtain the detection data of the interpolation channel.

[0073] In some embodiments, the first transformation result and the second transformation result can be fused by weighted fusion, and wavelet inverse transform can be performed based on the fused result to obtain the detection data of the interpolation channel.

[0074] In other embodiments, the low-frequency subband coefficients in the first and second transformation results can be fused using a weighted fusion method to obtain a low-frequency subband coefficient fusion result. The high-frequency subband coefficients in the first and second transformation results can be fused by taking the largest amplitude to obtain a high-frequency subband coefficient fusion result. Finally, an inverse wavelet transform is performed based on the low-frequency subband coefficient fusion result and the high-frequency subband coefficient fusion result to obtain the detection data of the interpolation channel.

[0075] Step 304: Based on the interpolation direction and detection data corresponding to the interpolation channel, and the detection data corresponding to each of the multiple initial detection channels, the interpolation radar data is obtained by combining them.

[0076] The combination process involves arranging the beam azimuth angles of each detection channel (including the initial detection channel and the interpolation channel) in an orderly manner from small to large, and then matching the interpolation data of the interpolation channel with the angular gaps between two adjacent initial detection channels to obtain an orderly multi-channel dataset, i.e., interpolated radar data.

[0077] It should be noted that the interpolated radar data corresponds one-to-one with the initial radar data, that is, one frame of initial radar data corresponds to one frame of interpolated radar data.

[0078] In this embodiment, the azimuth of the interpolation channel is determined based on the beam angle of the adjacent initial detection channels. Then, the dual-channel echoes are decomposed by wavelet transform and fused to generate interpolation data. Finally, all channels are integrated according to the beam azimuth to obtain interpolation radar data. This allows for the use of wavelet multi-scale decomposition to preserve disease details, reduce interpolation distortion, and improve the lateral resolution and signal continuity of the interpolation radar data.

[0079] In some embodiments, the detection data of the interpolation channel is obtained by fusing the first transformation result and the second transformation result, including:

[0080] The low-frequency subband coefficients of the interpolation channel are obtained by weighting the low-frequency subband coefficients in the first transformation result and the low-frequency subband coefficients in the second transformation result.

[0081] In some embodiments, the weights corresponding to the low-frequency subband coefficients in the first transformation result and the weights corresponding to the low-frequency subband coefficients in the second transformation result can be determined based on the directional deviation of the interpolation direction of the interpolation channel relative to the detection direction corresponding to the first initial detection channel, and the directional deviation of the interpolation direction of the interpolation channel relative to the detection direction corresponding to the second initial detection channel.

[0082] For example, if the interpolation direction of the interpolation channel is located at the center of the detection direction corresponding to the first initial detection channel and the second initial detection channel, then the weights of the low-frequency subband coefficients in the first transformation result and the low-frequency subband coefficients in the second transformation result are equal, that is, both are 0.5.

[0083] When the interpolation direction of the interpolation channel is more biased towards the detection direction of the first initial detection channel, the weight of the low-frequency sub-band coefficient in the first transformation result can also be increased accordingly; correspondingly, the weight of the low-frequency sub-band coefficient in the second transformation result decreases synchronously. The weight is dynamically adjusted according to the azimuth deviation between the interpolation channel and the original channels on both sides. The closer the distance, the greater the weight, thus conforming to the actual spatial distance distribution and making the transition of the echo amplitude obtained by interpolation more natural.

[0084] The high-frequency subband coefficients of the interpolation channel are determined based on the maximum value of the high-frequency subband coefficients in the first transformation result and the second transformation result.

[0085] Understandably, the high-frequency subband coefficient reflects the details of echo abrupt changes caused by minor defects such as cracks, voids, and interface abrupt changes in the tunnel lining. By taking the maximum value, the abnormal features of defects appearing on either side of the two original channels can be completely preserved, preventing the signals of minor defects from being diluted by average during the interpolation process, maximizing the retention of high-frequency details required for defect identification, and reducing the probability of missing minor defects in the tunnel.

[0086] The detection data of the interpolation channel is obtained by performing inverse wavelet transform based on the low-frequency subband coefficients and high-frequency subband coefficients of the interpolation channel.

[0087] Among them, the inverse wavelet transform is the inverse operation of wavelet decomposition, which can reconstruct and restore the low-frequency subband coefficients and high-frequency subband coefficients of the fused interpolation channel into a complete and continuous echo intensity sequence distributed along the depth, which is the detection data of the interpolation channel.

[0088] In the above embodiments, weighted fusion of low-frequency subband coefficients can eliminate overall amplitude jumps and achieve smooth transitions between channels. At the same time, taking the maximum value of the two sets of high-frequency subband coefficients can completely preserve the details of abrupt echoes caused by defects such as cracks and cavities. Finally, the complete echo data of the interpolated channel is restored by wavelet inverse transform, which ensures that the transverse signal is continuous and stable as a whole after interpolation, and does not lose the characteristics of minor defects.

[0089] In some embodiments, for each grid point within the grid range, bilinear interpolation is performed based on the detection data from the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point, including:

[0090] For each grid point within the grid range, determine the two target detection channels that are closest to the spatial location corresponding to the grid point from the channels contained in the interpolated radar data.

[0091] The spatial position corresponding to a grid point is the actual spatial position determined by converting the rectangular coordinates of the grid point to the actual spatial coordinates.

[0092] It is understandable that the two target detection channels closest to the spatial location corresponding to the grid point are the two target detection channels adjacent to the spatial location corresponding to the grid point; in other words, the spatial location corresponding to the grid point is located between the two target detection channels.

[0093] It is understandable that, since the two target detection channels are closest to the spatial location corresponding to the grid point, the detection data corresponding to the two target detection channels can match the real detection data around the grid point to the greatest extent, thereby effectively avoiding interpolation distortion.

[0094] Based on the local variance of the detection data of the two target detection channels in the interpolated radar data, the saliency factors of the detection data of the two target detection channels for grid points are determined.

[0095] The saliency factor is used to measure the importance of the target detection channel's detection data to the interpolated data of the grid points.

[0096] The local variance of the probe data refers to the variance of the probe data within a continuous depth range. The variance is used to characterize the stability of the probe data in that range. The larger the variance, the more signal noise and the lower the reliability. The smaller the variance, the more stable the medium echo and the higher the reference value.

[0097] In some embodiments, for the detection data of each target detection channel, since it contains the echo intensity of multiple detection points, for each detection point, the variance is calculated using the detection data within a fixed-length window containing that detection point to obtain the variance corresponding to that detection point. This process is repeated to calculate the variance corresponding to each detection point. At the same time, the detection depth r corresponding to the grid point is determined. Based on the variances corresponding to the multiple detection points, the variance corresponding to the detection depth r is determined by linear interpolation and used as the significance factor of the detection data of the target detection channel relative to the grid point. Similarly, the significance factor of the detection data of another target detection channel relative to the grid point is obtained.

[0098] Based on the detection data of the target detection channel in the interpolated radar data, the detection data of the two target detection channels are determined to be the basic interpolated data of the grid points.

[0099] In some embodiments, for the detection data of each target detection channel, since it contains the echo intensity of multiple detection points, the detection depth r corresponding to the grid point is determined based on this. According to the echo intensity of multiple detection points, the echo intensity corresponding to the detection depth r is determined by linear interpolation, which serves as the basic interpolation data of the detection data of the target detection channel relative to the grid point. Similarly, the basic interpolation data of the detection data of another target detection channel relative to the grid point is obtained.

[0100] Interpolation is performed based on the saliency factor and the basic detection data to obtain the detection data of the grid points.

[0101] The detection data of the grid points can be obtained by performing bilinear interpolation on the saliency factor of the grid points based on the detection data of the two target detection channels and on the basic interpolation data of the grid points based on the detection data of the two target detection channels.

[0102] In some embodiments, the detection data of the grid points can be determined by the following formula:

[0103] .

[0104] in, .

[0105] in, Representative target detection channel For the basic interpolation data of the grid points, Representative target detection channel For the basic interpolation data of the grid points, Representative target detection channel For the saliency factor of grid points, Representative target detection channel For the saliency factor of grid points, To preset the maximum ratio, This is the preset minimum ratio.

[0106] In the above embodiment, two target detection channels with adjacent spatial positions of grid points are first selected as interpolation benchmarks. The significance factor representing the importance of the data is dynamically calculated based on the local variance of the dual-channel detection data. At the same time, the basic interpolation data at the corresponding depth is extracted. Finally, the interpolation calculation is completed by weighting the significance factor and the basic detection data. By using the local variance to quantify the stability of the channel echo data, the accuracy of the interpolated radar data is effectively improved, and the misjudgment of defects is reduced.

[0107] In some embodiments, the detection data includes the echo intensity of multiple detection points; the two target detection channels include a first target detection channel and a second target detection channel; the saliency factors include a first saliency factor and a second saliency factor; based on the local variance of the detection data of the two target detection channels in the interpolated radar data, the saliency factors of the detection data of the two target detection channels for the grid points are determined, including:

[0108] Based on the detection data from the first target detection channel, determine the variance of each detection point in the first target detection channel within a preset window.

[0109] For example, for each detection point arranged sequentially along the depth of the first target detection channel, a depth window of fixed length is selected as the preset window. With the detection point whose variance is to be calculated as the center of the window, the echo intensity values ​​corresponding to all detection points in the window are extracted, the variance of the set of echo intensity is calculated, and the calculation result is used as the window variance corresponding to the current detection point. By traversing all detection points in the channel, the local variance sequence of each depth position of the entire channel can be obtained.

[0110] In some embodiments, if there are not enough detection points to fill the preset window for the detection points located at the beginning and end boundaries, the boundary value filling method can be used to repeatedly take the echo intensity of the first detection point to fill the missing data on the left side of the window, and then calculate the variance of the echo intensity in the filled window.

[0111] Linear interpolation is performed based on the variance of each detection point in the first target detection channel within a preset window to obtain the first saliency factor of the first interpolation point. The first interpolation point is determined based on the grid points and the first target detection channel.

[0112] In some embodiments, the first interpolation point is determined based on the spatial depth corresponding to the grid point and the detection direction corresponding to the first target detection channel.

[0113] For example, the spatial location corresponding to a grid point is (r, The detection direction of the first target detection channel is Then the spatial position of the first interpolation point can be (r, 1).

[0114] It is understandable that, since the first interpolation point is located in the first target detection channel, after determining the variance corresponding to each detection point in the first target detection channel, the variance of the first interpolation point can be determined by linear interpolation, and thus the variance can be used as the first significance factor of the first interpolation point.

[0115] Based on the detection data from the second target detection channel, determine the variance of each detection point in the second target detection channel within a preset window.

[0116] Similar to the first target detection channel, for each detection point arranged sequentially along the depth of the second target detection channel, a fixed-length depth window is selected as the preset window. Taking the detection point whose variance is to be calculated as the center of the window, the echo intensity values ​​corresponding to all detection points within the window are extracted, and the variance of this set of echo intensities is calculated. The calculation result is used as the window variance corresponding to the current detection point. By traversing all detection points in the channel, the local variance sequence of each depth position of the entire channel can be obtained. For detection points located at the beginning and end boundaries, if there are not enough detection points to fill the preset window, the boundary value filling method can be used. The echo intensity of the first detection point is repeatedly taken to fill the missing data on the left side of the window, and then the variance of the echo intensity in the filled window is calculated.

[0117] Linear interpolation is performed based on the variance of each detection point in the second target detection channel within a preset window to obtain the second significance factor of the second interpolation point. The second interpolation point is determined based on the grid points and the second target detection channel.

[0118] In some embodiments, the second interpolation point is determined based on the spatial depth corresponding to the grid point and the detection direction corresponding to the second target detection channel.

[0119] For example, the spatial location corresponding to a grid point is (r, The detection direction of the second target detection channel is... Then the spatial location of the second interpolation point can be (r, 2).

[0120] It is understandable that, since the second interpolation point is located in the second target detection channel, after determining the variance corresponding to each detection point in the second target detection channel, the variance of the second interpolation point can be determined by linear interpolation, and thus the variance can be used as the second significance factor of the second interpolation point.

[0121] In the above embodiments, the local variance of the dual-channel detection points is calculated point by point by sliding a preset window, which intuitively reflects the noise disturbance level of the echo signal at each depth position. Then, the matching significance factor is obtained by linear interpolation of the corresponding depth position of the grid point, which effectively avoids the interpolation deviation caused by the global fixed weight, and allows the channel with stable signal and low noise to occupy a higher weight in the calculation, effectively suppressing the interference of clutter on the grid point echo value.

[0122] In some embodiments, the basic interpolation data includes first basic interpolation data and second basic interpolation data; based on the detection data of the target detection channel in the interpolated radar data, the basic interpolation data of the two target detection channels for the grid points are determined, including:

[0123] Linear interpolation is performed based on the echo intensity of multiple detection points in the first target detection channel to obtain the first basic interpolation data of the first interpolation point.

[0124] Similar to the significance factor, since the first interpolation point is located in the first target detection channel, after determining the echo intensity corresponding to each detection point in the first target detection channel, the echo intensity of the first interpolation point can be determined by linear interpolation, and thus the echo intensity can be used as the first basic interpolation data of the first interpolation point.

[0125] Linear interpolation is performed based on the echo intensity of multiple detection points in the second target detection channel to obtain the second basic interpolation data of the second interpolation point.

[0126] Similarly, since the second interpolation point is located in the second target detection channel, after determining the echo intensity corresponding to each detection point in the second target detection channel, the echo intensity of the second interpolation point can be determined by linear interpolation, and the echo intensity can be used as the second basic interpolation data of the second interpolation point.

[0127] In the above embodiment, by performing linear interpolation on the detection data of the two target detection channels, the first basic interpolation data and the second basic interpolation data that match the depth of the grid points are determined, providing a data foundation for the subsequent determination of the detection data of the grid points.

[0128] In some embodiments, bilinear interpolation is performed based on a saliency factor and basic probe data to obtain probe data for grid points, including:

[0129] Based on the first significance factor and the second significance factor, the first weight of the first basic interpolation data and the second weight of the second basic interpolation data are determined.

[0130] In some embodiments, the first weight of the first basic interpolation data is The second weight of the second basic interpolation data is .

[0131] in, .

[0132] Representative target detection channel For the saliency factor of grid points, Representative target detection channel For the saliency factor of grid points, To preset the maximum ratio, This is the preset minimum ratio.

[0133] Obviously, within the preset range, the larger the significance factor of the first target detection channel relative to the significance factor of the second target detection channel, the smaller the weight of the first target detection channel corresponding to the first basic interpolation data; conversely, the smaller the significance factor of the first target detection channel relative to the significance factor of the second target detection channel, the larger the weight of the first target detection channel corresponding to the first basic interpolation data.

[0134] Based on the first weight and the second weight, the first basic interpolation data and the second basic interpolation data are weighted to obtain the detection data of the grid points.

[0135] It is understandable that by weighting, dynamic data allocation can be performed by combining the signal stability of the two target detection channels, thereby amplifying the echo contribution of the channel with low noise and high reliability, suppressing the numerical deviation caused by the channel with large clutter interference, and making the detection data of the grid points more consistent with the actual medium distribution.

[0136] In the above embodiments, by assigning weights through a saliency factor, and then using the assigned weights to weight and fuse the two sets of basic interpolation data to obtain grid point detection data, the echo values ​​corresponding to the final grid points can better match the actual medium distribution of the tunnel, thereby improving the accuracy of the interpolation results.

[0137] In some embodiments, before performing raster conversion on the interpolated radar data based on a preset resolution to determine the raster range corresponding to the interpolated radar data, the method further includes:

[0138] Based on the detection directions corresponding to each of the multiple initial detection channels in the initial radar data, the edge detection channels and detection direction intervals are determined.

[0139] For example, if the detection directions of multiple initial detection channels are 10°, 50°, 90°, 130° and 170° respectively, then the edge detection channels are the initial detection channels corresponding to 10° and 170°, with a detection direction interval of 40°.

[0140] Based on the detection data of the detection direction interval and the edge detection channel, auxiliary channel data for the edge detection channel is determined.

[0141] Specifically, the interval angle between the auxiliary channel and the adjacent edge detection channel is half of the channel detection direction interval. At the same time, the complete detection data corresponding to the center line of the edge detection channel is reused as the detection data of the newly added auxiliary channel, i.e., auxiliary channel data.

[0142] Auxiliary channel data is added to the interpolated radar data as supplementary radar data.

[0143] It is understandable that multiple initial detection channels are distributed in a fan shape. During the inter-channel interpolation process, the detection data outside the center line of the edge detection channels is not utilized. If the raster conversion is performed directly, there will be data loss. Therefore, by adding auxiliary channel data for the edge detection channels to the interpolated radar data, the blank data intervals on the left and right boundaries of the fan-shaped beam are filled in, eliminating the defect of missing effective sampling points at the edge of the map, ensuring that there is a complete data source when raster interpolating the edge areas on both sides of the tunnel, and avoiding data distortion in the boundary areas.

[0144] In the above embodiments, the edge detection channels are distinguished and the detection azimuth interval is calculated according to the detection direction of each original detection channel. Then, the auxiliary channel data is supplemented for the edge channels based on the interval, which makes up for the shortcomings of the lack of adjacent channels and the lack of boundary interpolation data on both sides of the multi-beam fan radar, and avoids data gaps and boundary distortion.

[0145] The following description will be based on specific application examples. Please refer to [link / reference]. Figure 4 , Figure 4 This application illustrates an application flow diagram according to an embodiment of the present application, including steps S1 to S6, wherein:

[0146] S1, Acquire multi-beam ground-penetrating radar (GPR) data. For example, when conducting GPR data acquisition, the direction of travel of the tunnel to be inspected is first determined as the positive direction of the survey line. The GPR system is mounted on a high-speed railway train. As the train travels at a constant speed along the designated survey line, the system simultaneously acquires radar data from each directional channel in real time and records mileage information. For any directional channel, each data point recorded by the GPR along the survey line is used as an A-scan signal. Continuously acquired A-scan signals are combined to form B-scan profile data. Taking the N-directional channel as an example, the relevant parameters for this channel are set as follows: the directional angle of the first channel is... 0, beamwidth is BeamAngle, and the spacing between adjacent beam centers is The number of radial sampling points is Samples.

[0147] S2, Radar Data Preprocessing. The acquired ground-penetrating radar echo data cannot be used directly and needs to undergo appropriate preprocessing operations depending on the specific situation, such as zero-bias removal, gain compensation, filtering to remove interference, background removal, etc.

[0148] S3, Pre-interpolation of directional channel data. Because multi-beam ground-penetrating radar has low directional resolution, direct data rasterization interpolation will cause data distortion. Therefore, pre-interpolation is used to improve the initial directional resolution. For example, polar coordinate beamline data interpolation is performed between the centers of two beams. Assuming the number of directional channels is N, after one polar coordinate data interpolation, the number of directional channels becomes 2N-1.

[0149] In some embodiments, the Db4 (Daubechies 4) orthogonal compactly supported wavelet is selected as the wavelet transform decomposition and reconstruction method, and the two A-scan beamline data (B) of the acquired adjacent directional channels are processed. i B i+1 Perform wavelet decomposition on B. i B represents the probe data for channel i. i+1 The probe data representing channel i+1, after decomposition, yields B. i One-dimensional vector cA_B of low-frequency subband coefficients of the beam i , and the one-dimensional vector cD1_B of the high-frequency subband coefficients i cD2_B i cD3_B i ...; at the same time for B i+1 The beam undergoes the same operation to obtain the corresponding one-dimensional vector cA_B of the low-frequency subband coefficients. i+1 , and the one-dimensional vector cD1_B of the high-frequency subband coefficients i+1 cD2_B i+1 cD3_B i+1 ...; Let the center beam A-scan to be interpolated between the two beams be denoted as B. i，i+1 .

[0150] Since low-frequency subband coefficients reflect background or local energy information, such as stratigraphic contours and large stratigraphic interfaces, a weighted average method is used for data fusion to maintain the consistency and smoothness of background or local energy. For example, the weighted average coefficients (w...) are used to... i =0.5, w i+1 =0.5) is a commonly used low-frequency data fusion weighting coefficient. The weighted average is used to obtain the fused center beam B to be interpolated. i,i+1 The low-frequency subband coefficient is cA_B i,i+1 =w i *cA_B i +wi+1 *cA_B i+1 .

[0151] Since high-frequency subband coefficients reflect the structural information of underground targets, such as hyperbolic reflections, noise, ringing, and small targets, the fusion of one-dimensional vector data of high-frequency subband coefficients can be achieved by taking the largest amplitude or the largest energy value to determine the final fused one-dimensional vector data of the high-frequency subband coefficients of the center beam to be interpolated, denoted as cD1_B. i,i+1 cD2_B i,i+1 ...; Finally, the high and low frequency subband coefficients of the center beam to be interpolated are subjected to inverse wavelet decomposition transform to reconstruct the one-dimensional vector signal B of the center beam to be interpolated. i，i+1 .

[0152] In other embodiments, inter-channel interpolation can be performed again as needed after one inter-channel interpolation.

[0153] S4, Outer Auxiliary Channel Interpolation. Since each beam has a beamwidth (Beam Angle), if raster interpolation is performed directly, data outside the outermost beam centerline will be missing. Therefore, an auxiliary beam centerline needs to be added at each end of the array. Due to a lack of prior knowledge, a copy of the outermost beam centerline can be used as the new auxiliary beam centerline, and its beam center angle adjusted. Specifically, the original beam center angles at both ends are increased or decreased by 0.5 * BeamAngle, as shown below. Figure 5 As shown, Figure 5 In the diagram, B1, B2, B3, and B4 are the center lines of the initial detection channels. The center line between any two initial detection channels is the center line of the interpolation channel. The dashed lines outside B1 and B4 are the auxiliary interpolation channels.

[0154] S5, Adaptive radial angle bilinear interpolation. For example, setting the spatial resolution res of the rasterized coordinates, and using the polar coordinates (r, ...) of all channel data. Convert the x-coordinate to rectangular coordinates and statistically obtain the minimum coordinate (x) of the raster matrix. min y min ), and the maximum coordinate (x) max y max The raster range to be interpolated has been determined.

[0155] For the data in the i-th channel B i The data in its left adjacent channel is B. i+1 Iterate through all grid coordinates between the i-th and (i+1)-th beams. Assuming a grid coordinate is (x, y), convert it to polar coordinates (r, y). First, radial linear interpolation is performed, with the right channel data B... iThe interpolation result at the radial coordinate r is the echo intensity v. i Left channel data B i+1 The interpolation result at the radial coordinate r is the echo intensity v. i+1 .

[0156] Calculate the azimuth coordinates of the point to be interpolated The azimuth difference between the interpolation point and the centers of two adjacent beams is used to calculate the azimuth difference relative to each combination of adjacent detection channels, i.e., azimuth difference ratio = ( - ) / ( - ),in, The angle of the left detection channel in an adjacent detection channel combination. The angle of the right-hand detection channel in the adjacent detection channel combination is denoted as 'angle'. If the azimuth difference is less than 1 and greater than or equal to 0, it means that the point to be interpolated is located between the two detection channels of the adjacent detection channel combination. In other words, the two detection channels of the adjacent detection channel combination are the target interpolation channels of the point to be interpolated.

[0157] Calculate B respectively i With B i+1 The local variance of the two-way channel data is used as the significance factor of the echo characteristics, and each significance factor vector Var is constructed according to the original beam echo vector. i With Var i+1 And the significance factor var at the radial coordinate r is obtained by linear interpolation. i =Var i (r) and var i+1 =Var i+1 (r).

[0158] Next, an adaptive azimuth interpolation factor is constructed. By combining radial interpolation, the final interpolated echo intensity at grid coordinates (x, y) is: To constrain the minimum and maximum variances and avoid data singularities, we will... The range of values ​​is limited to and Between, that is:

[0159] .

[0160] in, Representative target detection channel For the basic interpolation data of the grid points, Representative target detection channel For the basic interpolation data of the grid points, Representative target detection channel For the saliency factor of grid points, Representative target detection channel For the saliency factor of grid points, To preset the maximum ratio, This is the preset minimum ratio.

[0161] Finally, following the order of arrival from smallest to largest beam angle, the raster data interpolation is completed by traversing between each pair of adjacent beams.

[0162] S6, 3D stacking of rasterized data. For example, following a pre-planned uniform survey line, rasterized data interpolation is performed frame by frame. The interpolated frame raster data is then stacked along the survey line's forward direction to form 3D raster data. Finally, any B-scan data can be compared with the B-scan profile of the 3D data to verify the correctness of the results. Figure 6 As shown, Figure 6 This paper presents a schematic diagram of the 3D imaging effect of the rasterized data and a schematic diagram of the original B-scan. Figure 6 As can be seen from the comparison of the two images, the 3D imaging obtained by the method of this application can completely reproduce the outline of the underground target and the reflection signal in the original B-scan data.

[0163] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0164] Based on the same inventive concept, this application also provides a tunnel detection device based on multi-beam ground-penetrating radar (MDPR) for implementing the tunnel detection method based on MPR described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more MPR-based tunnel detection device embodiments provided below can be found in the limitations of the MPR-based tunnel detection method described above, and will not be repeated here.

[0165] In one exemplary embodiment, such as Figure 7 As shown, a tunnel detection device 500 based on multi-beam ground-penetrating radar is provided, comprising:

[0166] The acquisition module 501 is used to acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points.

[0167] The channel interpolation module 502 is used to perform inter-channel interpolation between two adjacent initial detection channels in each frame of initial radar data by wavelet transform, so as to obtain interpolated radar data.

[0168] The conversion module 503 is used to perform grid conversion on the interpolated radar data based on a preset resolution to determine the grid range corresponding to the interpolated radar data.

[0169] The interpolation module 504 is used to perform bilinear interpolation on each grid point in the grid range based on the detection data of the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point; the target detection channel is determined based on the position of the grid point.

[0170] The output module 505 is used to obtain the target radar data corresponding to the initial radar data based on the detection data of all grid points within the grid range.

[0171] Analysis module 506 is used to perform tunnel detection analysis based on the target radar data corresponding to each of the multiple initial radar data frames, and to obtain tunnel detection results.

[0172] In some embodiments, two adjacent initial detection channels include a first initial detection channel and a second initial detection channel; the channel interpolation module 502 is used to determine the interpolation direction corresponding to the interpolation channel based on the detection directions corresponding to the first initial detection channel and the second initial detection channel in the initial radar data; perform wavelet transform on the detection data of the first initial detection channel and the second initial detection channel respectively to obtain a first transformation result of the first initial detection channel and a second transformation result of the second initial detection channel; fuse the first transformation result and the second transformation result to obtain the detection data of the interpolation channel; and combine the interpolation direction and detection data corresponding to the interpolation channel, as well as the detection data corresponding to multiple initial detection channels, to obtain interpolated radar data.

[0173] In some embodiments, the channel interpolation module 502 is further configured to: weight the low-frequency subband coefficients in the first transformation result and the low-frequency subband coefficients in the second transformation result to obtain the low-frequency subband coefficients of the interpolation channel; determine the high-frequency subband coefficients of the interpolation channel based on the maximum value of the high-frequency subband coefficients in the first transformation result and the high-frequency subband coefficients in the second transformation result; and perform inverse wavelet transform based on the low-frequency subband coefficients and high-frequency subband coefficients of the interpolation channel to obtain the detection data of the interpolation channel.

[0174] In some embodiments, the interpolation module 504 is configured to, for each grid point within the grid range, determine the two target detection channels closest to the spatial location corresponding to the grid point from the channels contained in the interpolated radar data; determine the saliency factors of the detection data of the two target detection channels for the grid point based on the local variance of the detection data of the two target detection channels in the interpolated radar data; determine the basic interpolated data of the detection data of the two target detection channels for the grid point based on the detection data of the target detection channels in the interpolated radar data; and perform interpolation based on the saliency factors and the basic detection data to obtain the detection data of the grid point.

[0175] In some embodiments, the detection data includes the echo intensity of multiple detection points; the two target detection channels include a first target detection channel and a second target detection channel; the saliency factor includes a first saliency factor and a second saliency factor; the interpolation module 504 is further configured to, based on the detection data of the first target detection channel, determine the variance of each detection point in the first target detection channel within a preset window; perform linear interpolation based on the variance of each detection point in the first target detection channel within the preset window to obtain a first saliency factor of the first interpolation point, the first interpolation point being determined based on grid points and the first target detection channel; based on the detection data of the second target detection channel, determine the variance of each detection point in the second target detection channel within a preset window; perform linear interpolation based on the variance of each detection point in the second target detection channel within the preset window to obtain a second saliency factor of the second interpolation point, the second interpolation point being determined based on grid points and the second target detection channel.

[0176] In some embodiments, the basic interpolation data includes first basic interpolation data and second basic interpolation data; the interpolation module 504 is further configured to perform linear interpolation based on the echo intensity of multiple detection points in the first target detection channel to obtain the first basic interpolation data of the first interpolation point; and to perform linear interpolation based on the echo intensity of multiple detection points in the second target detection channel to obtain the second basic interpolation data of the second interpolation point.

[0177] In some embodiments, the interpolation module 504 is further configured to determine a first weight of the first basic interpolation data and a second weight of the second basic interpolation data based on a first significance factor and a second significance factor; and to weight the first basic interpolation data and the second basic interpolation data based on the first weight and the second weight to obtain the detection data of the grid points.

[0178] In some embodiments, the channel interpolation module 502 is further configured to determine the edge detection channel and the detection direction interval based on the detection directions corresponding to the multiple initial detection channels in the initial radar data; determine the auxiliary channel data for the edge detection channel based on the detection direction interval and the detection data of the edge detection channel; and add the auxiliary channel data as supplementary radar data to the interpolated radar data.

[0179] The modules in the aforementioned tunnel detection device based on multi-beam ground-penetrating radar can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0180] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores data related to tunnel detection. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a tunnel detection method based on multi-beam ground-penetrating radar.

[0181] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0182] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0183] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0184] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0185] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0186] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0187] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0188] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A tunnel detection method based on multi-beam ground-penetrating radar, characterized in that, The method includes: Acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points. For each frame of the initial radar data, inter-channel interpolation is performed between two adjacent initial detection channels in the initial radar data by wavelet transform to obtain interpolated radar data. The interpolated radar data is rasterized based on a preset resolution to determine the raster range corresponding to the interpolated radar data. For each grid point within the grid range, bilinear interpolation is performed based on the detection data from the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point; the target detection channel is determined based on the position of the grid point. Based on the detection data of all grid points within the grid range, the target radar data corresponding to the initial radar data is obtained; Based on the target radar data corresponding to each of the multiple frames of initial radar data, tunnel detection analysis is performed to obtain tunnel detection results.

2. The method according to claim 1, characterized in that, The two adjacent initial detection channels include a first initial detection channel and a second initial detection channel; The step of inter-channel interpolation between two adjacent initial detection channels in the initial radar data using wavelet transform to obtain interpolated radar data includes: Based on the detection directions corresponding to the first initial detection channel and the second initial detection channel in the initial radar data, the interpolation direction corresponding to the interpolation channel is determined. Wavelet transform is performed on the detection data of the first initial detection channel and the second initial detection channel respectively to obtain the first transformation result of the first initial detection channel and the second transformation result of the second initial detection channel; The detection data of the interpolation channel is obtained by fusing the first transformation result and the second transformation result. Interpolated radar data is obtained by combining the interpolation direction and detection data corresponding to the interpolation channel, as well as the detection data corresponding to each of the multiple initial detection channels.

3. The method according to claim 2, characterized in that, The process of fusing the first transformation result and the second transformation result to obtain the detection data of the interpolation channel includes: The low-frequency subband coefficients of the interpolation channel are obtained by weighting the low-frequency subband coefficients in the first transformation result and the low-frequency subband coefficients in the second transformation result. The high-frequency subband coefficients of the interpolation channel are determined based on the maximum value of the high-frequency subband coefficients in the first transformation result and the second transformation result. The detection data of the interpolation channel is obtained by performing inverse wavelet transform based on the low-frequency subband coefficients and high-frequency subband coefficients of the interpolation channel.

4. The method according to claim 1, characterized in that, For each grid point within the grid range, bilinear interpolation is performed based on the detection data from the two target detection channels in the interpolated radar data to obtain the detection data corresponding to the grid point, including: For each grid point within the grid range, determine the two target detection channels closest to the spatial location corresponding to the grid point from the channels contained in the interpolated radar data; Based on the local variance of the detection data of the two target detection channels in the interpolated radar data, the saliency factors of the detection data of the two target detection channels for the grid points are determined respectively. Based on the detection data of the target detection channel in the interpolated radar data, the detection data of the two target detection channels are respectively determined to be the basic interpolation data of the grid points; Interpolation is performed based on the saliency factor and the basic detection data to obtain the detection data of the grid points.

5. The method according to claim 4, characterized in that, The detection data includes the echo intensity of multiple detection points; the two target detection channels include a first target detection channel and a second target detection channel; the significance factor includes a first significance factor and a second significance factor; The determination of the saliency factors of the detection data of the two target detection channels for the grid points based on the local variance of the detection data of the two target detection channels in the interpolated radar data includes: Based on the detection data from the first target detection channel, determine the variance of each detection point in the first target detection channel within a preset window; Linear interpolation is performed based on the variance of each detection point in the first target detection channel within a preset window to obtain the first significance factor of the first interpolation point. The first interpolation point is determined based on the grid points and the first target detection channel. Based on the detection data from the second target detection channel, determine the variance of each detection point in the second target detection channel within a preset window; Linear interpolation is performed based on the variance of each detection point in the second target detection channel within a preset window to obtain the second significance factor of the second interpolation point. The second interpolation point is determined based on the grid points and the second target detection channel.

6. The method according to claim 5, characterized in that, The basic interpolation data includes first basic interpolation data and second basic interpolation data; The determination of the detection data of the two target detection channels based on the target detection channel detection data in the interpolated radar data, respectively, for the basic interpolated data of the grid points, includes: Linear interpolation is performed based on the echo intensity of multiple detection points in the first target detection channel to obtain the first basic interpolation data of the first interpolation point; Linear interpolation is performed based on the echo intensity of multiple detection points in the second target detection channel to obtain the second basic interpolation data of the second interpolation point.

7. The method according to claim 6, characterized in that, The step of performing bilinear interpolation based on the saliency factor and the basic detection data to obtain the detection data of the grid points includes: Based on the first significance factor and the second significance factor, determine the first weight of the first basic interpolation data and the second weight of the second basic interpolation data; Based on the first weight and the second weight, the first basic interpolation data and the second basic interpolation data are weighted to obtain the detection data of the grid points.

8. The method according to any one of claims 1-7, characterized in that, Before performing raster conversion on the interpolated radar data based on a preset resolution to determine the raster range corresponding to the interpolated radar data, the method further includes: Based on the detection directions corresponding to each of the multiple initial detection channels in the initial radar data, the edge detection channels and detection direction intervals are determined. Based on the detection direction interval and the detection data of the edge detection channel, auxiliary channel data for the edge detection channel is determined; The auxiliary channel data is added to the interpolated radar data as supplementary radar data.

9. A tunnel detection device based on multi-beam ground-penetrating radar, characterized in that, The device includes: The acquisition module is used to acquire multiple frames of initial radar data collected at multiple locations along the tunnel detection line; each frame of the initial radar data includes detection data corresponding to multiple initial detection channels; different initial detection channels are used to detect different directions; the detection data includes the echo intensity of multiple detection points. The channel interpolation module is used to perform inter-channel interpolation between two adjacent initial detection channels in each frame of the initial radar data by wavelet transform to obtain interpolated radar data. The conversion module is used to perform raster conversion on the interpolated radar data based on a preset resolution to determine the raster range corresponding to the interpolated radar data. An interpolation module is used to perform bilinear interpolation on each grid point within the grid range, based on the detection data from two target detection channels in the interpolated radar data, to obtain the detection data corresponding to the grid point; the target detection channels are determined based on the position of the grid point. The output module is used to obtain the target radar data corresponding to the initial radar data based on the detection data of all grid points within the grid range; The analysis module is used to perform tunnel detection analysis based on the target radar data corresponding to each of the multiple frames of initial radar data, and to obtain tunnel detection results.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images