Underground cavity identification method and system in combination with multi-scale fusion and convolutional neural network

Abstract

The invention discloses an underground cavity identification method and system in combination with multi-scale fusion and a convolutional neural network, and the method comprises the following steps: firstly, processing ground penetrating radar time series data through Hilbert transformation, extracting the instantaneous attributes of a ground penetrating radar signal: an instantaneous amplitude (IA), an instantaneous phase (IP) and an instantaneous frequency (IF), and carrying out the calculation of the instantaneous amplitude (IA), the instantaneous phase (IP) and the instantaneous frequency (IF); therefore, accurate local signal features are obtained; secondly, multi-scale feature fusion is carried out on IA, IP and IF through dual-tree complex wavelet transform, and the problem that geophysical feature information in a B-scan profile (BP) is insufficient is effectively solved; and finally, the fused data set is input into the convolutional neural network for training, and a training result shows that the fusion method can capture multi-scale and multi-direction features, effectively reduces BP signal ambiguity, has higher recognition accuracy and robustness, and provides an effective new strategy for underground cavity anomalous body detection.

Description

Technical Field

[0001] The present invention relates to the technical field of underground anomaly detection by ground penetrating radar, and in particular to an underground cavity recognition method and system combining multi-scale fusion and convolutional neural network in ground penetrating radar data processing. Background Art

[0002] Detecting underground cavities is crucial to preventing problems such as ground subsidence, structural instability, and potential catastrophic failure. The formation of subsurface cavities can weaken the support of overlying materials, thereby compromising the integrity of underground structures, leading to subsidence, collapse, or damage to upper buildings and infrastructure (Ma Changying et al., 2024). Therefore, non-destructive, efficient, and accurate detection of underground cavities is essential (Hariri-Ardebili et al., 2023). As an efficient geophysical method for non-destructive testing (NDT), ground penetrating radar (GPR) can provide high-resolution imaging information of underground anomalies by emitting electromagnetic waves and recording their reflections. It has been widely used in many fields such as geological exploration, environmental monitoring, and engineering construction (Yu Shaohuai et al., 2023; Feng Deshan et al., 2021; Sonkamble and Chandra, 2021). GPR plays a vital role in identifying and describing these hidden anomalies (Li Xiaofeng et al., 2022).

[0003] However, in actual detection, interpreting and analyzing GPR data remains a major challenge. Underground structural materials are usually composed of a heterogeneous mixture of soil, rock, gravel and other components. These materials vary greatly in volume proportion, spatial distribution and physical properties, showing multiphase, discrete and random medium characteristics (Guo Shili et al., 2021). When the high-frequency electromagnetic waves emitted by GPR propagate in the underground internal medium, the heterogeneity of the medium will cause the GPR signal propagation path to be complex, attenuated, scattered and dispersed, thereby reducing the resolution of the received signal and blurring the B-scan profile (Wang Hui et al., 2022). Experienced technicians are usually required to process complex GPR data to improve the signal-to-noise ratio and eliminate wavelet attenuation and dispersion. They use migration, deconvolution, filtering and time-varying gain to reduce noise and enhance target reflection (Du Yuchuan et al., 2023). The efficiency and accuracy of data interpretation depend largely on the expertise of the technicians. However, when processing large amounts of data, it becomes challenging to ensure timely and accurate interpretation (Pryshchenko et al., 2022). Therefore, the effective integration of deep learning technology is of great significance for the automatic identification of underground anomalies using GPR ( Ma Changying et al., 2024 ). Summary of the invention

[0004] The present invention aims to overcome the shortcomings of the prior art in ground penetrating radar for detecting underground anomalies and to provide a more efficient and accurate method for detecting anomalies.

[0005] The technical solution provided by the present invention is as follows:

[0006] In the first aspect, a method for identifying underground cavities combining multi-scale fusion and convolutional neural network comprises the following steps:

[0007] Step (1): extract the instantaneous properties of the ground penetrating radar signal from the ground penetrating radar time series data: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0008] Step (2): using dual-tree complex wavelet transform DT-CWT to perform feature fusion on instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0009] Step (3): Input the IAF-datasets and its labels after the instantaneous attribute features are fused into the convolutional neural network model for training, and use the trained convolutional neural network model based on multi-scale fusion to classify and identify the ground penetrating radar signals of underground cavities collected in real time;

[0010] During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to steps (1) and (2) to obtain instantaneous attribute feature fusion data, which is input into the trained convolutional neural network model based on multi-scale fusion.

[0011] Labels refer to data with no anomalies or empty anomalies;

[0012] Furthermore, the instantaneous attribute extraction process of the GPR signal is as follows:

[0013] Step A1: Let the GPR time series data signal be x(t), and construct the complex signal z(t) of x(t):

[0014]

[0015] Where: x(t) is the ground penetrating radar time series data signal, which contains information about the underground structure; is the Hilbert transform of x(t), representing the component related to the phase of x(t); i is the imaginary unit;

[0016] Perform Hilbert transform on x(t):

[0017]

[0018] Step A2: Based on the complex signal z(t) of the GPR time series data, the instantaneous properties of the GPR signal are calculated as follows:

[0019]

[0020] Among them: A(t), φ(t), They are the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF of the ground penetrating radar respectively.

[0021] Furthermore, the process of using dual-tree complex wavelet transform to fuse the features of IA, IP and IF is as follows:

[0022] Dual-tree complex wavelet transform uses a pair of complex filter banks to decompose and reconstruct signals. Compared with traditional wavelet transform, DT-CWT has superior performance and feature extraction capabilities. The signal is first decomposed into real and imaginary parts, and then processed by two tree-structured wavelet transforms. This multi-tree structure can better capture the multi-scale and multi-directional characteristics of the signal, thereby achieving more accurate signal analysis and feature extraction.

[0023] DT-CWT is mainly divided into two steps: decomposition and reconstruction. In the decomposition process, the signal is first filtered by a pair of complex filter banks, and then the wavelet transform of two tree structures is performed separately. In the reconstruction process, the results of each tree are inversely transformed and then merged into the final signal.

[0024] Step B1: Assuming that the ground penetrating radar time series signal x(t) is a real-valued signal, where t represents the time index, construct a complex wavelet expression of x(t);

[0025] x(t)=ψ h (t)+iψ g (t)

[0026] Among them, ψ h (t) and ψ g (t) are the real tree wavelet and the complex tree wavelet, respectively, i is the imaginary unit;

[0027] Step B2: Transform the wavelet coefficients and scale coefficients to the dual-tree complex wavelet transform through the real and imaginary tree transforms:

[0028] Wavelet coefficients and scaling coefficients of real dual-tree complex wavelet transform:

[0029]

[0030] in, and are the wavelet coefficients and scale coefficients of the real part dual-tree complex wavelet transform, j is the scale factor, J is the maximum scale factor, and k is the wavelet filter length;

[0031] Wavelet coefficients and scaling coefficients of the imaginary dual-tree complex wavelet transform:

[0032]

[0033] Step B3: Obtain the wavelet coefficients d that capture high-frequency details based on the wavelet coefficients and scale coefficients of the dual-tree complex wavelet transform j (t) and the scaling factor c representing the low-frequency trend J (t):

[0034]

[0035] Step B4: The wavelet coefficients d that capture high-frequency details j (t) and the scaling factor c representing the low-frequency trend J (t) is reconstructed, and the reconstructed signal is as follows:

[0036]

[0037] in, Represents the reconstructed signal.

[0038] After steps (1) and (2), the reconstructed signal Multi-scale feature fusion is performed mainly by calculating the difference in the corresponding wavelet coefficient positions in IA, IP, and IF to determine their fusion weights. Specifically, the fusion weight reflects the contribution of each instantaneous attribute at the wavelet coefficient position, and a new linear combination weight is formed by the difference in the three instantaneous attribute coefficients. Finally, these fusion weights are substituted into the above formula to calculate the fused wavelet coefficients. The fused signal is then subjected to the inverse dual-tree complex wavelet transform (IDT-CWT) to obtain the final reconstructed signal. The acquisition of this signal is a key part of this solution, which effectively integrates feature information from different scales and directions, laying the foundation for subsequent model training.

[0039] Furthermore, the convolutional neural network structure includes an input layer, a convolution-pooling combination module, a Flatten layer, a fully connected layer module and an output layer connected in sequence;

[0040] The convolution-pooling combination module consists of two parts. Each convolution combination contains two convolution blocks, a maximum pooling layer, a BatchNormalization layer, and a Dropout layer, and a ReLU activation function module is added after each convolution block. The number of convolution kernels in the first and second parts of the convolution block is 32 and 64 respectively.

[0041] The fully connected layer module includes two fully connected layer units, each of which includes a fully connected layer and a Dropout layer, and a ReLU activation function module is added after each fully connected layer;

[0042] The output layer uses the softmax function.

[0043] Furthermore, during the training process, the Adam optimizer was used, and the cross entropy loss function was used to measure the performance of the network in the classification task. The batch size was set to 64, and the initial learning rate was 0.001.

[0044] In the second aspect, an underground cavity recognition system combining multi-scale fusion and convolutional neural network includes:

[0045] Instantaneous attribute extraction module: Hilbert transform is used to process the GPR time series data to extract the instantaneous attributes of the GPR signal: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0046] Feature fusion module: Dual-tree complex wavelet transform DT-CWT is used to perform feature fusion on instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0047] Model training and identification module: The IAF-datasets and their labels after the fusion of instantaneous attribute features are input into the convolutional neural network model for training, and the trained convolutional neural network model based on multi-scale fusion is used to classify and identify the ground penetrating radar signals of underground cavities collected in real time;

[0048] During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to the instantaneous attribute extraction module and the feature fusion module, and the instantaneous attribute feature fusion data is input into the trained convolutional neural network model based on multi-scale fusion.

[0049] Furthermore, the extraction process of the instantaneous attribute extraction module is as follows:

[0050] Step A1: Let the GPR time series data signal be x(t), and construct the complex signal z(t) of x(t):

[0051]

[0052] Where: x(t) is the ground penetrating radar time series data signal, which contains information about the underground structure; is the Hilbert transform of x(t), representing the component related to the phase of x(t); i is the imaginary unit;

[0053] Perform Hilbert transform on x(t):

[0054]

[0055] Step A2: Based on the complex signal z(t) of the GPR time series data, the instantaneous properties of the GPR signal are calculated as follows:

[0056]

[0057] Among them: A(t), φ(t), They are the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF of the ground penetrating radar respectively.

[0058] Furthermore, the feature fusion module uses dual-tree complex wavelet transform to fuse the features of IA, IP and IF as follows:

[0059] Step B1: Assuming that the ground penetrating radar time series signal x(t) is a real-valued signal, where t represents the time index, construct a complex wavelet expression of x(t);

[0060] x(t)=ψ h (t)+iψ g (t)

[0061] Among them, ψ h (t) and ψ g (t) are the real tree wavelet and the complex tree wavelet, respectively, i is the imaginary unit;

[0062] Step B2: Transform the wavelet coefficients and scale coefficients to the dual-tree complex wavelet transform through the real and imaginary tree transforms:

[0063] Wavelet coefficients and scaling coefficients of real dual-tree complex wavelet transform:

[0064]

[0065] in, and are the wavelet coefficients and scale coefficients of the real part dual-tree complex wavelet transform, j is the scale factor, J is the maximum scale factor, and k is the wavelet filter length;

[0066] Wavelet coefficients and scaling coefficients of the imaginary dual-tree complex wavelet transform:

[0067]

[0068] Step B3: Obtain the wavelet coefficients d that capture high-frequency details based on the wavelet coefficients and scale coefficients of the dual-tree complex wavelet transform j (t) and the scaling factor c representing the low-frequency trend J (t):

[0069]

[0070] Step B4: The wavelet coefficients d that capture high-frequency details j (t) and the scaling factor c representing the low-frequency trend J (t) is reconstructed, and the reconstructed signal is as follows:

[0071]

[0072] in, Represents the reconstructed signal.

[0073] Furthermore, the convolutional neural network structure of the model training and recognition module includes an input layer, a convolution-pooling combination module, a Flatten layer, a fully connected layer module and an output layer connected in sequence;

[0074] Convolution-pooling combination module. The entire convolution module consists of two parts. Each convolution combination contains two convolution blocks, a maximum pooling layer, a batch normalization layer, and a dropout layer. A ReLU activation function module is added after each convolution block. The number of convolution kernels in the first and second parts of the convolution block is 32 and 64 respectively.

[0075] The fully connected layer module includes two fully connected layer units, each of which includes a fully connected layer and a Dropout layer, and a ReLU activation function module is added after each fully connected layer;

[0076] The output layer uses the softmax function.

[0077] In a third aspect, a computer-readable storage medium stores a computer program, wherein the computer program is called by a processor to execute: the steps of the above-mentioned underground cavity identification method combining multi-scale fusion and convolutional neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. Ordinary technicians in this field can also obtain other drawings based on these drawings without creative work.

[0079] Figure 1 It is a schematic diagram of the technical route of the instantaneous attribute feature fusion method of the ground penetrating radar in one embodiment of the present invention;

[0080] Figure 2 is a flow chart of ground penetrating radar data preprocessing in one embodiment of the present invention;

[0081] Figure 3 is a schematic diagram of ground penetrating radar data preprocessing in one embodiment of the present invention, wherein (a) is B-scanProfile, (b) is Preprocessed Profile;

[0082] Figure 4 is a schematic cross-sectional view of a ground penetrating radar B-scan in one embodiment of the present invention;

[0083] Figure 5 is a schematic diagram of the instantaneous amplitude of a ground penetrating radar in one embodiment of the present invention;

[0084] Figure 6 is a schematic diagram of the instantaneous phase of a ground penetrating radar in one embodiment of the present invention;

[0085] Figure 7 is a schematic diagram of the instantaneous frequency of a ground penetrating radar in one embodiment of the present invention;

[0086] Figure 8 is a schematic diagram of signal decomposition and reconstruction of dual-tree complex wavelet transform in one embodiment of the present invention;

[0087] Fig. 9 is a feature fusion flow chart of dual-tree complex wavelet transform in one embodiment of the present invention;

[0088] Fig.10 is a feature fusion result diagram of wavelet transform in one embodiment of the present invention;

[0089] Fig.11 is a feature fusion result diagram of dual-tree complex wavelet transform in one embodiment of the present invention;

[0090] Fig.12 is a schematic diagram of a convolutional neural network structure in one embodiment of the present invention;

[0091] Fig.13 is a diagram of the training result of a convolutional neural network using a traditional input method in one embodiment of the present invention;

[0092] Fig.14 It is a diagram of the convolutional neural network training result of the instantaneous attribute feature fusion input method in one embodiment of the present invention; DETAILED DESCRIPTION

[0093] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0094] The embodiment of the present invention provides an underground cavity recognition method combining multi-scale fusion and convolutional neural network, and its technical route is as follows: Figure 1As shown, it includes the feature extraction stage: forward modeling, data preprocessing, instantaneous attributes features extraction; feature fusion stage: applying DT-CWT, calculating wavelet coefficients, multi-scale features fusion; application stage: IAF-datasets creation, CNN model architecture buliding and training, evaluation, verification and application;

[0095] Specifically include:

[0096] Step (1): Use Hilbert transform to process the GPR time series data and extract the instantaneous properties of the GPR signal: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0097] Step (2): using dual-tree complex wavelet transform DT-CWT to perform feature fusion on instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0098] Step (3): Input the IAF-datasets and its labels after the instantaneous attribute features are fused into the convolutional neural network model for training, and use the trained convolutional neural network model based on multi-scale fusion to classify and identify the ground penetrating radar signals of underground cavities collected in real time;

[0099] During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to steps (1) and (2) to obtain instantaneous attribute feature fusion data, which is input into the trained convolutional neural network model based on multi-scale fusion.

[0100] The recognition method can enhance the detection and recognition of ground penetrating radar anomalies. The feature fusion method includes ground penetrating radar instantaneous attribute feature extraction and instantaneous attribute feature fusion. Finally, the fused data set is input into the convolutional neural network for training. The training results show that the feature fusion method can effectively classify and identify underground anomalies with high accuracy. In an embodiment of the present invention, the detection capability of underground anomalies can be effectively enhanced by performing feature fusion on ground penetrating radar time series data.

[0101] First, extract instantaneous attribute features from the ground penetrating radar signal, such as instantaneous amplitude, instantaneous phase and instantaneous frequency. These features can effectively capture the reflection characteristics and frequency response of the anomaly. Among them, instantaneous amplitude: represents the change in signal strength, which can highlight the reflection characteristics of the anomaly. Instantaneous phase: reflects the phase change of the signal, which helps to identify the interface of different materials and structures. Instantaneous frequency: shows the frequency change of the signal, which can distinguish anomalies with different frequency responses.

[0102] like Figure 2 As shown, a flow chart of ground penetrating radar data preprocessing in this embodiment is provided, which specifically includes the following steps:

[0103] First, the data module includes data loading (GPR Data loading) and data parameter setting (Data Parameter Setting).

[0104] In the data loading stage, the ground penetrating radar data file is loaded to obtain radar waveform data. The data parameter setting stage includes setting parameters such as the antenna center frequency (AntenaFrequency), relative dielectric constant (Relative Permittivity), time window size (TimeWindow), sampling interval (Sampling Interval), and the profile start position and profile end position (StartPoint and EndPoint).

[0105] The image processing module generates B-scan profiles, which are two-dimensional representations of radar waveform data in time and distance coordinates, to intuitively display the structural characteristics of underground media. The following preprocessing steps include: Remove Direct Wavelet, Remove Background, Gain, and 2-Dimensional Filtering, which help reduce noise and enhance the reflected signals of underground targets. The processed B-scan profiles (GPRB-Scan Profiles) will serve as the basic data for subsequent underground target detection and analysis, effectively improving data quality and underground target detection accuracy, and providing reliable data support and analysis basis.

[0106] However, in the context of inhomogeneous media, the B-scan profile of ground penetrating radar encounters some limitations, mainly reflection errors, signal attenuation and reduced resolution. The heterogeneity of the medium affects the propagation path of the electromagnetic wave, resulting in refraction and reflection effects, which may produce artifacts and errors in the imaging. In addition, scattering and absorption in the medium can also cause signal attenuation, reducing the detection depth and sensitivity of the ground penetrating radar, thereby limiting the detection of underground targets. These factors are largely responsible for the ambiguity of the B-scan profile, complicating data interpretation and analysis. Despite attempts to preprocess the ground penetrating radar data to mitigate these effects, the inherent complexity of the medium often causes the ambiguity to persist for a long time. Figure 3 (a) shows a schematic diagram of the original B-scan cross-section of the ground penetrating radar data in this embodiment. Figure 3 (b) shows a schematic diagram of the B-scan profile of the ground penetrating radar data preprocessing in this embodiment, where the horizontal axis is position (Position) in meters (m), and the vertical axis is depth (Deepth) in meters (m).

[0107] Scattering and absorption in the inhomogeneous medium background will reduce the spatial resolution of imaging, making it difficult to distinguish the boundaries and details of underground targets. In order to solve these limitations, the Hilbert transform is introduced to extract instantaneous features. Figure 4 The B-scan profile shown in the figure has the position on the horizontal axis in meters (m) and the depth on the vertical axis in meters (m), which shows the distribution of the ground penetrating radar data in the depth and position dimensions, revealing its inherent ambiguity and complexity.

[0108] The essence of using Hilbert transform to process GPR time series data is to convolve the time series x(t) with 1 / πt, which can extract the instantaneous characteristics of the electromagnetic wave such as amplitude, phase and frequency to obtain the instantaneous properties of the GPR signal.

[0109] Step A1: Let the GPR time series data signal be x(t), and construct the complex signal z(t) of x(t):

[0110]

[0111] Where: x(t) is the ground penetrating radar time series data signal, which contains information about the underground structure; is the Hilbert transform of x(t), representing the component related to the phase of x(t); i is the imaginary unit;

[0112] Perform Hilbert transform on x(t):

[0113]

[0114] Step A2: Based on the complex signal z(t) of the GPR time series data, the instantaneous properties of the GPR signal are calculated as follows:

[0115]

[0116] Among them: A(t), φ(t), They are the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF of the ground penetrating radar respectively.

[0117] The ground penetrating radar signal is processed by Hilbert transform to extract its instantaneous attribute features: IA, IP, IF. These features can better reflect the changes in time and frequency of the signal. Figures 5 to 7 As shown, the instantaneous amplitude diagram of the ground penetrating radar in this embodiment is provided, showing the distribution of the instantaneous amplitude in the depth and position dimensions; the instantaneous phase diagram of the ground penetrating radar shows the distribution of the instantaneous phase in the depth and position dimensions; the instantaneous frequency diagram of the ground penetrating radar shows the distribution of the instantaneous frequency in the depth and position dimensions, the horizontal axis is the position (Position), the unit is meter (m), and the vertical axis is the depth (Deepth), the unit is meter (m). Figure 4As shown, the B-scan profile of the ground penetrating radar in this embodiment is provided. It can be seen that under the inhomogeneous medium, the B-scan profile is seriously affected, and it is difficult to clearly display the morphology and distribution of the abnormal body. The inhomogeneity of the medium leads to a complex signal propagation path, and the signal attenuation and scattering phenomena aggravate the signal blur, making it difficult to distinguish the boundary of the target. At the same time, the beam expansion effect and the reduction in resolution further blur the image, making the detection and identification of the abnormal body difficult. The instantaneous attributes IA, IP, and IF obtained by Hilbert transform. Among them, IA reflects the reflection intensity of the ground penetrating radar signal, which is proportional to the square root of the total energy of the signal. It can be seen that due to the strong energy in the shallow layer and the weak energy in the deep layer, it is difficult to output and display the B-scan profile information. IP reflects the continuity of the phase axis in the ground penetrating radar signal. In an isotropic homogeneous medium, the phase of the high-frequency electromagnetic wave is continuous, but the underground medium is usually an inhomogeneous body, and a phase change will occur after passing through the abnormal body, which has nothing to do with the strength of the signal energy. IF reflects the time rate of change of IF. When high-frequency electromagnetic waves pass through changing lithology or changing strata, the frequency of electromagnetic waves will change significantly. Therefore, IA and IF can determine the approximate location of underground anomalies, and IP can roughly describe their contours.

[0118] The dual-tree complex wavelet transform (DT-CWT) is used to fuse the features of IA, IP and IF. The dual-tree complex wavelet transform is characterized by using a pair of complex filter banks to decompose and reconstruct the signal. Compared with the traditional wavelet transform, DT-CWT has better performance and feature extraction capabilities. The signal is first decomposed into two parts, the real part and the imaginary part, and then processed by two tree-structured wavelet transforms respectively. This multi-tree structure can better capture the multi-scale and multi-directional characteristics of the signal, thereby achieving more accurate signal analysis and feature extraction. Figure 8 As shown, a schematic diagram of signal decomposition and reconstruction of the dual-tree complex wavelet transform in this embodiment is provided, which decomposes the signal into processes of different scales and directions and finally reconstructs the signal.

[0119] DT-CWT is mainly divided into two steps: decomposition and reconstruction. In the decomposition process, the signal is first filtered by a pair of complex filter banks, and then the wavelet transform of two tree structures is performed separately. In the reconstruction process, the results of each tree are inversely transformed and then merged into the final signal.

[0120] Step B1: Assuming that the ground penetrating radar time series signal x(t) is a real-valued signal, where t represents the time index, construct a complex wavelet expression of x(t);

[0121] x(t)=ψ h (t)+iψ g (t) (4)

[0122] Among them, ψh (t) and ψ g (t) are the real tree wavelet and the complex tree wavelet, respectively, i is the imaginary unit;

[0123] Step B2: Transform the wavelet coefficients and scale coefficients to the dual-tree complex wavelet transform through the real and imaginary tree transforms:

[0124] Wavelet coefficients and scaling coefficients of real dual-tree complex wavelet transform:

[0125]

[0126] in, and are the wavelet coefficients and scale coefficients of the real part dual-tree complex wavelet transform, j is the scale factor, J is the maximum scale factor, and k is the wavelet filter length;

[0127] Wavelet coefficients and scaling coefficients of the imaginary dual-tree complex wavelet transform:

[0128]

[0129] Step B3: Obtain the wavelet coefficients d that capture high-frequency details based on the wavelet coefficients and scale coefficients of the dual-tree complex wavelet transform j (t) and the scaling factor c representing the low-frequency trend J (t):

[0130]

[0131] Step B4: The wavelet coefficients d that capture high-frequency details j (t) and the scaling factor c representing the low-frequency trend J (t) is reconstructed, and the reconstructed signal is as follows:

[0132]

[0133] in, Represents the reconstructed signal.

[0134] As an effective signal processing technology, dual-tree complex wavelet transform (DT-CWT) has been widely used in image compression, texture analysis, image restoration and other fields. By taking advantage of complex wavelet transform and dual-tree structure, DT-CWT can capture the time-frequency characteristics of the signal more accurately and has good directional selectivity. In GPR data processing, instantaneous attribute characteristics are crucial for the description of underground structure and disease characteristics. Traditional wavelet transform has limitations in extracting these features, while DT-CWT can overcome these problems. Utilizing the characteristics of complex wavelet transform, compared with WT, DT-CWT adopts a dual-tree structure to decompose the signal in different directions, which can capture the time-frequency characteristics of the signal more comprehensively, and is especially suitable for ground penetrating radar signals in inhomogeneous media.

[0135] When using dual-tree complex wavelet transform (DT-CWT) for feature fusion, the input images are first converted to grayscale images (Input images reading Convert to grayscale), and feature extraction is performed through DT-CWT (Execute DT-CWT). According to equations (5) and (6), wavelet coefficients of different scales and directions are extracted from each image. Then, for each wavelet coefficient position, the difference of the corresponding wavelet coefficients in the three images is calculated to determine their fusion weights (Execute fusion). Specifically, the fusion weight reflects the contribution weight of each image at that position through the coefficient difference of the three images, and then these weights are linearly combined to form new fusion weights. Finally, these fusion weights are substituted into equation (7) to calculate the fused wavelet coefficients. Next, the fused high-frequency wavelet coefficients are synthesized through the inverse dual-tree complex wavelet transform (Inverse DT-CWT), and the average value of the wavelet coefficients of the low-frequency part of the three images is used to reconstruct the final fused image (Low-frequency subband fusion). This method can effectively capture and fuse the detailed time-frequency features of multiple images, and is particularly suitable for application scenarios in complex media such as ground penetrating radar (GPR) data processing. Fig. 9 As shown, a feature fusion flow chart of the dual-tree complex wavelet transform in this embodiment is provided.

[0136] like Fig.10 As shown, the wavelet transform feature fusion of the instantaneous attributes of the ground penetrating radar in this embodiment is provided, such as Fig.11As shown, the dual-tree complex wavelet transform feature fusion of the instantaneous attributes of the ground penetrating radar in this embodiment is provided. The feature fusion image after wavelet transformation still has noise points, resulting in low resolution and information loss, while DT-CWT can fully and effectively fuse the extracted instantaneous attribute features to obtain a more comprehensive and accurate description of the underground structure. The response characteristics of the abnormal body can be better displayed. By comprehensively utilizing the fused image of each feature information, the complexity and diversity of the underground structure can be better reflected, and pseudo-features and noise interference can be well suppressed, thereby improving the effect of signal processing and the ability to identify underground diseases.

[0137] Convolutional Neural Network (CNN) is a feedforward neural network widely used in fields such as geomagnetic noise removal, natural blasting identification, and tunnel lining internal detection. CNN is usually composed of multiple layers of convolutional layers and pooling layers stacked alternately, as well as fully connected layers. In the embodiment of the present invention, a new convolutional neural network model is constructed, and training and effect analysis are performed on the model. The training process includes data preprocessing, model construction, loss function selection, and application of optimization algorithm, and evaluation is performed based on the obtained training results. Fig.12 As shown, a schematic diagram of the convolutional neural network structure in this embodiment is provided, including an input layer, a convolution-pooling combination, a fully connected layer and an output layer, which are used to process and classify ground penetrating radar data features.

[0138] The convolutional neural network structure includes an input layer, a convolution-pooling combination module (Conv-Pooling Combination), a Flatten layer, a fully connected layer module and an output layer connected in sequence;

[0139] The convolution-pooling combination module consists of two parts. Each convolution combination contains two convolution blocks, a maximum pooling layer, a BatchNormalization layer, and a Dropout layer, and a ReLU activation function module is added after each convolution block. The number of convolution kernels in the first and second parts of the convolution block is 32 and 64 respectively.

[0140] The fully connected layer module includes two fully connected layer units, each of which includes a fully connected layer and a Dropout layer, and a ReLU activation function module is added after each fully connected layer;

[0141] The output layer uses the softmax function.

[0142] Multiple convolution-pooling modules can extract multi-level abstract features of input data. At the input end of the network, the feature-fused data set is processed and enters the convolution layer, and then further processed by the activation function layer (ReLU) and the pooling layer, and finally output to the fully connected layer. In the fully connected layer, the output of the previous layer is used as input and calculated through the ReLU activation function to enhance the nonlinear expression ability of the model.

[0143] In order to prevent overfitting during training, Batch Normalization and Dropout techniques are added after the convolution layer and pooling layer to improve the training stability and generalization ability of the model in the underground cavity identification task. Specifically, Batch Normalization is applied after each convolution layer to speed up the training and improve stability by normalizing the activation value; while Dropout randomly discards some neurons to reduce the model's dependence on specific nodes, thereby enhancing the generalization ability of the model. As shown in Table 1, the convolutional neural network parameter information in this embodiment is provided.

[0144] Table 1 Convolutional neural network parameter information

[0145]

[0146] During the training process, the size of the input image is uniformly set to 128×128 pixels. This size design can effectively balance the demand for computing resources and the ability of feature extraction. The total sample size is 27,000 images, which cover the characteristics of a variety of underground cavities to ensure the generalization ability of the model in different situations. In order to evaluate the performance of the model, the dataset is divided into a training set and a validation set in an 8:2 ratio, where the training set contains 21,600 samples and the validation set contains 5,400 samples. With an 8:2 training set and validation set split, the model is trained using the Adam optimizer and the cross entropy loss function, with a batch size of 64, an initial learning rate of 0.001, and 80 epochs.

[0147] As shown in Table 2, the data set information in this embodiment is provided.

[0148] Table 2 Dataset information

[0149]

[0150] In the selection of optimization algorithms, the Adam optimizer is used because it combines the advantages of momentum and adaptive learning rate, and can converge faster and adapt to different gradient changes. In addition, the cross entropy loss function is used to measure the performance of the model in classification tasks, which is especially suitable for classification problems. The batch size is set to 64, which can effectively utilize the computing power of the GPU during training while avoiding memory overload.

[0151] The initial learning rate is 0.001, which is an optimal parameter verified by experiments and helps to maintain a stable learning process in the early stage of training. During the total training of 80 epochs, the performance of the validation set is constantly monitored to prevent overfitting and ensure the performance of the model on unseen data.

[0152] Finally, the accuracy of the model on the validation set reached 97.8%. This result shows that an underground cavity recognition method combining multi-scale fusion and convolutional neural network can capture multi-scale and multi-directional features, effectively reduce the ambiguity of B-scan profile (BP) signals, and has higher recognition accuracy and robustness. The constructed CNN model has high effectiveness and reliability in underground cavity recognition tasks, and can successfully extract useful features from complex GPR data, providing an effective new strategy for underground cavity anomaly detection. Fig.13 As shown, a convolutional neural network training result diagram of the traditional input method in this embodiment is provided, with the Acc Curve (accuracy curve) of training and verification on the left and the Loss Curve (loss curve) of training and verification on the right; Fig.14 As shown, a convolutional neural network training result diagram of the instantaneous attribute feature fusion input method in the embodiment is provided, with the Acc Curve (accuracy curve) of training and verification on the left and the Loss Curve (loss curve) of training and verification on the right;

[0153] As shown in Table 3, a classification report is provided in this embodiment. The indicators such as precision, recall, F1 score and support rate for each category listed in Table 3 are also very insightful and help to fully understand the performance of the model in different categories. These indicators are all crucial indicators that can test the ability of the model to accurately classify different categories of anomalies in detail. In the field of machine learning, it is crucial to carefully evaluate the performance of the classification model.

[0154] Table 3 Classification report

[0155]

[0156] In this process, precision, recall, F1 score, and support are commonly used key indicators. Precision refers to the proportion of samples correctly classified as positive among all samples classified as positive. That is, among all samples predicted as positive, how many are actually positive. Precision can provide important information about the accuracy of the classifier. Precision can be expressed by the following formula:

[0157]

[0158] Among them, TP stands for True Positives and FP stands for False Positives.

[0159] Recall rate refers to the proportion of samples that are correctly classified as positive among all samples that are actually positive. That is, how many samples that are actually positive are correctly predicted as positive. It measures the ability of the classifier to correctly identify positive samples and provides an important measure of the integrity of the classifier. Recall rate can be expressed by the following formula:

[0160]

[0161] Among them, TP represents true positives and FN represents false negatives.

[0162] The F1 score is the harmonic mean of precision and recall, which combines precision and recall and comprehensively considers the accuracy and completeness of the classifier. It provides a quantitative evaluation of the overall performance of the classifier through the harmonic mean, which is particularly suitable for cases with unbalanced class distribution. The F1 score can be expressed by the following formula:

[0163]

[0164] In addition, support indicates the actual number of occurrences of each category in the dataset, which is a direct measure of the sample size of the category in the dataset. The higher the support, the higher the frequency of the category in the dataset. The inclusion of these detailed indicators emphasizes the comprehensiveness of the evaluation process and provides significant inspiration for further improvement and optimization of the model.

[0165] The high accuracy achieved by the feature fusion method of this embodiment fully demonstrates its potential for application in anomaly detection tasks. These results highlight the effectiveness of fusion ground penetrating radar instantaneous attribute feature datasets (IAF-datasets), which outperform the separate B-scan profile datasets in various evaluation indicators. This approach improves accuracy and simplifies the training process, highlighting the value of multimodal data fusion technology in improving model capabilities.

[0166] This embodiment successfully improves the recognition accuracy and robustness of underground anomalies in non-destructive testing by applying a combination of dual-tree complex wavelet transform (DT-CWT) and convolutional neural network (CNN). By performing Hilbert transform on GPR data to extract the instantaneous properties of the signal, and using dual-tree complex wavelet transform for feature fusion, the challenges brought by the inhomogeneity of the underground medium are overcome, and more accurate local features of the signal are extracted. This method effectively solves the problem of fuzzy B-scan profile signals and provides a new way to accurately detect and identify underground anomalies. The results of this study show that compared with traditional methods, this method has higher accuracy and robustness in anomaly identification, providing important technical support and application prospects for safety assessment and management.

[0167] A method for identifying underground voids that combines multi-scale fusion with convolutional neural networks can better automatically identify and interpret underground anomaly detection, providing a reliable technical support for underground targets. Through local feature extraction, the instantaneous attribute features (IA, IP, IF) are extracted from GPR signals using Hilbert transform, which effectively solves the problem that radar signals are affected by underground inhomogeneous media, resulting in complex signal propagation paths, attenuation and scattering, making the B-scan profile (BP) signal blurred. In view of the lack of sample geophysical feature information, the instantaneous attributes are fused using dual-tree complex wavelet transform (DT-CWT), which can fully capture the shared information between IA, IP and IF, while retaining the local structural characteristics of GPR data. In addition, the fused IAF-dataset can obtain comprehensive and accurate anomaly feature information. This fused dataset can better complete the classification task after training with convolutional neural network (CNN). The results show that the proposed method is superior to the traditional method, with an identification accuracy of 97.8%. It provides a new way for the subsequent rapid and accurate identification, detection and interpretation of underground void anomaly hazards.

[0168] In some embodiments, the present invention further provides an underground cavity identification system combining multi-scale fusion and convolutional neural network, comprising:

[0169] Instantaneous attribute extraction module: Hilbert transform is used to process the GPR time series data to extract the instantaneous attributes of the GPR signal: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0170] Feature fusion module: Dual-tree complex wavelet transform DT-CWT is used to perform feature fusion on instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF;

[0171] Model training and identification module: The IAF-datasets and their labels after the fusion of instantaneous attribute features are input into the convolutional neural network model for training, and the trained convolutional neural network model based on multi-scale fusion is used to classify and identify the ground penetrating radar signals of underground cavities collected in real time;

[0172] During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to the instantaneous attribute extraction module and the feature fusion module, and the instantaneous attribute feature fusion data is input into the trained convolutional neural network model based on multi-scale fusion.

[0173] It should be understood that the specific implementation process of each module please refer to the above method content, the present invention will not be repeated here, and the division of the above functional modules is only for example illustration. In some embodiments, some functional modules can be merged, and some functional modules can be split. Each functional module can be implemented in software or hardware or a combination of software and hardware. Among them, the hardware and software equipment includes but is not limited to general-purpose computer equipment, programmable gate arrays, digital signal processors, microprocessors and their corresponding programming or burning software.

[0174] The present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a computer, the computer executes a method for identifying underground voids that combines multi-scale fusion and convolutional neural network as described in an embodiment of the present application.

[0175] For the specific implementation process of each step, please refer to the description of the above method.

[0176] The readable storage medium is a computer-readable storage medium, which may be an internal storage unit of the software and hardware device described in any of the aforementioned embodiments, such as a hard disk or memory of a controller. The readable storage medium may also be an external storage device of the controller, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the controller. Furthermore, the readable storage medium may also include both an internal storage unit of the controller and an external storage device. The readable storage medium is used to store the computer program and other programs and data required by the controller. The readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0177] The above contents are merely examples and explanations of the concept of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

[0178] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0179] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

[0180] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for identifying underground cavities by combining multi-scale fusion and convolutional neural network, characterized in that: The following steps are involved: Step (1): extract the instantaneous properties of the ground penetrating radar signal from the ground penetrating radar time series data: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF; Step (2): using dual-tree complex wavelet transform DT-CWT to perform feature fusion on instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF; Step (3): Input the IAF-datasets dataset and its cavity labels after the instantaneous attribute features are fused into the convolutional neural network model for training, and use the trained convolutional neural network model based on multi-scale fusion to classify and identify the ground penetrating radar signals of underground cavities collected in real time; During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to steps (1) and (2) to obtain instantaneous attribute feature fusion data, which is input into the trained convolutional neural network model based on multi-scale fusion.

2. The method according to claim 1, characterized in that The process of extracting instantaneous attributes of ground penetrating radar signals is as follows: Step A1: Let the GPR time series data signal be x(t), and construct the complex signal z(t) of x(t): Where: x(t) is the ground penetrating radar time series data signal, which contains information about the underground structure; is the Hilbert transform of x(t), representing the component related to the phase of x(t); i is the imaginary unit; Perform Hilbert transform on x(t): Step A2: Based on the complex signal z(t) of the GPR time series data, the instantaneous properties of the GPR signal are calculated, which are: Where: A(t) 、 φ(t) 、 They are the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF of the ground penetrating radar respectively.

3. The method according to claim 1, characterized in that The process of feature fusion of IA, IP and IF using dual-tree complex wavelet transform is as follows: Step B1: Assuming that the ground penetrating radar time series signal x(t) is a real-valued signal, where t represents the time index, construct a complex wavelet expression of x(t); x(t)=ψ h (t)+iψ g (t) Among them, ψ h (t) and ψ g (t) are the real tree wavelet and the complex tree wavelet, respectively, i is the imaginary unit; Step B2: Transform the wavelet coefficients and scale coefficients to the dual-tree complex wavelet transform through the real and imaginary tree transforms: Wavelet coefficients and scaling coefficients of real dual-tree complex wavelet transform: in, and are the wavelet coefficients and scale coefficients of the real part dual-tree complex wavelet transform, j is the scale factor, J is the maximum scale factor, and k is the wavelet filter length; Wavelet coefficients of imaginary dual-tree complex wavelet transform and scale factor j=1,2,L,J Step B3: Obtain the wavelet coefficients d that capture high-frequency details based on the wavelet coefficients and scale coefficients of the dual-tree complex wavelet transform j (t) and the scaling factor c representing the low-frequency trend J (t): Step B4: The wavelet coefficients d that capture high-frequency details j (t) and the scaling factor c representing the low-frequency trend J (t) is reconstructed, and the reconstructed signal is as follows: in, Represents the reconstructed signal.

4. The method according to claim 1, characterized in that: The convolutional neural network structure includes an input layer, a convolution-pooling combination module, a Flatten layer, a fully connected layer module and an output layer connected in sequence; The convolution-pooling combination module consists of two parts. Each convolution combination contains two convolution blocks, a maximum pooling layer, a BatchNormalization layer, and a Dropout layer, and a ReLU activation function module is added after each convolution block. The number of convolution kernels in the first and second parts of the convolution block is 32 and 64 respectively. The fully connected layer module includes two fully connected layer units, each of which includes a fully connected layer and a Dropout layer, and a ReLU activation function module is added after each fully connected layer; The output layer uses the softmax function.

5. The method according to claim 4, characterized in that During the training process, the Adam optimizer was used, and the cross entropy loss function was used to measure the performance of the network in the classification task. The batch size was set to 64, and the initial learning rate was 0.

001.

6. An underground cavity recognition system combining multi-scale fusion and convolutional neural network, characterized in that: include: Instantaneous attribute extraction module: Hilbert transform is used to process the GPR time series data to extract the instantaneous attributes of the GPR signal: instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF; Feature fusion module: Dual-tree complex wavelet transform DT-CWT is used to fuse the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF; Model training and recognition module: The IAF-datasets and their labels after the fusion of instantaneous attribute features are input into the convolutional neural network model for training, and the trained convolutional neural network model based on multi-scale fusion is used to classify and recognize the ground penetrating radar signals of underground cavities collected in real time; During the recognition process, the ground penetrating radar signal of the underground cavity collected in real time is processed according to the instantaneous attribute extraction module and the feature fusion module, and the instantaneous attribute feature fusion data is input into the trained convolutional neural network model based on multi-scale fusion.

7. The system according to claim 6, characterized in that The extraction process of the instantaneous attribute extraction module is as follows: Step A1: Let the GPR time series data signal be x(t), and construct the complex signal z(t) of x(t): Where: x(t) is the ground penetrating radar time series data signal, which contains information about the underground structure; is the Hilbert transform of x(t), representing the component related to the phase of x(t); i is the imaginary unit; Perform Hilbert transform on x(t): Step A2: Based on the complex signal z(t) of the GPR time series data, the instantaneous properties of the GPR signal are calculated as follows: Among them: A(t), φ(t), They are the instantaneous amplitude IA, instantaneous phase IP and instantaneous frequency IF of the ground penetrating radar respectively.

8. The system according to claim 6, characterized in that The feature fusion module uses dual-tree complex wavelet transform to fuse the features of IA, IP and IF as follows: Step B1: Assuming that the ground penetrating radar time series signal x(t) is a real-valued signal, where t represents the time index, construct a complex wavelet expression of x(t); x(t)=ψ h (t)+iψ g (t) Among them, ψ h (t) and ψ g (t) are the real tree wavelet and the complex tree wavelet, respectively, i is the imaginary unit; Step B2: Transform the wavelet coefficients and scale coefficients to the dual-tree complex wavelet transform through the real and imaginary tree transforms: Wavelet coefficients and scaling coefficients of real dual-tree complex wavelet transform: j=1,2,L,J in, and are the wavelet coefficients and scale coefficients of the real part dual-tree complex wavelet transform, j is the scale factor, J is the maximum scale factor, and k is the wavelet filter length; Wavelet coefficients and scaling coefficients of the imaginary dual-tree complex wavelet transform: j=1,2,L,J Step B3: Obtain the wavelet coefficients d that capture high-frequency details based on the wavelet coefficients and scale coefficients of the dual-tree complex wavelet transform j (t) and the scaling factor c representing the low-frequency trend J (t): Step B4: The wavelet coefficients d that capture high-frequency details j (t) and the scaling factor c representing the low-frequency trend J (t) is reconstructed, and the reconstructed signal is as follows: in, Represents the reconstructed signal.

9. The system according to claim 6, characterized in that The convolutional neural network structure of the model training and recognition module includes an input layer, a convolutional pooling combination module, a flatten layer, a fully connected layer module, and an output layer connected in sequence; Convolution-pooling combination module. The entire convolution module consists of two parts. Each convolution combination contains two convolution blocks, a maximum pooling layer, a batch normalization layer, and a dropout layer. A ReLU activation function module is added after each convolution block. The number of convolution kernels in the first and second parts of the convolution block is 32 and 64 respectively. The fully connected layer module includes two fully connected layer units, each of which includes a fully connected layer and a Dropout layer, and a ReLU activation function module is added after each fully connected layer; The output layer uses the softmax function.

10. A computer-readable storage medium, characterized in that: A computer program is stored, and the computer program is called by a processor to execute: the steps of the method according to any one of claims 1 to 5.

Images