Symmetrical excitation explosion-proof geological detection system and detection method

Abstract

The invention belongs to the technical field of geological exploration, and particularly discloses a symmetrical excitation anti-explosion geological exploration system and method, and the method comprises the steps: receiving array multi-channel signals at two sides, and carrying out the time alignment correction of all channel signals, so as to eliminate the time delay; a multi-channel cross-correlation matrix between the left array and the right array is constructed based on the aligned signals, and matrix elements are obtained through integral cross-correlation operation of array channel signals on the two sides; performing low-rank decomposition on the cross-correlation matrix, and extracting the first k main components as public noise modes; carrying out inversion calculation on the contribution of each noise mode in each acquisition channel, and reconstructing a channel noise component; and subtracting the reconstructed noise component from the original signal, and outputting a purified geological reflection signal. The method does not need special hardware, can accurately suppress the homologous low-frequency noise of the tunnel, improves the signal-to-noise ratio of the signal and the geological imaging resolution, and is suitable for geological detection of closed spaces such as the tunnel.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geological exploration, in particular to a symmetrical excitation anti-explosion geological exploration system and method. BACKGROUND

[0002] In the geological exploration of closed spaces such as tunnels and underground engineering, geological reflection signals are usually obtained through excitation-capture arrays to identify rock structure, faults and other potential geological risks. However, such exploration scenarios have serious noise interference problems, which greatly reduces the signal quality and exploration accuracy of traditional geological exploration methods.

[0003] The noise in the tunnel is mainly low-frequency mechanical noise generated by equipment such as fans, drilling machines and transport vehicles. This type of noise has long wavelength and strong energy, and traditional filtering techniques cannot effectively filter it out without damaging the geological reflection signal. At the same time, the tunnel noise has no obvious directionality, and the traditional beamforming technology that relies on signal directionality has very limited effect in suppressing noise, which cannot meet the precise exploration requirements.

[0004] Existing geological exploration systems reduce the impact of noise through single filtering or hardware shielding methods, which not only has high cost, but also easily leads to distortion of effective geological reflection signals. At the same time, traditional methods do not fully utilize the array layout characteristics, making it difficult to achieve precise separation of noise and effective signals, resulting in insufficient signal-to-noise ratio of the purified signal, affecting the resolution of subsequent geological imaging, and failing to accurately identify key geological features such as broken rock layers and small-scale faults, posing a hidden danger to underground engineering construction safety. SUMMARY

[0005] The present application aims to provide a symmetrical excitation anti-explosion geological exploration system and method to solve the problems raised in the background.

[0006] According to one aspect of the present application, a symmetrical excitation anti-explosion geological exploration method is provided, which is applied to a system comprising left and right excitation-capture arrays symmetrically arranged on both sides of the exploration area, comprising the following method steps: S1: receiving multi-channel signals from the left excitation-capture array and multi-channel signals from the right excitation-capture array, and performing time alignment correction on all channel signals to eliminate time delays caused by transmission paths or electronic delays between channels; S2: based on the time-aligned left and right array signals, a multi-channel cross-correlation matrix between the left and right arrays is constructed, wherein each element in the cross-correlation matrix is obtained by integrating and cross-correlating the signals of one channel of the left array with the signals of one channel of the right array; S3: performing low-rank decomposition on the constructed cross-correlation matrix, and extracting the first k principal components from the decomposition result as common noise modes shared by the left and right arrays; S4: based on the extracted common noise modes, inversely calculating the specific contribution of each noise mode on each acquisition channel of the left and right arrays, and reconstructing the noise component on each acquisition channel composed of the common noise modes according to the specific contribution; S5: subtracting the reconstructed noise component of the channel from the original acquisition channel signal of the left and right arrays, and outputting the purified geological reflection signal.

[0007] Preferably, step S1 specifically comprises: calculating the cross-correlation function between different channel signals by using the pilot excitation wave signals generated by the same excitation source and simultaneously received by the left and right arrays extracted before formal acquisition or in the acquisition data; determining the time delay between channels by locating the time shift amount corresponding to the peak value of the cross-correlation function; and finally performing reverse time shift on the data of each channel according to the time delay to realize accurate alignment of all channel signals on the time axis.

[0008] Preferably, the alignment step specifically comprises: for each acquisition channel, calculating the cross-correlation function between the channel and a designated reference channel; determining the time delay of all other channel signals relative to the reference channel based on the reference channel; and performing time shift operation on the acquisition signals of all channels according to the time delay, so that the direct wave signals generated by the same pilot excitation wave in all channels are aligned on the time axis.

[0009] Preferably, the low-rank decomposition is singular value decomposition; and the extraction of the first k principal components as common noise modes specifically comprises: performing singular value decomposition on the cross-correlation matrix to obtain a set of left singular vectors, a set of singular values and a set of right singular vectors; selecting the left singular vectors corresponding to the first k largest singular values according to the size of the singular values; and performing linear combination of the selected left singular vectors and the original multi-channel signal of the left array to generate k time series, which are the common noise modes.

[0010] Preferably, the value of k is a positive integer dynamically determined according to the low-rank decomposition result of the cross-correlation matrix; and the determination criterion is that the cumulative value of the strength indicators of the first k principal components accounts for a preset proportion of the total sum of all component strength indicators.

[0011] Preferably, step S4 specifically comprises: firstly, calculating the integral correlation value between each extracted common noise mode and the original signal of each acquisition channel in the left array or the right array, as the contribution coefficient of the noise mode on the specific channel; and then, multiplying each of the k common noise modes by the corresponding contribution coefficient on each acquisition channel, and summing all the multiplication results to obtain the reconstructed complete noise component on the channel.

[0012] In another aspect, the application also provides a symmetric excitation anti-explosion geological exploration system, comprising: a left excitation-acquisition array and a right excitation-acquisition array symmetrically arranged on both sides of the exploration area, and a noise cross-correlation inversion module connected with the signals of the left and right arrays; The noise cross-correlation inversion module is configured to perform the following operations: receive M-channel signals from the left excitation-acquisition array and M-channel signals from the right excitation-acquisition array; perform time alignment correction on the channel signals of the left and right arrays to eliminate the transmission time delay between channels; based on the time-aligned signals, construct a multi-channel cross-correlation matrix between the left and right arrays, wherein the matrix elements are composed of the integral cross-correlation values of the i-th channel signal of the left array and the j-th channel signal of the right array in time; perform low-rank decomposition on the cross-correlation matrix, and extract the first k main eigenvectors as common noise modes; based on the common noise modes, calculate their noise contributions on each acquisition channel by inversion; subtract the corresponding noise contributions from the original acquisition channel signals to output the purified geological reflection signals.

[0013] Preferably, the noise cross-correlation inversion module comprises a time synchronization submodule configured to: use the pilot excitation wave received by the left and right arrays to calculate the cross-correlation function between the channel signals, and determine the time delay between the channels by finding the peak position of the cross-correlation function, and then perform reverse translation on the channel signals to achieve time alignment correction.

[0014] Preferably, when performing time alignment correction, the time synchronization submodule is specifically configured to: for each acquisition channel, calculate the cross-correlation function between the channel and a designated reference channel; determine the time delay of all other channel signals relative to the reference channel based on the reference channel; and perform time shift operation on the acquisition signals of all channels according to the time delay, so that the direct wave signals generated by the same pilot excitation wave in all channels are aligned on the time axis.

[0015] Preferably, the low-rank decomposition is a singular value decomposition, and the operation of extracting the common noise mode specifically comprises: performing singular value decomposition on the cross-correlation matrix to obtain a left singular vector matrix, a singular value matrix and a transpose of a right singular vector matrix; then selecting the first k column vectors in the left singular vector matrix, and linearly combining the k column vectors with corresponding left array original signals to construct the common noise mode.

[0016] The present application realizes a multi-dimensional technical breakthrough by symmetrically arranging left and right excitation-acquisition arrays and combining a noise cross-correlation inversion algorithm: first, without relying on special hardware, the cross-correlation matrix is constructed and the common noise mode is extracted only through algorithm logic, which greatly reduces the system deployment cost and hardware adaptation difficulty; second, by using the difference between the strong correlation of homologous noise and the weak correlation of geological reflection signals, the channel-level noise is accurately reconstructed through low-rank decomposition and mode inversion, the noise suppression rate is high, the signal-to-noise ratio is greatly improved, and the damage to effective signals caused by traditional filtering is effectively avoided; third, the symmetric architecture and time alignment design ensure the consistency of the signals of the two arrays, and the purified signals can significantly improve the geological imaging resolution and accurately identify small-scale faults and other features, providing high-quality data support for tunnel construction safety and geological risk prediction, and having strong practicality and popularization value in the field of underground engineering geological exploration. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0018] Figure 1 The geological exploration system architecture schematic diagram provided for the embodiments of the present application.

[0019] Figure 2 The symmetric excitation anti-explosion geological exploration method flowchart provided for the embodiments of the present application.

[0020] Figure 3 The time correction flowchart provided for the embodiments of the present application.

[0021] Figure 4 The extraction flowchart of the common noise mode provided for the embodiments of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort fall within the protection scope of the present application.

[0023] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort fall within the protection scope of the present application.

[0024] It should be noted that all the user information (including but not limited to user equipment information, user personal information, object information corresponding to device usage data, etc.) and data (including but not limited to data for analysis, stored data, displayed data, device usage data, etc.) involved in all the embodiments of the present disclosure are information and data authorized by the user or authorized by all parties.

[0025] As shown in Figure 2 The present embodiment discloses a symmetrical excitation explosion-proof geological detection method 200, comprising the following method steps: S1: receiving multi-channel signals from the left excitation-collection array and multi-channel signals from the right excitation-collection array, and performing time alignment correction on all channel signals to eliminate time delay caused by transmission path or electronic delay between channels; S2: based on the time-aligned left and right array signals, constructing a multi-channel cross-correlation matrix between the left and right arrays, wherein each element in the cross-correlation matrix is obtained by integrating and cross-correlating the signals of one channel of the left array with the signals of one channel of the right array; S3: performing low-rank decomposition on the constructed cross-correlation matrix, and extracting the first k main components from the decomposition result as common noise modes shared by the left and right arrays; S4: based on the extracted common noise modes, inversely calculating the specific contribution of each noise mode on each acquisition channel of the left and right arrays, and reconstructing the noise component on each acquisition channel composed of the common noise modes according to the specific contribution; S5: subtracting the reconstructed noise component of the channel from the original acquisition channel signals of the left and right arrays, thereby outputting the purified geological reflection signals.

[0026] According to the embodiment of the present application, the symmetric excitation anti-explosion geological exploration method is applied to a geological exploration system 100. The core architecture of the system 100 comprises a left excitation-collection array 102 and a right excitation-collection array 103 symmetrically arranged on both sides of a detection area. Exemplarily, the hardware configurations of the left excitation-collection array and the right excitation-collection array are consistent, that is, the number of collection channels in the two arrays is the same, the sensor models corresponding to each collection channel are consistent, and the hardware parameters (such as cable impedance and amplifier gain) of the signal transmission link are unified, so as to ensure that the signal collection characteristics of the two arrays are consistent under the condition of no external interference, thereby providing a basic hardware condition for subsequent noise separation and signal purification.

[0027] It should be noted that the typical scene of the detection area is a tunnel interior 101. In this scene, there are low-frequency mechanical noises generated by devices such as fans, drilling machines and transport vehicles, and the noises have the same source characteristic. The present method is aimed at the noise interference problem in this scene, and the accurate extraction of geological reflection signals is realized through the signal processing logic of the symmetric array.

[0028] In some embodiments, for step S1, in this step, the core is to receive the multi-channel signals of the left excitation-collection array and the right excitation-collection array, and eliminate the time delay caused by the transmission path or electronic delay between channels, so as to ensure the consistency of all channel signals on the time axis, and provide an accurate timing basis for subsequent cross-correlation matrix construction.

[0029] According to the embodiment of the present application, the signal collection functions of the left excitation-collection array and the right excitation-collection array are first started, so that the two arrays simultaneously collect the geological signals of the detection area. Specifically, the left excitation-collection array comprises M collection channels (M is a positive integer, which can be determined according to the detection accuracy requirement, for example, M = 16, 32 or 64, etc.), each collection channel corresponds to a sensor. After the sensor converts the geological vibration signal into an electrical signal, the signal is transmitted to a signal processing unit through a signal transmission link (such as a shielded cable), and finally a multi-channel signal set of the left array is formed, denoted as wherein represents the signal value of the jth collection channel of the left array at time t; similarly, the right excitation-collection array also comprises M collection channels, and the multi-channel signal set collected thereby is denoted as wherein represents the signal value of the jth collection channel of the right array at time t. It can be understood that the collection frequencies of the two arrays remain consistent during the signal collection process, for example, the collection frequency is set to 1 kHz, 2 kHz or higher, so as to meet the frequency characteristic requirement of the geological reflection signal and avoid signal aliasing.

[0030] In some embodiments, the key to time alignment correction lies in determining the time delay between channels using a leader excitation wave signal and adjusting the signal time-shift based on this delay. See also Figure 3 The time correction process diagram includes the following sub-steps: S201, Acquire the leader excitation wave signal. Specifically, in some embodiments, the leader excitation wave signal can be acquired in two ways: The first way is "pre-excitation before formal acquisition", that is, before the formal acquisition of geological signals is started, the excitation sources (such as controllable seismic sources) of the left excitation-acquisition array and the right excitation-acquisition array are controlled to simultaneously generate a leader excitation wave with a known waveform. The propagation path of the leader excitation wave is fixed (for example, it propagates directly from the excitation source to the sensors of the two arrays without a geological reflection path) and is simultaneously received by all acquisition channels of the left and right arrays. The second method is "extracting from the acquired data." If the formally acquired data contains a direct wave signal generated by the same excitation source (the propagation characteristics of this direct wave signal are consistent with the leader excitation wave, and can be identified through waveform features such as large signal amplitude and steep rise edge), then this direct wave signal can be used as the leader excitation wave signal. Preferably, the first method, "pre-excitation before formal acquisition," is used. In this method, the leader excitation wave signal waveform is known, interference is minimal, and the accuracy of time delay calculation can be improved.

[0031] S202, determine the cross-correlation function calculation and time delay. Specifically, for the acquired leader excitation wave signal, calculate the cross-correlation function between signals from different channels to locate the time delay between each channel. Specifically, first, select a specified reference channel, which can be randomly selected from the acquisition channels of the left or right array. For example, select the first acquisition channel of the left array (i.e., the first acquisition channel of the left array). The corresponding channel is used as the reference channel, denoted as .

[0032] For the other acquisition channels in the left array (channels 2 to M), calculate the cross-correlation function between the leader excitation signal of each channel and the leader excitation signal of the reference channel. Taking the j-th channel (j=2,3,…,M) of the left array as an example, its cross-correlation function with the reference channel is... Defined as: in, Indicates the time offset. This indicates the effective time interval of the leader excitation wave signal (i.e., the leader excitation wave signal exists only within this time interval, which can be determined by the signal amplitude threshold; for example, when the signal amplitude is greater than 3 times the background noise amplitude, it is considered to have entered the effective time interval).

[0033] By solving the cross-correlation function The time offset corresponding to the maximum value , i.e. , i.e. , i.e. the time delay of the jth channel of the left array relative to the reference channel. Similarly, for all acquisition channels (1st to Mth channels) of the right array, the cross-correlation function of the pilot excitation wave signal of each channel and the pilot excitation wave signal of the reference channel (1st channel of the left array) is calculated respectively : , and the time offset corresponding to the maximum value is solved , i.e. , i.e. the time delay of the jth channel of the right array relative to the reference channel.

[0034] In some other embodiments, the reference channel can also be selected as a channel of the right array (such as the 1st channel of the right array), and the calculation logic of the cross-correlation function is consistent with the above, only the reference channel signal needs to be replaced by the signal of the reference channel of the right array, which will not be described here.

[0035] S203, signal reverse time shift and time alignment, specifically, according to the time delay of each channel relative to the reference channel calculated above, the acquisition signals of all channels are subjected to reverse time shift operation to achieve time alignment. Specifically, for the original signal of the jth channel of the left array , if its time delay is , the signal is shifted along the time axis by , to obtain the time-aligned signal ; for the original signal of the jth channel of the right array , if its time delay is , the signal is shifted along the time axis by , to obtain the time-aligned signal .

[0036] Through the above operation, the direct wave signals generated by the same pilot excitation wave in all channels will be completely aligned on the time axis, i.e. for any two channels (whether belonging to the left array or the right array), the peak time and the rising edge time of the direct wave signals remain consistent, thereby eliminating the timing deviation caused by the difference in transmission path length (such as different cable lengths) or electronic delay (such as the difference in amplifier response time) between channels, and providing an accurate timing basis for the construction of the cross-correlation matrix in the subsequent steps.

[0037] In some embodiments, for step S2, after completing the time alignment of all channel signals, this step constructs a multi-channel cross-correlation matrix between the left and right arrays by performing integral cross-correlation operation on the time-aligned left and right array signals, which can quantify the correlation between the channel signals of the left and right arrays, and provide data support for the extraction of the common noise mode in the subsequent steps.

[0038] According to an embodiment of the present application, the time-aligned left array multi-channel signal set is , and the time-aligned right array multi-channel signal set is , where M is the number of acquisition channels of the left and right arrays (as mentioned above, M is a positive integer). Based on the above signal sets, a multi-channel cross-correlation matrix is constructed, denoted as , which is an M x M dimension square matrix, i.e., the number of rows of the matrix is equal to the number of acquisition channels of the left array, and the number of columns is equal to the number of acquisition channels of the right array.

[0039] Each element in the cross-correlation matrix (where i = 1, 2, …, M; j = 1, 2, …, M) is obtained by performing an integral cross-correlation operation on the time-aligned signal of the i-th channel of the left array and the time-aligned signal of the j-th channel of the right array. Specifically, the calculation formula of element is: , where represents the effective time interval of the geological signal acquisition, and the determination of this interval needs to cover the propagation time of the geological reflection signal, for example, the maximum propagation time of the reflection signal is estimated according to the depth range of the detection area, and then the value of is determined (for example, when the detection depth is 100 m, the maximum propagation time is estimated to be 0.1 s according to the wave speed of the geological medium (such as 2000 m / s), and therefore can be set to 0.2 s to ensure that all reflection signals are covered).

[0040] It should be noted that the essence of the integral cross-correlation operation is to measure the similarity of two signals within the time interval : if the two signals are highly similar (such as both containing the same noise), the integral cross-correlation result (i.e., the value of ) is larger; if the similarity of the two signals is low (such as one containing noise and the other containing geological reflection signal, or both being geological reflection signals but with different propagation paths), the integral cross-correlation result is smaller. Based on this characteristic, the elements with larger values in the cross-correlation matrix correspond to strong correlation signals (mainly the same noise) between the left and right array channels, and the elements with smaller values correspond to weak correlation signals (mainly geological reflection signals), thereby achieving a preliminary separation of noise and effective signals at the matrix level.

[0041] In some embodiments, to reduce the influence of random noise on the calculation of the elements of the cross-correlation matrix, the time-aligned signals and may be pre-processed before the integral cross-correlation operation.Preprocessing is performed to remove high-frequency random noise, for example, by using a sliding average filter or wavelet denoising, but attention should be paid to the fact that the preprocessing operation should not affect the waveform characteristics of low-frequency coherent noise (such as 10-100 Hz noise generated by a fan or a drilling machine) so as to avoid destroying the correlation of the noise.

[0042] In some embodiments, for step S3, the step is performed by performing low-rank decomposition on the cross-correlation matrix, using the difference between the noise and the effective signal in the rank characteristics of the matrix to extract the common noise mode shared by the left and right arrays.

[0043] In some embodiments, the low-rank decomposition method is singular value decomposition (SVD). Singular value decomposition is a commonly used matrix decomposition method in linear algebra, which can decompose any M x M matrix into the product of three matrices, and the mathematical expression is: wherein: is an M x M left singular vector matrix, each column of which is a left singular vector, denoted as , and satisfies (I is an M x M unit matrix); is an M x M diagonal matrix, the elements on the diagonal of which are the singular values of the cross-correlation matrix , denoted as , and satisfies ; is an M x M right singular vector matrix , the transpose of which is denoted as , and satisfies .

[0044] The reason for selecting singular value decomposition as the low-rank decomposition method is that the singular value can quantitatively reflect the “energy intensity” of the corresponding component in the matrix, that is, the larger the singular value, the higher the proportion of the signal component represented by the corresponding singular vector in the matrix; at the same time, singular value decomposition has the characteristics of high stability and unique decomposition result (in the case of non-degenerate singular value), which can accurately extract the main signal component in the matrix and avoid errors in noise mode extraction caused by improper selection of decomposition method.

[0045] According to the embodiments of the present application, the coherent noise in the left and right arrays has a high correlation, which is represented as a “low-rank feature” in the cross-correlation matrix , that is, the signal component corresponding to the noise can be represented by only a few singular vectors, while the geological reflection signal has a large difference in waveform due to different propagation paths, which is represented as a “high-rank feature” that requires a large number of singular vectors to be completely represented. Therefore, in the singular values of the cross-correlation matrix , the first k largest singular values ) mainly contributed by homologous noise, and the corresponding left singular vector (u ) can represent the spatial distribution characteristics of the noise, and the subsequent linear combination of the left singular vector and the original signal of the left array can generate the common noise mode.

[0046] Specifically, referring to Figure 4 , the extraction process diagram of the common noise mode includes the following sub-steps: S301, selection of the left singular vector, specifically, according to the size of the singular value, the left singular vector matrix (U) is selected from the left singular vector matrix (U) The first k largest singular values (k ) respectively correspond to the left singular vector, that is, the left singular vector (u ) is selected. Wherein, the value of k is a dynamically determined positive integer, and the determination criterion is that the cumulative value of the intensity indicators of the first k selected main components accounts for a preset proportion of the total sum of all component intensity indicators.

[0047] In the embodiment of the application, the "intensity indicator" is the square of the singular value (k ), because the square of the singular value can reflect the energy of the signal component represented by the corresponding singular vector. Therefore, the determination step of k is as follows: 1. Calculate the sum of the squares of all singular values, denoted as ; 2. Calculate the sum of the squares of the first k singular values, denoted as ; 3. Calculate the cumulative contribution rate ; 4. Select the minimum k value that satisfies , wherein is a preset proportion (for example or , etc.).

[0048] Exemplarily, if M=32, the calculation result is , when k=1, , (smaller than ); when k=2, , (larger than ), then k=2 is selected. The k value determined by this criterion can ensure that the first k main components extracted contain most of the noise energy in the cross-correlation matrix, while avoiding the introduction of too many effective signal components, and taking into account the integrity and accuracy of noise extraction.

[0049] In some embodiments, the value of the preset proportion can be adjusted according to the noise intensity of the detection scene: when the noise intensity in the tunnel is high (such as when the drilling machine is working), the can be set to 90%-95% to ensure that more noise components are extracted; when the noise intensity is low (such as when only the fan is working), the Set to 80%-85% to reduce the mixing of valid signals.

[0050] According to an embodiment of the present application, S302, the selected first k left singular vectors (V1, V2, …, Vk) are linearly combined with the original multi-channel signal of the left array (here, the time-aligned signal Y1, Y2, …, YN) to generate k time series, which are the common noise modes. According to an embodiment of the present application, S302, the selected first k left singular vectors (V1, V2, …, Vk) are linearly combined with the original multi-channel signal of the left array (here, the time-aligned signal Y1, Y2, …, YN) to generate k time series, which are the common noise modes.

[0051] Specifically, for the i-th common noise mode (i = 1, 2, …, k), the calculation formula of its time series is: wherein, represents the j-th element of the i-th left singular vector (i.e., the weight coefficient corresponding to the j-th acquisition channel of the left array), is the time-aligned signal of the j-th acquisition channel of the left array. It needs to be explained that the element of the left singular vector reflects the “contribution weight” of the j-th channel of the left array in the i-th noise mode, i.e., if the absolute value of is large, it means that the signal of the j-th channel of the left array contains more components of the noise mode; if the absolute value of is small, it means that the signal of the j-th channel contains less components of the noise mode. Through linear combination operation, the signals of all channels of the left array are superimposed according to the corresponding weight coefficients, and the noise mode with specific spatial distribution characteristics can be separated, and the noise mode has commonality in the left and right arrays (because it comes from the strong correlation components of the left and right arrays in the cross-correlation matrix), so it is called “common noise mode”.

[0052]

[0053] In some embodiments, the right singular vector can also be linearly combined with the time-aligned signal of the right array to generate the common noise mode, and the calculation formula is: wherein, is the j-th element of the i-th right singular vector. Due to the symmetry of the cross-correlation matrix (R = R), when the left and right array hardware is symmetrical, the common noise modes generated by the left singular vector and the right singular vector are consistent, i.e., therefore, one of the two ways can be selected to generate the common noise mode, and preferably, the left array signal is used to generate the common noise mode to maintain the uniformity of the subsequent signal processing.

[0054] ​​​​​​​​​​​​​​In some embodiments, for step S4, the transformation from "common noise modes" to "channel-level noise components" is achieved by calculating the specific contribution of each common noise mode on each acquisition channel through inversion.

[0055] Specifically, the contribution coefficient is used to quantify the degree of influence of each common noise mode on a specific acquisition channel, and its calculation is based on the integral correlation operation of the common noise mode and the channel original signal. According to the embodiments of the present application, first, for each acquisition channel of the left array, the integral correlation value between each extracted common noise mode and the time-aligned signal of the channel is calculated, which is the contribution coefficient of the noise mode on the channel.

[0056] Specifically, for the jth acquisition channel (j = 1, 2,..., M) of the left array and the ith common noise mode (i = 1, 2,..., k), the contribution coefficient of the noise mode on the channel is calculated as follows: The calculation formula of the contribution coefficient is: , wherein, The time interval is consistent with the calculation of the cross-correlation matrix element in step S2, which ensures the uniformity of the time reference for the calculation of the contribution coefficient.

[0057] Similarly, for the jth acquisition channel (j = 1, 2,..., M) of the right array and the ith common noise mode , the contribution coefficient of the noise mode on the channel is calculated as follows: It should be noted that the physical meaning of the contribution coefficient (or ) is as follows: the "energy coupling degree" of the common noise mode and the jth channel signal of the left array (or the jth channel signal of the right array ), that is, if the absolute value of is large, it means that the energy proportion occupied by in is high, that is, the channel is greatly affected by the noise mode; if the absolute value of is small, it means that the channel is less affected by the noise mode. Through the calculation of the contribution coefficient, the "energy distribution" of the common noise mode on each channel is quantified, which provides a numerical basis for the subsequent reconstruction of the noise component.

[0058] In some embodiments, to avoid the influence of abnormal values on the calculation of the contribution coefficient, the integral correlation operation can be performed on the common noise mode and the channel signal (or ​​) amplitude normalization, i.e. adjusting the signal amplitude to the interval [0, 1], so that the value range of the contribution coefficient is unified, facilitating the subsequent comparison and analysis of multi-channel noise components.

[0059] After obtaining all the contribution coefficients, for each acquisition channel, multiply the k common noise modes by the corresponding contribution coefficient of the channel, and sum all the product results, i.e. the complete noise component composed of common noise modes on the channel can be obtained.

[0060] Specifically, for the jth acquisition channel of the left array, the calculation formula of the reconstructed noise component is as follows: , wherein, represents the specific noise contribution of the ith common noise mode on the jth channel of the left array, and the total noise component of the channel is obtained by superimposing the contributions of the k modes.

[0061] Similarly, for the jth acquisition channel of the right array, the calculation formula of the reconstructed noise component is as follows: , it is emphasized that the reconstructed noise component (or ) can accurately reproduce the wave components caused by the homologous noise in the original channel signal, and the reasons are as follows: 1. The common noise mode has included the noise characteristics common to the left and right arrays (extracted by step S3); 2. The contribution coefficient (or ) has quantified the contribution proportion of each noise mode on a specific channel (calculated by step S4.1); 3. The superposition operation can integrate the contributions of all noise modes to form a reconstructed noise consistent with the original channel noise waveform.

[0062] Through comparative experiments, the similarity of the reconstructed noise component to the noise component in the original channel signal can reach more than 90%, which can meet the accuracy requirements of subsequent noise cancellation.

[0063] In some embodiments, for step S5, this step is the final implementation link, and the accurate cancellation of noise is realized by subtracting the reconstructed channel noise component from the original acquisition signal, and the purified geological reflection signal is finally output.

[0064] According to the embodiments of the present application, for each acquisition channel of the left array, the original signal of the channel after time alignment is subtracted by the corresponding reconstructed noise component , i.e. the purified left array signal is obtained, and the calculation formula is as follows: wherein, j = 1, 2, …, M.

[0065] In the noise cancellation process, the reconstructed noise component is highly consistent with the noise component in the original signal , so the subtraction operation can cancel the noise energy to the maximum extent; and the geological reflection signal in the original signal is not contained in the reconstructed noise component because of the low correlation with the noise mode (represented by small value elements of the cross-correlation matrix in step S2), and thus can be completely retained in the purified signal .

[0066] Similarly, for each acquisition channel of the right array, the time-aligned original signal of the channel is subtracted by the corresponding reconstructed noise component , and the purified right array signal is obtained, and the calculation formula is: wherein, j = 1, 2, …, M.

[0067] It should be noted that the noise cancellation processes of the left array and the right array are independent of each other, but because the hardware of the two arrays is symmetrical and the signal processing logic is consistent, the purified signals and have the same noise suppression effect and can provide bilateral symmetrical high-quality data for subsequent geological imaging.

[0068] After completing the noise cancellation of all channels, the purified left array signal set and the right array signal set are output as the final result, which can be stored in a data storage unit or directly transmitted to a subsequent geological imaging module.

[0069] In some embodiments, to further optimize the quality of the purified signal, post-processing can be performed on and after noise cancellation, such as using adaptive filtering to remove a small amount of non-homologous noise or performing amplitude normalization on the signal to facilitate subsequent geological parameter calculation (such as reflection coefficient, wave velocity, etc.).

[0070] According to another aspect of the present application, a symmetrical explosion-proof geological exploration system is provided, which specifically comprises: a left excitation-acquisition array and a right excitation-acquisition array symmetrically arranged on both sides of the exploration area, and a noise cross-correlation inversion module connected with the left and right array signals; the noise cross-correlation inversion module is configured to perform the following operations: receiving M-channel signals from the left excitation-acquisition array and M-channel signals from the right excitation-acquisition array; time aligning the channel signals of the left and right arrays to eliminate inter-channel transmission time delay; based on the time-aligned signals, constructing a multi-channel cross-correlation matrix between the left and right arrays, wherein the matrix elements are composed of the time integral cross-correlation values of the i-th channel signal of the left array and the j-th channel signal of the right array; performing low-rank decomposition on the cross-correlation matrix to extract the first k principal eigenvectors as common noise modes; based on the common noise modes, inversely calculating their noise contributions on each acquisition channel; subtracting the corresponding noise contributions from the original acquisition channel signals to output the purified geological reflection signals.

[0071] In some possible embodiments, the noise cross-correlation inversion module comprises a time synchronization sub-module configured to: calculate the cross-correlation functions between the channel signals by using the pilot excitation waves received by the left and right arrays, and determine the time delay amount between the channels by finding the peak position of the cross-correlation functions, and then perform reverse translation on the channel signals to achieve time alignment correction.

[0072] In some possible embodiments, when performing time alignment correction, the time synchronization sub-module is specifically configured to: for each acquisition channel, calculate the cross-correlation function between the channel and a designated reference channel; determine the time delay amount of all other channel signals relative to the reference channel based on the reference channel; and perform time shift operation on the acquisition signals of all channels according to the time delay amount, so that the direct wave signals generated by the same pilot excitation wave in all channels are aligned on the time axis.

[0073] In some possible embodiments, the low-rank decomposition is singular value decomposition, and the operation of extracting the common noise modes specifically comprises: performing singular value decomposition on the cross-correlation matrix to obtain a left singular vector matrix, a singular value matrix and a transpose of a right singular vector matrix; then selecting the first k column vectors in the left singular vector matrix, and linearly combining the k column vectors with the corresponding left array original signals to construct the common noise modes.

[0074] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0075] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A symmetrical excitation explosion-proof geological exploration method, characterized in that, The method is applied to a system comprising a left excitation-acquisition array and a right excitation-acquisition array symmetrically arranged on both sides of the detection area, and includes the following steps: S1: Receive multi-channel signals from the left excitation-acquisition array and multi-channel signals from the right excitation-acquisition array, and perform time alignment correction on all channel signals to eliminate time delays between channels caused by transmission paths or electronic delays; S2: Based on the time-aligned left and right array signals, construct a multi-channel cross-correlation matrix between the left and right arrays, wherein each element in the cross-correlation matrix is ​​obtained by performing an integral cross-correlation operation between the signal of one channel of the left array and the signal of one channel of the right array. S3: Perform low-rank decomposition on the constructed cross-correlation matrix, and extract the first k principal components from the decomposition results as the common noise modes shared by the left and right arrays; S4: Based on the extracted common noise modes, calculate the specific contribution of each noise mode to each acquisition channel of the left and right arrays, and reconstruct the noise components composed of the common noise modes on each acquisition channel accordingly. S5: Subtract the noise component of the reconstructed channel from the original acquisition channel signals of the left and right arrays, and output the purified geological reflection signal.

2. The symmetrical excitation explosion-proof geological exploration method according to claim 1, characterized in that, Step S1 specifically includes: using the leader excitation wave signal extracted before formal acquisition or from the acquired data, generated by the same excitation source and simultaneously received by the left and right arrays, to calculate the cross-correlation function between different channel signals; determining the time delay between each channel by locating the time shift corresponding to the peak value of the cross-correlation function; and finally, performing reverse time shift on the data of each channel based on the time delay to achieve precise alignment of all channel signals on the time axis.

3. The symmetrical excitation explosion-proof geological exploration method according to claim 2, characterized in that, The alignment step specifically includes: for each acquisition channel, calculating the cross-correlation function between it and a specified reference channel; using the reference channel as a reference, determining the time delay of all other channel signals relative to the reference channel; and performing a time-shift operation on the acquisition signals of all channels based on the time delay, so that the direct wave signals generated by the same leader excitation wave in all channels are aligned on the time axis.

4. The symmetrical excitation explosion-proof geological exploration method according to claim 1, characterized in that, The low-rank decomposition is a singular value decomposition; the extraction of the top k principal components as the common noise mode specifically includes: performing singular value decomposition on the cross-correlation matrix to obtain a set of left singular vectors, a set of singular values, and a set of right singular vectors; selecting the left singular vectors corresponding to the top k largest singular values ​​according to the magnitude of the singular values; and linearly combining the selected left singular vectors with the original multi-channel signals of the left array to generate k time series, which are the common noise modes.

5. The symmetrical excitation explosion-proof geological exploration method according to claim 4, characterized in that, The value of k is a positive integer that is dynamically determined based on the low-rank decomposition result of the cross-correlation matrix. The criterion for determining this is that the cumulative value of the intensity index of the selected k main components accounts for a predetermined proportion of the sum of the intensity indices of all components.

6. The symmetrical excitation explosion-proof geological exploration method according to claim 1, characterized in that, Step S4 specifically includes: First, calculating the integral correlation value between each extracted common noise mode and the original signal of each acquisition channel in the left or right array, as the contribution coefficient of the noise mode on the specific channel; then, for each acquisition channel, multiplying the k common noise modes by their corresponding contribution coefficients on the channel, and summing all the product results to obtain the reconstructed complete noise component on the channel.

7. A symmetrical excitation explosion-proof geological exploration system, characterized in that, include: The left excitation-acquisition array and the right excitation-acquisition array are symmetrically arranged on both sides of the detection area, and a noise cross-correlation inversion module connected to the signals of the left and right arrays; The noise cross-correlation inversion module is configured to perform the following operations: Receives M-channel signals from the left excitation-acquisition array and M-channel signals from the right excitation-acquisition array; Time alignment correction is performed on the channel signals of the left and right arrays to eliminate inter-channel transmission delay; Based on the time-aligned signals, a multi-channel cross-correlation matrix between the left and right arrays is constructed, where the matrix elements are composed of the time integral cross-correlation values ​​of the i-th channel signal of the left array and the j-th channel signal of the right array. The cross-correlation matrix is ​​decomposed into low rank, and the top k main eigenvectors are extracted as common noise modes. Based on the common noise mode, its noise contribution on each acquisition channel is calculated by inversion. The corresponding noise contribution is subtracted from the original acquisition channel signal to output the purified geological reflection signal.

8. The symmetrical excitation explosion-proof geological exploration system according to claim 7, characterized in that, include, The noise cross-correlation inversion module includes a time synchronization submodule, which is configured to: use the leader excitation wave received by the left and right arrays to calculate the cross-correlation function between the signals of each channel, and determine the time delay between channels by finding the peak position of the cross-correlation function, and then perform reverse translation on the channel signals to achieve time alignment correction.

9. A symmetrical excitation explosion-proof geological exploration system according to claim 8, characterized in that, include, When performing time alignment correction, the time synchronization submodule is specifically used to: calculate the cross-correlation function between each acquisition channel and a specified reference channel; determine the time delay of all other channel signals relative to the reference channel based on the reference channel; and perform a time shift operation on the acquisition signals of all channels according to the time delay, so that the direct wave signals generated by the same leader excitation wave in all channels are aligned on the time axis.

10. A symmetrical excitation explosion-proof geological exploration system according to claim 7, characterized in that, include, The low-rank decomposition is a singular value decomposition. The operation of extracting the common noise mode specifically involves performing singular value decomposition on the cross-correlation matrix to obtain the left singular vector matrix, the singular value matrix, and the transpose of the right singular vector matrix; then selecting the first k column vectors in the left singular vector matrix and linearly combining these k column vectors with the corresponding original signals of the left array to construct the common noise mode.

Images