Method for enhancing coal rock minor fault response characteristics by fusing wavelet transform

Abstract

The invention discloses a method for enhancing coal and rock minor fault response characteristics by fusing wavelet transform, which belongs to a seismic signal processing and interpretation technology, and comprises the following steps of: acquiring seismic data and performing certain preprocessing; performing multi-resolution decomposition on the seismic signal by using discrete wavelet transform; screening the multi-stage detailed signals, and selecting the detailed signals of the seismic excitation main frequency band as characteristic signals; adding the characteristic signal and the low-frequency approximate signal, and reconstructing time-domain seismic data; and carrying out response characteristic attribute extraction on the reconstructed seismic data and the like. According to the invention, through multi-resolution analysis of wavelet transform, dominant frequency information of seismic data can be directly highlighted, data noise can be suppressed, response characteristics of minor faults can be enhanced, the accuracy of minor fault characteristic identification can be improved, and coal mining can be safer and more efficient.

Description

Technical Field

[0001] The present invention relates to a seismic signal processing and interpretation technology, and more specifically, to a method for enhancing the response characteristics of small coal-rock faults by fusing wavelet transform. Background Art

[0002] At present, with the vigorous development of coal technology, coal mining enterprises have higher requirements for the accuracy of fault identification. In order to further improve the safety of coal mining and improve the efficiency of coal resource mining, it is necessary to further improve the accuracy of fault detection and avoid safety problems in advance, especially for the identification of small faults. At present, the more common fault detection method is to obtain high-precision seismic data through seismic exploration, and then identify faults in combination with interpretation technology. In theory, small faults in coal mines can be identified using high-precision three-dimensional seismic exploration technology, but in actual production, we found that due to the influence of complex geological structures, the characteristics of many small faults will be masked by noise and the strong characteristics of large faults, so that small faults cannot be accurately identified, which in turn affects the safe and efficient mining of coal mines. Therefore, it is necessary to further study and improve the processing and interpretation methods of seismic data obtained by three-dimensional exploration. Summary of the invention

[0003] One of the purposes of the present invention is to address the above-mentioned shortcomings and provide a method for enhancing the response characteristics of small coal rock faults by fusing wavelet transform, in order to solve the technical problems in the prior art that due to the influence of complex geological structures, the characteristics of many small faults will be masked by noise and the strong characteristics of large faults, thus resulting in the inability to accurately identify small faults, affecting the efficient mining of coal mines, etc.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0005] The present invention provides a method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform, wherein the method comprises the following steps:

[0006] Step A: Obtain interpretable seismic data obtained from seismic exploration.

[0007] Step B, using discrete wavelet transform to perform multi-resolution analysis of the seismic signal from time domain to frequency domain, to obtain detailed signals and approximate signals in different frequency bands; the multi-resolution analysis from time domain to frequency domain is to perform continuous high-pass and low-pass filtering on the time domain signal, thereby decomposing the time domain signal into different frequency bands.

[0008] Step C: Select the detailed signal in the frequency band where the main frequency of the earthquake is located from among the detailed signals and approximate signals in different frequency bands, and combine it with the low-frequency approximate signal to form a seismic characteristic signal in the frequency domain.

[0009] Step D: Reconstruct the seismic characteristic signal in the frequency domain into new seismic data in the time domain through inverse wavelet transform. The formula of inverse wavelet transform is:

[0010]

[0011] Where W(a, b) represents the wavelet coefficients, a and b represent the scale parameter and translation parameter respectively. f(t) is the original signal, ψ(t) is the wavelet basis function, and dt represents the integral differential of the time variable t, that is, the integral of time. The principle of wavelet transform can be explained by its formula. For the inverse transform, its formula can be expressed as:

[0012]

[0013] Where f(t) is the original signal in time. W(a, b) is the coefficient obtained by wavelet transform, which depends on the scale parameter a and the translation parameter b. ψ(t) is the wavelet basis function. ψ is a normalization constant associated with the wavelet basis function ψ, ensuring that the wavelet transform is reversible. represents the integration over all possible scales a and translations b. is the scale factor, which is used to adjust the expansion and contraction of the wavelet function. da and db are integral differentials, corresponding to the integral of the scale parameter and the location parameter, respectively.

[0014] Step F, extracting fault response characteristic attributes of the new seismic data; the fault response characteristics of the new seismic data are characterized by calculating the variances of the following formulas.

[0015]

[0016] In the above formula, δ 2 The variance value of the current sampling point, t is the time position of the current sampling point, j is the spatial position of the current sampling point, and w j-t is the triangle weight function, w ij is the weight at the jth sampling point in the window, i is the index of the sampling point, x ij is the i-th amplitude value of the j-th sampling point, is the average amplitude value of the jth sampling point, L is the length of the time window, and I is the amplitude value of each sampling point. is the amplitude value x at the jth sampling point in the window ij The square of the corresponding weight w ij .

[0017] Preferably, a further technical solution is: in step A, the interpretable seismic data obtained from seismic exploration are also preprocessed, and the preprocessing is to perform high-resolution prestack migration seismic data processing on high-density seismic single-shot records collected in the field, so as to obtain processed interpretable seismic data; the high-resolution prestack migration seismic data processing is to enhance the signal-to-noise ratio and resolution of the seismic signal by signal processing, and eliminate interference factors other than geological factors in the seismic signal.

[0018] A further technical solution is that the discrete wavelet of the discrete wavelet transform in the above step B is a suitable mother wavelet, and the maximum order of the discrete wavelet transform decomposition is proportional to the length of the seismic signal.

[0019] A further technical solution is that the length of the seismic signal in the above step B is a power of 2.

[0020] A further technical solution is: the seismic characteristic signal in the frequency domain in the above step C is reconstructed into a new detailed signal and a new approximate signal of the new seismic data through the above inverse wavelet transform formula.

[0021] A further technical solution is that the new detailed signal is a detailed signal of a wavelet frequency band that is consistent with the main frequency of the earthquake, and the new approximate signal is an approximate signal at the maximum scale.

[0022] A further technical solution is: the variance calculation is centered on the sampling point, and the number of sample points of 1 / 2 time window length are selected around, and the sum of the variances of the amplitude value of the sampling point and the amplitude mean of the selected sample point is calculated.

[0023] A further technical solution is: the variance value in step F is obtained by multiplying the variance sum by the above weight function and then normalizing it.

[0024] Compared with the prior art, the beneficial effects of the present invention are at least one of the following: using wavelet transform to transform seismic data from time domain to frequency domain, and taking advantage of the multi-resolution analysis of wavelet transform, retaining the time information of the signal while obtaining the frequency characteristics.

[0025] By integrating and deleting the frequency information, the main frequency information of the earthquake is retained, thereby highlighting the characteristics of small faults and filtering out the characteristics of noise and other interfering signals.

[0026] By amplifying the characteristics of small faults in the frequency domain and suppressing the characteristics of noise in the frequency domain, the characteristics of small faults can be highlighted, and then combined with seismic attribute display to achieve the purpose of enhancing the characteristic response of small faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a method flow chart of an embodiment of the present invention.

[0028] Figure 2 The original seismic data before partial feature enhancement in one embodiment of the present invention.

[0029] Figure 3 for Figure 2 Seismic data after multiresolution analysis using discrete wavelet transform.

[0030] Figure 4 This is seismic data after feature enhancement after main frequency extraction and wavelet inverse transformation reconstruction in one embodiment of the present invention.

[0031] Figure 5 for Figure 4 Variance attribute extraction plot of . DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings.

[0033] One embodiment of the present invention is a method for enhancing the response characteristics of small faults in coal and rock by integrating wavelet transform. The general process of the method is as follows: Figure 1 As shown, the specific implementation steps are as follows:

[0034] P1: Obtain seismic data and perform certain preprocessing.

[0035] The high-density single-shot seismic records collected in the field are processed with high-resolution prestack migration seismic data to obtain interpretable seismic data after processing, such as Figure 2 As shown,

[0036] The high-resolution prestack migration seismic data processing mentioned above refers to the process of enhancing the signal-to-noise ratio and resolution of seismic signals through a series of signal processing methods. Through basic processing methods such as static correction, surface consistency processing, dynamic correction, amplitude compensation, etc., interference in the signal other than geological factors is eliminated as much as possible.

[0037] P2: Use discrete wavelet transform to perform multi-resolution decomposition of seismic signals, as follows: Figure 3 shown.

[0038] In the time domain, we can use different mathematical calculation methods to obtain the phase differences and relationships between data, so as to highlight the different properties and characteristics of geological bodies. The data obtained are collectively called seismic attributes, which further characterize the response characteristics of the fault.

[0039] The purpose of wavelet transform is to represent a changing signal (i.e., time or position with a certain value) to reorganize the information contained in the signal and reveal some properties that are not clearly displayed in the initial representation. It is worth noting that after the wavelet transform decomposes the time signal into a frequency signal, we can observe the main frequency of the signal change and how the main frequency changes with time in the time-frequency space, which is very suitable for approximating data with obvious discontinuities. In the variable time window of the wavelet transform, we can extract information from different frequency bands at the same time according to different needs.

[0040] Discrete wavelet transform analyzes signals in different frequency bands at different resolutions by decomposing them into approximate information and detailed information. Discrete wavelet transform mainly consists of scaling functions and wavelet functions, which are associated with low-pass filters and high-pass filters respectively. The time domain signal can be decomposed into different frequency bands by simply performing continuous high-pass and low-pass filtering on the time domain signal. The original signal x[n] first passes through a half-band high-pass filter g[n] and a low-pass filter h[n]. After filtering, half of the samples can be removed according to the Nyquist rule, and the highest frequency of the signal is π / 2 radians. Therefore, the signal can be sampled twice by simply discarding all other samples. This constitutes a level of decomposition, which can be mathematically expressed as follows:

[0041]

[0042] In the formula, y high[ k] and y low [k] are the outputs of the high-pass filter g[2k-n] and the low-pass filter h[2k-n] after downsampling. x[n] is the discrete time series of the original signal, where n represents the time point. k represents the output sequence y high [k] and y low [k] The time index.

[0043] After decomposition in this way, the current signal has only half the number of samples of the original signal, so the time resolution of the signal is halved, but the frequency resolution is doubled. Because the current frequency band now spans only half of the original frequency band, the uncertainty of the frequency is effectively reduced by half. This process is also called sub-band coding and can be repeated for further decomposition.

[0044] The dominant frequency information of the original signal will appear as high amplitudes in the DWT signal containing these specific frequencies, and the difference from the Fourier transform is that the temporal locality of these frequencies is not lost. However, the resolution of the temporal locality depends on the level at which they occur.

[0045] P3: Screen the multi-level detailed signals and select the detailed signals in the main frequency band of earthquake excitation as characteristic signals.

[0046] In the process of DWT decomposition, the number of samples decreases step by step with the decrease of frequency. The time resolution of the signal decreases with the decrease of frequency. At the same time, the frequency interval decreases with the decrease of frequency, so the frequency resolution of the signal increases. Of course, due to the continuous 2 times of downsampling, the length of the signal must be a power of 2, or at least a multiple of a power of 2, so that the scheme is effective. At the same time, the length of the signal also determines the number of layers into which the signal can be decomposed.

[0047] Assume that the original signal x[n] has 32 sampling points, spanning a frequency band of 0 to 32MHz. The original signal passes through a high-pass filter g[n] and a low-pass filter h[n], and its output is downsampled twice. The output of the high-pass filter is the first-level DWT coefficient, with 16 samples (hence half the time resolution) representing the signal in the frequency range of 16 to 32MHz (hence twice the frequency resolution). The output of the low-pass filter also has 16 samples, but it represents the other half of the signal in the frequency range of 0 to 16MHz. The signal will then continue to pass through the same low-pass and high-pass filters for further decomposition. The output of the second high-pass filter is the second-level DWT coefficient after downsampling, with 8 samples representing the signal in the frequency range of 8 to 16MHz. The output of the second low-pass filter after downsampling also has 8 samples, and it also represents the other half of the signal in the frequency range of 0 to 8MHz. At this time, the time resolution of this signal is half of that of the first-level signal, but the frequency resolution is twice that of the first-level signal, that is, compared with the original signal, the time resolution is reduced by 1 / 4 and the frequency resolution is increased by 1 / 4. The low-pass filter output is then further decomposed again. This process continues until only 1 DWT coefficient can be calculated at the 5th level.

[0048] P4: Add the characteristic signal to the low-frequency approximate signal to reconstruct the time domain seismic data, such as Figure 4 shown.

[0049] An important property of the discrete wavelet transform is that the high-pass and low-pass filters are not independent of each other, they can be related by the formula:

[0050] g[L-1-n]=(-1) n ·h[n]

[0051] Where g[n] is a high-pass filter, h[n] is a low-pass filter, L is the filter length (in samples), and n is the time step. These two filters are inverse versions of each other with alternating odd indices, with the low-pass to high-pass transition being (-1) n The terms are provided, and they are also called quadrature mirror filters (QMF). The formulas for these two filters and the downsampling operation are the same as above:

[0052]

[0053] Reconstruction is very easy in this case because the half-band filters form an orthonormal basis. The above steps are performed in reverse order. The signal at each level is upsampled twice and passed through the synthesis filter g′ [n] and h′ [n] (high pass and low pass respectively) and then add them together. Therefore, the reconstruction formula for each layer is:

[0054]

[0055] P5: Extract response feature attributes from reconstructed seismic data, such as Figure 5 shown.

[0056] There are many ways to express the fault response of seismic data. The most common and widely accepted one is to characterize the fault response characteristics by calculating the variance volume. The larger the variance value of a point, the greater the difference between it and the surrounding area, and the higher the possibility of being a fault. Therefore, in order to characterize the fault response characteristics of the reconstructed data, it is necessary to quantify it by calculating the variance. The calculation formula is as follows:

[0057]

[0058] In the formula, δ 2 The variance value of the current sampling point, t is the time position of the current sampling point, j is the spatial position of the current sampling point, and w j-t is the triangle weight function, w ij is the weight at the jth sampling point in the window, i is the index of the sampling point, x ij is the i-th amplitude value of the j-th sampling point, is the average amplitude value of the jth sampling point, L is the length of the time window, and I is the amplitude value of each sampling point. is the amplitude value x at the jth sampling point in the window ij The square of the corresponding weight w ij .

[0059] It can be seen from the above embodiments that the present invention can directly highlight the main frequency information of seismic data and suppress data noise by utilizing multi-resolution analysis of wavelet transform, thereby enhancing the response characteristics of small faults and improving the accuracy of small fault feature identification, thereby making coal mining safer and more efficient.

[0060] In addition to the above, it should be noted that "one embodiment", "another embodiment", "embodiment", etc. mentioned in this specification refer to the specific features, structures or characteristics described in conjunction with the embodiment included in at least one embodiment generally described in this application. The same expression appearing in multiple places in the specification does not necessarily refer to the same embodiment. Further, when describing a specific feature, structure or characteristic in conjunction with any embodiment, it is claimed that the realization of such feature, structure or characteristic in conjunction with other embodiments also falls within the scope of the present invention.

[0061] Although the present invention is described herein with reference to a number of illustrative embodiments of the present invention, it will be appreciated that those skilled in the art may devise many other modifications and implementations that fall within the scope and spirit of the principles disclosed herein. More specifically, within the scope of the present disclosure, drawings, and claims, a variety of variations and improvements may be made to the components and / or layout of the subject combination layout. In addition to variations and improvements made to the components and / or layout, other uses will also be apparent to those skilled in the art.

Claims

1. A method for enhancing the response characteristics of small faults in coal and rock by integrating wavelet transform, characterized in that The method comprises the following steps: Obtain interpretable seismic data from seismic exploration; Using discrete wavelet transform to perform multi-resolution analysis from time domain to frequency domain on the seismic signal, and obtain detailed signals and approximate signals in different frequency bands; the multi-resolution analysis from time domain to frequency domain is to perform continuous high-pass and low-pass filtering on the time domain signal, and then decompose the time domain signal into different frequency bands; Among the detailed signals and approximate signals in different frequency bands, the detailed signal in the frequency band where the main frequency of the earthquake is located is selected, and combined with the low-frequency approximate signal to form a seismic characteristic signal in the frequency domain; Reconstructing the seismic characteristic signal in the frequency domain into new seismic data in the time domain through inverse wavelet transformation; Extracting fault response characteristic attributes from the new seismic data; the fault response characteristics of the new seismic data are characterized by calculating the variances of the following formulas respectively; In the above formula, δ 2 The variance value of the current sampling point, t is the time position of the current sampling point, j is the spatial position of the current sampling point, and w j-t is the triangle weight function, w ij is the weight at the jth sampling point in the window, i is the index of the sampling point, x ij is the i-th amplitude value of the j-th sampling point, is the average amplitude value of the jth sampling point, L is the length of the time window, and I is the amplitude value of each sampling point. is the amplitude value x at the jth sampling point in the window ij The square of the corresponding weight w ij .

2. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 1 is characterized in that: The method also includes preprocessing the interpretable seismic data obtained from seismic exploration, wherein the preprocessing is to perform high-resolution prestack migration seismic data processing on high-density seismic single-shot records collected in the field, so as to obtain processed interpretable seismic data; the high-resolution prestack migration seismic data processing is to enhance the signal-to-noise ratio and resolution of the seismic signal by signal processing, and eliminate interference factors other than geological factors in the seismic signal.

3. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 1 is characterized in that: The discrete wavelet of the discrete wavelet transform is a suitable mother wavelet, and the maximum order of the discrete wavelet transform decomposition is proportional to the length of the seismic signal.

4. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 3 is characterized in that: The length of the seismic signal is a power of 2.

5. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 1, characterized in that: The seismic characteristic signal in the frequency domain is used to reconstruct a new detailed signal and a new approximate signal of new seismic data through an inverse wavelet transform formula.

6. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 5 is characterized in that: The new detailed signal is a detailed signal of a wavelet frequency band that is consistent with the main frequency of the earthquake, and the new approximate signal is an approximate signal at the maximum scale.

7. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 1 is characterized in that: The variance calculation is centered at the sampling point, and the number of sample points of 1 / 2 time window length is selected in all directions to calculate the sum of the variances of the amplitude value of the sampling point and the amplitude mean of the selected sample points.

8. The method for enhancing the response characteristics of small faults in coal and rock by fusing wavelet transform according to claim 1 or 7, characterized in that: The variance value is obtained by multiplying the variance sum by the above weight function and then normalizing it.

Images