An adaptive window length seismic quality factor determination method

Abstract

The application discloses a method for determining seismic quality factor with adaptive window length, which comprises the following steps: obtaining post-stack seismic data volume and interval velocity field of a target geologic body, and dividing the geologic body according to the interval velocity field; taking out a first seismic trace from the post-stack seismic data volume, processing the seismic trace by using short-time Fourier transform, calculating centroid frequency at each time sample, and performing smoothing filtering to obtain average centroid frequency of each layer; obtaining window length adjustment coefficient to adaptively determine the time window of each layer; performing time-frequency analysis on the seismic data by using generalized linear frequency modulation wavelet transform to obtain time-frequency spectrum of the seismic trace; calculating instantaneous energy spectrum of the seismic trace; extracting maximum peak energy at each time, and calculating seismic quality factor in the time domain of the seismic trace according to the relative attenuation amount of the maximum peak energy within a wavelength; sequentially extracting other seismic traces in the post-stack data volume, repeating the above steps, arranging the quality factors calculated for the seismic traces in sequence, and obtaining the seismic quality factor of the whole geologic body.

Description

Technical Field

[0001] This invention belongs to the field of reservoir prediction in petroleum geophysical exploration, and in particular relates to a method for determining seismic quality factors with adaptive window length. Background Technology

[0002] In seismic exploration, seismic signals are often non-stationary and undergo energy attenuation during propagation through the strata. This attenuation includes non-inherent attenuations unrelated to the elastic properties of the rock strata, such as reflection, transmission, and spherical diffusion, as well as intrinsic attenuations reflecting the microscopic properties of the rock medium, such as viscosity, thermal conductivity, and thermal relaxation. In practical applications, the quality factor Q is typically used to quantitatively describe the absorption attenuation related to the interaction between seismic waves and the two-phase medium.

[0003] In recent years, extracting stratigraphic quality factors from seismic information in the time-frequency domain has become a research hotspot. These time-frequency analysis methods can be broadly categorized into linear and quadratic methods. The classic linear time-frequency analysis method is based on the assumption that the signal is relatively stable over a very short time period, and its accuracy is limited by the Heisenberg uncertainty principle. Quadratic time-frequency analysis methods are limited to frequencies in the time-frequency domain, ignoring the time variable, and therefore have poor adaptability.

[0004] Generalized linear frequency modulated wavelet transform (GFM), as an advanced parametric time-frequency analysis method, eliminates the influence of modulation components in the time-frequency transform by introducing a time-varying demodulation factor, resulting in a more concentrated and instantaneous time spectrum. Window length, as one of the important parameters of the time-frequency transform, is a key factor affecting the analysis results. Previous GFM wavelet transform methods selected a fixed window length for the entire signal. A window that is too short leads to poor resolution of low-frequency components, while a window that is too long leads to poor time resolution of high-frequency components. This prevents the time-frequency resolution from changing with signal characteristics, resulting in large errors in the calculated seismic quality factor and failing to accurately characterize oil and gas reservoirs. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a seismic quality factor determination method that can use adaptive window length while possessing both high energy transientity and resolution, thereby effectively identifying thin reservoir and oil and gas layer boundaries.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] An adaptive window length seismic quality factor determination method includes the following steps:

[0008] S1: Obtain the post-stack seismic data volume and layer velocity field of the target geological body, and layer the geological body according to the layer velocity field;

[0009] S2, extract the first seismic trace from the post-stack seismic data volume, process it using short-time Fourier transform, and calculate the centroid frequency of each time sampling point;

[0010] S3, perform a smoothing filter on the calculated centroid frequency to obtain the average centroid frequency of each stratum;

[0011] S4, obtain the window length adjustment coefficient, and adaptively determine the time window of each layer based on the adjustment coefficient;

[0012] S5. Using the adjusted time window, the seismic data is analyzed by time-frequency analysis using generalized linear frequency modulated wavelet transform to obtain the time spectrum of the seismic trace.

[0013] S6. Using the time spectrum obtained from S5, calculate the instantaneous energy spectrum of the seismic trace;

[0014] S7. Extract the maximum peak energy at each time point, and calculate the seismic quality factor Q in the time domain of the seismic trace based on its relative attenuation within a wavelength.

[0015] S8. Extract other seismic traces from the post-stack data volume in sequence. Repeat steps S2-S7. Finally, arrange the quality factors Q calculated from each seismic trace in order to obtain the seismic quality factor Q of the entire geological volume.

[0016] Preferably, the process of calculating the centroid frequency of the seismic trace is as follows: first, the seismic trace is processed using a short-time Fourier transform:

[0017] S(j,k)=|STFT(j,k)| 2

[0018] In the formula, STFT(j,k) is the short-time Fourier transform result of the seismic record; S(j,k) is the spectrum of the seismic record; j is the time sampling point number; and k is the frequency sampling point number.

[0019] Furthermore, after obtaining the spectrum S(j,k) of the seismic trace, the centroid frequency at each time sampling point is calculated:

[0020]

[0021] In the formula F m (k) is the frequency vector; F s It is the sampling frequency; F real (j) is the centroid frequency.

[0022] Preferably, the extracted centroid frequencies are smoothed and filtered, and the average centroid frequencies of each layer are calculated:

[0023]

[0024] In the formula F aver(l) is the average centroid frequency of each layer; h l It represents the total number of time samples across all layers.

[0025] Preferably, in S4, the ratio of the centroid frequency of each time sample point to the average centroid frequency of each stratum is used as the window length adjustment coefficient:

[0026] C(j)=F real (j) / F aver (j)

[0027] In the formula, C(j) is the window length adjustment coefficient for each time sample point, where the high-frequency component has a larger window length adjustment coefficient and the low-frequency component has a smaller window length adjustment coefficient.

[0028] Furthermore, the time window for each discrete time sample is adaptively determined by adjusting the window length coefficient.

[0029] W(j) = Win / C(j)

[0030] In the formula, W(j) is the time window at each time point after adaptive adjustment; Win is the preset time window; W(j) enables the high-frequency components of the seismic signal to use a short time window to achieve higher time resolution, and enables the low-frequency components of the seismic signal to use a long time window to achieve higher frequency resolution, thereby improving the resolution of the seismic signal in the time and frequency domain.

[0031] Preferably, in S5, the time window adaptively determined by S4 is used, and the generalized linear frequency modulated wavelet transform is used to perform time-frequency analysis on the seismic trace to obtain the time spectrum of the seismic trace:

[0032]

[0033] In the formula, s(τ) is the seismic signal; w(τ-j) is the window function; τ is time; and GLCT(j,k,α) is the time spectrum obtained from the generalized linear frequency-modulated wavelet transform. The time-varying demodulation factor introduced is:

[0034]

[0035] In the formula, α is the rotation angle; F s It is the sampling frequency; T s That is the sampling time.

[0036] Generalized linear frequency modulated wavelet transform, as a parameterized time-frequency analysis method, can eliminate the influence of modulation components in time-frequency transformation by introducing a time-varying demodulation factor. The generated time spectrum has stronger energy concentration and instantaneity, providing a more powerful tool for achieving high-resolution time-frequency analysis of seismic signals.

[0037] Preferably, in S6, the time spectrum obtained in S5 is used to calculate the instantaneous energy spectrum of the seismic trace using the Tiger operator:

[0038] E R (j,k)={Re[GLCT(j,k,α)]} 2 -Re[GLCT(j-1,k,α)]×Re[GLCT(j+1,k,α)]

[0039] E I (j,k)={Im[GLCT(j,k,α)]} 2 -Im[GLCT(j-1,k,α)]×Im[GLCT(j+1,k,α)]

[0040] In the formula, Re[] is the real part of the spectrum when taking complex numbers; Im[] is the imaginary part of the spectrum when taking complex numbers; GLCT(j,k,α) is the time spectrum obtained by the generalized linear frequency modulated wavelet transform; E R (j,k) and E I (j,k) are the real and imaginary parts of the instantaneous frequency amplitude of a single frequency, respectively.

[0041] Preferably, the specific process of S7 is as follows: The formula for extracting the maximum peak energy at each moment is:

[0042]

[0043] In the formula, E(j,k)=E R (j,k)+E I (j,k) is the instantaneous energy spectrum; This involves taking the maximum instantaneous energy at each sampling point. Finally, based on the definition of the quality factor Q, the relative attenuation of the instantaneous peak energy at each time point is calculated. The formula for estimating the seismic quality factor Q is:

[0044]

[0045] In the formula, E0 is the energy at the reference point; E j is the instantaneous seismic energy at each time sample point, where j = 1, 2, ..., N, and N is the total number of time samples.

[0046] Preferably, the specific process of S8 is as follows: extract other seismic traces in the post-stack data volume in sequence, repeat S2-S7, and at this time the seismic quality factor Q field of the entire geological volume is obtained. Attached Figure Description

[0047] Figure 1 This is a flowchart of a method for determining the seismic quality factor with an adaptive window length, provided in an embodiment.

[0048] Figure 2These are single-track post-stack seismic data provided in the embodiment;

[0049] Figure 3 This is the layer velocity curve at the location of the seismic trace provided in the embodiment;

[0050] Figure 4 It is an adaptively determined time window provided in the embodiment;

[0051] Figure 5 This is a schematic diagram of the adaptive time window generalized linear frequency modulated wavelet transform provided in the embodiment;

[0052] Figure 6 This is a seismic trace time spectrum diagram provided in the embodiment;

[0053] Figure 7 This is the instantaneous energy spectrum of the seismic trace provided in the embodiment;

[0054] Figure 8 The example provides the 1 / Q value of the seismic trace calculated using the method of the present invention;

[0055] Figure 9 This is the actual post-stack seismic data volume provided in the embodiment;

[0056] Figure 10 The seismic quality factor Q is calculated based on a fixed time window, as provided in the embodiment.

[0057] Figure 11 This is the actual seismic Q-value profile calculated using the method of the present invention, provided in the embodiment. Detailed Implementation

[0058] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0059] like Figure 1 As shown, this invention proposes a method for determining the seismic quality factor with an adaptive window length, comprising the following steps:

[0060] (i) Obtain the post-stack seismic data volume and layer velocity field of the target geological body, and divide the geological body into layers according to the layer velocity field.

[0061] (ii) Use short-time Fourier transform to process and calculate the centroid frequency at each time sampling point.

[0062] The short-time Fourier transform formula used is as follows:

[0063] S(j,k)=|STFT(j,k)| 2 (1)

[0064] In the formula, STFT(j,k) is the short-time Fourier transform result of the seismic record; S(j,k) is the spectrum of the seismic record; j is the time sample number; and k is the frequency sample number.

[0065] The formula for calculating the centroid frequency is:

[0066]

[0067] In the formula F m (k) is the frequency vector; F s It is the sampling frequency; F real (j) is the centroid frequency.

[0068] (iii) Smooth the calculated centroid frequencies to obtain the average centroid frequencies of each stratum.

[0069] The specific formula for calculating the average centroid frequency of each stratum is as follows:

[0070]

[0071] In the formula F aver (l) is the average centroid frequency of each layer; h l It represents the total number of time samples across all layers.

[0072] (iv) Obtain the window length adjustment coefficient and adaptively determine the time window of each layer based on the adjustment coefficient.

[0073] The ratio of the centroid frequency at each time point to the average centroid frequency of each stratum is used as the window length adjustment coefficient:

[0074] C(j)=F real (j) / F aver (l) (4)

[0075] In the formula, C(j) is the window length adjustment coefficient for each time sample point, where the high-frequency component has a larger window length adjustment coefficient and the low-frequency component has a smaller window length adjustment coefficient.

[0076] Furthermore, by adjusting the window length coefficient, the time window for each discrete-time sample in the generalized linear frequency modulated wavelet transform is adaptively determined:

[0077] W(j)=Win / C(j) (5)

[0078] In the formula, W(j) is the adaptively adjusted time window; Win is the preset time window; W(j) enables the high-frequency components of the seismic signal to use a short time window to achieve higher time resolution, and enables the low-frequency components of the seismic signal to use a long time window to achieve higher frequency resolution, thereby improving the resolution of the seismic signal in the time and frequency domain.

[0079] (v) Using the adjusted time window, the seismic data is analyzed in time and frequency using generalized linear frequency modulated wavelet transform to obtain the time spectrum of the seismic trace.

[0080] The formula for time-frequency analysis using generalized linear frequency-modulated wavelet transform is as follows:

[0081]

[0082] In the formula, s(τ) is the seismic signal; w(τ-j) is the window function; τ is time; and GLCT(j,k,α) is the time spectrum obtained from the generalized linear frequency-modulated wavelet transform. The time-varying demodulation factor introduced is:

[0083]

[0084] In the formula, α is the rotation angle; F s It is the sampling frequency; T s That is the sampling time.

[0085] Generalized linear frequency modulated wavelet transform, as a parameterized time-frequency analysis method, can eliminate the influence of modulation components in time-frequency transformation by introducing a time-varying demodulation factor. The generated time spectrum has stronger energy concentration and instantaneity, providing a more powerful tool for achieving high-resolution time-frequency analysis of seismic signals.

[0086] (vi) Calculate the instantaneous energy spectrum of the seismic trace.

[0087] Calculating the instantaneous energy spectrum of seismic traces using the Tiger operator:

[0088] E R (j,k)={Re[GLCT(j,k,α)]} 2 -Re[GLCT(j-1,k,α)]×Re[GLCT(j+1,k,α)]

[0089] E I (j,k)={Im[GLCT(j,k,α)]} 2 -Im[GLCT(j-1,k,α)]×Im[GLCT(j+1,k,α)] (8)

[0090] In the formula, Re[] is the real part of the spectrum when taking complex numbers; Im[] is the imaginary part of the spectrum when taking complex numbers; GLCT(j,k,α) is the time spectrum obtained by the generalized linear frequency modulated wavelet transform; E R (j,k) and E I (j,k) are the real and imaginary parts of the instantaneous frequency amplitude of a single frequency, respectively.

[0091] (vii) Extract the maximum peak energy at each time point and calculate the seismic quality factor Q in the time domain of the seismic trace based on its relative attenuation within a wavelength.

[0092] Combining the instantaneous spectral peak extraction method in spectral decomposition, the method for extracting the maximum peak principal energy at each time sample point is as follows:

[0093]

[0094] In the formula, E(j,k)=E R (j,k)+E I (j,k) is the instantaneous energy spectrum; This involves taking the maximum instantaneous energy value at each time sampling point. Finally, based on the basic definition of the quality factor Q, the relative attenuation of the instantaneous peak energy at each time sampling point is calculated. The specific formula for estimating the seismic quality factor Q is as follows:

[0095]

[0096] In the formula, E0 is the energy at the reference point; E j is the instantaneous seismic energy at each time sample point, where j = 1, 2, ..., N, and N is the total number of time samples.

[0097] (viii) Extract other seismic traces from the post-stack data volume in sequence, repeat steps (ii) to (vii), arrange the quality factors Q calculated from each seismic trace in order, and finally obtain the seismic quality factor Q of the entire geological body.

[0098] The process of obtaining the seismic quality factor of this invention is shown below based on a practical example.

[0099] Example 1. The method for determining the seismic quality factor with an adaptive window length according to the present invention is characterized by comprising the following steps:

[0100] In specific implementation, step 1: Obtain the post-stack seismic data volume and layer velocity field of the target geological body, and layer the geological body according to the layer velocity field, including:

[0101] like Figure 2 The image shows the input single-channel post-stack seismic data, with a duration of 7 seconds. The locations of oil and gas reservoirs are at 4s and 5s.

[0102] like Figure 3 The image shows the layer velocity curves at the location of the input seismic trace, and the geological body is divided into h0, h1, ..., h based on the layer velocity. 15 It has 15 floors in total.

[0103] In practice, step 2 involves processing the seismic data using a short-time Fourier transform, with a window length of 100.

[0104] S(j,k)=|STFT(j,k)| 2 (1)

[0105] In the formula, STFT(j,k) is the short-time Fourier transform result of the seismic record; S(j,k) is the spectrum of the seismic record; j is the time sampling point number; and k is the frequency sampling point number.

[0106] Furthermore, after obtaining the spectrum S(j,k) of the seismic trace, the centroid frequency at each time sampling point is calculated:

[0107]

[0108] In the formula F m (k) is the frequency vector; F s It is the sampling frequency; F real (j) is the centroid frequency.

[0109] In practice, step 3 involves smoothing the extracted centroid frequencies and calculating the average centroid frequencies of each layer.

[0110]

[0111] In the formula F aver (l) is the average centroid frequency of each layer; h l It represents the number of time samples at each layer.

[0112] In practice, step 4: use the ratio of the centroid frequency of each time sample point to the average centroid frequency of each stratum as the window length adjustment coefficient.

[0113] C(j)=F real (j) / F aver (l) (4)

[0114] In the formula, C(j) is the window length adjustment coefficient for each time sampling point.

[0115] Based on the adjustment coefficient in formula (4), the time window for each time sampling point is adaptively determined, and the formula is as follows:

[0116] W(j)=Win / C(j) (5)

[0117] In the formula, W(j) is the adaptively adjusted time window; Win is the preset time window.

[0118] like Figure 4 The figure shows the time window adaptively determined based on the window length adjustment coefficient. In this case, the pre-set window length is 100. As can be seen from the figure, the window length changes adaptively with the frequency of the seismic signal. Within 3 seconds, the seismic signal frequency is relatively high, so a short window length is used to achieve high time resolution in the time spectrum. However, within 3 seconds to 7 seconds, the seismic signal frequency gradually decreases after absorption and attenuation by the strata, so a long window length is used to achieve high frequency resolution in the time spectrum, thereby improving the problem of poor time-frequency resolution with a fixed window length.

[0119] In specific implementation, step 5: Using the time window determined adaptively in step 4, perform time-frequency analysis on the seismic trace using generalized linear frequency modulated wavelet transform to obtain the time spectrum of the seismic trace:

[0120]

[0121] In the formula, s(τ) is the seismic signal; w(τ-j) is the window function; τ is time; and GLCT(j,k,α) is the time spectrum obtained from the generalized linear frequency-modulated wavelet transform. The time-varying demodulation factor introduced is:

[0122]

[0123] In the formula, α is the rotation angle; F s It is the sampling frequency; T s That is the sampling time.

[0124] like Figure 5 The diagram illustrates the adaptive window-length generalized linear frequency modulated wavelet transform provided in this embodiment of the invention. Solid lines represent the actual instantaneous frequency, dashed boxes represent the analysis window, and gray areas represent the distribution of time-frequency energy. It can be seen that by selecting high-energy-density portions within multiple frequency-modulated windows and adaptively changing the window length with frequency, the adaptive window-length generalized linear frequency modulated wavelet transform provided in this case can achieve higher energy density than any fixed-window-length frequency-modulated window, thus providing more accurate time-frequency analysis results.

[0125] like Figure 6 The figure shows the time spectrum obtained after analyzing the post-stack seismic trace using adaptive window length generalized linear frequency modulated wavelet transform. It can be seen from the figure that strong amplitude values ​​appeared in the 4th to 6th seconds.

[0126] In specific implementation, step 6: Using the time spectrum obtained in step 5, the instantaneous energy spectrum of the single-channel seismic data is calculated using the Tiger operator. The Tiger operator can accurately estimate the single-frequency instantaneous energy of the signal. It can accurately map the time-frequency information into a two-dimensional distribution reflecting the instantaneous energy characteristics of the seismic signal. The formula for the discrete Tiger operator is as follows:

[0127] E R (j,k)={Re[GLCT(j,k,α)]} 2 -Re[GLCT(j-1,k,α)]×Re[GLCT(j+1,k,α)]

[0128] E I (j,k)={Im[GLCT(j,k,α)]} 2-Im[GLCT(j-1,k,α)]×Im[GLCT(j+1,k,α)] (8)

[0129] In the formula, Re[] is the real part of the spectrum when taking complex numbers; Im[] is the imaginary part of the spectrum when taking complex numbers; GLCT(j,k,α) is the time spectrum obtained by the generalized linear frequency modulated wavelet transform; E R (j,k) and E I (j,k) are the real and imaginary parts of the instantaneous frequency amplitude, respectively.

[0130] like Figure 7 The figure shows the time spectrum obtained in step 6. The instantaneous energy spectrum of the seismic signal was further calculated using the Tiger operator. The figure shows that the Tiger operator effectively identified the target reservoir, and that the seismic reflection frequency at 5 seconds significantly decreased after attenuation. This demonstrates that the Tiger operator exhibits strong concentration and instantaneous accuracy in calculating seismic signal energy, enabling accurate reservoir location.

[0131] In specific implementation, step 7: Combining the idea of ​​instantaneous spectral peak extraction in spectral decomposition, the formula for extracting the instantaneous peak energy of seismic waves, i.e., the instantaneous maximum value of energy distribution, is as follows:

[0132]

[0133] In the formula, E(j,k)=E R (j,k)+E I (j,k) is the instantaneous energy spectrum of Tiger; This involves taking the maximum instantaneous energy at each time sampling point. Finally, based on the definition of the quality factor Q, the relative attenuation of the instantaneous peak energy at each time sampling point is calculated, and the formula for estimating the seismic quality factor Q is:

[0134]

[0135] In the formula, E0 is the energy at the reference point; E j is the instantaneous seismic energy at each time sample point, where j = 1, 2, ..., N, N is the total number of time samples, and Q is the seismic quality factor. The reciprocal of the seismic quality factor Q, i.e., 1 / Q, can be taken as needed. This formula links reservoir characteristics with the energy properties of seismic signals. By calculating the Q value of each time sample point using formula (10), the seismic quality Q of the entire target geological body can be obtained.

[0136] like Figure 8 The figure shows 1 / Q of the post-stack seismic trace calculated based on the method of the present invention. It can be seen from the figure that Q anomalies appeared near the 4s and 5s, and there were no obvious anomalies at other positions. This shows that the Q value calculated based on the method of the present invention can accurately locate the target reservoir, thus proving the effectiveness of the method of the present invention.

[0137] Example 2. Further, the effectiveness and practicality of the adaptive window length seismic quality factor calculation method provided by this invention in identifying reservoir hydrocarbon content will be illustrated below through specific application examples.

[0138] The seismic quality factor determination method with adaptive window length of the present invention was tested and analyzed using seismic data from an oilfield in the Tarim Basin, and good results were achieved. Figure 9 These are actual earthquake data. As can be seen from the map, the geological features of this area are relatively simple, mainly consisting of slopes. Figure 10 The seismic quality factor Q profile is calculated based on a fixed window length. The result roughly reflects the distribution pattern of the seismic quality factor Q, but the resolution is poor and it cannot accurately locate the target reservoir. Figure 11 The seismic quality factor Q profile calculated based on the adaptive window length seismic quality factor determination method of the present invention is shown in the figure. It can be seen from the figure that its instantaneous energy is stronger and it clearly depicts the spatial distribution and boundary of the reservoir. This proves that the seismic quality factor Q calculated by the method of the present invention has a stronger ability to identify thin reservoirs and a higher lateral resolution, and further verifies the effectiveness, stability and accuracy of the method of the present invention.

Claims

1. A method for determining the seismic quality factor with an adaptive window length, characterized in that, Includes the following processes: S1: Obtain the post-stack seismic data volume and layer velocity field of the target geological body, and layer the geological body according to the layer velocity field; S2, extract the first seismic trace from the post-stack seismic data volume, process it using short-time Fourier transform, and calculate the centroid frequency at each time sampling point; S3, perform a smoothing filter on the calculated centroid frequency to obtain the average centroid frequency of each stratum; S4, obtain the window length adjustment coefficient, and adaptively determine the time window of each layer based on the adjustment coefficient; In S4, the ratio of the centroid frequency of each time sample point to the average centroid frequency of each stratum is used as the window length adjustment coefficient: ， In the formula It is the window length adjustment coefficient for each time sampling point; Based on the window length adjustment coefficient, the time window for each time sampling point is adaptively determined using the following formula: ， In the formula, W(j) is the time window at each time point after adaptive adjustment; Win is the preset time window; S5. Using the adjusted time window, the seismic data is analyzed by time-frequency analysis using generalized linear frequency modulated wavelet transform to obtain the time spectrum of the seismic trace. S6. Using the time spectrum obtained from S5, calculate the instantaneous energy spectrum of the seismic trace; S7, extract the maximum peak energy at each time point, and calculate the seismic quality factor Q in the time domain of the seismic trace based on its relative attenuation within a wavelength. S8. Extract other seismic traces from the post-stack data volume in sequence. Repeat steps S2-S7. Finally, arrange the quality factors Q calculated from each seismic trace in order to obtain the seismic quality factor Q of the entire geological volume.

2. The method for determining the seismic quality factor with an adaptive window length according to claim 1, characterized in that: In S1, the geological body is divided into layers based on the layer velocity field. Layer, i.e. , where h l It represents the total number of time samples contained in each layer.

3. The method for determining the seismic quality factor with an adaptive window length according to claim 1, characterized in that: The spectrum of the short-time Fourier transform in S2 is as follows: ， In the formula, STFT(j, k) is the short-time Fourier transform result of the seismic record; S(j, k) is the spectrum of the seismic record; j is the time sampling point number; and k is the frequency sampling point number. The formula for calculating the centroid frequency is: ， In the formula F m (k) is the frequency vector; F s It is the sampling frequency; F real (j) is the centroid frequency.

4. The method for determining the seismic quality factor with an adaptive window length according to claim 2, characterized in that: In S3, a smoothing filter is applied to the centroid frequency, and the average centroid frequency of each layer is calculated: ， In the formula It is the average centroid frequency of each layer; h l It represents the total number of time samples across all layers.

5. The method for determining the seismic quality factor with an adaptive window length according to claim 4, characterized in that: The formula for time-frequency analysis using the generalized linear frequency-modulated wavelet transform in S5 is as follows: ， In the formula, s(τ) is the seismic signal; w(τ-j) is the window function; and τ is time. It is the time spectrum obtained by the generalized linear frequency modulated wavelet transform; The time-varying demodulation factor introduced is: ， In the formula, α is the rotation angle; F s It is the sampling frequency; T s That is the sampling time.

6. The method for determining the seismic quality factor with an adaptive window length according to claim 5, characterized in that: In S6, the time spectrum is converted into an instantaneous energy spectrum using the Tiger operator. The discrete implementation method of the Tiger energy operator is as follows: ， ， In the formula, Re[] is the real part of the spectrum when taking complex numbers; Im[] is the imaginary part of the spectrum when taking complex numbers; It is the time spectrum obtained from the generalized linear frequency modulated wavelet transform; E R (j, k) and E I (j, k) are the real and imaginary parts of the instantaneous frequency amplitude of a single frequency, respectively.

7. The method for determining the seismic quality factor with an adaptive window length according to claim 6, characterized in that: In S7, the method of extracting instantaneous spectral peaks in spectral decomposition is combined to determine the maximum peak principal energy of each time sample point, i.e.: ， In the formula It is Tiger's instantaneous energy spectrum; The maximum instantaneous energy value is taken at each time sampling point; finally, according to the definition of the quality factor Q, the relative attenuation of the instantaneous peak energy at each time sampling point is calculated, and the formula for estimating the seismic quality factor Q is: ， In the formula, E0 is the energy at the reference point; E j is the instantaneous seismic energy at each time sample point, where j=1, 2, … N, and N is the total number of time samples.

Images