CN-122017989-A - Rayleigh wave double-component fusion method and system based on model prediction

CN122017989ACN 122017989 ACN122017989 ACN 122017989ACN-122017989-A

Abstract

The application belongs to the technical field of exploration geophysics and seismic signal processing, and relates to a Rayleigh wave double-component fusion method and system based on model prediction. Aiming at the problems that fusion decision depends on an experience threshold value and the effect of prediction cannot be quantified, the method comprises the steps of obtaining seismic data continuously collected by station pairs in a target area, calculating the signal-to-noise ratio of a ZZ cross-correlation function and the signal-to-noise ratio of an RR cross-correlation function of each station pair, calculating the ratio of the signal-to-noise ratio of the RR cross-correlation function of each station pair to the signal-to-noise ratio of the ZZ cross-correlation function, calculating a similarity index according to the ratio, predicting fusion gain by adopting the similarity index, intelligently deciding whether to perform double-component fusion according to the comparison result of the prediction gain and the target threshold value, and outputting a final effective signal. The application improves the fusion decision from experience judgment to a model driving process which can be calculated and optimized, can stably and reliably improve the overall signal-to-noise ratio of the data, and directly improves the quality of the seismic data for imaging.

Inventors

PANG GUANGHUA
DONG JIZHE
JIANG HAIYU
ZHANG NIAONA

Assignees

长春工业大学

Dates

Publication Date: 20260512
Application Date: 20260210

Claims (10)

1. A Rayleigh wave bi-component fusion method based on model prediction is characterized by comprising the following steps: acquiring continuously acquired seismic data of a station pair in a target area; Calculating the signal-to-noise ratio of the ZZ cross-correlation function and the signal-to-noise ratio of the RR cross-correlation function of each station pair; Calculating the ratio of the signal-to-noise ratio of the RR cross-correlation function to the signal-to-noise ratio of the ZZ cross-correlation function of each station pair; Calculating a similarity index according to the ratio; predicting fusion gain by adopting a similarity index; comparing the predicted fusion gain with a quality safety threshold; And selecting and outputting the final effective signal according to the comparison result.
2. The method for model-based prediction of Rayleigh wave bi-component fusion according to claim 1, wherein, The similarity index is calculated according to the ratio: , Is used as an index of the similarity degree, Is the ratio.
3. The method for model-based prediction of rayleigh bi-component fusion according to claim 2, wherein the method for predicting the fusion gain by using the similarity index comprises the steps of: Calculating a fusion gain by adopting a quantitative relation model: , wherein, And In order to fit the constants, Is the fusion gain.
4. A method of rayleigh bi-component fusion based on model prediction as claimed in claim 3, wherein the quantitative relation model is obtained by linear regression using least square method with a plurality of station pairs with different signal to noise ratio levels as sample data sets, comprising: Calculating the signal-to-noise ratio of the ZZ cross-correlation function and the signal-to-noise ratio of the RR cross-correlation function of the sample data set; Calculating the signal-to-noise ratio of RR cross-correlation function the ratio of signal to noise ratio of the ZZ cross-correlation function; Calculating a similarity index according to the ratio; After linearly superposing the ZZ cross-correlation function and the RR cross-correlation function, calculating the signal-to-noise ratio after superposition; Calculating a fusion signal-to-noise ratio gain according to the superimposed signal-to-noise ratio; and taking the fused signal-to-noise ratio gain as a dependent variable, taking a similarity index as an independent variable, and adopting a least square method to carry out linear regression.
5. The method for combining the rayleigh wave bi-components based on model prediction according to claim 4, wherein the calculation formula of the combined signal-to-noise ratio gain is as follows: , Wherein, the In order to integrate the signal-to-noise ratio gain, For the signal-to-noise ratio after the superposition, For the signal-to-noise ratio of the ZZ cross-correlation function, SNR (RR) is the signal-to-noise ratio of the RR cross-correlation function, To take the maximum value.
6. A method for model-predicted rayleigh bi-component fusion according to claim 3, If the predicted fusion gain is smaller than the quality safety threshold, comparing the signal to noise ratio of the ZZ cross-correlation function and the RR cross-correlation function of the station pair, and selecting the cross-correlation function with a large signal to noise ratio as the final effective signal.
7. A Rayleigh wave double-component fusion system based on model prediction is characterized in that, The system comprises a prediction module, a calculation module, a fusion gain prediction module, a calculation module and a fusion gain prediction module, wherein the prediction module is used for acquiring seismic data continuously acquired by station pairs in a target area, calculating the signal-to-noise ratio of the ZZ cross-correlation function and the signal-to-noise ratio of the RR cross-correlation function of each station pair, calculating the ratio of the signal-to-noise ratio of the RR cross-correlation function of each station pair and the signal-to-noise ratio of the ZZ cross-correlation function, calculating a similarity index according to the ratio, and adopting the similarity index to predict the fusion gain; And the decision module is used for comparing the predicted fusion gain with a quality safety threshold value, and selecting and outputting the final effective signal according to the comparison result.
8. The model prediction-based rayleigh two-component fusion system according to claim 7, wherein the similarity index is calculated according to the ratio as follows: , Is used as an index of the similarity degree, Is the ratio.
9. The model prediction-based rayleigh bi-component fusion system according to claim 7, wherein predicting the fusion gain using the similarity index comprises: Calculating a fusion gain by adopting a quantitative relation model: , wherein, And In order to fit the constants, Is the fusion gain; the quantitative relation model is obtained by taking a plurality of station pairs with different signal to noise ratio levels as a sample data set and adopting a least square method to carry out linear regression, and comprises the following steps: Calculating the signal-to-noise ratio of the ZZ cross-correlation function and the signal-to-noise ratio of the RR cross-correlation function of the sample data set; Calculating the signal-to-noise ratio of RR cross-correlation function the ratio of signal to noise ratio of the ZZ cross-correlation function; Calculating a similarity index according to the ratio; After linearly superposing the ZZ cross-correlation function and the RR cross-correlation function, calculating the signal-to-noise ratio after superposition; Calculating a fusion signal-to-noise ratio gain according to the superimposed signal-to-noise ratio; and taking the fused signal-to-noise ratio gain as a dependent variable, taking a similarity index as an independent variable, and adopting a least square method to carry out linear regression.
10. The model prediction-based rayleigh wave bi-component fusion system according to claim 9, wherein the obtained similarity index is used as a quality safety threshold by setting a target gain to reversely solve the similarity index.

Description

Rayleigh wave double-component fusion method and system based on model prediction Technical Field The application belongs to the technical field of exploration geophysics and seismic signal processing, and particularly relates to a Rayleigh wave double-component fusion method and system based on model prediction. Background When using seismic background noise for surface wave imaging, the rayleigh wave signal exists mainly in both the vertical (Z) and radial (R) components. The current mainstream method generally only adopts Z component data with higher signal-to-noise ratio to process, and directly discards R component data with lower signal-to-noise ratio, so that the data utilization rate is low, and the potential of dual component data cannot be fully utilized. Although the R component also carries an effective signal, practice shows that if Z and R components are directly superimposed without discrimination (i.e., blind fusion), noise in the low quality signal can contaminate the high quality signal, which in turn results in a decrease in the overall signal-to-noise ratio of the fused signal. To overcome the drawbacks of blind fusion, several fusion decision methods based on signal quality assessment have been developed in the art. For example, there is a class of methods to determine whether to fuse by comparing the two-component signal-to-noise ratio similarities. The method can avoid the worst condition to a certain extent, and improves the fusion decision from completely random to a rule-based level. However, the applicant research finds that such an empirical criterion method based on signal-to-noise ratio similarity still has limitations in principle. First, its decision core relies on a static, empirical quality safety threshold δ. The determination of this threshold lacks theoretical guidance and typically relies on trial and error and manual debugging of a particular data set. This not only results in inefficient parameter optimization processes, poor portability, but the more essential problem is that it simplifies the continuous scientific decision of "whether to fuse" or not to a binary judgment of "not to do so" (whether to fall into or not to fall into a fixed interval), failing to make a refined assessment of cases in a critical state. A further limitation is that the prior art (including the above methods) lacks systematic quantitative research and theoretical guidance on the fundamental question of how the fusion effect varies with the two-component mass difference. Practice in the field has realized that excessive signal-to-noise ratio differences can lead to fusion failure, but the key problems of how to quantitatively influence gain, whether a predictable revenue inflection point exists and the like for specific differences still remain in a qualitative cognition or experience intuition level. Because the exact quantitative relation model (Y=F (X)) between the fusion gain (Y) and the signal-to-noise ratio difference (X) can not be established all the time, the prior art has the inherent defects that (1) the fusion effect of unprocessed data can not be predicted in a reliable way, (2) the decision parameters can not be reversely optimized according to the target effect to realize self-adaptive control, and (3) the automatic processing flow requiring accurate quality control is difficult to embed. The present application is therefore directed to providing a more accurate, predictable, and optimizable fusion decision scheme. Disclosure of Invention In order to solve the technical problems, the application discovers that in Rayleigh wave bi-component fusion, a stable linear negative correlation relationship exists between fusion gain Y and two-component signal-to-noise ratio similarity index X, namely Y= -kX+C. Based on the finding, the application provides a Rayleigh wave double-component fusion method and a Rayleigh wave double-component fusion system based on model prediction. The application provides a Rayleigh wave bi-component fusion method based on model prediction, which comprises the following steps: acquiring continuously acquired seismic data of a station pair in a target area; Calculating the signal-to-noise ratio of the ZZ cross-correlation function and the signal-to-noise ratio of the RR cross-correlation function of each station pair; Calculating the ratio of the signal-to-noise ratio of the RR cross-correlation function to the signal-to-noise ratio of the ZZ cross-correlation function of each station pair; Calculating a similarity index according to the ratio; predicting fusion gain by adopting a similarity index; comparing the predicted fusion gain with a quality safety threshold; And selecting and outputting the final effective signal according to the comparison result. Further, the similarity index is calculated according to the ratio as follows:, Is used as an index of the similarity degree, Is the ratio. Further, predicting the fusion gain using the similarity index