Annihilation Filter for Seismic Cross-Talk Noise Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for deblending seismic data from simultaneous sources often damage the signal during the process of removing cross-talk noise, necessitating a method that can process blended seismic data without such limitations.
Innovation Solution
The application of an annihilation filter to estimate cross-talk noise and convolve it with an operator to form a signal estimate, allowing for the generation of a subsurface image without degrading the signal, as opposed to traditional coherency filters which remove incoherent energy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional coherency filters are used to remove cross-talk noise from simultaneous source data, then cross-talk noise is attenuated, but the signal is damaged
Solution Approach 1:
Instead of directly removing cross-talk noise as in traditional coherency filtering, the patent applies an annihilation filter that inverts the approach by estimating the signal first and then removing it from the blended data. This indirect method prevents signal damage because the estimation process inherently preserves signal characteristics while the subtraction removes both cross-talk and estimated signal, leaving primarily the cross-talk attenuated data.
Solution Approach 2:
The patent introduces an intermediate step of signal estimation using an annihilation filter before the final cross-talk removal. This intermediary estimation acts as a mediator that guides the subtraction process, ensuring that cross-talk is removed while minimizing signal damage. The estimated signal serves as a reference that prevents direct aggressive filtering of the original signal.
2Productivity
If simultaneous source acquisition is used to reduce acquisition time, then productivity increases, but cross-talk noise is introduced
Solution Approach 1:
The patent extracts cross-talk noise from the blended simultaneous source data through the annihilation filter process. By estimating the signal component and subtracting it from the blended data, the method effectively takes out the cross-talk noise while preserving the desired signal, enabling productive simultaneous source acquisition with cleaned output data.
3Object-affected harmful factors
If aggressive filtering is applied to remove cross-talk noise, then cross-talk attenuation improves, but signal fidelity deteriorates
Solution Approach 1:
The patent inverts the traditional filtering approach by estimating the signal first and then removing it through subtraction. This inversion allows for aggressive cross-talk removal because the signal estimation guides the process to preserve signal characteristics. The method achieves high cross-talk attenuation while maintaining signal fidelity by working backwards from signal estimation rather than direct filtering.
Data Source
AI summary
A device, medium and method for deblending seismic data associated with a subsurface of the earth. The method includes receiving an input dataset generated by first and second sources S1 and S2 that are operating as simultaneous sources; arranging the input dataset based on the firing times of source S1; applying with a computing system an annihilation filter to the arranged input dataset to estimate cross-talk noise; convolving the cross-talk noise estimate with an operator to form a signal estimate using the firing times of S1 and S2; and generating an image of the subsurface based on the signal estimate.