Fourier Anti-Leakage Seismic Interpolation With Efficient Impulse Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anti-leakage Fourier transform (ALFT) methods for seismic data interpolation and regularization are computationally costly, especially for large seismic datasets, necessitating a more efficient implementation to reduce computational resources and improve processing efficiency.
Innovation Solution
The method involves generating a window function based on the seismic dataset, transforming it to a specific domain range, iteratively estimating a weight function, identifying and removing the strongest attribute from the input spectrum, and combining the output spectrum to generate an efficient output seismic dataset.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional ALFT methods are used for seismic data interpolation and regularization, then high-quality seismic images can be generated, but computational resources and processing time increase significantly
Solution Approach 1:
The patent segments the seismic data processing into distinct phases: initial regularization using a subset of traces, followed by iterative impulse removal on the regularized data. This segmentation allows the computationally intensive ALFT to be applied only once on reduced data, while subsequent processing works on already-regularized data, significantly reducing overall computational burden while maintaining image quality
Solution Approach 2:
The patent extracts and removes impulsive artifacts from the seismic data through iterative identification and subtraction of spike components in the frequency-wavenumber domain. By separating the impulse removal as a distinct post-processing step from the initial ALFT regularization, the method avoids redundant computations and focuses computational resources only where needed, improving processing efficiency without compromising seismic image quality
2Measurement precision
If ALFT is applied to large seismic datasets with fine sampling, then accurate subsurface imaging is achieved, but computational cost increases
Solution Approach 1:
The patent performs preliminary regularization on a subset of seismic traces before applying full ALFT to the complete dataset. This preliminary action creates an initial regularized dataset that serves as a foundation for subsequent processing, reducing the computational burden of applying ALFT to the entire large dataset while maintaining the accuracy needed for precise subsurface imaging
Solution Approach 2:
The patent applies different processing strategies to different portions of the data: initial regularization on representative subsets, followed by iterative impulse removal tailored to local characteristics of the regularized data. This localized approach allows accurate subsurface imaging by adapting processing intensity to data-specific needs rather than uniformly processing entire large datasets, thereby reducing overall computational resource consumption
Data Source
AI summary
Methods and systems for improving the efficiency of ALFT are disclosed. The methods include obtaining a seismic dataset of a subterranean region, generating a window function, transforming the seismic dataset from a first domain to a second domain using a transform function, and transforming the window function from the first domain to the second domain using the transform function. The methods also include generating an input spectrum from the transformed seismic dataset, iteratively, or recursively, until a condition is met, for each slice, estimating a weight function, generating a weighted input spectrum, identifying a strongest attribute, determining an output spectrum based on the identified strongest attribute, and updating the input spectrum by removing the identified strongest attribute. The methods further include determining an output spectrum volume by combining the output spectrum for each slice, and generating an output seismic dataset using an inverse transform function and the output spectrum volume.