Seismic Data High-Frequency Restoration via Time-Frequency Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Seismic data processing methods, such as nonlinear beamforming and supergrouping, enhance signal-to-noise ratio (SNR) but often damage higher frequencies, leading to reduced frequency band and vertical resolution in prestack seismic data, making it challenging to maintain high-frequency content essential for detailed subsurface imaging.
Innovation Solution
The proposed method involves generating time-frequency spectra of original and enhanced seismic traces, calculating noise estimates, and constructing time-frequency masks to restore high-frequency content by recombining amplitude and phase spectra, using techniques like Modified Ideal Rationale Mask (MIRM) and Modified Optimal Ratio Mask (MORM), to generate output traces that preserve high-frequency information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If SNR enhancement procedures (nonlinear beamforming and supergrouping) are applied to prestack seismic data, then signal-to-noise ratio is improved, but higher frequencies are damaged leading to reduced frequency band and vertical resolution
Solution Approach 1:
The patent segments the seismic data processing into distinct stages: first applying SNR enhancement to improve signal quality, then separately restoring high-frequency content through spectral processing. This segmentation allows each stage to optimize for its specific goal without compromising the other.
Solution Approach 2:
The patent changes processing parameters at different stages - using stacking parameters optimized for SNR improvement in the first stage, then applying deconvolution and spectral processing parameters optimized for high-frequency restoration in the second stage. This parameter adaptation resolves the contradiction by allowing each parameter set to serve its specific function.
2Reliability
If local stacking is performed to enhance weak seismic signals, then signal-to-noise ratio improves, but vertical resolution deteriorates due to loss of high-frequency content
Solution Approach 1:
The patent maintains continuous processing where the output of local stacking feeds directly into high-frequency restoration operations. This continuity ensures that signal enhancement and resolution preservation work together in an unbroken processing chain, with each stage building on the previous stage's results.
Solution Approach 2:
The patent temporarily accepts the loss of high-frequency content during the local stacking stage, then recovers it in a subsequent stage through spectral processing and deconvolution. This discarding and recovering approach allows the stacking to focus on signal enhancement while a dedicated later stage restores the frequency content.
3Manufacturing precision
If frequency band is widened to preserve high-frequency content, then vertical resolution improves, but signal-to-noise ratio may deteriorate due to inclusion of noisy high-frequency components
Solution Approach 1:
The patent performs preliminary SNR enhancement through local stacking before attempting to widen the frequency band. This preliminary action improves the quality of the data entering the frequency-widening stage, ensuring that when high-frequency components are restored or enhanced, they have a solid signal foundation that maintains good SNR.
Solution Approach 2:
The patent replaces direct mechanical filtering approaches with spectral processing and deconvolution techniques. Instead of using simple frequency filters that might compromise SNR, the patent uses sophisticated spectral methods that can selectively restore high-frequency content while maintaining the signal-to-noise ratio achieved in earlier processing stages.
Data Source
AI summary
Methods, systems, and computer-readable medium to perform operations including: generating a first time-frequency spectrum of a first seismic trace from an original seismic dataset; generating a second time-frequency spectrum of a second seismic trace from an enhanced seismic dataset, where the second seismic trace corresponds to the first seismic trace; calculating a difference between the first time-frequency spectrum and the second time-frequency spectrum to generate a noise estimate in the first seismic trace; characterizing the initial noise estimate as White Gaussian Noise (WGN); calculating, based on the characterization of the initial noise estimate, a third time-frequency spectrum of a refined noise estimate; constructing, based on the first time-frequency spectrum, the second time-frequency spectrum, and the third time-frequency spectrum, a time-frequency mask (TFM); and using the constructed TFM to generate a fourth time-frequency spectrum of an output trace that corresponds to the first and second seismic traces.