Vector Denoising for Multicomponent Seismic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for denoising multicomponent seismic data struggle to effectively separate effective signals from noise within the same time window, leading to loss of vector amplitude information and poor signal-to-noise ratios, especially for horizontal components.
Innovation Solution
A vector denoising method that calculates a first mean wave vector using a moving window, applies median filtering to obtain the true modulus of the ground roll wave, and then uses a second moving window for mean filtering and further median filtering within the same receiving line to suppress random noise, thereby isolating the purified effective signal vector without relying on ellipticity or directionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional denoising methods are applied to multicomponent seismic data, then signal-to-noise ratio is improved, but vector amplitude information is lost
Solution Approach 1:
The patent segments the denoising process into distinct stages: first separating ground roll waves from effective signals using polarization filtering, then denoising the effective signal component. This segmentation allows each stage to focus on specific aspects, preserving vector amplitude information while improving signal-to-noise ratio.
Solution Approach 2:
The patent extracts and removes ground roll wave components from the multicomponent seismic data before applying denoising operations. By taking out the interfering ground roll signals first, the subsequent denoising process can effectively enhance the effective signals without being compromised by the presence of coherent noise, thus maintaining vector amplitude information.
2Reliability
If polarization filtering is used to separate effective signals from noise, then signal separation is improved, but decomposition of signals and noise in the same time window remains difficult
Solution Approach 1:
The patent employs dynamic polarization filtering that adapts to the varying polarization characteristics of seismic waves in the time-frequency domain. By analyzing polarization attributes dynamically across different time windows and frequency bands, the method can effectively separate effective signals from noise even when they arrive in the same time window, overcoming the limitations of static polarization filtering approaches.
Data Source
AI summary
The present disclosure provides a vector denoising method for the multicomponent seismic data, including: obtaining a wave vector of the multicomponent seismic data; calculating a first mean wave vector for the multicomponent seismic data by applying a first moving window, performing a median filtering for the first mean wave vector to obtain a true modulus of the ground roll, subtracting the wave vector of the ground roll from the multicomponent seismic data to obtain a vector time-series; performing a mean filtering for the vector time-series by using a second moving window to obtain a second mean wave vector, performing a median filtering and performing a median the median filtering within the same receiving line for the second mean wave vector to suppress the wave vector of the random noise in the vector time-series, thereby obtaining a wave vector of a purified effective signal.