Adaptive Seismic Denoising via Hankel Matrix Rank Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional denoising methods are inadequate for removing erratic noise from seismic data, leading to severe degradation in subsurface structure imaging and elastic property inversion due to high amplitude and frequency-wavenumber domain overlap, and require significant human intervention.
Innovation Solution
An automated process that partitions seismic data into index sets and data blocks, forms Hankel matrices, determines optimal ranks, and separates clean signals from erratic noise using adaptive robust singular spectrum analysis, reducing signal distortions and improving noise suppression with minimal manual intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional denoising methods are used, then processing simplicity is maintained, but noise removal effectiveness deteriorates due to inadequate handling of erratic noise with high amplitude and frequency-wavenumber overlap
Solution Approach 1:
The seismic data is divided into multiple frequency slices, each processed independently through Hankel matrix formation and rank determination. This segmentation allows targeted noise removal at different frequency components while maintaining overall signal integrity, resolving the contradiction between effective noise removal and processing complexity.
Solution Approach 2:
The patent transforms the denoising problem from the time-space domain to the frequency domain through Fourier transformation, then further to a Hankel matrix structure for rank analysis. This dimensional transformation enables separation of signal and erratic noise that overlap in traditional domains, improving noise removal effectiveness while providing a systematic processing framework.
2Adaptability or versatility
If manual intervention is used for denoising, then adaptability to local characteristics is improved, but automation level deteriorates requiring significant human input
Solution Approach 1:
The system performs self-adaptation through automated rank determination for each Hankel matrix, where the optimal rank is selected based on local signal characteristics without manual intervention. The algorithm automatically adjusts to different frequency slices and data blocks, achieving both adaptability and high automation level simultaneously.
Solution Approach 2:
The patent dynamically changes processing parameters including frequency slice boundaries, Hankel matrix dimensions, and rank thresholds based on local data characteristics. These parameter adaptations are performed automatically through computational algorithms, maintaining adaptability to local signal and noise characteristics while eliminating the need for manual parameter tuning.
3Productivity
If large-scale data sets are processed using traditional methods, then data storage requirements are maintained, but computational efficiency deteriorates due to large multi-level matrices
Solution Approach 1:
The large-scale seismic data is partitioned into smaller data blocks and further divided into frequency slices, with each block processed through Hankel matrix formation. This segmentation reduces the size of individual matrices that need to be stored and manipulated, improving computational efficiency while the systematic processing framework ensures complete data coverage.
Solution Approach 2:
The patent discards redundant information by determining the optimal rank of Hankel matrices, keeping only the essential signal components while eliminating noise. This rank-based compression reduces storage requirements for processed data while preserving the critical information needed for subsurface imaging, achieving both computational efficiency and reduced storage needs.
Data Source
AI summary
Seismic survey data is received, indexed into index sets, and each index set partitioned into data blocks. For each particular data block of a particular index set, the particular data block is sliced into frequency slices. For each particular frequency slice of the particular data block, the particular frequency slice is processed to remove random and erratic noise by: forming a Hankel matrix from the particular frequency slice: determining an optimal rank for the Hankel matrix, determining a clean signal and erratic noise from the ranked Hankel matrix, and returning the clean signal and erratic noise for the particular frequency slice. A clean signal is assembled from the index sets.