Weighted 4D Difference Stack for Seismic Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional seismic data processing methods fail to effectively suppress 4D-noise, which is non-repeatable noise present in seismic surveys, especially in deepwater environments, leading to noisy images despite using up- and down-going wave-fields.
Innovation Solution
A method involving the creation of a weighted stack of up- and down-going 4D differences using a similarity-based weighting function, such as NRMSD, to suppress noise while preserving signals, by aligning and scaling seismic data from multiple datasets collected at different times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional stacking of up- and down-going wave-fields is performed, then the fold of the survey is doubled and noise is statistically suppressed by a factor of 1/â2, but the desired level of noise enhancement is not always achieved
Solution Approach 1:
The patent applies different weighting factors to different portions of the stacked image based on local noise characteristics. The weighting function is calculated separately for each spatial location, allowing optimal noise suppression at each location while preserving signal integrity. This resolves the contradiction by making the noise suppression adaptive rather than uniform, achieving the desired enhancement level across different regions.
Solution Approach 2:
The patent introduces dynamic weighting factors that adapt to local conditions rather than using a fixed stacking approach. The weighting function is computed based on local signal-to-noise ratios and adjusts the contribution of up- and down-going wave-fields accordingly. This dynamic adaptation achieves superior noise suppression compared to conventional static stacking.
2Object-affected harmful factors
If up- and down-going wave-fields are stacked to suppress noise, then the fold increases and statistical noise reduction occurs, but 4D-noise from different sources (multiple contamination, swell noise, etc.) persists
Solution Approach 1:
The patent changes the parameter of noise suppression by introducing a weighting function that operates in the signal domain rather than relying solely on statistical averaging. By computing weights based on local signal-to-noise ratios and applying them before final stacking, the method achieves more complete suppression of various 4D-noise sources including multiple contamination and swell noise, while maintaining reliability.
Solution Approach 2:
The patent creates a composite processing approach by combining weighted stacking with traditional stacking methods. The weighted stack integrates information from both up- and down-going wave-fields with optimized contributions, effectively combining the benefits of both approaches while suppressing diverse noise sources more completely than either method alone.
3Ease of manufacture
If a simple stack of up- and down-going 4D differences is performed, then processing is straightforward and fold is doubled, but the quality of seismic images is not optimally improved
Solution Approach 1:
The patent performs preliminary computation of weighting functions based on local signal-to-noise ratios before the final stacking operation. This preliminary action prepares the data in an optimized state, allowing the subsequent stacking to achieve superior image quality without significantly increasing processing complexity. The weighting calculation is performed once and then applied, maintaining ease of implementation.
Solution Approach 2:
The patent replaces the simple mechanical averaging of conventional stacking with a weighted combination approach. Instead of treating all data points equally, the method substitutes a more sophisticated weighting mechanism that considers local signal characteristics, thereby improving image quality while maintaining computational efficiency through algorithmic optimization.
Data Source
AI summary
A method for generating an image of a subsurface feature, comprises providing seismic data containing information about the feature, comprising i) a first dataset comprising a first up-going wave-field and a first down-going wave-field, ii) a second dataset comprising a second up-going wave-field and a second down-going wave-field, wherein the second dataset is collected at a time that is later than the first dataset by a time interval, creating an up-going 4D difference dataset U by subtracting one of the first and second up-going wave-fields from the other and creating a down-going 4D difference dataset D by subtracting one of the first and second down-going wave-fields from the other, creating a weighting function W that is a function of the similarity of the difference datasets, and creating an image of the feature by generating a 4D similarity stack, where the 4D similarity stack is defined as (U+D)*W/2.