Seismic Noise Attenuation via Dip Map Data Structure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Geophysical survey data is often contaminated with noise, making it challenging to accurately analyze geological structures for oil and gas exploration, as existing noise reduction methods are imperfect and require adaptation to specific survey conditions.
Innovation Solution
The use of complex-valued, directional, multi-resolution (CDM) transforms, such as the complex curvelet transform, to adapt noise templates in the transform domain, combining global and local adaptation techniques to improve noise reduction, and the application of dip maps to control the adaptation process and minimize signal damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If noise templates are adapted to fit specific survey conditions, then noise reduction effectiveness is improved, but signal damage increases
Solution Approach 1:
The patent applies local adaptation by adjusting noise template parameters specifically in regions where multiples are present, rather than uniformly across the entire data set. This allows aggressive noise reduction in multiple-prone areas while preserving signal integrity in regions where primaries dominate, thus resolving the contradiction between noise reduction effectiveness and signal damage.
Solution Approach 2:
The adaptation process dynamically adjusts noise template parameters based on local data characteristics. The system iteratively refines the noise model by comparing predicted multiples with actual data, automatically modifying template parameters to optimize the balance between noise removal and signal preservation in different spatial and temporal regions.
2Measurement precision
If complex-valued directional multi-resolution transforms are used, then noise reduction precision is improved, but computational complexity increases
Solution Approach 1:
The patent segments the seismic data into different frequency bands and directional components using CDM transforms. By processing data in segmented frequency-wavenumber bins rather than as a whole, the method achieves high precision noise reduction while managing computational complexity through divide-and-conquer processing of transformed coefficients.
3Reliability
If noise models are heavily adapted to match actual noise, then noise attenuation is improved, but risk of removing signal components increases
Solution Approach 1:
The patent implements feedback mechanisms where the adapted noise model is continuously evaluated against the data, and adaptation is adjusted based on residual analysis. The system monitors the matching between predicted and actual multiples, providing feedback to control the degree of adaptation, thus preventing over-adaptation that would lead to signal removal while maintaining effective noise attenuation.
Data Source
AI summary
Techniques are disclosed relating to reducing noise in geophysical marine survey data through use of a dip map data structure. Such techniques may include adapting a model of multiple noise to seismic data. A multiple dip map may be generated by convolving the adapted model of multiple noise with a set of directional filters. A modified record of the seismic data may be generated through adaptive subtraction that is sufficient to remove most multiple events but may also damage at least some primary events. A primary dip map may then be determined from the modified record by interpolation dependent on the multiple dip map. A noise template may then be adaptively subtracted from the seismic data. Prior to the adaptive subtraction, the noise template is adapted to the seismic data by a degree of adaptation that is determined dependent upon relative amplitudes of the primary and multiple dip maps.