Dip Angle-Steering Median Filtering for Seismic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for calculating event dip angles in seismic data, especially in complex geological environments with intersecting events, are inefficient and inaccurate, particularly in three-dimensional data processing where existing algorithms require uniform sampling and are sensitive to noise.
Innovation Solution
The dip angle-steering median filtering method employs a niche differential evolution algorithm to directly obtain true three-dimensional dip angles without separate calculations along spatial directions, using a neighborhood-based crowding differential evolution algorithm, hill-valley detection function, and Nelder-Mead simplex algorithm to improve accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional dip angle calculation methods (local slant stack, PWD, iterative f-k) are used in complex geological environments with intersecting events, then the calculation can be performed, but the accuracy of dip angle determination deteriorates due to multiple dip angle values at one data location
Solution Approach 1:
The patent divides the time-space domain into multiple overlapping time-space windows, allowing dip angle calculation to be performed locally in each window. This segmentation enables the method to handle complex geological structures by treating each local region independently while maintaining overall coherence through the overlapping windows.
Solution Approach 2:
The patent introduces a new dimension by calculating dip angles in both inline and crossline directions separately, then combining these two-dimensional measurements to obtain three-dimensional dip angle information. This dimensional approach allows accurate characterization of intersecting events that traditional single-direction methods cannot resolve.
2Productivity
If uniform sampling is required along spatial directions (as in traditional three-dimensional dip angle methods), then the calculation framework is simplified, but the computational efficiency deteriorates due to the large number of calculations required
Solution Approach 1:
The patent performs dip angle calculations only at selected event positions identified through energy curves, rather than uniformly across the entire spatial domain. This partial action approach significantly reduces the number of calculations required while maintaining accuracy where it matters most - at actual seismic events.
Solution Approach 2:
The algorithm automatically identifies event positions and adapts the calculation grid to the actual data characteristics, eliminating the need for predetermined uniform sampling schemes. The method self-adjusts to perform calculations only where seismic events are present, improving computational efficiency without requiring complex external control.
3Measurement precision
If separate dip angle calculations are performed along inline and crossline directions, then the three-dimensional dip angle can be obtained, but the computational time increases due to multiple separate calculations
Solution Approach 1:
The patent merges the inline and crossline dip angle calculation processes by performing both calculations within the same time-space window framework and combining results to obtain three-dimensional dip angles. This integrated approach reduces redundant calculations and optimizes computational time while maintaining three-dimensional accuracy.
Data Source
AI summary
A dip angle-steering median filtering method based on a niche differential evolution algorithm, comprising the following steps: dividing a data to be processed into a series of overlapping time-space windows; obtaining an event energy curve in a time-space window and obtaining an event position according to a local maximum value of the event energy curve; obtaining event dip angles and coherence values of the event dip angles through the niche differential evolution algorithm at the event position; filtering the event dip angles according to the event dip angles and the coherence values of the event dip angles; and performing a median filtering sequentially along a filtering dip angle. The disclosure can simultaneously obtain all dip angles of intersecting events and a true three-dimensional feature enable the present disclosure to obtain a better filtering effect.