First-break point detection in noisy seismic data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous first-break picking algorithms in seismic exploration face challenges with low signal-to-noise ratios and strong reflections, leading to inaccurate determination of first-break points, which are crucial for subsurface geological modeling and drilling operations.
Innovation Solution
The method combines 1-dimensional edge detection with a random sample consensus (RANSAC) algorithm to filter potential first-break points, using line regression and amplitude weighting to improve accuracy, particularly in datasets with low signal-to-noise ratios and strong noises.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If current autonomous first-break picking algorithms are used, then the process is automated, but accuracy deteriorates due to low signal-to-noise ratios and strong reflections
Solution Approach 1:
The algorithm segments the seismic trace analysis into multiple stages: initial potential first-break point identification, FB line generation through clustering, iterative filtering of incorrect points, and final precision determination. This multi-stage segmentation allows the system to handle noisy data progressively rather than attempting single-step accurate picking.
Solution Approach 2:
The patent transitions from 1-dimensional edge detection to 2-dimensional FB line analysis by plotting arrival times against offset distances. This dimensional expansion allows the algorithm to identify linear patterns in the data and filter out outliers that don't conform to expected seismic wave propagation geometry.
2Measurement precision
If iterative filtering and line regression are applied, then measurement precision improves, but device complexity increases
Solution Approach 1:
The algorithm is self-correcting through iterative processes where each cycle identifies and removes outliers, then regenerates FB lines with improved accuracy. The system automatically adjusts its own parameters and filtering thresholds based on the data characteristics without external intervention, allowing complex processing to emerge from simple repeated operations.
Solution Approach 2:
The iterative nature of the algorithm incorporates feedback loops where the output of one iteration (filtered PFB points and FB lines) becomes the input for the next iteration. The system continuously refines its estimates by comparing predicted versus actual arrival times and adjusting accordingly, with the process terminating when convergence criteria are met.
Data Source
AI summary
The present disclosure describes methods and systems, including computer-implemented methods, computer program products, and computer systems, for determining first-break (FB) points. One computer-implemented method includes: selecting, by a hardware processor, potential first-break (PFB) points based on seismic data obtained by plurality of seismic receivers in a geological location; determining, by the hardware processor, a first plurality of FB lines based on the PFB points; selecting, by the hardware processor, a first FB line among the plurality of FB lines; filtering, by the hardware processor, the PFB points based on the first FB line; determining, by the hardware processor, a second plurality of FB lines based on the filtered PFB points; selecting, by the hardware processor, a second FB line among the second plurality of FB lines; and determining, by the hardware processor, FB points based on the second FB line.