Statistical Edge Detection for Seismic Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting faults and fractures in seismic data, such as those used in the oil and gas industry, face challenges in automated edge detection and enhancement, particularly in handling computational costs and outlier values, and often rely on integral-based approaches like Radon and Hough transforms.
Innovation Solution
A statistical approach is employed for line and edge detection, using significance statistics and P-value calculations to identify significant edges, which involves defining a window around points of interest and applying non-parametric or parametric statistical tests to determine the presence of lines or planes, thereby minimizing computational costs and the impact of outliers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If integral-based approaches like Radon and Hough transforms are used for edge detection, then comprehensive line detection is achieved, but computational cost increases significantly
Solution Approach 1:
The patent segments the image processing task by focusing computations only on candidate regions identified by gradient magnitude thresholds, rather than applying integral transforms to the entire image. This segmentation reduces the search space and computational burden while maintaining detection accuracy in relevant areas.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image based on local characteristics. High-gradient regions receive detailed line detection analysis, while low-gradient regions are processed more simply or skipped, optimizing computational resources according to local information quality.
2Reliability
If statistical tests are applied to all points in the volume, then detection robustness improves, but processing time increases
Solution Approach 1:
The patent performs preliminary filtering using gradient magnitude calculations before applying the more computationally intensive statistical tests. This preliminary action identifies promising candidate regions, allowing the robust statistical analysis to be focused only where it is most likely to yield positive results, thereby reducing overall processing time.
Solution Approach 2:
The patent applies statistical tests selectively to regions that meet certain criteria rather than to all points in the volume. This partial action approach achieves sufficient detection robustness in critical areas without the excessive computational burden of universal application.
Data Source
AI summary
Embodiments of the present disclosure include one or more of a method, computing device, computer-readable medium, and system for statistical line and edge detection and/or enhancement. An example embodiment of the present disclosure may include a method that includes identifying a plurality of data values related to a first object defined by a first plurality of points within the volume, wherein the first object intersects a second object defined by a second plurality of points within the volume; calculating a statistical significance statistic related to the second object; interpolating a P-value related to the statistical significance statistic; and determining a significant P-value taken over the second object, wherein the significant P-value comprises a minimum P-value that provides a maximum negative log(P(statistical significance statistic)).