Borehole Image Log Autopicking for Faster Bedding Interpretation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for interpreting well data, such as image logs, are labor-intensive, subjective, prone to interpreter bias, and time-consuming, leading to underutilized resources and inaccurate results.

Innovation Solution

Implementing machine learning techniques for automated interpretation of well data, including breakout detection, autopicking of bedding, and facies classification, using image segmentation and object classification to reduce manual intervention and subjectivity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual interpretation methods are used for image logs, then interpretation accuracy can be maintained through expert judgment, but interpretation time and labor costs increase significantly

Engineering Contradiction:

Improveinterpretation accuracyVSAvoidinterpretation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical interpretation processes with automated machine learning systems. Neural networks and computer vision algorithms process image logs automatically, substituting human interpreters' visual analysis with computational models that can analyze images rapidly while maintaining consistent accuracy standards through trained algorithms.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service interpretation where the machine learning models autonomously perform bedding picking, breakout detection, and facies classification without requiring continuous human intervention. The automated workflows process multiple image logs independently, generating interpretations that reduce dependency on manual expert analysis for routine tasks.

Inventive Principle:

Principle #25Self-service

2Productivity

If automated solutions are implemented for image log interpretation, then interpretation time decreases, but accuracy and reliability may be compromised

Engineering Contradiction:

Improveinterpretation speedVSAvoidinterpretation accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent incorporates feedback mechanisms where machine learning models are trained on datasets annotated by expert interpreters, using their interpretations as ground truth. The systems continuously refine their algorithms through feedback from validation against known standards and allow for iterative improvement, ensuring automated results maintain high reliability while achieving rapid processing speeds.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary automated interpretation to identify potential features and anomalies, which then guide subsequent focused manual review only where needed. This preliminary action filters out routine cases that can be handled automatically, reserving expert human judgment for complex or uncertain interpretations, thereby maintaining accuracy while improving overall efficiency.

Inventive Principle:

Principle #10Preliminary action

3Ease of operation

If multiple interpreters are switched for a project, then workload distribution is improved, but interpretation consistency deteriorates due to different interpretation styles

Engineering Contradiction:

Improveworkload distributionVSAvoidinterpretation consistency

Core Design Contradiction:

Ease of operationVSStability of the object's composition

Solution Approach 1:

The patent applies homogeneity by standardizing the interpretation methodology across all projects through a unified machine learning system. All interpreters use the same automated algorithms and processing workflows, ensuring consistent application of interpretation criteria regardless of which person performs the analysis, thereby eliminating variability introduced by different individual interpretation styles.

Inventive Principle:

Principle #33Homogeneity

Solution Approach 2:

The machine learning system serves as a universal interpreter that can handle multiple types of image log analysis (bedding, breakouts, facies) with consistent methodology. This universal system can be applied across different projects and by different users without requiring customization or retraining, ensuring interpretation consistency is maintained while allowing flexible workload distribution.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Loss of information

If manual bedding picking is performed over entire imaged intervals, then comprehensive coverage is achieved, but labor intensity and costs increase

Engineering Contradiction:

Improvedata coverageVSAvoidlabor intensity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the image log processing into distinct analytical components: automated bedding detection, breakout identification, and facies classification. Each segment is handled by specialized machine learning modules that process specific features independently, reducing the overall labor intensity while maintaining comprehensive coverage of all interpretive elements across the entire imaged interval.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20260030883A1Machine learning based borehole data analysis

Publication Date: 2026.01.29 CGG SERVICES SAS

AI summary

A method for autopicking of bedding in a well includes receiving image logs associated with the well, eliminating tool marks from the image logs, performing a grid search for (1) a vertical amplitude and (2) a horizontal shift of the bedding at plural sampling depths to obtain a predicted bedding, calculating an azimuth and a dip of the predicted bedding, and generating an image of the predicted bedding, wherein the image includes structural features of the well.