Geological Feature Search Using Machine-Labeled Seismic Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for oil and gas exploration face challenges in efficiently processing large datasets of geological information, making it difficult to identify relevant geological features and areas of interest, which can lead to inefficiencies and safety risks.

Innovation Solution

A computer-implemented method using machine learning to identify geological features in seismic data, employing a search system that trains a neural network to recognize and rank seismic images, allowing for rapid identification of features of interest and hazardous areas, and providing curated search results based on user preferences and collaborative filtering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional methods are used to process large datasets of geological information, then data processing can be performed, but processing efficiency and accuracy are insufficient

Engineering Contradiction:
Improvedata processing efficiencyVSAvoidgeological feature identification accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces manual geological analysis methods with machine learning algorithms and automated processing systems. The machine learning model automatically identifies geological features from seismic data, substituting the mechanical/manual process of human analysis with an automated computational system that achieves both higher efficiency and improved accuracy in feature identification.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If manual analysis of geological datasets is performed, then detailed examination is possible, but the process is time-consuming and labor-intensive

Engineering Contradiction:
Improvegeological feature detection capabilityVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The machine learning system performs self-service by automatically processing seismic data and identifying geological features without requiring continuous human intervention. The system trains on historical data, then independently analyzes new datasets, generating results that would traditionally require extensive manual review by geologists, thereby dramatically reducing analysis time while maintaining detection capability.

Inventive Principle:
Principle #25Self-service

3Quantity of substance

If comprehensive geological data is collected, then more information is available, but data management and analysis become more difficult

Engineering Contradiction:
Improvevolume of geological dataVSAvoiddata processing system complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent transforms the complexity of managing large volumes of geological data by changing the processing parameters through machine learning. Instead of attempting to manually organize and analyze the vast dataset, the system converts the data into a format suitable for algorithmic processing, using trained models to automatically extract relevant features and patterns, thereby managing data complexity through parameter transformation rather than structural organization.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3803471B1Geological feature search engine
Publication Date: 2025.07.30 SERVICES GASOLINEIERS SCHLUMBERGER SPS
  • EP3803471B1 patent drawingFigure 1A~1D
  • EP3803471B1 patent drawingFigure 2
  • EP3803471B1 patent drawingFigure 3A

AI summary

A computer-implemented method includes receiving a geological feature search query identifying one or more geological features, executing, based on receiving the geological feature search query, a search of database storing a plurality of seismic data images. The seismic data images are labeled with geological features present in each of the plurality of seismic data images as part of a machine learning process. The method further includes determining, based on executing the search, search results, wherein the search results identify one more of the plurality of seismic data images having the one or more geological features identified in the geological feature search query, and outputting information regarding the search results.