Interactive labeling and training for fault detection
Interactive labeling and training for channel detection
Interactive labeling and training for salt dome detection
Interactive labeling and training for predicting Seismic Sequence
Predict faults, channels, and volcanos
>20x speed up of Seismic Interpretation Process
InterpretationAI
What is InterpretationAI
InterpretationAI is the next generation seismic interpretation software powered by artificial intelligence (AI) algorithms.
Click here to view the modules in the full package of InterpretationAI™
Why use InterpretationAI
1) >20x productivity increase in the seismic interpretation tasks than the legacy system, powered by GPU and the latest AI innovations. Using the InterpretationAI, you can complete an interpretation project of a large 3-D seismic survey within one week instead of five months
2) More accurate and consistent interpretation results. With the InterpretationAI, you achive unprecedented accuracy and consistentcy by utilizing other interpreter’s experiences captured by the AI models.
Testimonial
“The integrated attribute analysis and interpretation workflows in AttributeStudio allow me to delineate reservoir areas that would otherwise have been missed using a traditional interpretation package. AttributeStudio is indispensable to me as an interpretation geophysicist”
Bruce GawalkoGeophysical consultant at Whitecap Resources Inc
Workflow
- 1) Digitize training labels (faults, horizons, channels or geobodies) on an inline, crossline section, or a stratal slice on base map.
- 2) Train the AI model and view the prediction results on the label section or slice
- 3) Apply the trained model to predict on non-label sections or slices.
- 4) If prediction results on the non-label sections or slices are not satisfying, digitize the labels on the section or slice and re-train the AI model; then move to step 3. If the prediction results on the non-label sections or slices are -satisfying, generate AI interpretation volumes.
- 5) Extract faults, horizons, and geobodies from AI interpreted volumes.
Key Features
- 1) An interactive training process that is familiar to interpreter who is using legacy interpretation software
- 2) Fast in training because of transfer learning and base models for faults, channels, and horizons.
- 3) Training on RGB volumes
- 4) Digitize labels on stratal slices
- 5) Extract fault sticks and horizon surfaces from AI interpreted volumes