Ready-to-Go AI Workflows — Apply Pre-Trained Models
Run dGB's library of pre-trained Machine Learning models on your own data with a few clicks. No training, no Python — just load the model, point it at a 3D volume, and generate predictions.
AI / Machine Learning
AI-Powered Geoscience Interpretation
Artificial Intelligence is transforming how geoscientists interpret the subsurface. At dGB Earth Sciences, AI is not an add-on — it's embedded throughout our technology.
Our Origin Story
Three Decades Head Start on the Competition
dGB has such a huge head start on its competition because of a 1990's Statoil (now Equinor) blowout that occurred when drillers hit a gas pocket, causing its rig to collapse into the ocean. During an investigation aimed at preventing future disasters, it was determined that faint hydrocarbon migration trails (chimneys) were visible in the seismic record around the gas pocket. Statoil decided that accurate mapping of these hydrocarbon trails would minimize failures and save millions of dollars. As hand-mapping these patterns took six months, Statoil funded dGB to develop a trained neural network approach to identifying patterns and hydrocarbon trails.
The software that evolved from the Statoil work — a seismic pattern recognition processing engine — is OpendTect. This early neural network and Machine Learning development led to the deep learning algorithms and the complete environment now in place.
As the only fully integrated seismic interpretation platform with Machine Learning capability, dGB has no full-on competition. Although other vendors are working on Machine Learning features, only OpendTect can be used as an off-the-shelf solution as well as an R&D environment. Its SynthRock plugin lets users generate synthetic data through stochastic modeling, which can then be used to train new Machine Learning models — a superior simulator that trains models in the absence of real data examples.
Seismic Facies Classification
AI-driven pattern recognition
Reservoir Characterization
ML-powered property prediction
U-Net Salt Detection
Deep learning image segmentation
Channel Tracking
AI-assisted seismic geomorphology interpretation
Why Clients Choose Us
We accelerate the transition from research to deployment — significantly reducing time and cost.
Our technology accelerates interpretation cycles through AI-driven automation.
We improve accuracy with Machine-Learning-based pattern recognition.
Our solutions integrate seamlessly with enterprise data systems and workflows.
We deliver results proven in commercial operations around the globe.
AI Solutions for Every Type of Geoscientist
Whether you're an exploration geoscientist, data scientist, or researcher, our AI platform adapts to your workflow.
Operational Geoscientists
Deploy trained models — no coding required. Use ready-to-apply Machine Learning models for seismic and well interpretation.
- Apply trained Neural Networks directly to your seismic data.
- Save valuable time while increasing accuracy and repeatability.
Experimental Geoscientists
Create, test, and refine new AI workflows to extract more value from your data.
- Train and evaluate custom models using your own datasets.
- Combine seismic, well log, and attribute data in new AI workflows.
- Prototype and validate innovative interpretation methods.
Researchers & Data Scientists
Our Python API provides a robust environment for rapid experimentation.
- Automated data I/O, QC, and preprocessing.
- Ready-to-use visualization tools.
- Access to proven dGB data management libraries.
AI Workflows
Ready-to-Use AI Workflows in OpendTect
A curated set of OpendTect Machine Learning workflows — each one a complete, repeatable recipe for solving a real subsurface problem. Apply pre-trained models out of the box, fine-tune them on your data, or extend the workflow with your own Python code.
Quick UVQ Waveform Segmentation
Unsupervised waveform clustering for fast seismic facies analysis. Generate a segmentation volume of distinct waveform classes in minutes to highlight depositional patterns and geomorphology.
Fast & Simple Seismic Facies Analysis — 3D UVQ
Full-volume 3D Unsupervised Vector Quantization for high-resolution seismic facies. Cluster waveforms across the entire cube to map channels, fans and stratigraphic units automatically.
Supervised AI Seismic Facies
Train a supervised model on interpreter-labelled examples to classify seismic facies across an entire 3D volume. Combines geological expertise with deep learning for explainable, high-confidence results.
Using AI for Salt Detection
Train a 2D U-Net (128×128 samples) to segment salt bodies from 3D seismic. Transforms manual salt picking into an automated, repeatable workflow — the Seismic Image Segmentation method.
De-Risking Charge & Seal — Neural Network Chimney Cube
The original OpendTect ML workflow: a neural network trained on directive attributes to detect vertical gas chimneys and fluid migration paths — a proven tool for de-risking charge and seal in exploration.
Seismic Inversion Using AI
Machine-driven seismic inversion that predicts well-log properties directly from seismic. Train on real or synthetic data to build acoustic impedance, porosity or lithology volumes without a traditional low-frequency model.
Integrated ML Rock Property Prediction
End-to-end workflow that transforms 1D well models into a predicted 3D rock property cube. Combines the OpendTect Volume Builder with Machine Learning to integrate seismic attributes, wells and low-frequency guidance.
A Continuous Innovation Loop
Every workflow above follows the same three-stage cycle, so models stay fresh as your data and interpretations evolve.
01
Research
Prototype and train models using Python and your data.
02
Validation
Evaluate model performance inside OpendTect.
03
Deployment
Apply trained models operationally to new datasets — no coding required.
Built on Proven Technology
Integrated within OpendTect Pro — no external dependencies.
Compatible with industry-standard data formats.
GPU-accelerated for high-performance training and inference.
Scalable from desktop to cluster environments.
Case Studies & Results
Machine Learning Publications
Peer-reviewed papers, conference proceedings, and technical articles documenting commercial results and case studies achieved with dGB's AI and Machine Learning workflows — spanning three decades of innovation in geoscience.
Improve consistency and quality across teams.
Lower project costs and de-risk exploration decisions.
Unlock hidden value in existing seismic and well data.
Explore the Possibilities
Join the growing number of companies using AI from dGB Earth Sciences to achieve faster, smarter decisions.
Innovation That Leads the Industry
As the only fully integrated seismic interpretation platform with built-in Machine Learning capabilities, OpendTect continues to set the standard for intelligent geoscience software.
Our SynthRock plugin exemplifies this flexibility — enabling users to generate synthetic training data through stochastic modeling when real data is limited.
From Technology to Business Advantage
dGB doesn't just deliver software — we deliver competitive advantage through data-driven decisions and AI-powered geoscience workflows.
Explore Our AI Services