Reservoir property image slices from the Groningen gas field for image translation and segmentation

2026-05-05

AI Summary

This study addresses the scarcity of geological image data, which has hindered reproducible benchmarking of machine learning methods for reservoir characterization. To this end, the authors present the first publicly released, structured, multi-attribute-aligned, high-resolution 2D reservoir image slice dataset derived from the static geological model of the Groningen gas field, encompassing lithofacies, porosity, permeability, and water saturation. Accompanying this dataset is an open-source, reproducible workflow for data processing and augmentation. Through 3D grid slicing, PNG image generation, mask construction, and paired-image assembly, the framework supports image segmentation and translation tasks, substantially enhancing data transparency and reusability. This resource establishes a reliable benchmark for investigating cross-domain relationships among reservoir properties and for developing robust AI models in subsurface characterization.

Abstract

Reservoir characterization workflows increasingly rely on image-based and machine-learning/deep learning or even generative AI approaches, but openly available geological image datasets suitable for reproducible benchmarking remain limited. Here we describe a high-resolution dataset of reservoir-property image slices derived from the Groningen static geological model. The dataset contains aligned two-dimensional PNG images representing facies, porosity, permeability, and water saturation, generated from three-dimensional reservoir grids and prepared for downstream visualization, segmentation, and image-to-image translation tasks. In addition to the deposited original image corpus, we provide an archived software workflow for reproducing augmentation, mask generation, paired-image construction, and example baseline experiments. The resource is designed to support benchmarking of geological image analysis methods and the study of cross-domain relationships among reservoir properties. By separating the fixed image dataset from the reproducible processing workflow, this work provides a transparent foundation for reuse in geoscience, reservoir modeling, and machine-learning applications.

Problem

• reservoir characterization
• geological image dataset
• benchmarking
• image-based methods
• reproducibility

Innovation

• reservoir property imaging
• image-to-image translation
• reproducible workflow
• geological benchmark dataset
• deep learning for geoscience