Well Log-Guided Synthesis of Subsurface Images from Sparse Petrography Data Using cGANs

TL;DR

This paper introduces a cGAN-based method to generate realistic subsurface images from sparse petrography data, enabling continuous pore-scale visualization for reservoir analysis and energy applications.

Contribution

It presents a novel cGAN framework conditioned on well log porosity data to synthesize subsurface images, bridging gaps in traditional pore-scale imaging.

Findings

Achieved 81% accuracy within 10% of target porosity values.

Generated geologically consistent images across a wide porosity range.

Enabled continuous visualization along wellbores for reservoir characterization.

Abstract

Pore-scale imaging of subsurface formations is costly and limited to discrete depths, creating significant gaps in reservoir characterization. To address this, we present a conditional Generative Adversarial Network (cGAN) framework for synthesizing realistic thin section images of carbonate rock formations, conditioned on porosity values derived from well logs. The model is trained on 5,000 sub-images extracted from 15 petrography samples over a depth interval of 1992-2000m, the model generates geologically consistent images across a wide porosity range (0.004-0.745), achieving 81% accuracy within a 10\% margin of target porosity values. The successful integration of well log data with the trained generator enables continuous pore-scale visualization along the wellbore, bridging gaps between discrete core sampling points and providing valuable insights for reservoir characterization and energy transition applications such as carbon capture and underground hydrogen storage.