A deep learning-driven three-dimensional geological modeling method using sparse borehole sampling data

doi:10.1016/j.measurement.2025.118461

A deep learning-driven three-dimensional geological modeling method using sparse borehole sampling data

2025-07-23

0

PRE

AI

DOI：10.1016/j.measurement.2025.118461

Original

Original request for help

Save

Abstract

En 中文

• A unique approach is developed to self-organize sparse borehole sampling data into a structured dataset. • An autoencoder (AE) is used to extract deep and meaningful features from the sparse borehole data. • A residual capsule network (ResCapsNet) is employed to predict lithological distributions in discrete units.

Keywords:

sparse borehole data

autoencoder

residual capsule network

lithological prediction

self-organization

A deep learning-driven three-dimensional geological modeling method using sparse borehole sampling data

Abstract

Journal

Organization

Cited Papers