Three-dimensional geological modeling method based on DGCR-GNN architecture

The 3D geological modeling method based on the DGCR-GNN architecture solves the problems of feature homogenization and gradient vanishing when sparse geological sampling data is transformed into dense geological bodies, achieving high-precision 3D geological body reconstruction and improving the automation and intelligence level of the model.

CN122454083APending Publication Date: 2026-07-24CHINA UNIV OF GEOSCIENCES (WUHAN)
0 Cites 0 Cited by

Patent Information

Application Number
CN202610494975.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-15
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing GNN-based 3D geological modeling suffers from problems such as homogenization of node features, oversmoothing, and gradient vanishing when sparse geological sampling data is transformed into dense 3D geological bodies. This makes it impossible to take into account the overall characteristics of the geological body, resulting in low model accuracy and fitting distortion.

Method used

A 3D geological modeling method based on the DGCR-GNN architecture is adopted. By constructing a connected graph of known semantic feature points and a fully connected graph, and combining a single-layer graph attention mechanism, a multi-level message-passing residual convolutional layer and semi-supervised classification, feature encoding, message passing and decoding are realized, which solves the problem of high-precision reconstruction of sparse data into dense geological bodies.

Benefits of technology

It significantly improves the classification accuracy and geological body continuity of 3D geological modeling, reduces the number of model parameters, improves training efficiency and generalization ability, and achieves high-precision reconstruction from sparse sampled data to dense geological bodies.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a three-dimensional geologic modeling method based on a DGCR-GNN architecture, and relates to the field of three-dimensional geologic modeling, and the method comprises the steps: constructing a known semantic feature point connected graph and a full-node connected graph through geologic sampling data and a three-dimensional regular grid of a target research region; based on a single-layer graph attention mechanism, adaptive weighted transfer is carried out on node features in the known semantic feature point connected graph, and high-dimensional geological prior features are generated; high-dimensional geological prior features are injected into the full-node connected graph, and deep topological feature extraction is carried out through a plurality of cascaded message passing residual convolutional layers; and for unmarked encrypted grid nodes in the full-node connected graph, adopting a semi-supervised classification strategy, outputting geological category probability distribution of each node, and completing reconstruction of the three-dimensional geological model. According to the method, high-precision reconstruction from sparse geological sampling data to a dense three-dimensional geologic body is realized, and the modeling precision and the automation level are improved.
Need to check novelty before this filing date? Find Prior Art