A geological constraint-based multi-modal fusion prospecting method and system

By employing a multimodal fusion prospecting method with geological constraints in mineral resource distribution prediction, and utilizing a multi-branch neural network model and geological constraints, the problem of lack of professional knowledge constraints in neural network models is solved, thus achieving more accurate mineral resource distribution prediction.

CN122244686APending Publication Date: 2026-06-19CHINA GEOLOGICAL SURVEY XIAN MINERAL RESOURCES SURVEY CENT
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA GEOLOGICAL SURVEY XIAN MINERAL RESOURCES SURVEY CENT
Filing Date
2026-05-12
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing neural network models lack effective constraint mechanisms based on domain-specific knowledge in mineral resource distribution prediction, leading to inaccurate prediction results.

Method used

A multimodal fusion prospecting method based on geological constraints is adopted. By acquiring multi-source geological data of mined areas, hard and soft constraints are generated. The training set is split into channel branch bundles and spatial branch bundles. Convolutional neural networks and Transformer branches are used for training. A multi-branch neural network model is generated by combining geological constraints. The prospecting target area is adjusted by calculating the correlation index of regional features.

Benefits of technology

This improves the accuracy of mineral resource distribution prediction and ensures that the prediction results have a high degree of similarity to the actual geological conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122244686A_ABST
    Figure CN122244686A_ABST
Patent Text Reader

Abstract

This application provides a multimodal fusion mineral exploration method and system based on geological constraints, relating to the field of mineral exploration technology. The method includes: acquiring first multi-source geological data and first regional features of an already mined area; preprocessing the data and dividing it into training and testing sets; generating constraints based on the first regional features; constructing and training a multi-branch neural network model using the constraints to generate a target network model; when the model performance verification meets a preset accuracy threshold, acquiring second multi-source geological data and second regional features of an unmined area; substituting the second multi-source geological data into the target network model to generate an initial mineral exploration target area; calculating the correlation index between the first and second regional features; adjusting the initial mineral exploration target area based on the correlation index to generate a target mineral exploration target area. The technical effect of this application is: improving the accuracy of prediction results.
Need to check novelty before this filing date? Find Prior Art