Scholarlink

Abstract <p>Today, an urgent trend in geology is the development of approaches to applying neural network technologies at different stages of geological exploration. The article considers the architecture of the AlexNet neural network, which has already been tested in various territories. AlexNet makes it poddible to conduct training on a relatively small amount of data with sufficient accuracy to solve problems. To carry out operations with the selected neural network, a technique has been developed that makes it possible, based on prepared geological and spatial features (criteria) that indirectly or actually control ore objects, to train a neural network model with its further application to the studied territory. This approach allows one to analyze and obtain an expert assessment of the studied area in the form of a spatial predictive search model that predicts the location of the most promising sites for further study. In the current article, an example of using the developed methodology for forecasting hydrothermal massive sulfide deposits in the territory of Southeastern Transbaikalia is demonstrated.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Approach to Creating Spatial Predictive Prospecting Models of Deposits Based on Convolutional Neural Networks (A Case Study of the Territory of Southeastern Transbaikalia)

  • 期刊论文

  • 发表日期：

    2025年4月1日

查看全文
  • G. A. Grishkov,
  • I. O. Nafigin,
  • S. A. Ustinov,
  • V. A. Minaev,
  • V. A. Petrov

摘要

Abstract

Today, an urgent trend in geology is the development of approaches to applying neural network technologies at different stages of geological exploration. The article considers the architecture of the AlexNet neural network, which has already been tested in various territories. AlexNet makes it poddible to conduct training on a relatively small amount of data with sufficient accuracy to solve problems. To carry out operations with the selected neural network, a technique has been developed that makes it possible, based on prepared geological and spatial features (criteria) that indirectly or actually control ore objects, to train a neural network model with its further application to the studied territory. This approach allows one to analyze and obtain an expert assessment of the studied area in the form of a spatial predictive search model that predicts the location of the most promising sites for further study. In the current article, an example of using the developed methodology for forecasting hydrothermal massive sulfide deposits in the territory of Southeastern Transbaikalia is demonstrated.