Scholarlink

<p>The geoelectrical survey method generates subsurface cross-section images based on physical properties but requires a solution to an inverse problem with potential ambiguities in model interpretation and substructure uncertainties. The purpose of this study is to use an alternative machine learning computational approach to the traditional electrical resistivity tomography (<i>ERT</i>) method in order to reduce these ambiguities and uncertainties as well as the labour-intensive nature of conventional computational methods. Exploring a relationship between the apparent and true resistivity data in the training samples, our innovative method directly inverts the resistivity pseudo-section into the resistivity section (parameters). In this study, samples are drawn from a set of data collected from landfill locations in Nigeria and inverted using the conventional geophysical method of interpretation utilizing the <i>RES2DINV</i> software. The inverted data (true resistivity tomography images) along with the source data (apparent resistivity images) are used as training samples to develop predictor models based on the <i>Pix2Pix</i> conditional generative adversarial networks (<i>Pix2Pix-cGAN</i>). Initial results with a small number of training samples reveal about 89% structural similarity between the true resistivity tomography obtained by the standard inversion method and those predicted by the <i>Pix2Pix</i> translator.</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

Application of Generative Adversarial Networks in Geoelectrical Field Data Processing: Innovative Approach to Solving Inverse Problems

  • 期刊论文

  • 发表日期：

    2025年4月10日

查看全文
  • M. H. Shahnas,
  • O. M. Alile,
  • R. N. Pysklywec

摘要

The geoelectrical survey method generates subsurface cross-section images based on physical properties but requires a solution to an inverse problem with potential ambiguities in model interpretation and substructure uncertainties. The purpose of this study is to use an alternative machine learning computational approach to the traditional electrical resistivity tomography (ERT) method in order to reduce these ambiguities and uncertainties as well as the labour-intensive nature of conventional computational methods. Exploring a relationship between the apparent and true resistivity data in the training samples, our innovative method directly inverts the resistivity pseudo-section into the resistivity section (parameters). In this study, samples are drawn from a set of data collected from landfill locations in Nigeria and inverted using the conventional geophysical method of interpretation utilizing the RES2DINV software. The inverted data (true resistivity tomography images) along with the source data (apparent resistivity images) are used as training samples to develop predictor models based on the Pix2Pix conditional generative adversarial networks (Pix2Pix-cGAN). Initial results with a small number of training samples reveal about 89% structural similarity between the true resistivity tomography obtained by the standard inversion method and those predicted by the Pix2Pix translator.