Software Open Access

Predicting Magnetization Directions Using Convolutional Neural Networks

Felicia Disa Nurindrawati; Jiajia Sun

DataCite XML Export

<?xml version='1.0' encoding='utf-8'?>
<resource xmlns:xsi="" xmlns="" xsi:schemaLocation="">
  <identifier identifierType="DOI">10.5281/zenodo.3931029</identifier>
      <creatorName>Felicia Disa Nurindrawati</creatorName>
      <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="">0000-0003-1115-9589</nameIdentifier>
      <affiliation>University of Houston</affiliation>
      <creatorName>Jiajia Sun</creatorName>
      <affiliation>University of Houston</affiliation>
    <title>Predicting Magnetization Directions Using Convolutional Neural Networks</title>
    <subject>magnetization directions</subject>
    <subject>machine learning</subject>
    <subject>convolutional neural networks</subject>
    <date dateType="Issued">2020-07-05</date>
  <resourceType resourceTypeGeneral="Software"/>
    <alternateIdentifier alternateIdentifierType="url"></alternateIdentifier>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsVersionOf">10.5281/zenodo.3931028</relatedIdentifier>
    <rights rightsURI="">Creative Commons Attribution 4.0 International</rights>
    <rights rightsURI="info:eu-repo/semantics/openAccess">Open Access</rights>
    <description descriptionType="Abstract">&lt;p&gt;Magnetic data have been widely used for understanding basin structures, mineral deposit systems, formation history of various geological systems, and many others. Proper interpretation of magnetic data requires an accurate knowledge of total magnetization directions of the source bodies in the study area.&amp;nbsp;Existing approaches for estimating magnetization directions involve either unstable data processing steps such as reduction-to-pole, component conversions in wavenumber domain, or computationally intensive processes such as 3D inversions. &amp;nbsp;In this study, we developed a new method of automatically predicting the magnetization direction of a magnetic source body using Convolutional Neural Networks (CNN). CNNs have achieved great success in many other applications such as computer vision and seismic image interpretation, but have not been used to extract parameters from magnetic data. We simulated many magnetic data maps with different magnetization directions from a synthetic source body, all subject to the same background field. Two CNNs were trained separately, one for predicting the inclination and the other for predicting declination. We systematically trained and compared 13 different CNN architectures and selected one based on accuracy statistics.&amp;nbsp; In addition, we investigated the effect of having different parameters such as magnetization magnitude and source body shape and location, on the performance of our predictive models. We also tested the method with field data from Black Hill norite, Australia, and Yeshan region, China, for which prior research results are available for comparison. Our study shows that machine learning provides an effective means of automatically predicting magnetization directions based on magnetic data maps. The files provided contain the training data maps used in the manuscript, as well as the Python scripts and Jupyter Notebooks to generate the training set and predict the magnetization inclination and declination.&amp;nbsp;&lt;/p&gt;</description>
    <description descriptionType="Other">{"references": ["Simonyan, K., &amp; Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.", "Lecun, Y., &amp; Bengio, Y. (1995). Convolutional networks for images, speech, and time-series. In M. A. Arbib (Ed.), The handbook of brain theory and neural networks MIT Press.", "Nurindrawati, F. D., &amp; Sun, J. (2019). Predicting Magnetization Direction Using Convolutional Neural Networks. AGUFM, 2019, GP42A-09."]}</description>
All versions This version
Views 622622
Downloads 454454
Data volume 880.1 GB880.1 GB
Unique views 527527
Unique downloads 155155


Cite as