Estimating total magnetization directions using convolutional neural networks
Felicia Nurindrawati, Jiajia Sun · 2019
Proper interpretation of magnetic data requires an accurate knowledge of total magnetization directions of the source bodies in an area of study. In this study, we examined the use of machine learning, specifically Convolutional Neural Network (CNN), to automatically predict the magnetization direction of a magnetic source body based on a magnetic map. We simulated magnetic data maps with varying magnetization directions from a cubic source body, all subject to the same inducing field. Two CNNs were trained separately, one for predicting magnetization inclinations and the other for predicting magnetization declinations. We also investigated various CNN architectures and determined the optimal architectures for predicting inclinations and declinations. For the optimal architectures, we achieved 98% and 100% test accuracy for our declination and inclination predictors, respectively. Furthermore, this method was tested on magnetic field data from Black Hill Norite, Australia, with encouraging results. Our study shows that machine learning holds great promise for automatically predicting magnetization directions based on magnetic data maps. Presentation Date: Wednesday, September 18, 2019 Session Start Time: 8:30 AM Presentation Time: 8:55 AM Location: 301B Presentation Type: Oral