Details of the Abstract
| Title of paper | Machine learning methods for the classification of unexploded ordnance from electromagnetic data in marine settings |
| List of authors | Heagy, L.J., Lopez-Alvis, J., Oldenburg, D.W., Billings, S., Song, L.P. |
| Affiliation(s) | University of British Columbia Geophysical Inversion Facility, University of British Columbia Geophysical Inversion Facility, University of British Columbia Geophysical Inversion Facility, Black Tusk Geophysics, Black Tusk Geophysics |
| Summary | Electromagnetic induction (EMI) methods are commonly used to classify unexploded ordnance (UXO). Modern time-domain systems used for classification are multicomponent and acquire many transmitter-receiver pairs at multiple time-channels. Traditionally, classification is done using a physics-based inversion approach where polarizability curves are estimated from the EMI data. These curves are then compared with those in a library to look for a match based on some misfit. In this work, we develop a convolutional neural network (CNN) that classifies UXO directly from EMI data. Analogous to an image segmentation problem, our CNN outputs a classification map that preserves the spatial dimensions of the input. We train the CNN using synthetic data generated with a dipole model considering relevant UXO and clutter objects. We use a two-step workflow. First, we train a CNN to detect metallic objects in field data. From this, we extract patches of data that contain only background signal and use these to generate a new training data set adding this background noise to our synthetic data. A second CNN is trained with these data to perform the classification. We test our approach using field data acquired with the UltraTEMA-4 system in the Sequim Bay marine test site. |
| Session Keyword | 1.0 Instrumentation, data acquisition and processing |
| File upload |
1.0_machine_learning_methods_heagy_02.pdf
|