| Summary: | Mount Erebus, Antarctica, is a volcano with frequent lava-lake eruptions known as strombolian eruptions. The larger of these eruptions create strong seismic waves and have a characteristic seismic signature that can be analyzed through three-component cross-correlation to distinguish smaller strombolian eruptions from the background noise of the volcano. The addition of an infrared camera on the rim of Mount Erebus allows for the confirmation of strombolian eruptions as opposed to unrelated seismic activity. This research finds that eruption events can also be detected categorizing the images using machine learning. Machine learning in seismology is now a commonly used technique, yet to date, no research using machine learning has ever been used in volcanology. Image categorization along with cross-correlation can improve automatic detection of strombolian eruptions.
|
|---|