Can we develop an effective early warning system for volcanic eruptions using `off the shelf' webcams and low-light cameras?
Abstract
An effective early warning system to detect volcanic activity is an invaluable tool, but often very expensive. Detecting and monitoring precursory events, thermal signatures, and ongoing eruptions in near real-time is essential, but conventional methods are often logistically challenging, expensive, and difficult to maintain. Our investigation explores the use of `off the shelf' webcams and low-light cameras, operating in the visible to near-infrared portions of the electromagnetic spectrum, to detect and monitor volcanic incandescent activity. Large databases of webcam imagery already exist at institutions around the world, but are often extremely underutilised and we aim to change this. We focus on the early detection of thermal signatures at volcanoes, using automated scripts to analyse individual images for changes in pixel brightness, allowing us to detect relative changes in thermally incandescent activity. Primarily, our work focuses on freely available streams of webcam images from around the world, which we can download and analyse in near real-time. When changes in activity are detected, an alert is sent to the users informing them of the changes in activity and a need for further investigation. Although relatively rudimentary, this technique provides constant monitoring for volcanoes in remote locations and developing nations, where it is not financially viable to deploy expensive equipment. We also purchased several of our own cameras, which were extensively tested in controlled laboratory settings with a black body source to determine their individual spectral response. Our aim is to deploy these cameras at active volcanoes knowing exactly how they will respond to varying levels of incandescence. They are ideal for field deployments as they are cheap (0-1,000), consume little power, are easily replaced, and can provide telemetered near real-time data. Data from Shiveluch volcano, Russia and our spectral response lab experiments are presented here.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFMNH53B1998H
- Keywords:
-
- 4316 Physical modeling;
- NATURAL HAZARDSDE: 4341 Early warning systems;
- NATURAL HAZARDSDE: 8419 Volcano monitoring;
- VOLCANOLOGYDE: 8488 Volcanic hazards and risks;
- VOLCANOLOGY