AI and ML Incorporated with MOUNTS Positioned to Monitor Volcanoes and Analyse Satellite Data

by July 16, 2019

Monitor Volcanoe

The Technical University of Berlin and the GFZ German Research Centre for Geosciences in Potsdam have developed a new monitoring platform to analyses volcanoes through satellite imaging using AI. Sébastien Valade who is leading the research project along with his colleagues demonstrated that MOUNTS (Monitoring Unrest from Space), the volcano monitoring platform, can integrate varied sets of different types of data for comprehensive monitoring of volcanoes. The results to this study got published in the Remote Sensing.

Interestingly, satellites can provide significant data while the ground-based monitoring is confined. The long term observations from space are majorly important to better identify the signals of volcanic unrest.

In most of the cases, volcanic eruptions are predated by precursory signals which can last a few hours to certain years. Such signals are likely to involve seismic behavior, ground deformation, gas emission, temperature increase, and many others. Besides seismicity, other traits can be monitored from space by using various wavelength across the electromagnetic spectrum.

Reportedly, the research project is funded by GEO.X which is a research network for Geosciences in Berlin and Potsdam founded in 2010. Employing the MOUNTS system, the research team is able to exploit various satellite sensors in order to identify and scale the changes around volcanoes.

A significant part of the project was to examine if AI algorithms can be successfully integrated into the data analysis process. These algorithms were basically developed by Andreas Ley from the Technical University of Berlin. The man has applied artificial neural network automatically identify large deformation events. The algorithms were trained by computer-generated images, imitation of real satellite images. From this style of machine learning training, the software understood to identify large deformation events in real satellite data.

Andreas Ley said – “For us, this was an important ‘test balloon’ to see how we can integrate machine learning into the system. Right now, our deformation detector just solves a single task. But our vision is to integrate several AI tools for different tasks. Since these tools usually benefit from being trained on large amounts of data, we want to make them learn continuously from all the data the system gathers on a global scale.”

Currently, MOUNTS system monitors 17 volcanoes across the world which include the Popocatepetl in Mexico and Etna in Italy. Additionally, the platform’s website is freely accessible and can easily incorporate new data.