Discovering innovations that matter since 2002

AI analysis | Photo source Pixabay

Scientists use AI to analyse ash particles

Science & Environment

Researchers are using AI to help response teams know how to react to volcanic eruptions.

At Springwise, we have seen artificial intelligence systems designed for an increasingly wide range of uses. These have included marketing, branding and music composition. Now, AI is being used in the study of eruptions. Volcanologists have devised an AI application that can analyze the shape of volcanic ash. There are many different types of volcanic eruption – from explosive eruptions to lava flows. By studying the ash produced, volcanologists can gain insight into the type of eruption that occurred. This can help response teams, for example, by letting them know how large an area to evacuate or whether there will be more eruptions in the near future.

Until now, categorizing ash has largely been done by eye. This is time-consuming and relies on highly-trained experts. Now, scientists from the Earth-Life Science Institute at the Tokyo Institute of Technology have developed a way to use a convolutional neural network (CNN), a type of AI often used to analyze images, to categorise volcanic ash particles. The researchers first trained the system by generating thousands of two-dimensional images of ash particles. Then they classified the images into one of four basic shapes: blocky, vesicular, elongated, or rounded. They then fed these images into their CNN software to teach it what to look for. Additionally, the resulting AI program is 92 percent successful in accurately categorising the shape of a particle.

In order to increase the usefulness of the program, advanced magnification techniques, such as an electron microscopy, can be used to add colour and texture to the results. This could provide even more information into the type of eruption behind the ash. The volcanology program has demonstrated that CNN can be usefully applied to analysing a wide ranges of images. In the future, this technique could be used to analyze other types of particles. How might this type of CNN help scientists understand other microscopic phenomena?

Email: information@elsi.jp

Website: www.elsi.jp

Contact: www.elsi.jp/en/contact

Download PDF

Springwise Services:
Our expertise in spotting the latest innovations is the best resource to empower your team’s future planning.

Find out More