Fossil extraction robot can do a tedious job

Researchers have developed and demonstrated a robot that can extract, manipulate and identify microscopic marine fossils.

New technology automates a tedious process that has played an important role in improving our understanding of the world’s oceans and climate, both today and in the prehistoric past.

“The beauty of this technology is that it’s made using relatively inexpensive, off-the-shelf components, and we open source both designs and AI software,” says Edgar Lobaton, co-author of a paper on the study. Associate professor of electrical and computer engineering at North Carolina State University.

“Our goal is to make this tool widely accessible so that it can be used by as many researchers as possible to improve our understanding of the oceans, biodiversity and climate.”

The technology, called forabot, uses robotics and artificial intelligence to physically manipulate the remains of organisms called foraminifers or forams so that these remains can be isolated, viewed and identified.

Forams are protists, neither plant nor animal, and have been common in our oceans for over 100 million years. When forams die, they leave behind tiny shells, most of them less than a millimeter. These shells give scientists insight into the characteristics of the oceans as they existed when forams were alive. For example, different types of forams thrive in different types of ocean environments, and chemical measurements can tell scientists about everything from the chemistry of the ocean to its temperature when the crust forms.

However, evaluating foram shells and fossils is both tedious and time-consuming. That’s why a team of engineering and paleocinography experts developed Forabot to automate the process.

“At this point, Forabot can identify six different types of forams and process 27 forams per hour, but it never gets bored and never gets tired,” Lobaton says. “This is a proof-of-concept prototype, so we will increase the number of foram types it can identify. We are also optimistic that we can improve the number of forams it can process per hour.

“Also, at this point, Forabot has a 79% accuracy rate in identifying forams, which is better than most educated people.”

“Once the Forabot is optimized, it will become valuable research equipment, allowing students’ ‘foram collectors’ to better spend their time learning more advanced skills,” says Tom Marchitto, co-author of the paper and professor of geosciences. University of Colorado, Boulder. “By using community-sourced taxonomic knowledge to train the robot, we can also improve the uniformity of foram identification across research groups.”

Here’s how Forabot works. First, users have to wash and sift hundreds of foram samples. This leaves users with a pile that looks like sand. The foram sample is then placed in a container called an isolation tower. A needle at the bottom of the isolation tower then projects upward through the sample, lifting a single foram from which it is removed from the tower by suction. The suction pulls the forum into a separate container called the viewing tower, equipped with an automatic, high-resolution camera that captures multiple views of the foram. After the images are taken, the needle raises the forum until it is sucked up and left in the corresponding container at a sorting station.

“The idea is that our AI can use the images to identify what type of foram it is and sort it accordingly,” Lobaton says.

“We publish in an open-source journal and include the blueprints and artificial intelligence software in the supplemental materials for this article,” adds Lobaton. “I hope people take advantage of this. The next step for us is to work on expanding the types of forams the system can identify and optimizing operational speed.”

The paper appears in the magazine Geochemistry, Geophysics, Geosystems. The study received support from the National Science Foundation.

Source: NC State

Original Study DOI: 10.1029/2022GC010689

Leave a Reply

Your email address will not be published. Required fields are marked *