Ancient drawings — including 250-foot-long pair of legs — discovered in Peru desert

Reseachers discovered several ancient geoglyphs — massive designs etched into the ground — in Peru using a pioneering new technique.

Archaeologists harnessed the power of deep learning, a form of artificial intelligence, to uncover the designs, which date back thousands of years and are part of a UNESCO World Heritage Site, according to a study recently published in the Journal of Archaeological Science.

The geoglyphs, known as the Nazca Lines — named after the desert they were found in — can be hundreds of feet in size and often depict animals, such as hummingbirds and monkeys. Though their exact purpose is not known, it is believed they were created for some sort of ritual activity, such as ceremonial walking, researchers said.

They were first discovered about 100 years ago, and over the ensuing decades, more have been found, often via aerial photographs taken from balloons or planes.

But despite the fact that the ancient designs stretch across large swaths of the desert, finding new ones is no easy feat, researchers said. This is because from the ground, they’re often impossible to see, and the process of scanning aerial photographs is incredibly time consuming.

To speed up the process, researchers at Yamagata University in Japan developed a deep learning technique that was trained to spot the designs in satellite imagery.

The four geoglyphs discovered depict a humanoid with a club, a pair of legs, a fish and a bird, researchers said.
The four geoglyphs discovered depict a humanoid with a club, a pair of legs, a fish and a bird, researchers said.

Deep learning “is a particular kind of machine learning that achieves great power and flexibility in pattern analysis of images, speech, language, and more,” researchers said. “In archaeology, (it) is used in the analysis of the iconography, text, and writing of excavated objects.”

Though archaeologists have enlisted deep learning before, the Yamagata University researchers faced several unique challenges while searching for new Nazca Lines. The variability in geoglyph designs and the small number of them available to train the technology posed a problem.

To solve this problem, researchers developed a new approach that involved training the deep learning technology to detect geoglyphs based on small similarities to other designs rather than perfect matches.

Thanks to this method, four new geoglyphs were detected. They depict a humanoid holding a club, a pair of legs (or potentially hands) stretching over 250 feet, a fish and a bird. The designs, which are “quite unique,” were later verified during on-site visits.

“We could identify new geoglyph’s candidates approximately 21 times faster than with the naked eye alone,” researchers said. “The approach would be beneficial for the future of archaeology in a new paradigm of combining field survey and AI.”

The Nazca Lines are in southern Peru, about 260 miles southeast of Lima.

Watchful cat, slithering snake among 2,000-year-old drawings found in Peru. Take a look

Stone penis found in medieval Spanish ruins had a violent purpose, experts say

‘Cryptic’ mountain creature — with 9 babies on its back — discovered as new species