Discovered over 300 new Nazca lines
About 300 new Nazca Lines have been recently discovered thanks to the use of artificial intelligence, a number almost equivalent to what archaeologists had identified in 70 years of research. The result was published in the journal of the American Academy of Sciences (Pnas) and is the work of a team of Japanese researchers from Yamagata University, in collaboration with IBM and the University of Paris 1 Pantheon-Sorbonne.
The Nazca Lines are ancient geoglyphs — enormous drawings etched into the ground of the Peruvian plateau desert by an ancient civilization — dating back around 2,000 years and first documented in the 16th century by Spanish explorers. The uniqueness of these extraordinary works of art is that they are only visible from above, making them ideal for study using modern technologies such as aerial photography, drone footage, satellite images, and Lidar scans.
To facilitate the search for these intricate works, researchers developed a machine learning algorithm trained to recognize geoglyphs in high-resolution images of the Nazca region. As a result, 303 new drawings were identified, adding to the 430 already known.
Artificial intelligence not only proved effective in detecting linear geoglyphs but also excelled in identifying smaller and more complex relief geoglyphs, which often escape traditional visual analysis. Specifically, “81.6% depict domestic animals or decapitated human heads,” while “64% of linear-type geoglyphs depict wild animals.”
These relief geoglyphs are, on average, located 43 meters from pedestrian paths, suggesting they were created to convey cultural concepts related to domestic animals and human sacrifices to those walking the paths. On the other hand, the positioning of the linear geoglyphs suggests they were an integral part of communal rituals.
