A compilation of the most famous Nazca Geoglyphs. Credit: Machu Travel Peru

AI System Helps Discover New Nazca Lines

Artificial Intelligence has helped scientists discover four new Nazca Lines.


The Nazca Lines, situated in southern Peru, are an exquisite collection of giant geoglyphs etched into the desert floor. Encompassing a diverse array of simple lines, geometric shapes, and complex animal and human figures, these impressive creations stretch across approximately 520 square kilometers. Best appreciated from an aerial perspective, they form an intricate tapestry of art that has fascinated researchers and tourists alike.

Created by the ancient Nazca culture between 500 BC and 500 AD, the lines’ purpose remains mysterious, with theories ranging from astronomical calendars to religious ceremonial pathways. This UNESCO World Heritage Site continues to intrigue us, epitomizing the ingenuity and artistic flair of a civilization long past.

Now, the ever-evolving landscape of deep learning artificial intelligence (AI) promises to revolutionize our understanding and preservation of the mysterious Nazca geoglyphs nestled in the Peruvian desert.

A Skyward Perspective on Ancient Artwork

Spanning an impressive 520 square kilometers, the Nazca lines, first unearthed archaeologically in 1932, are best appreciated from the heavens. Now, advanced AI systems inspired by multilayered neural networks bring new finds to light, continually expanding our catalog of these magnificent relics.


Harnessing AI for Discovery

The innovation stems from the diligent research team at Yamagata University, who tapped into deep learning-based object detection to reveal figurative geoglyphs in high-quality aerial snapshots. They expressed, “We launched a feasibility study with AI to expedite discovery, focusing on the northern region of Pampa de Nasca – a hub for line-style figurative geoglyphs.”

Overcoming Challenges with Smart Solutions

Four Nazca lines found with the help of AI.
Four Nazca lines were found with the help of AI.

The task of training an effective deep-learning object detection model with scant, poor-quality data required exceptional problem-solving skills. While deep learning ideally needs thousands of training data points, the researchers ingeniously expanded their training data to 307 by dissecting geoglyph features and capturing images at varied scales.

Testing and Validating the AI’s Efficiency

Using this creatively assembled detection model, the team proceeded to discover geoglyphs in aerial photographs. Confirmations came when the system flagged known geoglyphs that were not in the training data, validating their approach. Following an in-depth review of the detections, they pinpointed new geoglyph candidates for subsequent field research.

Exciting Finds: A Humanoid, Legs, a Fish, and a Bird

This endeavour led to the identification of four geoglyphs in the northern part of Pampa de Nasca: a 5-meter humanoid figure found in November 2019, a pair of legs (78 m), a fish (19 m), and a bird (17 m). Although the latter three were among the 358 geoglyphs studied by 2022, their photographs and sketches are now published for the first time in this enlightening study.


The Power of AI: Speed and Efficiency

By harnessing deep learning technology, the researchers demonstrated that geoglyphs can be identified approximately 21 times faster than human detection. The team is planning a comprehensive AI-driven geoglyph survey across Pampa de Nasca with IBM T.J. Watson and is also working with Peru’s Ministry of Culture to safeguard the newfound geoglyphs.

Have something to add? Visit Curiosmos on Facebook. Join the discussion in our mobile Telegram group.

Written by Ivan Petricevic

I've been writing passionately about ancient civilizations, history, alien life, and various other subjects for more than eight years. You may have seen me appear on Discovery Channel's What On Earth series, History Channel's Ancient Aliens, and Gaia's Ancient Civilizations among others.

Write for us

We’re always looking for new guest authors and we welcome individual bloggers to contribute high-quality guest posts.

Get In Touch