The recent application of artificial intelligence (AI) in archaeology has revolutionized the way we uncover ancient mysteries, as evidenced by a groundbreaking discovery in the Nazca Desert of Peru. This cutting-edge technology has not only accelerated the pace of archaeological research but has also provided unprecedented insights into the ancient Nazca culture.
The Nazca Desert, known for its enigmatic geoglyphs, has long fascinated archaeologists. These ancient drawings, etched into the desert floor, depict various figures and patterns, some of which are visible only from the air. Traditionally, archaeologists relied on manual surveys, using planes, cameras, and human effort to map these geoglyphs. However, the process was slow and labor-intensive, with only a few new figures discovered each year.
The turning point came when a team of archaeologists led by Masato Sakai from Yamagata University collaborated with data scientists from IBM Research. They employed AI techniques to analyze high-resolution images of the Nazca Desert, training the model on everyday photographs and fine-tuning it with a few hundred known Nazca figures. This approach, while challenging due to the limited training data, proved to be a game-changer.
The AI system generated a continuous geoglyph probability map, with a resolution of five meters, indicating areas where geoglyphs might exist. This map identified 47,410 candidate patches, each linked to precise coordinates, allowing field teams to focus their efforts efficiently. Human specialists then screened these candidates, narrowing them down to 1,309 high-potential sites, which were further investigated using drones for ultra-detailed imagery.
The field survey revealed an astonishing 303 new figurative Nazca geoglyphs and 42 additional geometric designs, bringing the total to 683 examples. These geoglyphs are categorized into relief-type and line-type figures, each with distinct characteristics. Relief-type figures, made by piling or removing stones, are smaller and often depict humans and domesticated animals. Line-type figures, on the other hand, are larger and represent wild animals and other creatures.
One fascinating aspect of this discovery is the distribution of these geoglyphs. Relief-style figures are found closer to informal, winding trails, suggesting they were meant for people walking local paths. Line-type figures, however, are positioned along more formal networks, possibly anchoring long-distance processions. This distinction provides valuable insights into the cultural and social practices of the Nazca civilization.
The impact of AI on archaeology is profound. The study, published in the Proceedings of the National Academy of Sciences, demonstrates a sixteen-fold acceleration in the discovery of figurative Nazca geoglyphs compared to traditional methods. AI not only speeds up the process but also enhances accuracy, as evidenced by the 968 high-potential sites and 248 AI-proposed figures that still require verification.
This project has set a precedent for AI archaeology and drone-based surveys worldwide. By documenting their methods and sharing their findings, the researchers have created a valuable resource for other archaeological teams. The use of AI and drones has not only revolutionized the Nazca Desert exploration but also opens up new possibilities for uncovering ancient mysteries across the globe.
