Artificial intelligence is revolutionizing archaeological research, as demonstrated by the recent discovery of 303 new geoglyphs in Peru's Nazca Desert. This breakthrough, achieved by Japanese scientists from Yamagata University in collaboration with IBM, nearly doubles the known figurative designs in the region. By leveraging AI models trained to detect faint outlines in satellite and drone imagery, researchers are uncovering ancient artifacts and structures at an unprecedented rate, offering new insights into civilizations like the Nazca. The AI-assisted approach has proven to be 20 times faster than traditional human analysis, opening up possibilities for similar discoveries in other archaeological sites worldwide and transforming our understanding of ancient cultures.
AI-Assisted Geoglyph DiscoveryÂ
The groundbreaking collaboration between Yamagata University and IBM resulted in the development of an AI model capable of detecting faint geoglyph outlines in satellite and drone imagery. This innovative approach proved to be 20 times faster than human analysis, enabling the identification of 303 previously unknown figurative geoglyphs in just six months of fieldwork. The AI system's effectiveness was particularly notable in detecting smaller "relief-type" figures, which are challenging to discern with the naked eye. To achieve this, researchers trained the model using deep learning techniques on a dataset of known geoglyphs, allowing it to recognize subtle differences in shape and structure that distinguish geoglyphs from natural features.
Characteristics of New Geoglyphs
The 303 newly discovered geoglyphs exhibit distinct characteristics that provide insights into Nazca artistic preferences and cultural practices. Predominantly smaller "relief-type" figures, these geoglyphs are more intricate and complex than the larger "line-type" designs, with 81.6% depicting human motifs or elements modified by humans, such as domesticated animals and decapitated heads. Located within an average of 43 meters from ancient trails, these figures were likely constructed and viewed at an individual or small-group level. The diversity of subjects includes abstract humanoids, camelids, and a peculiar 72-foot-long "killer whale holding a knife," showcasing the rich imagination and artistic skills of the ancient Nazca people.
Cultural Significance of Nazca FindingsÂ
The newly uncovered geoglyphs, dating back to 200 BC - 700 AD, offer valuable insights into the ancient cultures of the Nazca Pampa and their artistic practices. Archaeologist Masato Sakai suggests that ceremonies involving the destruction of high-quality pottery may have taken place around the linear-type geoglyphs, based on fragments found in their vicinity. This discovery not only expands the known catalog of Nazca figures but also provides a foundation for further analysis of their motifs and distribution across the region. The findings shed light on the purpose behind the geoglyphs' creation, which had remained elusive for decades due to a lack of comprehensive information about their distribution and types.
Future Geoglyph ResearchÂ
Ongoing research promises to uncover even more ancient geoglyphs in the Nazca region. With 1,309 potential candidates initially proposed by the AI system, researchers anticipate identifying an additional 250 geoglyphs within the next few years using their AI-assisted method. This continued exploration has the potential to revolutionize our understanding of the ancient Nazca civilization, providing deeper insights into the purpose and cultural significance of these mysterious desert etchings. The team plans to refine their AI application further, enhancing its ability to detect and analyze faint outlines in satellite imagery, which could lead to discoveries in other archaeological sites worldwide.
If you work within a wine business and need help, then please email our friendly team via admin@aisultana.com .
Try the AiSultana Wine AI consumer application for free, please click the button to chat, see, and hear the wine world like never before.
Comments