IBM’s AI Helps to Unearth Hidden Nasca Lines
The Nasca Lines, a group of giant, mysterious designs in the earth in southern Peru that date back thousands of years, have mystified explorers and researchers since their rediscovery in the 1920s. Some think the lines, which depict animals, shapes, plants and more, were solstice-related; others think they were meant as tributes to deities. Now, IBM Research has announced that Yamagata University is deploying its AI tools to help uncover more Nasca Line formations – and that they’ve already discovered one.
The Nasca Lines, which are best viewed from the air, are not all evident upon first glance. Researchers from Yamagata University themselves uncovered more than 100 geoglyphs just in the last 13 years. To date, this has primarily been done through satellite- or drone-based hyperspectral sensing and imagery over the 500 square kilometer area occupied by the Nasca Lines. These efforts have been obscured by “white noise” (such as roads and flood trails) that surround the geoglyphs, making it difficult to discern relevant formations.
IBM’s PAIRS Geoscope platform, which launched in February, is a cloud-based AI technology that allows researchers to scale their geospatial analytics to massive, complex datasets for tasks like crop identification and irrigation management. IBM Research, together with Yamagata University, are leveraging this technology to better understand the Nasca Lines – and hopefully, discover new ones.
PAIRS is aggregating and analyzing disparate datasets describing the Nasca Lines: LiDAR elevation data, drone imagery, satellite imagery, survey information and more. “Using traditional approaches, it would require a significantly longer time to integrate these types of data volumes, potentially adding months to the discovery process,” IBM Research wrote. “With PAIRS, these same tasks and analyses are expected to take minutes.”
IBM Research and Yamagata University are not entering into this partnership sight unseen. Before committing to the collaboration, they tested their method of assessing the Nasca Lines by training a deep neural network to identify them, then feeding it more images to see if the algorithms could detect geoglyphs that the researchers couldn’t. And, in fact, it could: the model was able to rediscover a new geoglyph (a humanoid figure) among the Nasca Lines.
Now, the teams say they “look forward” to using PAIRS to assist with further Nasca Lines analyses. Their hope, they wrote, is that by training AI and deep learning models on these volumes of unique and unstructured information, they will be able to gain valuable insights that could lead to greater discoveries and information surrounding the Nasca Lines.