Artificial intelligence discovers new geoglyphs such as the Nazca lines.
Archaeologists first discovered the mysterious Nazca lines in southern Peru while traveling on foot in the late 1920s.
For nearly a century, new technologies such as detection and remote hyperspectral imagery, based on satellites or drones, have helped researchers discover hundreds of these figures in an area spanning over 500 square kilometres.
Researchers at Japan's Yamagata University have discovered more than 100 new geoglyphs since 2006.
These scientists are now looking to work more efficiently and improve their ability to find and study new geoglyphs through the use of the IBM PAIRS platform and artificial intelligence (AI).
Created between 500 B.C. and 500 A.D., Nazca lines represent shapes of varying complexity.
They range from simple geometric shapes and plants, to zoomorphic designs of animals; some of which are several hundred meters in length, engraved on the ground.
The exact purpose of the geoglyphs is still unknown. Many theories have emerged about why ancient Nazca cultures created them, from marking solstice points to offering art to deities in the sky.
By discovering more of these mysterious formations, archaeologists hope to gather clues about their existence.
Large-scale geospatial analysis
Hoping to discover new formations and getting a better understanding of the complexity and culture of the Nazca lines, archaeologists at IBM Research and Yamagata are jointly implementing IBM PAIRS Geoscope.
This is IBM's cloud-based artificial intelligence technology to scale large-scale geospatial analysis with very complex datasets.
Until now IBM's PAIRS was used for crop identification and irrigation management, as well as to monitor vegetation growth around assets such as power lines to reduce the risk of outages.
For Yamagata research, PAIRS offers the unique ability to analyse massive and disparate geospatial and temporal data sets from various sources.
This includes analysing the layers of LiDAR data, which are used to detect and examine the Earth's surface, along with drone images, satellite images and geographic survey information, to help reveal new lines and formations.
Such integration is often a difficult challenge given the scale and heterogeneity of these data sources.
Using traditional approaches, would require significantly longer time to integrate these types of data volumes, which could add months to the discovery process.
With PAIRS, these same tasks and analyses are expected to take just a few minutes.
Before Yamagata adopted and deployed PAIRS, researchers at the university and IBM spent the last few months researching the feasibility of artificial intelligence to help locate and better understand new formations.
Discovering new formations of the Nazca Line has historically been difficult due to the amount of "white noise" surrounding them, including roads and flood trails.
The initial objective of the researchers was to explore whether AI could help examine huge amounts of data to identify relevant clues that could lead to the discovery of new figures.
To test this theory, IBM and Yamagata used IBM Watson Machine Learning Accelerator (WMLA) at IBM Power Systems to help researchers quickly analyse images of drones and satellites to identify possible new geoglyphs.
After training a deep neural network to identify the Nazca lines, the researchers fed more images into the system to see if the AI could detect marks that the researchers had not detected.
The process was a success, as the artificial intelligence model discovered a humanoid-like figure in those images that the researchers could not detect, resulting in the discovery of a new formation of Nazca lines.
IBM Research and Yamagata hope to integrate large, multimodal and disparate datasets using PAIRS.
Their hope is that by training artificial intelligence models and deep learning in these volumes of unique and unstructured information, they can obtain valuable information that could lead to further discoveries and information surrounding the Nazca Lines.
