Revolutionary method based artificial intelligence allowed archaeologists to find hidden traces of ancient buildings in the northern regions of Chile. Technology is radically changing the approach to searching for historical monuments.
In the Altos de Arica region of northern Chile, scientists have used advanced technology to search for ancient settlements. They analyzed satellite images taken by the WorldView-2 apparatus. A feature of these images is their high resolution of 0.5 meters per pixel, which allows you to see even small stone structures remaining from ancient dwellings and animal pens.
To process a huge amount of data, the researchers used modern neural network ConvNeXt v2. This artificial intelligence system was specially trained on eight hundred examples and learned to accurately distinguish man-made stone circles from natural rock formations.
The results of testing the method exceeded expectations. Over an area of almost nine square kilometers, the intelligent model identified 85% of buildings already known to science and discovered previously unknown objects. At the same time, the percentage of errors turned out to be minimal – less than 0.1% of the total research area.
New AI search technology brings practical benefits to archaeologists. Instead of spending many days studying each image, specialists can now immediately proceed to checking the areas marked by the neural network. This significantly saves time and resources, making the process of discovering ancient settlements in the inaccessible mountains of Chile as efficient as possible.
The material was prepared with the support of the Russian Ministry of Education and Science as part of the Decade of Science and Technology.
Source: @twrussia
