Polish scientists are the first in the world to want to use artificial intelligence and artificial neural networks to detect pre-historic monuments, including cemeteries, castles, and settlements.
The artificial neural network (ANN) is the computing system created to replicate the way the human brain works, analyzes, and processes information. It’s built like a human brain and solves problems that would be impossible or difficult for humans. It has self-learning capabilities that enable them to gain better outcomes because more data are available.
Piotr Wroniecki, archaeologist and project initiator said that they want to use artificial neural networks to detect ancient traces of human-made structures. In contrast to some popular opinion, even in Poland, there are still thousands of unknown relics.
He received a grant from the National Science Center for that purpose and his team, apart from archaeologists, includes geographers, historians, IT specialists, and experts from the commercial sector.
Wroniecki claims if the system works as expected, it could successfully revolutionize archeology.
According to him, to make ANN work in an appropriate manner, it needs a very large amount of data for processing so they will “feed” it terabytes of satellite imaging. For several decades, scientists have been analyzing this imaging for archeology, but people are browsing the photos and it takes a long time. Additionally, people view them subjectively so that’s why old traces of human activity are often overlooked. He hopes that artificial intelligence will be able to mark them quickly, efficiently, and objectively.
The neural network will be programmed in a way to be focused on the so-called plant distinguishing features visible in satellite or aerial photographs. In areas where there are remains of old buildings, grain or grass grows in a slightly different way than in the surrounding area. The outlines of prehistoric houses are indicated, for example, by the presence of darker vegetation.
Wroniecki explains that satellite images will be uploaded to the artificial intelligence system. The system will find plant characteristics, which it will classify as archaeological remains, mark them, outline them and apply them on the map. Normally it would take weeks, maybe months, to get the job done by hand, depending on the area. An artificial neural network will do it in a few or several hours for hundreds of thousands of square kilometers.
Artificial neural networks - as Wroniecki says - are a hit in various areas of life. In archeology, this will not be the first use of them, because they are used to e.g. classify fragments of ceramics. This is one of the most common categories of monuments discovered by archaeologists. However, no one has used artificial neural networks to analyze plant characteristics in satellite imagery in search of traces of ancient human activity.
Wroniecki emphasizes that an in-depth analysis of satellite images will enable the analysis of human presence traces in a comprehensive, holistic manner.
Not only individual settlements will emerge, but also cemeteries, roads, and remains of other structures located next to them. We can expect a very large increase in data - even greater when people began to use satellite imaging on a larger scale several decades ago.
The project will last three years. The archaeologist hopes that it will result in a prototype of a tool that can be used throughout Europe. He emphasizes that it will not be a finished product because the tool will require improvement.
We can’t wait to see what will be discovered in our beautiful country.