Cienciaes.com: Satellites and Artificial Intelligence for Archaeology. We speak with Hector A. Orengo.

by time news

2020-08-22 19:19:26

In southeastern Pakistan there is now a vast dry region, known as the Cholistan Desert, which forms the western part of the Great Indian or Thar Desert. Currently it is a large expanse of yellowish land where fossilized sand dunes abound among which sparse shrub vegetation grows, with a few patches of trees growing around artificial ponds or tuffs created by its few inhabitants to store rainwater for their consumption and that of their livestock.

4,500 years ago, on the other hand, the region was green and fertile, it was furrowed by a large river with numerous tributaries through which meltwater from the Himalayas flowed abundantly. Many human settlements flourished on its shores, some of which, due to their size, are among the first cities in Asia. Cholistan became one of the most productive regions of the Indus Valley civilization. Later, the climate gradually changed and the inhabitants had to adapt their way of life, abandoning old settlements and moving to other places. Finally, the river stopped flowing and those human settlements, which keep thousands of years of history inside, were forgotten and many of them were buried under the sand.

At the beginning of the century XIX Europeans who arrived in the area frequently heard stories in which some mounds scattered across the desert were identified with the remains of the ancient settlements of the lost civilization. Various investigations carried out during the 20th century, especially those carried out by Mohammed Mughal, from Boston University, provided information on more than 400 places that have been associated with the different periods that go from the primitive settlements of the Indus to the arrival of Islam. .

Now, a group of scientists, led by our guest on Talking to Scientists, Hector A. Orengoresearcher of Catalan Institute of Classical Archeologyhas carried out an investigation of the Cholistan region from space, using the information provided by space satellites and using analysis procedures generated by artificial intelligence.

Hèctor explains during the interview how his team has used images of an area that covers some 36,000 square kilometers of the desert, obtained by satellites, in different bands or colors of the visible spectrum, in the infrared and by synthetic aperture radar. To carry out the analysis of the enormous amount of data provided by satellite observations, the researchers have developed a machine learning algorithm. The team fed a part of the collection of known settlements to the artificial intelligence program so that it could “learn” to differentiate the mounds that hide the remains of ancient human settlements from the natural ones. Once the learning period was over, the researchers provided the algorithm with data from the rest of the known mounds and the algorithm was able to identify them all. After the learning period, Héctor Orengo and his colleagues contributed the images and data obtained by satellites to the program and it discovered more than 300 new archaeological sites, unknown until then. The new places spread over a larger area than previously recognized and some of them, due to their large size, belong to cities.

This research not only provides fundamental information for the study of the Cholistan desert region, but also provides a very effective tool that can be applied to other locations and dry regions of the Earth that hide archaeological remains.

I invite you to listen to Hector A. Orengoresearcher of Catalan Institute of Classical Archeology

Reference:
Orengo, Conesa, Garcia-Molsosa, Lobo, Green, Madella and Petrie. Automated detection of archaeological mounds using machine-learning classification of multisensor and multitemporal satellite data

A system of automatic learning (“machine learning”) allows to locate archaeological sites in remote areas

A machine learning system makes it possible to locate archaeological sites in remote areas

#Cienciaes.com #Satellites #Artificial #Intelligence #Archaeology #speak #Hector #Orengo

You may also like

Leave a Comment