Artificial Intelligence

The artificial intelligence (AI) revolution is already beginning to make its presence felt in the natural stone sector in Portugal. Engineers from the Department of Mineral Resources and Energy Engineering (DER) at the Instituto Superior Técnico in Lisbon have developed a scanner for marble blocks weighing up to 25 tons which, using AI, will predict their interior texture, anticipating the colors and vein designs that the slabs may present once they have been cut. A feat unattainable by humans that promises to raise the competitiveness of the natural stone industry to another level - both indoors and outdoors.

Currently, you only know what the slabs - the slices that the stone block will give you - will look like when they are cut. Cutting is based on experience, human observation that gives you an idea of what it will look like inside. “But, obviously, we always have surprises, and mostly negative ones - it's never as good as we think,” says Paulo Diniz, director of the Galrão Group. However, he says that this “day in the life” seems to be coming to an end, thanks to the digitizer developed by DER engineers, which is already being tested on a laboratory scale. These tests in an industrial environment are taking place at the Galrão Group's factory in Pero Pinheiro, Sintra, the headquarters of the family business with 70 years of history. “We made mini-blocks and photographed all the faces. Then those plates were sliced and photographed, feeding the algorithm so that it could be more reliable,” says Paulo Diniz. In addition to “saving on raw materials, since there are fewer errors”, AI will offer “almost instantaneously, a giant virtual catalog”.

DER vice-president Gustavo Paneiro says that “AI will make it possible, in a more efficient way, to create mathematical models that predict the interior of blocks, based on what can be seen on their outer faces, which are large blocks”. Gustavo Paneiro also says that the use of AI to “predict visual characteristics that are so important in the context of civil construction”, if not pioneering, is at least rare in the industry worldwide: “From the state of the art that we analyzed, including with AI, we don't know of any technique that proposes a solution of this type.”

Read more