Robot independently reassembles destroyed Pompeii frescoes
An AI analyzes fresco fragments, and a robot then assembles them. Years of puzzle work by archaeologists can thus be accelerated.
(Image: Universität Bonn/Screenshot)
Scientists at the University of Bonn have taught a two-armed robot to reassemble archaeological fragments of frescoes. To achieve this, the fragments must be recognized, analyzed by AI, and put back together like a puzzle. The Bonn researchers tested the system together with partner institutes using heavily fragmented ceiling paintings in the Archaeological Park of Pompeii.
It would take archaeologists a long time to reassemble fragmented frescoes: the individual fragments must be sorted, their contents and fracture edges recognized, and arranged correctly. The problem: parts of the fresco may be heavily weathered or missing. Furthermore, the pictorial template is missing.
Due to this expected lengthy puzzle work, many frescoes are still stored in depots. This is set to change with the robot system from the University of Bonn, which is part of the now completed RePAIR project (Reconstructing the Past: Artificial Intelligence and Robotics Meet Cultural Heritage), funded by the European Union. The project was coordinated by the Ca’ Foscari University of Venice. Further project partners included the Italian Intitute of Technology (ITT), the Archaeological Park of Pompeii, Ben-Gurion University in the Negev, and the Instituto Superior Técnico Lisbon.
AI analyzes, robot puzzles
The process begins with a 3D scan and the digitization of the fragments of a fresco. An artificial intelligence (AI) determines which fragments might belong together based on the fracture edges and image elements. This then leads to suggestions for possible combinations to reconstruct larger image segments.
Empfohlener redaktioneller Inhalt
Mit Ihrer Zustimmung wird hier ein externes YouTube-Video (Google Ireland Limited) geladen.
Ich bin damit einverstanden, dass mir externe Inhalte angezeigt werden. Damit können personenbezogene Daten an Drittplattformen (Google Ireland Limited) übermittelt werden. Mehr dazu in unserer Datenschutzerklärung.
“The challenge is that, unlike a normal puzzle, we don't have a picture on the box,” says Maren Bennewitz, Professor of Humanoid Robots and Vice-Rector for Digitization and Information Management. “Many pieces are heavily damaged or missing entirely, and fragments from different works are often mixed. It is therefore all the more important that robotics and AI work closely with archaeological expertise.”
The possible composition calculated by the AI is then carried out by a two-armed robot equipped with five-fingered soft hands. It then assembles the pieces in the suggested combination.
“Our algorithms calculate the movements for the two robot arms with which the fragments are reliably picked up and carefully placed at the calculated position,” says doctoral student Nils Dengler.
Videos by heise
The researchers tested the system on heavily fragmented ceiling paintings and frescoes in the Archaeological Park of Pompeii. They come from the “House of the Painters at Work” and the Schola Armaturum, which were heavily destroyed during the eruption of Vesuvius in 79 AD by bombing raids in World War II and by a collapse in 2010. They could be partially restored.
The researchers are confident that the system, based on AI and robotics, can also be applied to other fragments of wall paintings, architectural elements, and ceramics stored in depots worldwide. The reconstruction of archaeological objects could thus be greatly accelerated.
(olb)
