Sunday, January 11th 2026
As the weather worsened, rescue teams in the Piedmont region of northern Italy were trying to solve a difficult puzzle.
66-year-old Italian mountaineer and orthopedic surgeon Nicola Ivaldo did not show up for work on Monday morning and the alarm was immediately raised.
Ivaldo had set off alone on a Sunday morning in September 2024, without informing friends or relatives of his destination. his. The only clue rescuers had was his car, which was found in the village of Castello di Pontechianale, in the Varaita Valley.
Based on his last cellphone signal, rescue teams estimated the seasoned climber was likely headed for one of the two most iconic peaks in the Cottian Alps: the 3,841-meter Monviso, or the neighboring Visoloto.
However, both they have dozens of routes, from different sides, in a landscape full of streams, rock faces and hundreds of kilometers of trails.
On the day of the disappearance, good weather had drawn crowds of climbers to the popular passes. However, no one reported sightings, leading to the conclusion that Ivaldo moved to more isolated and difficult routes.
More than 50 rescuers searched the area on foot for almost a week while helicopters flew overhead.
With the arrival of the first snow in late September, hopes of finding him alive faded and the operation was suspended.
In July 2025, after the snow on the steep streams melted, the search began again – this time with a new weapon: artificial intelligence.
Drones and algorithms in the service of rescue
The Piedmont Mountain and Cave Rescue Service used artificial intelligence software capable of analyzing thousands of high-resolution photos taken by drones. In just five hours, two drones covered 183 hectares of difficult terrain and recorded more than 2,600 images.
The algorithm analyzed each image pixel by pixel, looking for colors or shapes that “escape” from the natural environment. Although it identified dozens of possible points of interest, the final assessment was made by humans, considering the logical route an experienced hiker could follow.
Among the possible points, a small red object stood out. When the drones approached again the next day, it turned out to be Ivaldo’s skull. His body was found in a stream on the north side of Monviso, at an altitude of about 3,150 meters, and was retrieved by helicopter.
According to rescue service spokesperson Simone Bobbio, the software was able to detect the red color of the cocoon, even though it was in the shade and partially covered by snow. Without this data, the body may never have been found.
Similar technology has been used in the past in rescue operations in Poland, Austria and Scotland.
However, experts point out that these systems have limitations: drones have difficulty in wooded areas or in low visibility, while algorithms can produce false positives.
Continuous training of machine learning systems is considered essential to improve the accuracy of them, while ethical and legal issues are also raised regarding the use of aerial images and the recognition of people.
Research teams in Italy and the United Kingdom are already working on new applications, from algorithms that predict the behavior of a missing person to systems that will analyze real-time images directly on drones.
For rescue services, which often face time constraints, weather and limited resources, artificial intelligence is not a replacement for people. But it can be essential. And in some cases, it can save lives. /tesheshi.com/
Source: prizrenpost
