A digital nose for sick trees and infected pigs – Diepresse.com

There would be a lot to hear in a forest – if we understand the language of the trees. A lot quoted, but not yet scientifically adequately occupiedis hers Communication via root mushrooms (Mycorrhiza). On the other hand, more is known from the lively exchange about a variety of chemical compounds. Plants use such odorless molecules transmitted by the air to attract their pollinators or to warn each other of pest infestation. The latter in particular increasingly moves into the research focus in view of the climate change, which demands its toll in the forest.

Sensor smells dry stress

« What if we could understand the language of the trees? » Asks the biologist, which specializes in plant diseases Claudia Probst from the Research Center Wels of the University of Applied Sciences (FH) Upper Austria for years. Together with an computer scientist and an automation technician, she made her seriously out of the mind: in the project sponsored by the Waldfonds Republic of Austria « Diginase“The researchers have developed a small device with gas sensors that can use artificial intelligence to capture stressful, volatile organic compounds of trees and pheromones (fragrances) of insects (olfactory perception).

The former provide information about the growth, the immune system and the state of health of plants – a good way to monitor the plants’ vitality in an uninvious way. Seconds provide insights into the activity and pairing behavior of the beetle population.

Drone flight would dilute fragrances too much

« In addition, a visual artificial intelligence is running that takes a picture of the state of the bark, » explains Probst. « The combination of all of this information, i.e. barking beetle, stressed trees and visible changes, results in an early warning system for forest workers and forester, which can thus be seen whether there is a risk of bark infestation. »

The practical: The first prototype is only 180 grams and should be able to be mounted on the collar of a dog in the future. The team (Georg R. Schneider, Gregor Maier-Lehner, Thomas Penkner, Eva O. Huber) is currently working on a second prototype that will be even easier. An alternative to this is a small autonomous vehicle robot, which not only registers illnesses on its journey through the forest, but also creates a digital map of forest sections with the help of a laser scanner. Why not a drone? « A lot of air flows come into play with a drone, which dilutes the fabrics very much, which is not ideal. »