Rostock (dpa/mv) - Willy and Maja can fly, but they are no lightweights. The two Agras T30 drones with the familiar bee names belong to the Fraunhofer Institute for Computer Graphics Research (IGD) in Rostock and weigh 26 kilograms empty and up to 76 kilograms with full fertilizer tanks. They are actually only used at the very end, when the highly complex, AI-supported graphical recognition work over the field is done. The IGD researchers bring the terms "smart farming" and precision agriculture to life. Their goal: to drive forward the digital revolution in the agricultural sector.

The young team around Philipp Wree, consisting of agricultural scientists and computer scientists, sees itself as a partner for farmers, industry and science. "We ask the farmers whether they really need what we want to develop," says Wree. IGD is all about applied research. High-resolution optical data technology is used to create digital screens of plants, land and even animals. A lot is possible with AI, but not everything is necessary. "Just because we can do it doesn't mean we have to do it," says Wree, who heads the Smart Farming department. The focus of the research is on the benefits.

Across the field with hyperspectral cameras

The doctor of agricultural sciences moved to IGD in Rostock in 2022. Wree previously worked in industry and was also a management consultant specializing in the agricultural and food sector. The IGD wants to provide farmers with digital expertise. Whether for rapeseed, maize or other crops - the small drones equipped with highly sensitive and expensive hyperspectral cameras fly over the fields and collect information. "They see what humans certainly can't see," says IGD agronomist Florian Männer.

The camera lenses scan the plants on various wavelengths and 160 channels. For example, stress symptoms in plants caused by drought, moisture and disease are detected early on and data on nutrient supply and protein content is collected. In addition, weeds and pests in the field are identified. However, the algorithm-based systems first have to be taught all of this, like a trainee. The information can later be used to apply herbicides or fertilizers by drone and with pinpoint accuracy using the spot spraying method.

Optical data acquisition and AI can also be used to determine the condition of moors or to confirm compliance with biodiversity requirements for grassland funding guidelines by recognizing species. In the agriculture of the future, farmers will have one or two drones in a garage on a charger, which will then fly in and out independently and scan fields and plants: "That will happen at some point," says Wree, even if many regulatory issues still need to be clarified before then - as with autonomous driving.

AI as a lever for efficiency and resource planning

A fundamental question for other sectors is whether artificial intelligence should replace people or human knowledge in certain areas of work. The IGD believes that the digital expert created is more likely to assist the farmer. "It is a tool to increase efficiency, save costs and make resource consumption sustainable," says Wree. How open to technology and digital are farmers? "The affinity comes with a certain benefit."

The farmers at the 1300-hectare Papendorf agricultural cooperative laid the foundations for precision farming, a kind of precursor to smart farming, back in 2014. "It all started with site-specific basic fertilization," says the chairman of the cooperative, Steven Hirschberg. This involves taking specific soil sample grids for five hectares. The data is digitized, fed into computers and evaluated. "Then I can load it all onto the USB stick. I take it over to my tractor and load it onto the terminal," says Hirschberg, explaining the procedure.

With the help of GPS, the tractor knows which plot of land it is on, and the attached fertilizer spreader is designed to recognize where it should spread how much nutrient. "That's how it all started." Depending on the soil, this saved 30 to 40 percent fertilizer. Away from the blanket principle and towards targeted use. This now also applies to plant protection products at the cooperative, although the data basis is different.

Farmers open to technology

Robotics, computer-controlled systems and automation have been finding their way into farms and stables for decades. Politicians are also pushing and promoting this development. Agriculture 4.0 is already a reality in stables and fields, as Federal Agriculture Minister Cem Özdemir (Greens) assured the Bundestag Digital Committee at the end of last year. He also referred to 36 projects in the field of AI. AI-based milking robots that can simultaneously record how the animal is doing offer great opportunities for animal health.

Agricultural scientist Wree attests that farmers in Mecklenburg-Vorpommern are "very open-minded and have a great spirit of innovation". Steven Hirschberg, who belongs to the younger generation, is also not afraid of technological developments for professional reasons. For him, digital solutions must be as individual as agriculture itself. The farmer has the final say, he says, deciding which numbers and quantities to use. "At the end of the day, it's still people. That's actually the most important thing."