Environmentally friendly farming with the help of AI

It is estimated that around ten billion people will be living on earth in 2050. Global agriculture is therefore confronted with the challenging question: How can we feed humanity without harming nature? Agricultural scientists from the University of Hohenheim, the Georg-August-University of Göttingen (UGOE) and the Julius Kühn Institute (JKI) are working on a cultivation system that combines organic and conventional farming in a meaningful way with the help of artificial intelligence.

The photo shows a crossed-out tractor spreading pesticides in the field.
Alternatives to the existing chemical-synthetic crop protection are being researched in the NOcsPS project. © University Hohenheim

Hundreds of thousands of animal and plant species worldwide are threatened with extinction due to the use of chemical pesticides. A gentler form of agriculture would be desirable, as purely organic farming will not be able to feed the world population in the foreseeable future.

The strategy of an Agriculture 4.0 relies on automated and digitally networked technologies and follows biological principles. It completely dispenses with the use of chemical synthetic pesticides (csPSM) and instead uses mineral fertilizers to ensure soil fertility to produce the required yields.

Field robot distinguishes weeds from crops

In the NOcsPS (Agriculture 4.0 without Chemical-Synthetic Plant Protection) research project, scientists have developed the AI-supported field robot Phoenix, which can be equipped for various tasks in the interests of sustainable agriculture. A setup enables the field robot to autonomously travel on fields using intelligent sensor technology and distinguish crops from weeds.

To do this, the field robot records the plants with camera and laser sensors. It evaluates the data with Artificial Intelligence methods onboard and in real time. The machine system removes the weeds mechanically using attached tools and without the use of pesticides. Useful accompanying flora, which promotes the growth of the crop and provides a habitat for insects, is largely spared.

Real-time data on field and plant condition

"In terms of data economy, the machine system must be integrated into an IT structure to provide data, for example on use and field and plant condition," explains Prof. Dr. Hans W. Griepentrog, head of the Department of Process Engineering in Plant Production at the University of Hohenheim. The field robot also sometimes needs to obtain data from the IT system in real time to optimize data analysis and AI methods.

Due to its light weight (500kg), the field robot is gentle on the soil compared to conventional heavy agricultural machinery. The electrically powered machine system can be operated with renewable electricity and thus does not emit any climate-relevant harmful gases. "Phoenix can make a significant contribution to sustainability in crop production by significantly reducing the use of pesticides," says Griepentrog.

Further applications explored

Numerous companies from production, processing and consulting are involved in the NOcsPS research project. Cultivation systems are thus developed as an interplay of innovative agronomic and other technical measures (e.g. sensor technology, robotics). In the future, Phoenix could also be further developed in terms of nature conservation, according to Prof. Dr. Enno Bahrs of the University of Hohenheim, who heads the joint project. To this end, research groups are testing applications that can use AI to distinguish "good" from "bad" accompanying flora with associated fauna. Useful accompanying flora is preserved if necessary and provides a better habitat for insects.

The Hohenheim scientists see opportunities for the technology in all types of farming. An arable farming system such as NOcsPS could inspire organic farming with regard to alternative fertilization strategies and modified crop rotations. Agriculture 4.0 without chemical-synthetic crop protection thus stands for an independent path in arable farming. The resulting products could occupy a price category that lies between conventionally and organically produced products.

Application facts


Technology field
Image Recognition and Understanding
Data Management and Analysis
Robotics and Autonomous Systems
Application industry
Agriculture
Field of application
  • Miscellaneous
Value-added activity
Research and Development [R&D]
Funding
Bundesministerium für Bildung und Forschung (BMBF)
AI Developers

Universität Hohenheim

University/Research Institution
Website

AI Development partner

Universität Göttingen
Julius Kühn-Institut
K.U.L.T. Kress Umweltschonende Landtechnik GmbH

Video


Artificial Intelligence in modern agriculture