Tracking down spores

A box suspended from a thin blue steel construction on rubber wheels films sugarbeets. Another contraption stands on a small vehicle the size of a serving trolley and directs a laser onto the leaves. These apparatuses in a field near the Bavarian town of Plattling are helping to find a solution to a problem with serious repercussions: infection of sugarbeet leaves with the harmful leaf spot disease Cercospora.

If farmers don’t detect and combat this fungus early enough, they soon face the threat of losing a large part of their sugarbeet in a field, according to Dr. Ulrike Willer, a physicist from Clausthal University of Technology. Several companies and research institutes have joined forces to prevent that. Their goal is “to automate testing of a large number of plants in a field and detect infection quickly and take countermeasures in good time,” explains the physicist and project manager Dr. Christoph Bauer from KWS.

The method used by Clausthal University of Technology

Clausthal University of Technology is focusing on a special laser that uses a non-visible infrared beam. An automated assessment method is to enable early detection of Cercospora. In this approach:

• The sugarbeet leaves are illuminated with the laser. “At specific wavelengths in the mid-infrared range, the infected leaf areas warm differently than healthy ones,” explains Willer.
• An infrared camera films the plant’s illuminated leaf, measures the temperature in this way and stores the values.
• The ultimate goal is for a thermal imaging camera to recognize whether a plant is infected – before the human eye can.

The method used by Jülich Research Center

Jülich Research Center is pursuing a different approach: Its LIFT system focuses on photosynthesis in the leaves or, to be more precise, what is termed chlorophyll fluorescence, explains its phenotyping expert Dr. Onno Muller:

• Photosynthesis, i.e. the conversion of solar energy and carbon dioxide into chemical energy in the form of glucose, begins in the plant by the light being caught by the pigment chlorophyll.
• In this process, part of the energy captured by the plant is re-emitted and radiated back. That “glow” is called chlorophyll fluorescence.
• The researchers from Jülich record it with the LIFT sensor. The assumption is that if an area of the leaf is infected with Cercospora, that influences photosynthesis in that area. Since chlorophyll fluorescence correlates closely with the efficiency of photosynthesis, the change in fluorescence on the affected leaf areas could indicate infection by the harmful fungus.

Computer scientists train artificial intelligence

During this first measurement campaign in the project on around a hectare of a field, the scientists from the two institutes will collect about two terabytes of data with their two experimental setups – not to mention the weather data that is important to growth: moisture, wind, wind direction, precipitation, air pressure and temperature. The Technical University of Dortmund analyzes and processes all those data sets.

In the case of Clausthal’s measurements, the computer scientists from Dortmund identify, for example, the laser wavelength at which the temperature difference measured on the diseased plant is the most striking. They tell the software that such a difference is an indicator of disease in order to teach it how to detect a sick plant. They repeat that step over and over again and train the artificial intelligence algorithm so that it can later identify by itself whether a plant is infested or not.

The upcoming analyses will likewise show how the data of the scientists from Clausthal and Jülich can best be combined.