Detecting high-yield varieties of plants

The consequences of climate change are extremely complex and they affect developing countries more than most. For instance, rising temperatures can make regions uninhabitable and cut people off from vital access to drinking water. Even richer countries cannot escape the impacts of climate change and are being forced to change their way of thinking – particularly when it comes to agriculture. Modern plant cultivars cannot adapt quickly enough to the effects of climate change, so farmers need to grow plants that have adapted especially well to the prevailing conditions. This is why we are focusing on non destructive monitoring and plant analysis at the Development Center X-ray Technology EZRT.


Many plant varieties (such as potatoes, wheat, rice, and cassava) are struggling to cope with the world’s shifting climatic conditions. In order to find appropriate ways of dealing with the changing circumstances, we are analyzing how different varieties of plants react to these environmental impacts. Phenotyping is one way of identifying plants that can, for instance, still produce sufficient yields at high temperatures.

Realistic environment for plant analysis

“Theoretically, you can assess plants simply visually in the field. However, this approach is subjective and therefore inaccurate. If a person gave hundreds of plants a school grade, one after the other, a trend would be visible, but the result will always vary. This is why we use nondestructive monitoring systems,“ says Oliver Scholz, who leads the Systems group at the Development Center X-ray Technology. To produce meaningful data, we are analyzing multiple series for dozens of plants. Our site in Fürth has a greenhouse for this purpose as well as several environmental chambers that can realistically simulate defined climatic conditions.


© Photo Fraunhofer IIS/David Hartfiel

Environmental chamber for simulating different climatic conditions.

© Photo Fraunhofer IIS/David Hartfiel

LASER LIGHT SHEET: A laser with a special widening lens projects a line of light onto the surface of an object. The light line follows the contours of the surface so that the shape of the curve corresponds to the profile of the surface. By analyzing the line’s position as the object moves, the system can measure the entire surface and represent it as a 3D dataset. A complete 360 view can be produced by using multiple sheet-of-light sensors. Image processing algorithms can use this data as a basis for analysis of various plant features.

A software program cal culates the key parameters of a leaf.

Identifying high-yield varieties

Plants consist of overground and underground organs. Important indicators of a plant’s wellbeing and fertility are above ground. We can glean valuable information from their leaves (a plant’s “solar panels”) in particular. Optical monitoring technologies, such as 3D laser processes, are well suited for observing the leaves and their development.

»We are applying our 3D plant scanner essentially to take three-dimensional photographs of a plant. A laser projects a narrow line onto the surface of the leaf. As the line travels down the leaf, a camera records the displacement of the line. In just a few seconds, this produces millions of 3D coordinates that describe the surface of the leaf,« says Scholz.

Since our work involves large test series of plants that we observe over a long period of time, this approach produces are large amount of 3D data. To help us compare data from the plants’ individual leaves, we developed a special software program that uses a sophisticated process to calculate key parameters of a leaf and then provides us with those parameters in much smaller data packets. This enables us to directly read and accurately analyze the leaf’s size, surface area, incline, and curvature. Biologists take these phenotypic data and link them with microbiological knowledge so that they can identify the biological mechanisms that allow a certain plant variety to flourish and produce sufficient yields even under extreme conditions.

Underground X-ray vision: 3D CT in minutes

The parts of a plant found underground, such as their root structures and infructescence, can also provide important information about aspects such as the plant’s biomass. Optical monitoring technologies reach their limits here, which is why we are applying X-rays instead. X-ray imaging and microscopy have made enormous advances in recent decades. The technology can now easily examine even large test objects made of steel or other alloys. Tiny material defects, for instance in aluminum tire rims or cylinder head casings, show up clearly on today’s systems and are easily classifiable. However, researchers who work with phenotyping face very different challenges. »Unlike with many industrial and laboratory applications, phenotyping is not first and foremost about razor-sharp image quality. Our limiting factor is the imaging time,” says Dr. Stefan Gerth, head of the Innovative System Design group. “We have developed our own laboratory systems that aim to strike a balance between valid image data and much shorter measurement times,“ he says.

The measurement time is significant because we generally measure a whole series of plants. Longer measurement times are not economically feasible, and keeping a plant in the X-ray machine for a long time involves »ripping« it out of its familiar climatic environment, which can seriously affect the validity of the results. This is why our work at the Development Center X-ray Technology involves optimizing our X-ray systems so that they can fully scan a plant in roughly five to seven minutes. In addition to specially adapted hardware components, the software that we use also plays a key role. Due to the short imaging time, the source data contain noise and is therefore difficult to process. Intelligent algorithms largely compensate for this and can fully automatically separate the plant‘s organs from surrounding soil.

In the next processing step, the software automatically identifies the aspect ratio of the fruit and root structures, and the weight of the plant’s organs. “To be able to make reliable statements, we observe the test series over several weeks and months. Using a diagram over time, we can work out at the end of the experiment how the plant developed in terms of underground growth,“ explains Joelle Claussen, who has so far measured thousands of plants at the Development Center X-ray Technology. “Although we achieve an exceptionally high success rate with our test series, we can never fully simulate real environmental impacts in a greenhouse environment. This is why biologists verify the research fi ndings under real environmental conditions,“ says Claussen.

With the support of national and international partners from business and research, we are confi dent that our nondestructive monitoring systems can help deliver appropriate responses to the consequences of climate change.

X-rays allow us to see underground. These are potato tubers at various stages of development

THE PRINCIPLE OF 3D COMPUTED TOMOGRAPHY: 3D computed tomography (CT) produces multiple X-ray images (known as projections) from various directions. Unlike with medical CT scanners, an object to be scanned by an industrial CT system is often placed on a rotary table and positioned between the X-ray tube and the detector. The projections are recorded as the object rotates around its own axis.