Longread

Kooistra inaugurated as Professor of Remote Sensing: 'Working on optimal agricultural systems’


Lammert Kooistra will be inaugurated as a personal professor of Remote Sensing at Wageningen University & Research's Laboratory of Geoinformatics and Remote Sensing on October 5th. Within this broad field, he will focus on optimising agriculture and making it more sustainable.

Kooistra and his team are working with remote sensing techniques that provide detailed information at the millimetre scale to gather data at the field and plant level regarding production and harvest. "The Earth has the capacity to feed ten billion mouths by 2050, but our agricultural systems need to become more efficient for this to happen," Kooistra stated in his address. "Think about precision fertilisation, efficient use of chemical substances, and dealing with drought. That's what I would like to work on."

The Earth has the capacity to feed ten billion mouths by 2050, but our agricultural systems need to become more efficient for this to happen

Within this field, Kooistra emphasises the importance of not imposing technology but rather thinking from the perspective of the user's needs. "There is a lot of interesting technology, and we aim to innovate with it. However, I believe it's crucial to make targeted choices. When working with students, I encourage them to reason backward from an application to the user's questions. Sometimes, it's sufficient for a farmer to divide the field into zones using satellites, which require different management approaches. In other situations, more technology and sensor work may be necessary."

Research on plant stress

An example of a high-tech project in which Kooistra is involved is the European
collaboration project I-Seed. In this project, sensors as small as seeds are being developed, and they are biodegradable. These sensors can be scattered in hard-to-reach places to measure temperature, moisture, or mercury concentration in the soil. "We call this soft robotics. It's a rather experimental study, but we're planning to test it in the field soon. The part I'm primarily working on is measuring active fluorescence. With a fluorescence camera, we aim to fly over an area with these scattered mini-sensors and measure the fluorescence signal. This is a completely new sensor principle."

Fluorescence is a part of the photosynthesis process for plants. "If we can gain more insight into photosynthesis processes at the plant level, we can identify limitations and determine how to address them. What stress conditions affect photosynthesis, such as drought? One outcome might be that you see a group of plants requires more irrigation to achieve higher production. This leads toward precision agriculture."

Fluorescence research extends beyond the I-Seed project and is a significant research goal for Kooistra. "It's also interesting for breeders. Which plants are more drought-resistant? Breeders increasingly want to utilize these techniques to gather data. This allows them to combine genetics from the lab with field processes and make decisions about the best varieties."

Infrared, 3D, and nitrogen measurements

Remote Sensing also involves developing cameras that can operate in different parts of the spectrum, such as ultraviolet (UV). "Insects use UV to locate flowers. When we observe this too, we learn more about the routes insects take." Additionally, there are cameras like LiDAR that can create 3D images to detect diseases in the field. "We're training models to detect visual signs of disease from
camera observations, similar to machine learning. This could be a viable solution for labour shortages in the field, especially on large plots."

Kooistra aims to understand how plants and insects interact with each other, how plants grow, and how to detect pests with all of these technologies. Agronomists and farmers can then use this information to take appropriate measures. "The goal is to gain better insights into plant growth in agricultural systems. For this, we need both high-level information - quite literally, using satellites - and detailed information from drones."

The goal is to gain better insights into plant growth in agricultural systems

"To give one more example: we discovered that growing broad beans with cabbage is a beneficial combination. We measured that nitrogen is released when broad beans are harvested, which boosts the growth of cabbage. This is valuable information for strip farming, an emerging form of agriculture where a wide variety of crops are grown alongside each other."

"Open discussion and sustainable AI use are essential"

Kooistra, who started as a researcher at Wageningen University & Research (WUR) in 2003 after studying Soil, Water, and Atmosphere and completing a Ph.D. at Radboud University in Nijmegen, began to see the potential of drones in research in 2012. He now observes that drones are increasingly being used alongside traditional methods. "WUR was a pioneer in this field ten years ago. Now, at least ten research groups within WUR use drones for data collection. For example, the Marine Research group uses them to study the seal population on the Wadden Sea. They also want to use 3D camera techniques to assess the seals' health."

To effectively organize drone usage, Kooistra has been coordinating the Unmanned Aerial Remote Sensing Facility (UARSF) since 2012. "Within this group, we discuss innovation possibilities, safe drone usage, and data collection across all WUR departments." The group also exchanges ideas about the use of AI. How does Kooistra view AI? "I want to understand how it works precisely. It can't remain a black box for me. Why can AI recognize weeds? How does the system do it? Another area I want to work on is the efficient use of training data for machine learning. Collecting field data is expensive, and reuse is important. Algorithms are often trained on specific datasets, such as detecting weeds in sandy soil. Generalizing such models requires attention, as does fair data usage and data reuse. It's crucial to continue meeting and collaborating with each other. That's what I consider being sustainable."

What Kooistra finds most fascinating about Remote Sensing is that it allows spatial information to be obtained with a piece of technology, which can then be used to establish spatial relationships. "I'm not even a super-technician, but the questions it raises and the opportunity to understand how plants grow precisely are incredibly interesting to me. That's why my group consists of diverse individuals, from agricultural researchers to optical physicists. We continually question each other, and I learn new things every day. Maintaining an open mind is essential, even for a professor."