Ard Nieuwenhuizen carries out field research in Wageningen (Photo: WUR)
Every farmer will soon be thinking in terms of precision agriculture
5 May 2026
Every two months, we introduce a specialist. We offer a glimpse of the person behind the research and the work they do. This time: WUR researcher at Vision+Robotics Ard Nieuwenhuizen. He talks about a revolutionary spot sprayer, the importance of timing, and why technology and agroecology complement each other perfectly: “Multipurpose vehicles are the future.”
When the second week of June brings a few dry days in a row and the sun looks set to stay, WUR researcher Ard Nieuwenhuizen gets hay fever. Then there is only one thing on his mind: making hay. He strikes out his diary appointments and heads to Zevenhoven, where the grassland around his family’s farm is waiting.
Making hay is no nostalgic picture-postcard scene. It means trusting the sun, keeping an eye on the sky and acting at exactly the right moment. To mow and dry grass into hay, and to bale it, you need three to five sunny days in succession. If you wait too long in June, the grass starts to lodge, says Nieuwenhuizen. The result: “The growth goes out of it and the grass becomes stale and musty at the bottom, close to the ground. That spoils the taste of the hay for the cows, which then miss out on a healthy food supply for the winter.”
So he goes. Out into the fields with five to seven others. Making thousands of hay bales, each weighing 15 to 20 kilos. In the background, the rattle of the tedder, the smell of dry grass and the tractor radio blaring across the land. It is not work. It is an event. It is togetherness.
Hoogmade
Nieuwenhuizen grows up in Hoogmade, a Dutch village in the polder landscape, with straight ditches, broad stretches of water and large farming families. In winter, you can step straight onto the ice from almost any meadow. Skating for hours, seemingly without end.
His grandparents run a farm. His father’s eldest brother takes it over. His father chooses a life as a civil servant in The Hague. Nieuwenhuizen himself has more farming blood in him. As a boy, he loves tinkering in his bedroom. He takes radios and other electrical devices apart, only to screw them back together again afterwards. But he would rather go to the farm of his uncle Jos, his mother’s brother, who has started a farm of his own. In Zevenhoven, eighteen kilometres away.
As soon as he can, he gets on his bike. Pouring concrete, caring for cows, making hay, driving tractors. “It was wonderful,” he says, his eyes lighting up. “I helped build that farm.”
Green and technology both pull equally strongly at Nieuwenhuizen. So when he chooses to study Agricultural Engineering in Wageningen in 1998, no one is really surprised.
Potato volunteers
After graduating, he begins his PhD research in 2004, as one of the first PhD students of Eldert van Henten, who has only just been appointed professor. Under his leadership, robotics and high-tech gain a firm foothold in agricultural engineering. Nieuwenhuizen investigates how cameras can identify weeds among sugar beet, with special attention to potato volunteers: “Tubers and potato remnants left behind in the soil after the potato harvest. Potential carriers of disease. You can hoe them off, but they will simply grow back.”
Global innovation
Detecting weeds with a camera. Today it sounds like business as usual, but in 2005 it is pioneering work. Mobile phones are not yet pocket-sized cameras and computers. AI does not extend much beyond autonomous robot vacuum cleaners for consumers. “For three years I carry out measurements in the field. First walking behind a small vehicle fitted with cameras, later for kilometres behind a tractor. Collecting image data, carrying out precision spraying trials.” The result is a device that recognises weed plants at walking pace and sprays them with centimetre precision. “That was globally innovative.”
Technology as an accelerator
In 2010, Nieuwenhuizen more or less lays the foundations for the spot sprayer. “The prototype works, but the market is not yet ready to develop it further.” Too early, too expensive, perhaps also too daunting. But more than ten years later, the first commercial spot sprayers do in fact come onto the market, helped by technological progress. “What was not possible then is possible now.”
The arrival of AI and the further development of mobile phones have made image recognition more reliable and spraying technology more refined. “The real-time camera technology and the algorithms needed for it have greatly accelerated the arrival of the spot sprayer.”
He also points to growing social and political pressure on the use of crop protection products, and to the shortage of labour available for manual weeding. “That increases the desire to develop something,” says Nieuwenhuizen. That ‘something’ is sitting on a shelf in Wageningen. “Suddenly, there is room for what we developed fifteen years earlier.”
The importance of timing
After his breakthrough, the world lies open to Nieuwenhuizen. But he stays in Wageningen, as a researcher in precision agriculture. There he continues to work on targeted crop protection: lowering dosages, deploying sensors and spraying only where a disease or pest occurs. Not by field, but by plant, by row, by spot. Thanks to the rise of sensors, this is also becoming increasingly data-driven.
A second development is timeliness: “doing something precisely at the right moment.” And that is difficult. The weather is becoming more erratic. Sometimes it is too wet for days on end to go onto the land, followed by a few days in which everything has to happen at once. “But then you need enough people, and they are not always available.”
More than technology alone
That has everything to do with the increase in farm scale. Farms are taking on more hectares, but the number of dry days is not increasing with them. So the pressure grows, while at the same time there are fewer hands available to switch quickly. Technology can help here: practical, user-friendly robots and machines that do what is needed at the right moment, as soon as nature allows it.
He knows the criticism from the field well. Fine machines are developed in factories, but on the farmyard they often prove not yet usable enough. He tries to tackle that by putting usability, rather than technology, at the centre of his research. That means sitting down with farmers and other users. How can they deploy a field robot? What knowledge do they need to use it as efficiently as possible? What maintenance does such a machine require, and what does it deliver? And just as important: “How multifunctional is the robot? Ideally, you want the camera system to recognise weeds among carrots, but also among onions or potatoes.”
Multipurpose
According to Nieuwenhuizen, multipurpose vehicles are the future. A machine that can only harvest or weed remains, however advanced it may be, a magician with just one trick. For practical use, a robot that can carry different implements is far more interesting: a machine that can sow, weed, hoe and measure. “That requires the intelligence to sit not only in the vehicle, but also in the implements.”
That ranges from small robots the size of a wheelbarrow to 250-horsepower tractors fitted with a kit that allows them to drive across the field without a driver. “They are not yet allowed to drive on the road by themselves. But that is where we are heading.”
Sensors for drivers
Manufacturers still mainly sell implements for use behind tractors. These can often still do without the extensive sensor systems needed for autonomous work. At the same time, that technology is also becoming increasingly important for drivers themselves. “Because of the shortage of experienced workers, you more often see drivers with less knowledge of machines, crops and soil,” says Nieuwenhuizen. “That is another reason why sensors and feedback systems are becoming more important. Not only for driverless robots, but also to support drivers. To provide them with information they themselves overlook.”
Listening, looking, feeling
Robots are still mainly designed for use behind a tractor with a driver, Nieuwenhuizen observes. The driver presses buttons, keeps an eye on whether everything is running properly, notices from vibrations whether something has play in it, and hears when a machine starts to falter. He looks, listens and feels.
In principle, a robot can take over all those human observations, he says. But that still requires a great deal of research. With cameras at the front and rear of the machine. With sensors that record what a person would normally detect with their hands and the rest of their body. With systems that continuously check whether a machine is functioning properly. “If you are hoeing: fifty carrot plants with weeds at the front. And then at the back you want fifty carrot plants without weeds. Not forty, because then you have damage.”
3D simulations
Together with his team, he develops algorithms to analyse all the research data collected. Repeating tests in this field often takes a great deal of time, Nieuwenhuizen explains. After all, you can only test a crop once a year at the same growth stage. If something goes wrong, or indeed goes well, you can only go back into the field a year later to repeat the same trial. And then again the year after that.
Because they do not want to wait a full year every time, they also carry out those tests in 3D simulations. A kind of game environment in which you can place cameras in different positions to observe a crop. From above, from the side. Can I count plants? Can I see damage? It saves time, but in the end the field decides. Real mud remains in charge.
Real plants rather than pixels
As soon as spring bursts into life and crops begin to grow and flower, they all head outside together. Taking the robot into the field, or an agricultural implement behind an autonomous tractor. Cameras mounted front and rear to film and photograph. And sometimes they deliberately set the hoeing machine incorrectly, just to see what happens. Nieuwenhuizen says: “You can stay behind your computer, but nothing grows there.”
His colleagues enjoy it too: carrying out field measurements outdoors, the wind on your face, the smell of the soil. Among real plants, instead of pixels on a screen.
Agroecology
Technology is not only about efficiency or solving labour shortages, but also about agroecology. About restoring and preserving soil, nature and biodiversity. “When, as a farmer, you want to make your business operations more sustainable, that initially costs extra labour,” Nieuwenhuizen explains. You work in a more fine-grained and selective way, go into the field more often to check what is happening where, and weed mechanically more often. That requires more time and attention per hectare.
But here too technology can help. “A robot not only works faster and more efficiently, but also more specifically than an ordinary machine. Where a traditional hoeing machine behind a tractor removes everything, you can instruct a robot to leave weeds standing that an arable bird likes to eat.” This is not gadget talk, but a different way of working: more precise, with less damage and with greater awareness of everything happening in a field. “That is how you work towards future-proof agriculture that is more efficient, more precise and less dependent on manual labour, while at the same time promoting biodiversity.”
Every farmer a precision farmer
He is certain of it: before long, every farmer will work and think in terms of precision agriculture. Nieuwenhuizen makes no distinction here between organic and conventional farming. “Precision agriculture means doing the right thing, precisely and purposefully, in the right place and at the right time.” Whether someone farms organically or not, every farmer knows the size of their fields and how much manure is needed where. He measures, calculates and doses. “Every farmer will become a precision farmer. Not because they have to, but because it makes sense.”
Social pressure is increasing and the room for manoeuvre is shrinking. That is precisely why technology should not be seen as a luxury or a toy, but as a tool to make agriculture workable, productive and cleaner.
Celebrating
No, as a boy he had no idea that he would one day develop agricultural robots. He worked on the land because he enjoyed it, and that is still the case. “I want to mean something to the world, to society. I want users to be pleased with what we all develop here together.”
That feeling only grows stronger the longer he works at WUR. Because of his colleagues, but also because of the results. It lies in concrete moments: an implement that does what it is supposed to do in a field test. A project that is completed. A machine that does not end up on a shelf, but in a field. “Those are moments you have to celebrate.”
