Last week, Nederlandse Vereniging Techniek in de Landbouw (Dutch association for technology in agriculture, NVTL) organised their annual Studiedag, a day filled with interesting tours and talks by people at the forefront of agricultural technology. This was an opportunity not to be missed. There were four themes – greenhouses, open field farming, livestock production, and AI/data/blockchain. I decided to spend the day looking at greenhouse technology. Here’s what I saw.
Tours in the morning
Gerrit Polder (Wageningen University and Research)
We were shown around the hyperspectral lab. There were all sorts of cameras as well as a spectrometer.
It was interesting to see the kinds of images that could be made by taking an image for every wavelength. Every pixel had its own unique spectrum. With this, they could segment the plant into stem, leaves, and fruit. Hyperspectral imaging can also help detect disease or estimate the lycopene content in tomatoes, using big data and AI.
Jochen Hemming (WUR)
Jochen showed us some of the robots being developed at Wageningen University: the Sweeper and the Trimbot. The Sweeper is a harvesting machine for sweet peppers. One of the interesting parts are its end-effectors, in other words what it uses to grab the fruit. It grabs the fruit with fin-ray grippers, a very clever solution. The Trimbot trims hedges.
Lydia Meesters (WUR)
Lydia showed us around the post-harvest technology centre, where they are working on non-invasive ways to assess the quality of crops in order to segment them. With technologies such as spectral imaging, they could determine the hardness and ripeness of fruits. They could even make a digital 3D reconstruction of a plant, more useful perhaps for harvesting. One thing that’s still difficult to measure is the Brix, which relates to the sugar content of fruits.
Talks in the afternoon
Luca Locatelli
You may have read National Geographic’s article on the Netherlands, the ‘tiny country that feeds the world’. There’s some breathtaking photography in there. Luca is the man behind it all. A photographer with no particular knowledge in agriculture, he shared some of his observations with us.
What most impressed him were the high-tech (to him futuristic) greenhouses in the Westland. He also learned that technology is often portrayed as the enemy by the media. Dramatisation wins, yet we don’t have a practical alternative to tech. The familiar childhood image of a farmer is lovely, but it’s a luxury and it isn’t sustainable. It was interesting to hear this from a photographer, because this is the conclusion of most biosystems engineers as well.
Peter Hendriks (LetsGrow)
LetsGrow’s business model revolves around collecting data from growers, adding value through analytics, and then using this for consulting and decision support. It’s a data sharing platform, not only for growers, but also for suppliers and researchers, who can help improve the technology being used in greenhouses. To enable the Internet of Things (IoT), wireless sensors have been crucial. They allow for far more data points in a greenhouse.
Some of the next steps Peter mentioned included allowing growers to monitor their crops from a control room. Another interesting idea is to act on predictions of the future – for example, if certain weather is predicted, that the grower can already optimally prepare. Rather than figuring out what is happening, the next step is to figure out the best that could happen and acting on it.
Andreas Hofland (HortiKey)
How do we make greenhouses more sustainable? More data. HortiKey has created the CHIMP – the crop health information monitoring platform. This involves an autonomous vehicle that goes around the greenhouse and takes measurements. As more and more measurements get made, the grower will be able to respond better and better, with tailored responses. For example, if a plant is affected by disease, a robot will be able to mix the cure (probably a concoction of pesticides) on-site, to give the plant exactly what it needs.
Raymond van den Berg (ISO Group)
ISO Group has come up with a number of machines. Their proudest creation is the Cutting Planter, which plants cuttings into their growing medium. The video of the robot doing this was quite a sight. The robot also detects the quality of the cuttings used and rejects bad ones, leading to a guaranteed quality of cuttings. It’s a serious contender to manual labour. Interestingly, robots find it easier to pick up small things like cuttings. It certainly seems so, looking at the effort that goes into grabbing a sweet pepper.
Future developments include doing more with data and AI (surprise surprise), as well as making the robot do more tasks, like giving the cuttings rooting hormone.
Jochen Hemming (WUR)
In this presentation, Jochen went more into detail about the robots he showed us in the morning. Amazingly, the Sweeper’s software sees far more peppers in an image than a human being could. Seriously. Take a look and see how many you spot in the original image. Probably not eight. This is the power of neural networks. Soon he hopes to expand this to selective broccoli harvesting, the detection of disease in seed potatoes, and animal welfare (something done the with ‘classical’ computer vision by my friend Jaap).
Closing remarks
That was my experience of NVTL’s Studiedag. It was interesting and fascinating to see what was being done. Many thanks to NVTL for organising this wonderful day.