|
VDL CropTeq Robotics
© VDL Groep 2026
At the moment, the focus is on cutting leaves from cucumber plants in high-wire cultivation. It is clear that applying this solution to other high-wire crops, such as tomatoes, will follow. We have also started further development for harvesting. Using the same CropTeq robot platform, we are adding new functionality for harvesting. Specifically, this means an additional harvesting end effector and the logistical facilities for removing the produce. The leaf-cutting robot can therefore be upgraded to a leaf-cutting & harvesting robot.
Cuts with predictable reliability, you know what you’re getting.
Cuts with consistent quality, always neatly along the trunk.
Requires no training or instruction; your entire fleet of robots is equally capable.
Enjoys doing your low-value work, has no headache, or needs accommodation.
Cuts at a constant speed, with a sharp hourly rate.
You gain more control over potential virus spread
Over time, a single robot can do more and more, such as harvesting, sorting, etc.
Artificial intelligence (AI) will bring even more new possibilities
Unlock your data per stem; light, humidity, temperature, CO2, shape, weight, etc.
More and more innovative end effectors.
The robot, together with the end effector, has all the required degrees of freedom. So you can reach everywhere. You can even remove a leaf behind the stem.
We can remove the leaf using several methods: cutting, snipping or breaking. These are configuration choices. Cutting along the stem gives a very neat cut, a "clean cut". Snipping leaves slightly more stem than cutting.
To reduce virus pressure, we have equipped the cutting unit with a system that continuously disinfects the parts that come into contact with the plant. This is a major advantage, as you are not dependent on human behaviour.
The robot is actually a cobot. A collaborative robot that can work safely alongside people without guarding. If it unexpectedly comes into contact with, for example, a person, the system stops without causing injury. We use the maximum speeds permitted for a cobot and achieve good performance as a result. Including the careful avoidance of fruit, leaves and stems, this means we can certainly cut 1,000 leaves per hour. That is a top performance. With only a doubling of the operating hours, we are already faster than a person. So you do not need to operate 24/7, but the robot can be deployed for as long as you want.
The CropTeq robot is not an ordinary machine, but a true combination of a hardware and software platform. Without robust Artificial Intelligence and robot path planning, the robot would be unable to find or reach a leaf. But without our smart end effectors, the robot cannot cut leaves or harvest cucumbers.
Besides the robot’s primary function, many other software applications can be envisaged that would be useful in a greenhouse. This can be from a static position, but in the case of precision agriculture, something or someone will need to move the application tools through the greenhouse. CropTeq’s robots are such a platform on which third parties – other parties – can also integrate their applications.
The simplest answer to this is: Yes. VDL technology has several key features that make it a game changer. It is about more than simply replacing labour.
For a grower, the payback period and the problem you solve with it are crucial. We believe that robotics is only viable if the business case stacks up. Choosing carefully what to automate is essential.
Our platform concept is fundamentally a modular concept with interfaces. It uses standard components and functionality, with a roadmap for more to come.
Artificial Intelligence (AI) is a technology that has, in recent years, grown into a disruptive technology. Breakthroughs in a technique called deep learning are driving new levels of performance for various AI applications, including computer vision.
AI and deep learning have enabled a first wave of successful consumer applications, driven by leading companies such as Google, Uber, Tesla, Facebook and Netflix.
A second wave of AI-powered industrial applications is now within reach, along with predicting quality checks and smart automation.
Smart automation in horticulture means the following to us: using autonomous robotics and integrating vision and AI, combined with the grower’s expertise.
For software programmers, machine learning is another way of programming.
Traditionally, a software programmer, often an expert, writes the rules and conditions in their programme that translate input into a defined output.
With machine learning, it works quite differently. By providing it with a dataset containing the desired output, the machine (learning algorithm) learns for itself how to generate an output for new data.
As an engineer, you train (modify) your deep learning algorithm (development environment) with your dataset, which you then use to generate your application software.
Using the generated software, your industrial product can autonomously associate new data with the trained output.
Machine learning can be applied to almost all industry sectors, from finance to manufacturing. Predictive models help reduce costs, improve quality and enable new applications.
This technology also has its limitations, but its implementation across a broad range of industrial applications will increase significantly.
Deep Learning is a technology introduced by a Google team in 2012, which is increasingly being applied and optimised for specific applications. We naturally use the available technology to build our application.
Convolutional neural networks – AI algorithms specialised in learning from image data, used among other things to detect objects and people, process photos, and analyse medical or scientific images.
Reinforcement learning – AI algorithms specialised in learning to take optimal sequences of actions. Usually used for process optimisation tasks and (robotic) automation.
Network segmentation to provide even better control of our robot movements using point clouds.
A robot’s capabilities are often determined by the performance of the end effector. Creativity and smart mechatronic constructions are crucial in an environment with living products.