“We use robotics, image processing and artificial intelligence”

For more than two years, researchers at the ETH (Eidgenössische Technische Hochschule or Federal Institute of Technology) in Zurich have been working on the automation of tomato harvesting. The goal of the project is to develop, based on robotics, image processing and artificial intelligence (AI), a modern picking robot that can be used for other products in the future. At a seminar earlier this year, researcher Till Karbacher gave an overview of current advances and results achieved so far.


In the tomato greenhouse, there are, among other things, two major workstations: picking and packaging, the fruit then being ready to be marketed, and the stripping of the leaves from the plants. “On average, farming one hectare requires a total of 2,815 hours of work per year. A robot is operational 4,246 hours a year, and works at night if necessary. By extrapolation, a picking robot alone can provide the work needed for 1.5 hectares, which would allow it to replace 3.5 employees. »

This innovation is scientifically based, describes Karbacher. Behind the technology are two years of research at ETH that have allowed the robot to learn how to detect thin objects adequately. This intensive research was largely funded by Innosuisse R&D. “The research work is undoubtedly bearing fruit: thanks to image processing and AI, the picking robot already manages to harvest the tomatoes by the peduncle, not letting the fruit be crushed by the gripper. In addition, we have also succeeded in developing technology for the various tasks in the greenhouse. The next step would be to clip the rod onto the support rope. »


AI and advanced control system
For a full-fledged installation in the greenhouse, an extension of the prototype for Trolley is planned, continues Karbacher. For this, some optimizations would be necessary.robotersaries, such as the arm and the tool, the cameras and the swivel system as well as the computer and the software. According to Karbacher, the packaging system also needs some adaptations to ensure optimal performance of the robot.

Right: Overview of the picking robot workflow

The researcher also draws attention to the key role of AI and strong datasets, such as RGB color photos. “The highly optimized and integrated technology enables performance that is measured in milliseconds. The advanced control system results in particular in eligibility checks and a robust system. »

Suspended harvester arm for arboriculture and viticulture
In addition to the tomato picker, ETH researchers have also developed a harvesting robot for viticulture and arboriculture. Thanks to a suspended arm, the process is able to harvest grapes and fruit almost autonomously, says Karbacher. “It’s a whole new way to get around. The manipulator arm is suspended using four posts and has access to the orchard from above. The process allows CO neutrality2, thanks to the solar drive, and can operate on slopes of up to 45°. It is therefore optimal for hilly orchards.

The fruit picker with hanging arm

Pictures: ETH Zurich – Robotic Systems Lab (RSL)

For more information :eth
Till Karbacher
ETH Zurich – Robotic Systems Lab (RSL)
Such. : +41 446338493

Publication date:

We want to thank the author of this post for this awesome web content

“We use robotics, image processing and artificial intelligence”

Our social media profiles here and additional related pages here.https://yaroos.com/related-pages/