Intelligente Unkrautregulierung im Beet

With the support and guidance of all project partners, two systems are being developed in parallel that can be scaled to fit both mounting frames and field robots:

1. Researchers at the Leibniz Institute for Agricultural Engineering and Bioeconomy (ATB) in Potsdam and the Technical University of Berlin are jointly developing a laser portal whose actuator is based on parallel robot kinematics and is equipped with a blue 5W diode laser (445 nm). An image-recognition-based control system positions the diode laser precisely and dynamically over the weed plant. Using a stop-and-go motion, the laser is precisely aligned with the apical meristem—the growth center—of the weed plant. To achieve this, an additional camera was mounted directly on the actuator to detect the apical meristem using image recognition and artificial intelligence and to position the actuator with high precision. This ensures the laser’s maximum effectiveness, shortening the required irradiation time and reducing the necessary laser power.
    
2. The automated mechanical weed control modules developed in JaetRobi are based on an existing system from nAIture that uses optical sensors to scan carrot and onion fields, detect weeds, and then remove them with narrow rotating tools. HYDRIVE, in collaboration with researchers from the Technical University of Dresden, has developed a hoeing module for mechanical weed control. A magazine holding 15-mm-wide, low-cost tools allows multiple tools to operate in parallel. The goal was to enable a travel speed of up to 1 m/s.