AI assists in mechanical hoeing
Fendt and PTx Trimble have jointly developed RowPilot, an Artificial Intelligence (AI)-based row management system for mechanical hoeing devices. AI-based systems meet proven standards such as ISOBUS to achieve high precision with maximum compatibility. The difference to conventional systems is that AI does not recognize the rows on the basis of their leaf mass, but […]
Fendt and PTx Trimble have jointly developed RowPilot, an Artificial Intelligence (AI)-based row management system for mechanical hoeing devices. AI-based systems meet proven standards such as ISOBUS to achieve high precision with maximum compatibility.
The difference to conventional systems is that AI does not recognize the rows on the basis of their leaf mass, but at the exit point of the plants at soil level. This is particularly important when working on the side slope and wind and thus leads to high robustness and precision. The drift of the hoe on the slope is additionally compensated by a position sensor.
In contrast to conventional systems, AI relies not only on the color of the crops, but also on its geometric properties. The RowPilot achieves a reliable work result even with dense weed and inhomogeneous crops.
The comprehensive training data of the AI significantly facilitate the application of the RowPilot. Drivers set the position of the camera and the row spacing on the terminal. Information on the height, width or color of the crop is therefore obsolete. This also allows a quick change between different crops. The system is fully integrated into the ISOBUS of the hoe. Functions such as individual row lift can also be carried out easily and automatically via Section Control or controlling the lateral-shift frame.
