Bildanalytische Farberkennung zur Geschwindigkeitsregulierung beim Jäteflieger im ökologischen Möhrenanbau

Summary

Weed control in organic carrot production is time and cost intensive. There are several practicable techniques available for inter-row weed control, but the most sensitive area - ted by manual weeding. Weed regulation is usually done in uniform operations for one field, but weeds appear in a huge spatial variability, reflecting heterogeneous soil and climate conditions. Following the idea of precision farming, we developed a concept for camera-based weed detection, which displays the density as recommendation for
velocity adaptation for the driver. The system was tested against the usual weeding with one constant velocity in a field trial with randomized plots. By using the new
technology it was possible to increase the area performance by higher average velocities without reducing the weeding quality compared to the usual system. The next step of this low-cost component approach would be a complete implementation into the tractor system with the possibility of automatically velocity regulations.