Detection and Localizing of Rumex Seedlings for Robotic Weeding

Abstract

Weeds are causing substantial losses for livestock farming and decrease land cultivability in farming. Treating weeds with herbicides is harmful for the environment and we are targeting automatic mechanical weeding of Rumex longifolius in pastures using mobile field robot technologies. We are relying on 2D/3D computer vision technologies and have developed an algorithmic solution for detecting seedlings of Rumex weeds. We show promising test results on weed detecting and lozalizing for robotic removal.