Mechanical Control with a Deep Learning Method for Precise Weeding on a Farm

This paper presents a mechanical control method for precise weeding based on deep learning. Deep convolutional neural network was used to identify and locate weeds. A special modular weeder designed, which can be installed the rear of mobile platform. An inverted pyramid-shaped tool equipped in shovel out weeds without being contaminated by soil. The weed detection implemented an embedded system with high-speed graphics processing unit integrated weeder. experimental results showed that even if speed platform reaches 20 cm/s, still accurately detected position located system. Moreover, mechanism successfully roots proposed has been tested field, its performance coverage have verified weeding.