Robotics for weed control – a classification of the potential and the market

Abstract

Agriculture is undergoing a significant change, with robots playing an increasingly important role. Among these robots, weeding robots are particularly helpful in automating the weeding process on fields and have recently been introduced to the market. These robots operate using a GNSS (Global navigation satellite system) or vision-based approach, which is explained and evaluated in this context. In contrast to the relatively easy method of storing the crop position in the GNSS-based process, vision-based approaches require a sophisticated image analysis, usually based on deep learning routines and extensive sets of training data. This concept has been successfully applied to two market-ready robots, and a market overview following the introduced classification is also provided. The paper concludes with a discussion of the challenges involved in using robots in the field and how these robots can support existing agriculture workflows and interaction models between remote sensing and field robots.