Answers are grounded in this paper and cite their sources.
Labour demands and the need for more precise use of agricultural inputs have encouraged the development of robots for repeated field operations. In row crops, their practical value depends on reaching the assigned row, travelling within the crop corridor and placing the implement at the intended location. A global task reference specifies the route and assigned row; global localisation relates the robot to that reference, while visual guidance supplies current crop geometry for local alignment. This narrative review organises evidence around assigned-row task execution, linking global localisation and visual crop-relative guidance to vehicle motion and implement placement. Classical image processing, learned row representations and geometric projection are discussed in terms of their outputs, operating requirements and limitations. Colour and planting regularity can support explicit row extraction, whereas learned representations accommodate more varied appearance with performance dependent on training coverage and deployment conditions. Global localisation and visual crop-relative guidance have different availability constraints during canopy obstruction, crop gaps and row-end transitions. Field studies illustrate these trade-offs and the importance of matching local alignment to the assigned row. Spraying and weeding connect vehicle motion to implement geometry and response delay. Research priorities include confidence assessment, transfer across crop conditions, recovery during reference degradation and evaluation from perception to application. The synthesis provides a basis for selecting and comparing guidance methods according to the crop, vehicle and action-placement requirements.
No takes yet. Share an insight, caveat, or question.
Han et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: