Vegetable production imposes a wide variety of farming operations and among these operations, an early weeding is necessary to avoid competition between weed and crop. If the competition is too strong, the farmer can have huge yield loss.
The classical solution is the chemical weeding. However, consumers’ demand is for food quality with less phytosanitary residues now. To answer this demand, the farmers must change their cultural practices and mechanical weeding can be a part of this change. In practice, this solution is hard to implement, especially when the weeding occurs within the rank with dense crop (like carrot or maize) because this requires a data analysis in real time with a high accurate crop or weed recognition and a very precise positioning to extract the weed without damaging the crop. To date, there is no solutions that concern intra-row weeding with dense crop. To create a solution available to a wide range of culture and planting parameters,
Bordeaux-INP, in collaboration with CEA and STGNB, develops an autonomous weeding of vegetables system embedded on an electrical tractor, based on AI vision adaptable to various cultural methods.