Machine learning models as an alternative to determine productivity losses caused by weeds.

The best performing ANNs can indicate the beginning of weed control since they can accurately estimate losses due to competition. These results encourage future studies implementing ANNs based on computer vision to extract information about the weed community.