Site-specific weed management appears as an innovative way of saving herbicides in crop while maintaining yield. This can potentially lead economic and ecological benefits. However, it was reported in the literature that savings range from 1 % to 94 % from one field to the other. This implies that certain fields present good potential for profitable application of site-specific weed management while others do not. It is thus important to identify fields with higher potential to make a profitable use of herbicide sprayer equipped with canopy sensors and variable rate application hardware. This paper presents an approach based on imagery to estimate the potential of a field, based on both level of infestation and spatial structure of the weed infestation. Using the proportion of the field that is not infested by weeds and Pielou’s index of segregation measured at 9 m along the crop rows, it was possible to build a linear model (R2 = 0.91) estimating the potential of a field on a 1 to 5 scale. Subsequent analysis demonstrated that the potential of a field could be estimated in about 5 minutes per hectare using an ATV equipped with a GPS and cameras at the front.
