The analysis of both sampling strategies (a), (b), demonstrated that there are
significant differences in the resulting weed maps (Table 1, Table 2). The sparse
sampling design overestimates the occurrence of weeds compared with the dense
sampling grid (Fig. 4).