Fig. 5. Interpolated weed maps from sparse sampling (a), dense sampling (b) and five shifted
starting points (shifted by 1.5 m) for sampling (S1-S5). S0 is equal to (a) with weed quantity
information from (b). The weed mapped is ‘Common Chickweed’ (Stellaria media L.).
The cross-correlation values that have been calculated for the different maps prove a
weak relation between the maps which should ideally be close to a value of 1 (Table
1, Table 2) in case the weed maps were invariant to shifting. Concerning the
‘clustered’ weeds the relation between the datasets appears to be either random or
neglectable. The ‘scattered’ weeds prove to have significantly higher levels of
correlation.
As a further outcome of the analysis, the differences between the maps (a) and S0
should be stressed. S0 is supposed to be a simulation of (a) because the information of