Image-based modelling of the effect of s-metolachlor plus atrazine on the soaking kinetics of maize seeds

Pre-emergent herbicides can have negative effects on maize seeds. The objective of this study was to model seed soaking curves through the processing of red-green-blue imagery of maize seeds under the influence of concentrations of s-metol-achlor + atrazine on both the soaking kinetics and primary root emission. Seeds were placed to soak for 114 h in Petri dishes con-taining aqueous solutions of a herbicide containing s-metolachlor (290 g l-1) + atrazine (370 g l-1) with the following concentrations: 0% (water only), 2, 5,10, 20 and 50%, based on the recommended dose (4.0 l of the commercial product per hectare). The images were systematically taken from a flatbed scanner with artificial light control. The red excess index was adapted to improve image segmentation. From the binary masks applied, the soaking curves for each herbicide concentration were obtained using estimates of seed intumescence over time. The soaking curves were described by fitting Peleg's model. The herbicide concentration has signifi-cant effects on both the absorption rate and primary root emission; the absorption rate was reduced by 50%. A concentration of s-metolachlor (290 gl-1) + atrazine (370 gl-1) in aqueous solution that is above 20% can fully inhibit seed germination.