Proxi-detection to monitor the growth status of wheat in the presence of weeds using low-cost and simple digital tools to track the emergence of stress.

This study combines proximal sensing technology based on a mobile visible imaging device and a dynamic ecophysiological model to estimate the growth status of two winter wheat crops (Apache and Rubisko cultivars) in the presence of uncontrolled broadleaf weeds. Working at early growth stages and on four different dates, each plot was photographed and then destructive biomass samples were collected. After calibration, crop aboveground biomass is inferred from an image indicator, the fractional vegetation cover (FVC). A supervised image classification algorithm was used to calculate FVC for crop (FVCc) and for weeds (FVCw). Analysis of FVC measurements focused on stand heterogeneity by comparing their variability and the impact of weeds on crop growth. Over time, the discrepancy between plant biomass derived from the image and data simulated from a simple ecophysiological model increased for most areas of the plot, indicating the presence of crop stress. At the same time, the weed pressure (WP) study concluded that weeds did not have a major influence on crop growth, although locally some areas showed a negative impact on crop growth. Therefore, weeds were not the major stress observed on the later dates of the study. These low-cost technologies aim to determine stress in wheat crop and support farmers in their transition to agroecology for highly accurate plot monitoring. [ABSTRACT FROM AUTHOR]