Green Harvesting in Vineyard: A Deep Learning Empowered Methodology Under Low Budget Constraints

Green harvesting, also known as crop thinning, is a crucial practice in viticulture that aims to remove selected unripe grape clusters so that the remaining harvest can achieve better ripeness with less competition for resources. To help the automation of this process, we propose a Computer Visionassisted methodology for automatically positioning and counting growing grape clusters on either side of a vineyard row. Having a focus on small-medium enterprises (SMEs) winemakers, we developed our method to take into account the constraint of a low-budget output.