Accurate yield estimation is a critical aspect of both viticulture and winemaking as it affects the entire supply chain. Industry standard practice in yield estimation can take substantial resources yet still provide inaccurate results. In order to automate yield estimation methods we developed a method for generating high-resolution relative maps of visible vine parameters such as shoot or bunch density. The method is based on low-cost cameras (i.e. GoPros) as sensors that can be mounted on vehicles that are already traveling through the vineyard. Camera data is processed to give a geo-referenced map without requiring an expensive GPS unit. The relative yield maps can be generated from when the first leaves separate this means that the maps can be used to adjust management practices such as trimming, thinning or mulching during the season. The processed images can also be used to map non-bearing sections of canopy and to identify missing vines, which has the potential to help in the detection of trunk diseases such as Eutypa. These methods along with recent work on detecting water stress and sensing bunch maturity in vineyards will be presented along with discussion of some of the challenges facing viticulturists.
Dr. Mark Whitty is a Lecturer in the School of Mechanical and Manufacturing Engineering at UNSW Australia and the leader of the Smart Robotic Viticulture (SRV) group. His research interests include precision agriculture, mobile robots, 3D mapping, indoor localisation and image processing. These have been applied in field robotics including current work on yield estimation in viticulture. He is the Chief Investigator for a Wine Australia project titled "Improved Yield Estimation for the Australian Wine Industry" and is engaged in projects for water stress detection and maturity estimation using mobile and lightweight sensors. For more information on Dr. Whitty's work, please visit http://www.robotics.unsw.edu.au/srv/