Farmers regardless of size and scale face known challenges each season dealing with the uncertainty of seasonal variability of fruit yields. These include both operational challenges like harvest labour planning and scheduling post-farm logistics, and financial challenges like revenue forecasting and managing forward supply agreements. This is particularly the case for major commercial viticultural entities where seasonal fluctuations can have millions of dollars’ worth of consequences if not adequately accounted for ahead of time.
This project was created from the foundations of The Yield's AI prediction services, and looked to explore how in-field observations of grape vines and high-resolution remote sensing data can complement this prediction service with even greater accuracy and precision. The combination of both ground-level observations from cameras driven through the vines, with high resolution overhead imagery sourced from satellite and airborne platforms meant that we were able to gain a comprehensive understanding of canopy and fruit development over the course of the growing season. Ground observations from both cameras and LIDAR sensors monitored fruit growth features such as bunch counts/vine and canopy leaf density, whilst overhead observation assisted with monitoring vine canopy health and total leaf area.
Project outcomes highlight the utility of in-field metrics, derived via the processing of Lidar and camera data, for use in advanced analytics and A.I applications. Models developed using the data captured by the project across the Australian and Californian vineyards suggest an 8% improvement to block level predictive accuracies are possible from fruit-set onwards.
This is the third collaborative research partnership between The Yield (now Yamaha Agriculture Australia), Food Agility and University of Technology Sydney (UTS) focussing on using advanced data analytics and machine learning to enhance the accuracy and quality of harvest yield predictions in horticulture. This was also the first major initiative commencing the partnership between The Yield and Yamaha Motor Co and had an explicit focus on exploring potential synergies between Yamaha's robotics and automation capabilities with The Yield's AI and machine learning expertise.
