A virtual tomato training arena for harvesting robots
Peer-Reviewed Publication
Updates every hour. Last Updated: 17-Jun-2026 04:16 ET (17-Jun-2026 08:16 GMT/UTC)
Researchers have developed a new way to generate realistic synthetic training data for agricultural robots by recreating tomato cultivation environments in a virtual world. The approach could help overcome one of the biggest barriers in agricultural robotics: the time and labor required to collect and manually label real-world data under challenging greenhouse conditions such as changing lighting, dense foliage, and fruit occlusion.
The use of digital tools and better coordination between different organisations can help the UK significantly optimise its first line of defence against ecological degradation, new research shows.
Virtual reality is entering the agricultural science classroom, and University of Tennessee Institute of Agriculture researchers are helping teachers incorporate and create their own virtual reality resources.
With a $500,000 grant from USDA’s National Institute of Food and Agriculture, a faculty team in the Department of Agricultural Leadership, Education and Communications is building on their findings from a previously funded USDA grant project aimed to further develop efforts to train secondary school teachers and bring virtual reality (VR) to the classroom through the Agriscience Metaverse Academy.