A $20 million grant has helped launch 16 new research projects in areas such as crop breeding, greenhouse technology and micronutrient research.
The Artificial Intelligence Institute for Food Systems got the grant. The new institute is led by the University of California-Davis.
Funded by the USDA and the National Science Foundation, the institute is a collaboration between UC-Davis, Cornell University and the University of Illinois, Urbana-Champaign.
“AI is a growing strategic focus area for the federal government, and as a leading ag university we decided we need to be at the forefront of this initiative,” said Gabe Youtsey, chief innovation officer at the UC Agricultural and Natural Resources Division.
Current projects being explored include a focus on crop breeding, greenhouse technology and recognizing mineral and micronutrient content in food using AI and imagery. Some of these projects were already conceptualized so when AIFS put out a call for projects, it received a lot of proposals that aligned with its mission.
The institute is looking at how to quickly integrate students and technology with the industry to begin solving real world problems as soon as possible, Youtsey said.
One noteworthy project is digital twins. The concept is to create digital copies of the farm or factory, to enable informed decision-making to reduce the amount of irrigation water, pesticides or fertilizer applied to crops.
“Digital twinning is talked about a lot but often without the right use of sensors and AI,” Youtsey said. “Once all this field data is put into a digital twin, you can create more automated solutions.”
With food safety, the continuous need for food processors to detect and trace sources of contamination is critical, and the digital twin model can simulate how pathogens move through food facilities through machine learning data.
In greenhouse technology, the focus will be predicting how fast a lettuce variety will grow and what the sales numbers and profit will be at the end of the season. By extension, AI can eventually be used to predict harvest trends for fruits, nuts and vegetable crops.
For orchard crops such as stone fruits, customers don’t want bruised fruit or nail impressions, which makes hand harvesting challenging. It’s also not easy to automate, but AIFS researchers have begun by identifying the number of fruit on trees, their quality, when they need to be harvested and how many harvesters are needed.
AI can give growers and the supply chain decision-making tools for hiring labor, scheduling machines and timing to market. It can also help enable driverless technology for navigating through an orchard, with help from machine learning to distinguish between fruits, flowers and buds.
“We don’t have common data infrastructure between universities or companies to use AI to recognize fruit, so each outfit has to do this over and over and over again,” Youtsey said. “So what we are doing is creating common datasets that will be open source so everyone can use it and contribute their own datasets and imagery to the repository.”