This spring, a small army of Cornell University-developed PhytoPatholoBots (PPB) will be deployed to four grape breeding programs across the U.S. on a mission to guide the global grape and wine industry into the 21st century.

These autonomous robots will roll through vineyards, using computer vision to gather data on the physiological state of each grapevine. By combining this data with a decade of grape breeding breakthroughs, Cornell researchers are refining the PPB to allow breeders and growers to evaluate their vineyards — leaf by leaf, in real time, down to the chemical level.

The PPB rollout is happening in the first year of a new four-year project at Cornell funded through a nationwide $10 million grant from the National Institute of Food and Agriculture, Specialty Crops Research Initiative (NIFA-SCRI), and led by the University of Minnesota. The grant extends NIFA-SCRI’s previously funded VitisGen1 and 2 projects, a decade-long collaboration whose national team of Cornell-led scientists discovered many of the genes that control important traits in grapevines, such as disease resistance, insect resistance, and fruit and wine quality. Armed with these valuable new genetic resources, grape breeders across the country have been able to stock their pipelines in record time with new varieties combining high quality and high disease resistance.

The new Cornell project focuses on bringing VitisGen’s genetic and technological innovations into the vineyard by combining plant pathology, computer vision, AI, and robotics. This work is crucial for encouraging growers to embark on widespread plantings of new disease-resistant grape varieties made possible by VitisGen. Nearly all grape varieties grown today are highly susceptible to powdery mildew and downy mildew — which, for the last 140 years, growers worldwide have managed using multiple applications of chemical fungicides.

“Adoption of these new varieties alone has the potential to reduce pesticide use by 90%,” said Lance Cadle-Davidson, co-project director and research plant pathologist at the USDA-ARS Grape Genetics Research Unit at Cornell AgriTech. “Now that breeders have introduced natural disease resistance into soon-to-be commercialized varieties, growers need updated guidance.”

To develop this guidance holistically, Cadle-Davidson has tapped Katie Gold, co-principal investigator, and assistant professor of grape pathology at Cornell AgriTech, and applied roboticist Yu Jiang, assistant professor, Horticulture Section in the School of Integrative Plant Science. The team also includes longtime VitisGen leader Bruce Reisch, professor of plant breeding and genetics in the School of Integrative Plant Science, Horticulture Section, and collaborator Qi Sun, co-director of the Cornell Bioinformatics Facility in Ithaca.

Gold, who specializes in using imaging spectroscopy for disease detection, will conduct field trials to design low-input disease management programs for new varieties in VitisGen’s pipeline. Spectroscopy measures how matter interacts with light and other electromagnetic radiation, with different types of matter producing different spectral signatures. Traditional cameras measure spectral signatures within the visible (RGB) light spectrum. But imaging spectroscopy — originally pioneered by NASA to study the solar system — produces data covering a range of electromagnetic radiation seven times larger than what the human eye can see.

The PPB took shape during VitisGen2, after Jiang had helped Cadle-Davidson speed the process of evaluating thousands of seedlings in the lab by developing AI-based models to detect and quantify powdery and downy mildew on grapes using digital imaging. Jiang used those models to drive Cadle-Davidson’s automated phenotyping RGB microscopy robot, named BlackBird. With BlackBird, Cadle-Davidson’s lab saw a 60-fold increase in the number of seedlings they could evaluate and found that AI also quantified disease more accurately than humans.

For this project, Jiang is expanding the PPB’s scope with imaging spectrometers (also known as hyperspectral sensors) to scale those VitisGen lab investigations to the vineyard and enable breeders to phenotype grapevines in their natural environment. Gold also will use the newly equipped hyperspectral PPB, or HyperPPB, to ‘see’ plants at the chemical level. She’ll apply this data in the field to detect disease before visible symptoms appear, and in the lab to begin characterizing the mechanisms behind disease resistance — which is ultimately determined by the complex interplay between a plant’s genetic makeup and its environment.

“Much of the true variation in foliage is captured in wavelengths we can't see, which primarily correspond to chemistry and physiology,” Gold said. “Hyperspectral has truly shown its ability in detecting and differentiating the multifaceted, less cut-and-dry aspects of disease resistance and infection dynamics.”

Jiang hopes to commercialize the PPB robot family so growers can monitor disease as well as many other aspects of vine development in their vineyards on a larger-than-ever scale. The collaboration is a prime example of how transdisciplinary research can accelerate the design process for next generation agri-food systems, he said.

“We’re helping breeders target fundamental improvements while enabling them to respond more rapidly to changes, whether predicted or not,” Jiang said.

Many of those unpredictable changes will be due to climate change, which Gold said is only going to increase disease and pest pressure in New York state, where it’s already higher than in other major U.S. grape growing regions. For the state’s $15 billion industry to continue to thrive, growers will have to adopt new disease-resistant varieties and precision management methods.

“Over the course of VitisGen, more than 65 co-investigators have worked together at the cutting edge of technology,” said Cadle-Davidson. “But what I once thought was cutting edge is nothing compared to what Katie and Yu are bringing to the table. What we can accomplish now is going to be so exciting, powerful, and probably revolutionary.”

Source