Researchers in Geneva have built an imaging system that can identify signs of grapevine mildew days before damage becomes visible, a step they say could speed breeding for disease resistance and help growers target treatments sooner.
The research system, called HyperBird, analyzes tiny leaf samples under a hyperspectral microscope. Unlike an ordinary camera's three visible-light color bands, it records hundreds of narrow bands that can reveal changes inside a leaf before symptoms appear on the surface, according to Cornell.
What the imaging system can detect
The researchers at Cornell AgriTech are studying powdery and downy mildew, diseases that Cornell says cause billions of dollars in vineyard losses worldwide each year. HyperBird can process hundreds of leaf samples in a few hours, allowing scientists to compare diseased spots with healthier areas of the same leaf and assess traits relevant to resistance breeding.
Katie Gold, an assistant professor of grape pathology at Cornell AgriTech and a senior scientist on the project, said the system now provides detection accuracy on day three comparable to what earlier methods achieved on day six or nine. She said throughput has increased two- to threefold. Those are research-performance claims, not a guarantee that a vineyard can eliminate disease or fungicide use.
HyperBird extends an earlier Cornell and U.S. Department of Agriculture prototype called Blackbird. That 2021 system used red, blue and green imaging to help examine thousands of grape leaves for infection, removing a bottleneck in research on powdery-mildew-resistant varieties. HyperBird retains microscope-level spatial detail while supplying about 200 times more spectral resolution per pixel, Cornell said.
Yu Jiang, a Cornell AgriTech systems engineer and another senior scientist, said resolution matters because a tiny diseased area can be difficult to distinguish from surrounding healthy tissue. The more detailed spectral reading helps separate those signals and follow how an infection progresses.
Three studies test different uses
Three papers trace the technology's development. In a May 20 Plant Disease study led by former Cornell postdoctoral researcher Saeed Hosseinzadeh, the team collected nine terabytes of hyperspectral images in one day as proof of principle for detecting infection before visible symptoms. The system also identified fungicide residues and optical characteristics tied to grape lineages being bred for resistance.
A Sept. 8 Plant Phenomics paper led by doctoral student Jinhong Yu describes the high-throughput microscopic system and its ability to compare small diseased areas with the rest of a leaf. A June 29 Plant Disease study led by former visiting scientist Lorenzo Pippi tested leaf samples taken from field vines treated with conventional fungicides, biofungicides or no spray. The imaging distinguished early optical changes associated with downy mildew and residue patterns from the treatments, helping researchers assess their performance despite uneven infections across a vineyard.
The work does not yet remove the need to collect physical samples. Technicians must cut dime-sized pieces of leaf for the machine, which Cornell identified as the most time-consuming part of the process. Robotics students are working to automate that step.
The researchers are working with Moblanc Robotics on manufacturing HyperBird. Cornell said the approach could eventually be used for other plants and diseases, but did not announce a commercial release date or a grower deployment plan. The team has also begun discussions with a company in the cheese industry about another possible application.
HyperBird is part of the federally supported VitisGen3 grape-breeding project. The research also received support from Cornell federal capacity funds, the President's Council of Cornell Women Research fund, the New York Wine and Grape Foundation and the USDA Grape Genetics Research Unit.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to news@fingerlakes1.com.