From the Sky, NASA Detects Early Signs of Grape Disease

To validate their findings, the researchers compared the camera results with ground observations. Surprisingly, they discovered that healthy grapes from 2020, which showed signs of the virus the following year, were actually infected all along. Through machine learning, the system was trained to differentiate between healthy and diseased vines. The computer-trained camera achieved an 87% success rate in detecting hidden viruses and accurately identified 85% of visibly sick vines. This breakthrough means that previously hidden changes in grapevines can now be detected, potentially paving the way for future satellite surveillance of large vineyards.

Study senior author Katie Gold from Cornell University explains that early detection is the biggest challenge facing growers, as sick plants may not exhibit immediate symptoms. Fernando Galvan, a co-author of another study focusing on machine learning and virus detection, emphasizes the exciting opportunities presented by remote sensing and plant disease detection. He believes scalable solutions can empower growers to make data-driven, sustainable decisions for crop management. Furthermore, these advancements could lead to more accessible and affordable wine for consumers. (Read more scientific study stories.)