drone-sensing
Drone-sensing uses unmanned aircraft systems equipped with cameras and sensors to capture aerial imagery of crops and fields without the need for a human pilot on board. This technology allows researchers to monitor plant growth, detect weed infestations, and assess crop health across large areas quickly and with high spatial precision. By enabling efficient, non-invasive data collection, drone-sensing supports more targeted agricultural management and deepens understanding of plant-environment interactions at scale.
open_in_new WikipediaOpenAlex · 2026-08-01
Researchers in Utah built a drone-and-AI system that pinpoints weeds in commercial corn fields, enabling targeted herbicide application only where weeds actually grow. The lightest AI model tested ran accurately enough for real-time use on drone-mounted hardware, and the dataset they built is the first public corn weed image library for the Intermountain West.
YOLOv9s outperformed 27 other AI models at detecting weeds in drone imagery, achieving the best balance of accuracy and processing speed for real-time onboard use
USU-CornWeedDB, containing 800 labeled images across three weed species plus 8,000 unlabeled images, is the first public corn weed dataset for the Intermountain West region
Simpler semi-supervised learning methods outperformed complex ones under real field conditions, reducing the manual labeling burden without sacrificing reliability
