Omsk Scientists Develop Neural Network to Protect Crops From Diseases and Pests
The computer-vision algorithm will analyze aerial images collected by drones.
Researchers at Omsk State Agrarian University are developing a new neural-network system designed to identify plant diseases using images captured by unmanned aerial vehicles.
The project is currently in the data-collection phase, during which specialists are assembling the materials needed to train the computer-vision model. The dataset includes both drone imagery and photographs captured directly on the ground. All images are uploaded into a centralized storage system and then labeled in vector format with detailed attribute information, including weed species, pathology type, plant growth stage, and other parameters.
The end result will be a neural network capable of detecting plant diseases, weeds, and agricultural pests across farmland. Researchers say the AI system could help agricultural producers take preventive action earlier and reduce crop losses.
