By Andi Anderson
Farmers often struggle to identify insects on their crops. InsectNet, a new AI-powered tool, provides real-time insect identification, helping farmers determine if an insect is a pest or a beneficial species.
Developed by researchers at Iowa State University, InsectNet uses a database of 12 million images to identify over 2,500 insect species with 96% accuracy.
Farmers can upload photos of insects using a smartphone or computer, and the system predicts their classification, whether they are pests, pollinators, predators, or invasive species.
“InsectNet provides accurate predictions and helps farmers make informed decisions,” said Baskar Ganapathysubramanian, a lead researcher. The tool can be fine-tuned for different regions using local datasets, making it useful for farmers worldwide.
Beyond agriculture, InsectNet can help border agents detect invasive species and assist researchers studying ecosystems. The app works across different life stages of insects, from eggs to adults, and handles various image qualities and species look-alikes.
Currently, InsectNet is not available for download but can be accessed via a web-based platform hosted at Iowa State University. The researchers demonstrated the tool at the Farm Progress Show, gaining interest from the agricultural community.
Unlike existing insect identification apps, InsectNet offers a large-scale, open-source solution. Its advanced AI capabilities allow for global and regional insect classification, setting a foundation for future agricultural tools, including weed and plant disease detection.
Photo Credit: gettyimages-v_zaitsev
Categories: Iowa, Rural Lifestyle, Farm Safety