On May 13th, the project team AI4Leafhopper presented the new application iCountPests, an innovative solution that uses Artificial Intelligence (AI) to detect and count green leafhoppers in chromotropic traps - quickly, accurately and in real time.
Developed as an intuitive mobile application, the iCountPests was designed to facilitate the monitoring of various agricultural pests. With a simple and accessible interface, it allows users to record the evolution of pests in their crops by submitting photographs of the traps installed in the field.
The application uses advanced computer vision models to automatically detect and count insects, delivering results in around a minute - a much faster and more practical process than traditional manual counting.
In its first version, the app already has a model for detecting the green leafhopper (Jacobiasca lybica), achieving an average accuracy of approximately 90%. Functionalities will soon be added to identify other relevant pests, such as curl moth (Cryptoblabes gnidiella) and grape moth (Lobesia botrana).
In addition to automatic counting, the iCountPests makes it possible to monitor the evolution of pest populations over time, making it easier to identify trends and plan more effective interventions.
This project is the result of a multidisciplinary team combining skills in ecology, entomology, artificial intelligence, computer vision, remote sensing and software development, with the aim of making pest monitoring simpler, more accurate and more accessible.
During the presentation session, it was possible to hear the opinions and suggestions of future users of the application. These contributions are essential if we are to continue to improve the tool and ensure that it responds in a practical way to the real needs of farmers and technicians in the sector. We want to develop solutions that evolve with agriculture!
Innovation is at the heart of everything we do and our motto is clear:
“Innovate together, protect better.”
If you want to know more about iCountPests, contact us by email:
📩 apps@iplantprotect.pt
Image credits: InnovPlantProtect
Português
English