- PROJECT CODE: BI-HU/26-27-007
- PROJECT TITLE: Pest insect control in the forestry and timber sector using non-traditional grid-based alarm systems
- PROJECT TEAM: Miklos Kresz (leader), PhD; Jakub Michal Sandak, PhD; Niki Hrovatin, PhD; Edit Foldvari-Nagy; Balazs David, PhD
- PERIOD: 1. 1. 2026 – 31. 12. 2027
- BUDGET: 7.000,00 EUR
- FINANCING: Slovenian Research and Innovation Agency (ARIS)
- PROJECT COORDINATOR: University of Primorska, Andrej Marušič Institute (Slovenia)
- PARTNERS: Eszterházy Károly Catholic University (Hungary)
The aim of this project is to develop an intelligent, automated system to predict insect invasions, thereby protecting plants in the forestry and timber sectors. Pest insects are a significant threat, as even a few can devastate an entire crop. Climate change and global trade have led to an increase in invasive insect species, which spread rapidly and cause substantial damage. This project seeks to enable timely interventions against these pests, preventing major economic losses. The system will work by collecting data in forestry environments using insect traps and cameras. An AI-based program will analyze the images to identify pest insects, predict potential invasions, and alert farmers so they can take timely action. This approach not only saves crops but also reduces pesticide use, which benefits the environment. A unique aspect of the project is the active involvement of young researchers and PhD students. The team is also exploring creative ideas, such as using hexagonal patterns like honeycombs in image analysis, which helps the AI system detect even the smallest insects more accurately. The research results can benefit not only the forestry and timber sectors but also other fields like medical imaging and environmental protection. In the long term, the project aims to protect nature, support farmers, provide international experience for young researchers, and help maintain economic competitiveness. The project is pioneering the research and practical testing of non-conventional grids (e.g. triangular, hexagonal, irregular topology) to increase the efficiency of convolutional neural networks. Such lattice structures offer the potential to achieve better performance than conventional square-gridbased image processing in detecting small objects, such as insect pests.
