Collaborative project GoodWeedBadWeed - Damage threshold 4.0: Artificial intelligence for the treatment of weeds taking into account the ecological value
'GoodWeedBadWeed' aims to accurately assess 40 weed species in an agricultural context. Implementation involves various parameters such as plant protection, agroecological aspects, and socioeconomic factors. The main objective is to develop control thresholds and evaluation indices specifically designed for modern agricultural practices that differentiate weeds according to their ecological value and damage potential. These newly developed standards are intended to improve agricultural land management sustainably, optimize pesticide use, and protect biodiversity. The outcomes of this research will be publicly accessible with the intention of fostering greater efficiency and environmental sustainability in agriculture. Low-cost drones with RGB cameras will be used for data collection, while state-of-the-art AI models will be developed for precise weed detection and mapping. In conjunction with the developed indices and automatic deviation of damage thresholds, these technologies will enable precise weed control while saving pesticides. The project holds great potential for science and industry offering a technology to harmonize the competing objectives of preserving agricultural productivity while fostering agrobiodiversity.
Funding reference: 281D205A22
Project duration: 15.10.2024 - 14.05.2028
Project team:
- Rebecka Dücker (Principal investigator, Georg-August Universität Göttingen)
- Friedrich Bartels (PhD student, Georg-August Universität Göttingen)
- Corinne Jampou (external PhD student, Leibniz-Institut zur Analyse des Biodiversitätswandels)
- Tim Baroth (technical assistant)
Project partners:
Funding / Project management agency:
