NemDetect: Early detection of quarantine nematodes in potatoes using remote sensing.

Plant-parasitic nematodes cause significant economic losses in crop production, with root-knot nematodes (RKN) and potato cyst nematodes (PCN) being the top two in terms of agronomic impact. Effective pest management requires accurate detection and monitoring, especially methods for early (pre-symptomatic) detection, as nematode damage can take years to become visible and lead to full field infestations if unmanaged. Remote sensing using hyper- and multispectral sensors enables pre-symptomatic detection at field scale, but it needs to reliably
distinguish between biotic stress from nematodes and abiotic stress such as drought, and account for variety heterogeneity. Nemdetect focused on four nematode species and included drought stress in experiments to ensure accuracy. The project was divided into four parts: literature review, identifying implementation opportunities, pilot case studies, and knowledge transfer. The
literature review analysed the current state-of-the-art in remote sensing for plant health assessment and provided guidelines for future research. Identifying implementation opportunities assessed the potential of remote sensing for nematode detection and monitoring. Pilot case studies tested the effectiveness of remote sensing in detecting nematode infestations and distinguishing
between biotic and abiotic stress, for above-ground parts of potato plants and tubers. Knowledge transfer disseminated the findings to stakeholders. The pilot case study used hyperspectral imaging to identify infested tubers and plants in a greenhouse setting. In addition to accurate - detection of infected plants, ncluding identification of nematode species, latent infections in tubers were accurately detected. Greenhouse-scale models were validated using UAV ultispectral and airborne hyperspectral imaging at field scale in Belgium. Overall, remote sensing methods show promise for early detection and management, but additional data are needed for more robust predictive models.
The research was supported by funds from the Slovenian Research Agency (ARIS) (MR 38128, MR 54720, P4-0072) and the European Food Safety Authority, P/EFSA/ALPHA/2018/02.