Crop stress detection from UAVs: best practices and lessons learned for exploiting sensor synergies

• DOI: 10.1007/s11119-024-10168-3
• Publication Date: 2024-08-11T10:01:49Z
• Authors: Erekle Chakhvashvili; Miriam Machwitz; Michal Antala; Offer Rozenstein; Egor Prikaziuk; Martin Schlerf; Paul Naethe; Quanxing Wan; Jan Komárek; Tomáš Klouek; Sebastian Wieneke; Bastian Siegmann; Shawn Kefauver; Marlena Kycko; Hamadou Balde; Veronica Sobejano...; Jose A. Jimenez-B...; Henning Buddenbaum; Lorenz Hänchen; Na Wang; Amit Weinman; Anshu Rastogi; Nitzan Malachy; Maria-Luisa Bucha...; Juliane Bendig; Uwe Rascher
• Keywords: Precision Agriculture

Introduction Detecting and monitoring crop stress is crucial for ensuring sufficient sustainable production. Recent advancements in unoccupied aerial vehicle (UAV) technology provide a promising approach to map key traits indicative of stress. While using single optical sensors mounted on UAVs could be monitor status general sense, implementing multiple that cover various spectral domains allow more precise characterization the interactions between crops biotic or abiotic stressors. Given novelty synergistic sensor detection, standardized procedures outlining their optimal use are currently lacking. Materials methods This study explores aspects acquiring high-quality multi-sensor data, including importance mission planning, characteristics, ancillary data. It also details essential data pre-processing steps like atmospheric correction highlights best practices fusion quality control. Results Successful acquisition depends timing, appropriate calibration, such as ground control points weather station information. When fusing different it should conducted at level physical units, with flags used exclude unstable biased measurements. The paper checklists, considering illumination conditions conducting test flights detection potential pitfalls. Conclusion Multi-sensor campaigns require careful planning not jeopardise success campaigns. provides practical information how combine UAV-mounted discuss proven scientific image post-processing context monitoring.