Pixel to practice: multi-scale image data for calibrating remote-sensing-based winter wheat monitoring methods.
Jonas Anderegg, Flavian Tschurr, Norbert Kirchgessner, Simon Treier, Lukas Valentin Graf, Manuel Schmucki, Nicolin Caflisch, Camille Minguely, Bernhard Streit, Achim Walter
Author Information
Jonas Anderegg: Plant Pathology Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland. jonas.anderegg@usys.ethz.ch. ORCID
Flavian Tschurr: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland.
Norbert Kirchgessner: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland. ORCID
Simon Treier: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland.
Lukas Valentin Graf: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland.
Manuel Schmucki: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland.
Nicolin Caflisch: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland. ORCID
Camille Minguely: School of Agricultural, Forest and Food Sciences HAFL, Bern University of Applied Sciences, Zollikofen, 3052, Switzerland.
Bernhard Streit: School of Agricultural, Forest and Food Sciences HAFL, Bern University of Applied Sciences, Zollikofen, 3052, Switzerland.
Achim Walter: Crop Science Group, Department of Environmental System Science, ETH Zurich, Zurich, 8092, Switzerland.
Site-specific crop management in heterogeneous fields has emerged as a promising avenue towards increasing agricultural productivity whilst safeguarding the environment. However, successful implementation is hampered by insufficient availability of accurate spatial information on crop growth, vigor, and health status at large scales. Challenges persist particularly in interpreting remote sensing signals within commercial crop production due to the variability in canopy appearance resulting from diverse factors. Recently, high-resolution imagery captured from unmanned aerial vehicles has shown significant potential for calibrating and validating methods for remote sensing signal interpretation. We present a comprehensive multi-scale image dataset encompassing 35,000 high-resolution aerial RGB images, ground-based imagery, and Sentinel-2 satellite data from nine on-farm wheat fields in Switzerland. We provide geo-referenced orthomosaics, digital elevation models, and shapefiles, enabling detailed analysis of field characteristics across the growing season. In combination with rich meta data such as detailed records of crop husbandry, crop phenology, and yield maps, this data set enables key challenges in remote sensing-based trait estimation and precision agriculture to be addressed.
