Combined use of uncrewed aerial vehicle and satellite remote sensing data to gain crop insights within Colombia

Drone (or more scientifically termed UAV) and satellite remote sensing data were captured over rice fields and palm plantations within Colombia. The satellites used included Sentinel-2 and Landsat-8, while the UAV platforms carried a MicaSense RedEdge 3 camera. In addition, ground-based phenological measurements were taken using instruments such as a Licor LAI-2200 (LI-COR, Lincoln, NE, United States) plant canopy analyzer. The results showed the difficulty in performing comparisons for highly heterogeneous crops. The phenological measurements were essentially point-based, while the UAV data had a spatial resolution of circa 0.25 m, and the satellite sensors varied in spatial resolution from 10 to 30 m. Both the rice crop fields and palm plantation plots have high spatial and temporal variability, which is differently captured when sensors have different spatial and temporal resolutions. Vegetation products were derived using band ratio approaches. Initial comparisons, not shown, focused on the satellite and UAV data correlation, but it was difficult to understand the causes of spatial variability. Difficulties in having comparable temporal matchups were caused by the satellite data being limited by cloud cover alongside the phenological and UAV data captured at discrete points in time due to data capture taking considerable resources. Therefore time-series plots were generated that had the advantage of allowing relationships to be seen even when measurements did not occur on the same day. The time-series plots demonstrated a relationship between the three measurement methodologies, showing the potential for monitoring crop growth using UAV and satellite data. Still, there were specific crop cycles where the results did not align. Therefore further work is needed to quantify why these differences occur, and the uncertainty in the final remote sensing–derived products.