A Mestre-Quereda1, JM Lopez-Sanchez1, C Mosquera2, C van der Sande3, D Hoekman4 and M Noort5
1Institute for Computer Research (IUII), University of Alicante, P.O. Box 99, 03080 Alicante, Spain
2AgroAP: C.I. AGROAP SAS, Ka 100 #11-60, ofi. 819. Cali, Valle, Colombia
3eLEAF: Stadsring 65B, 3811 HN Amersfoort, The Netherlands
4SarVision: Agro Business Park 10, 6708 PW Wageningen, The Netherlands
5HCP International: Oostelijke Handelskade 831, 1019 BW Amsterdam, The Netherlands
Earth observation satellites provide useful information on the crop status which help organising cultivation practices, harvest planning, and yield optimisation. A project funded by the European Space Agency was carried out to establish an operational monitoring platform of sugarcane in the Cauca Valley. That project, called CostCutting4Sugarcane, served to demonstrate the potential of using Sentinel-2 images, acquired in the optical part of the spectrum, for this purpose. However, due to the presence of clouds during long periods of time, optical images are missing at key dates for cultivation. To overcome this issue, a new project funded by the European Commission has started, named DINOSAR, in which a methodology for integrating images from Sentinel-1 (a radar satellite, unaffected by clouds) with optical images is being developed for improving sugarcane monitoring. In the DINOSAR project, a one-year long in situ campaign is being carried out in distributed fields to gather relevant reference data on the plants (number of stems, plant height, fresh biomass, etc.) which is being used to develop a physical model that links the sugarcane features with the radar observations (backscatter and coherence). Then, leveraging this model and the available methodology with optical images, an integrated algorithm based on dynamic systems is implemented for combining both image types and for ensuring a valid and accurate input source of information to the decision-making tools developed for sugarcane. The integration methodology is aimed at increasing productivity and reducing the costs associated with fertilisation, also decreasing the environmental impact. This methodology will be transferred to other geographical locations and other crop types.