Using C-Band SAR and Temperature to Monitor Tropical Agricultural Fields

Abstract

This manuscript presents the analysis and a methodology for monitoring asparagus crops from remote sensing observations in a tropical region, where the meteorological conditions change considerably between production cycles. We use data provided by the Sentinel-1 satellite and temperature from a ground station to show how particularly the VH polarisation can be used for crop monitoring in order to visualise the canopy formation, the growth rate and canopy biomass, revealing high dependencies on temperature. We also present a multi-output machine learning regression algorithm trained on a rich spatio-temporal dataset in which each output estimates the number of asparagus stems that are present in each of the pre-defined crop phenological stages. We present the results of two separate scenarios: Using a single SAR image plus temperature as input for the algorithm and using multitemporal SAR data. Results show that the methodology presented is able to retrieve each individual monitored variable when using temperature as predictor with coefficients of determination (R2) above 0.85. Further research is currently investigating the added value of multitemporal SAR data to complement the predictions and potentially replace the temperature feature.