Amazon Rainforest Mapping using Sentinel-1 Short Time Series

Abstract:
The present work aims at showing the application of Sentinel-1 Short-Time-Series for the constant monitoring of forests at large scale. In particular, here we present how the interferometric capabilities of the Sentinel-1 satellite constellation can be exploited for the monthly mapping of the Amazon rainforest. We model the interferometric coherence temporal trend as an exponential decay and retrieve interferometric parameters that are eventually exploited as input features to a Random Forests classifier.