SAR data for the analysis of forest features: current Brazilian experiences

Abstract

This article presents some applications of airborne polarimetric and/or interferometric microwave data to improve the knowledge of forest structures. Three airborne SAR (Synthetic Aperture Radar) experiments were done in the Amazon tropical forest: (a) to study the spatial distribution of very large trees (VLTs) in the primary forest using local maximum filtering and a series of Markov processes; (b) to model the estimation of biomass variations in primary and secondary forests; (c) to analyze the retrieval timber volume over selective logging areas. Another experiment (d) was to investigate the relation among SAR data and the volumetric configuration in stands of Eucalyptus sp done by an airborne SAR imaging mission in SE-Brazil. To perform the objectives (b), (c) and (d) we carry out regression techniques, using variables got from multipolarimetric and/or interferometric SAR attributes and biophysical parameters from the forest cover. All data from the experiments were calibrated radiometrically to extract information during digital processing, besides an exhaustive field survey which was done simultaneously to SAR imaging, to know the physiognomy/structure of forest typology and to support the models produced for each case. The results of this series of experiments show advances at the techniques to treat SAR data, focusing on models of stand architecture and forest stock density. This will be helpful to increase the regional inventory and surveying procedures of forest conversion in the Brazilian territory in the near future.