The proposed Sentinel-1 backscatter data pre-processing encompasses a number of standard SAR pre-processing steps in order to derive geo-coded intensity backscattering images. The pre-processing starts from ground detected (GRD) data provided by ESA. These operations are performed using the tools included in SNAP (Sentinel Application Platform) version 5.0 and some custom tools developed in Python by eurac researchers. In detail, the Sentinel-1 backscatter pre-processing operations are the following (S indicates SNAP tool, C indicates custom tool):
- Application of the precise Sentinel orbit to the data (S);
- Removal of the thermal noise present in the images (S);
- Removal of the noise present at the border of the images (C);
- Beta nought calibration (S);
- Assembly of the S1-tiles coming from the same track (S);
- Co-registration of the multi-temporal images (S);
- Application of the multi-temporal filtering (C);
- Application of the gamma-MAP spatial filtering (S);
- Geocoding and sigma nought calibration (S);
- Masking of the layover and shadow (C).
It is worth noting that we use the multi-temporal filter proposed in . This filter, which is suited for long time-series, allows a suppression of the speckle noise by preserving at the same time the geometrical detail. The pre-processed Sentinel-1 backscatter data uses the ETRS89 Lambert Azimuthal Equal-Area projection coordinate reference system and they have a spatial resolution of 20 m.
 Quegan, T. L. Toan, J. J. Yu, F. Ribbes and N. Floury, “Multitemporal ERS SAR Analysis Applied to Forest Mapping”, IEEE Transactions on Geoscience and Remote Sensing, vol. 38, no. 2, March 2000.
For visualizing the time series, you can use the Time Series Visualizer Tool.