A Python project for classification of forest structure (healthy vs. defoliated trees) using satellite remote sensing data.

The primary data source is quad-polarimetric RADARSAT-2 scenes. The added value of Sentinel-2 multispectral optical data is also studied.

Dependencies: Python 3.4 (to be compatible with SNAPPY) numpy (numeric computations, array handling, etc.) GDAL 2.1.0 (reading GeoTIFF images) tk 8.5.18 (tkinter for file dialouge) pandas (for reading Excel/CSV ground-truth data files) xlrd (requiered by pandas for reading Excel files) xlwt 1.3.0 (requiered by pandas for write Excel files, note that xlwt 1.3.0 does not officially support Python 3.4) scikit-learn 0.18.1 (for kNN and other Machine Learning methods)