The goal of this project is to develop a reproducible validation framework to assess the accuracy and temporal consistency of multi-temporal forest cover maps using independently collected ground-truth data. A harmonised multi-temporal reference dataset (points and polygons) of forest and related classes will be compiled through satellite photo-interpretation and forest inventories. The dataset will be used to validate ESA CCI LC (300 m) and GLC_FCS30 (30 m) products over selected areas in Italy and Spain. The project will deliver a harmonised validation dataset, a Python toolbox with demo notebooks in a GitHub repository, and a comprehensive accuracy assessment report.
Language: Python
Suggested Software/Libraries:
Collect Earth (Google Earth), Jupyter Notebook, pandas, NumPy, xarray, SciPy, GeoPandas, rasterio, matplotlib, GitHub