The specific objectives of this thesis dissertation will be to: (a) explore and process multi-sensor Earth Observation data to generate a spatially continuous mangrove forest cover map in an area of very high cloud cover persistence using the Copernicus Sentinel-1 and 2 imagery; (b) combine optical and radar data to evaluate whether classification accuracy is improved (c) explore transferability and reproducibility in other locations d) present the results of the analysis in the form of a web map.
To attain the objectives the following specific tasks will be carried out:
- Investigate, evaluate, and apply an existing mangrove forest cover mapping methodological approach.
- Classify mangrove forest cover from Sentinel-1 and Sentinel-2 satellite images by applying a machine-learning classification technique.
- Assess accuracy and validate results.
- Produce a web map application to visualize and access the classification outputs.
Acces to the GEE repository containing the scripts used to classify the mangrove cover. A GEE account is required
Access to the web map application presenting mangrove cover ditribution along the Colombian Pacific coastline.
Access to the output products from the classification, the thesis text, and poster