R and Google Earth Engine code for interpretable seagrass percent cover estimation from Google Satellite Embeddings, using a two-stage Random Forest framework (morphology classification + SPC regression) evaluated via leave-one-location-out cross-validation across five Indonesian coastal sites.
random-forest remote-sensing indonesia google-earth-engine interpretability seagrass satellite-embeddings percent-cover
-
Updated
Sep 3, 2026 - R