‘deforestation detection’
Spatio-Temporal Change Point Characterization in Tropical Forests using a Hybrid Pipeline
A hybrid pipeline including a modified BFAST, convolutional neural network embeddings and a XGBoost achieving a ~97% accuracy on a synthet-ic tropical-forest time series and classify into 5 major classes.
Posted by buchanle on Tuesday, May 19th, 2026 in May 2026, BFAST, convolutional neural network, deforestation detection, synthetic data, tropical forests