Anomaly Detection via Machine Learning
- Madeline Lambert
Both the SLC and S1-GUNW data products can undergo additional processing to meet scientific needs to track anomalies caused by various natural disasters. SLC stacks can be run through either the flood mapping anomaly detection pipeline (to produce an L3 flood proxy map) or the landslide anomaly detection pipeline (to produce a landslide anomaly map). S1-GUNWs can be run through the volcano anomaly detection pipeline to produce a displacement time series and detect potential anomalies within the time series signal.
Main Processing Pipeline
The figure above shows the Standard Product S1-GUNW Processing Pipeline.
The red dashed outline shows where the Volcanic Anomaly Detection pipeline branches off. The light blue dashed outline indicates where both the Flood Mapping Anomaly Detection and Landslide Anomaly Detection pipelines stem from.