...
Data Pipeline
...
The pipeline in the figure above stems from the portion of the larger product pipeline outlined with a red dashed line in the figure on the Anomaly Detection via Machine Learning pageStandard Product Pipeline figure.
The AOI-Track Evaluator PGE MintPy PGE (Displacement Time-Series using MintPy), is triggered by the production of an S1-GUNW within a pre-defined volcanic AOI. The evaluator then publishes a JSON file upon receiving all GUNWs within the specified AOI track. The production of the JSON file triggers the Displacement Time-Series using MintPy PGE , which calculates the displacement time-series for the defined AOI, and publishes an HDF5 file to S3.
The time-series h5 file then triggers the a Volcano Anomaly Detection ML PGE, which outputs a Volcano Anomaly map. Currently two separate ML codes are under development, indicated by the diagram. Any given MintPy PGE output will only go through one of the ML options. (format/contents TBD)
The Volcano Anomaly Map is then run through the TBD “Production Rules for Detected Anomalies” PGE. If the map contains (TBD) value over (XX) threshold, then (something happens; probably notify the user who submitted the job).
AOI track Evaluator
...
MintPy PGE
Trigger product: S1-GUNW within a mintpy polygon
Inputs: S1-GUNWs, polygon, track number, start and end dates
Outputs: time series data (standard time-series product structure TBD): includes timeseries_demErr.h5, velocity.h5, maskTempCoh.h5,demErr.h5, avgSpatialCoherehnce.h5, temporalCoherence.h5, and inputs/geometrygeo.h5
Volcano Anomaly Detection ML PGE:
Trigger product:
Inputs: Time series data products from MintPy PGE
Outputs:
Notes: Need to incorporate reference point selection code again before running through the ML code
Use Cases
Use case 1: Maintaining forward keep-up production of S1-GUNW products for a volcanic AOI.
...