We takes inspiration from TREAD* (Traffic Route Extraction and Anomaly Detection)[1], that learns a statistical model for maritime traffic from AIS data in an unsupervised way
National Oceanic and Atmospheric Administration: https://marinecadastre.gov/accessais
The information contained in the AIS data (generally) are of dynamic and static type
Area of interest: Hawaii
A bounding box is selected and it corresponds to the (chosen) specific area under surveillance
Waypoints (WPs): identify either stationary points (ports, offshore platforms, etc.), entry points and exit points.
WPs identification is based on DBSCAN (Density-Based Spatial Clustering of Applications with Noise) method
Note: The hyperparameters of DBSCAN are tuned manually.
Once the waypoints are learned, a route can be built by clustering the extracted vessel flows, which connect:
- Two ports
- Entry point and port
- Port and Exit point
- Entry point and Exit point (i.e., transit routes)
It is necessary to remove outliers from the routes to identify normal traffic. In TREAD, it is used the stochastic method Kernel Density Estimation (KDE).
[1] *Pallotta, Giuliana, Vespe, Michele and Bryan, Karna. "Vessel pattern knowledge discovery from AIS data: A framework for anomaly detection and route prediction." Entropy 15.6 (2013): 2218-2245.