In this paper the authors have investigated the applicability of trajectory clustering for anomaly detection in a coastal surveillance scenario. Some weaknesses have been identified, which questions the usefulness of trajectory clustering in this type of setting. This paper has presented spline-based clustering as a potential extension to alleviate some of these problems. In spline-based clustering, clusters are represented by splines, which effectively compress the main trajectory of clusters, as well as possibly reducing the computational complexity coupled with matching. Finally, also some minor algorithmic problems are addressed as whole trajectory matching, creation trajectory being outrun, and splits occurring during creation.
Trajectory clustering and coastal surveillance
Fahrstraßen-Clusterbildung und Küstenüberwachung
Informatik trifft Logistik, INFORMATIK, Jahrestagung der Gesellschaft für Informatik e.V., 37 ; 125-129
2007
5 Seiten, 1 Bild, 7 Quellen
Aufsatz (Konferenz)
Englisch
Moving object trajectory clustering method in intelligent surveillance video
British Library Online Contents | 2009
|Sea Sentinel: Undersea Coastal Surveillance System
British Library Online Contents | 1996
|