Persistent Surveillance in an Urban Environment

Automatically detecting threatening behavior of dismounts would be an important capability in providing security to both civilians and miltiary personnel. Numerous environments could benefit from improved protective security monitoring, including military bases, governmental offices or complexes, schools, public gatherings, and industrial plants. A system designed to address such a problem requires a collection of distributed sensors that monitor an area-of-interest (AOI) to provide persistent surveillance. Persistent surveillance does not simply mean having enough sensors to provide good coverage of an AOI, but the ability to process the sensor data and provide situation assessment. Toyon Research proposes to extend our work in the areas of EO/IR detection and tracking of dismounts as well as feature-aided tracking algorithms used for track stitching to the problem of continuously tracking dismounts in urban surveillance environments as they pass through multiple, disjoint sensor field-of-views (FOVs). In Phase I Toyon will investigate different features (e.g., facial, motion characteristics) that can be extracted from EO/IR imagery and used to uniquely characterize a dismount detected and tracked in a video sequence. Experiments will be carried out to demonstrate the ability to identify a dismount as it passes through several disjoint sensor FOVs from features collected on the dismount.