Spatial Navigation for Context-Aware Video Surveillance

摘要

User interfaces for video surveillance networks are traditionally based on arrays of video displays, maps, and indirect controls. To address such interfaces' usability limitations when the number of cameras increases, some video surveillance systems aim to improve context awareness. However, interactive spatial navigation is still difficult: unconstrained, free 3D control is too complex, and a predefined camera path limits flexibility. Especially for live tracking of complex events along cameras, operators must make navigation decisions quickly and accurately on the basis of the actual situation. The novel spatial-navigation interface described in this article facilitates such video surveillance tasks. Users directly navigate in the visible video via the mouse, which lets them maintain attention on the action instead of an external navigation interface. While users track the action with the mouse, interactive 3D widgets augmented in the video provide visual updates regarding available camera transitions. Optimized visual transitions between individual videos ensure context awareness while users focus on the action. Several surveillance datasets, along with results from a pilot user evaluation, demonstrate this interface's effectiveness.