Value from the seemingly unrelated
The global war on terror presents unique challenges in its intelligence requirements, and there exists a need to monitor at-risk individuals, groups and installations. To detect these potential threats and create actionable intelligence to support expeditionary warfighting, there is a requirement for an array of sensors which can provide the necessary information, and a system which can ingest data from multiple sources (such as cameras and network analyzers) to detect complex threatening activities.
ObjectVideo is researching solutions to combine seemingly unrelated pieces of information gathered from different sensors and automatically learning the relationships between them to infer higher-level complex target interactions.
ObjectVideo's R&D Services has developed algorithms for tracking targets across multiple sensors with overlapping and non-overlapping fields of view. This framework integrates detection, classification and tracking of the target from each sensor, and solves the global data association problem to find site-wide target trajectories across multiple sensors. Our analysis algorithms can process video from sensors mounted on static or moving platforms such as unmanned aerial vehicles (UAVs), unmanned ground vehicles (UGVs), smart phones, etc. The software is capable of handling multi-modal data such as electro-optic (EO), infrared (IR), short-wave IR, thermal etc.
ObjectVideo is also conducting pioneering research in developing video analysis algorithms for pan-tilt-zoom (PTZ) cameras. Based on the specific requirements of our customers, we have developed multi-sensor solutions for various PTZ camera configurations such as leader-follower cameras, scanning cameras and active tracking cameras.