The use of video surveillance on roadways, railroads, buildings and other structures is becoming pervasive in the U.S. With the preponderance of video security networks, it is logical to suspect that the capability of existing video security solutions can be enhanced by integrating information from other sensor modalities that detect and provide actionable information about invisible threats, such as radiological and nuclear (RN) weapons. For this project, Deep Analytics LLC (DA) proposes the RadioNuclide Threat Tracking System (RNTTS). The RNTTS solution utilizes an existing state-of-the-art radiation sensor with integrated IP cameras, and cutting-edge automated target classification algorithms that correlate RN threat detections with the carriers (person or vehicle) in video streams. The overarching project goal is to supply DNDO and their partners with a low-cost and robust RN sensor technology that readily integrates with existing enterprise video management systems (VMSs) to track RN carriers. During Phase I, DA will develop and demonstrate cutting-edge machine learning algorithms to identify and track RN carriers in video streams. DA has partnered with Applied Research Associates, Inc. (ARA) in Phase I to leverage their decades of experience in RN sensor technology and to utilize their unique RN sensor for this effort. DA has begun establishing relationships with high-volume VMS vendors to assist with Phase I development of interface control documents that describe how RN carriers will be tracked on video streams with select VMSs. The completed Phase II prototype will enable drop-in integration of the RN technology with select VMS vendors.
