Researchers are looking to develop an intelligent image system that can monitor large areas, perhaps miles wide, identify potential threats based on the correlation of events and anomalies it detects, and issue timely alerts with few false alarms.
Such a surveillance system is at the heart of what researchers at the Defense Advanced Research Projects Agency calls a Persistent Stare Exploitation and Analysis System (PerSEAS) that can automatically and interactively discover intelligence from optical or infra-red devices in the air on drones, for example, or spread over urban, suburban, and rural environments.
DARPA said it envisions two major applications for such a system. Perhaps most important, the first would use the system in a near real-time mode to receive alerts and warnings to react to and avert disasters. For example, if it notices a number of activities that were out of the usual, such as the gathering of lots of soldiers and trucks it could alert local authorities.
The second would be to use the data gathered from the system to use archived data from the system to analyze events, such as an attack to determine the movements and origins of the entities involved in the event, DARPA said. For both types of applications DARPA said the PerSEAS system ideally could receive or generate cues from/to other sensor systems to identify places or people of interest for additional details.
Overall the challenge is to identify potential threats based on the accumulation and correlation of multiple events and anomalies, and issue alerts so military folks in the field can take quick action or other officials can alert the public of problems, DARPA said.
Specifically the PerSEAS system will gather data from sensors and feed the data into an intelligent software engine supporting algorithms that discover relationships and anomalies that are indicative of suspicious behavior, match previously learned threat activity, or match user defined threat activity should also be incorporated, DARPA stated.
Read more here.
No comments:
Post a Comment